The second phase of a genetic study of patients with schizophrenia identified seven genome-wide significant regions of interest that need additional exploration and validation, new research shows.

Lead author Tiffany A. Greenwood, PhD, noted that clinically diverse schizophrenia patients often are grouped together to get large sample sizes for genome-wide association studies, which can miss specific features of the heterogeneous disorder. Instead, in phase 2 of the Consortium on the Genetics of Schizophrenia (COGS) study, Dr. Greenwood and associates sought to connect features clustered into endophenotypes with certain genetic regions of interest.

“As stable biomarkers of the underlying brain dysfunctions, endophenotypes hold promise for parsing clinical heterogeneity of schizophrenia and refining the genetic signal,” they wrote. The study was published in JAMA Psychiatry.

Using this approach among 1,533 participants, they found seven regions exceeding the conventional genome-wide association significance of P less than 5 x 10^–8, including regions associated with the endophenotypes of face memory (chromosome 3p21; effect size, –0.72; P = 4.2 x 10^–8), antisaccade task (chromosome 9q31; effect size, –0.24; P = 3.5 x 10^–8), and abstraction and mental flexibility (chromosome 10q23; effect size, –0.56; P = 1.5 x 10^–8).

Those endophenotypes and genes intersect theoretical molecular and biological processes that have been identified in other research and could explain underlying mechanisms of schizophrenia. For example, research has suggested that NRG3, which is near the region associated with abstraction and mental flexibility, and affects certain cellular signaling pathways, could be a locus of susceptibility; in particular, some variants of NRG3 have been associated with cognitive and psychotic symptom severity in previous research.

“Although shared genetic substrates appear likely, this is not a study of schizophrenia but rather a study of neurophysiological and neurocognitive deficits that occur in the general population but are more pronounced in the context of schizophrenia and have implications for treatment,” wrote Dr. Greenwood, of the University of California, San Diego, and associates.

One limitation of the study is that, as investigators have demonstrated elsewhere, P values can prove highly variable over the course of replication studies, so the significance thresholds shown in this study still could run the risk of hiding false positives or negatives.

“As many of the 11 endophenotypes have been endorsed as targets for the development of novel treatments for schizophrenia, a better understanding of the corresponding cellular and molecular processes may pave the way for precision-based medicine in schizophrenia and perhaps other psychiatric illnesses with a shared genetic liability,” Dr. Greenwood and associates concluded.

[email protected]

SOURCE: Greenwood TA et al. JAMA Psychiatry. 2019 Oct 9. doi: 10.1001/jamapsychiatry.2019.2850.

The second phase of a genetic study of patients with schizophrenia identified seven genome-wide significant regions of interest that need additional exploration and validation, new research shows.

Lead author Tiffany A. Greenwood, PhD, noted that clinically diverse schizophrenia patients often are grouped together to get large sample sizes for genome-wide association studies, which can miss specific features of the heterogeneous disorder. Instead, in phase 2 of the Consortium on the Genetics of Schizophrenia (COGS) study, Dr. Greenwood and associates sought to connect features clustered into endophenotypes with certain genetic regions of interest.

“As stable biomarkers of the underlying brain dysfunctions, endophenotypes hold promise for parsing clinical heterogeneity of schizophrenia and refining the genetic signal,” they wrote. The study was published in JAMA Psychiatry.

Using this approach among 1,533 participants, they found seven regions exceeding the conventional genome-wide association significance of P less than 5 x 10^–8, including regions associated with the endophenotypes of face memory (chromosome 3p21; effect size, –0.72; P = 4.2 x 10^–8), antisaccade task (chromosome 9q31; effect size, –0.24; P = 3.5 x 10^–8), and abstraction and mental flexibility (chromosome 10q23; effect size, –0.56; P = 1.5 x 10^–8).

Those endophenotypes and genes intersect theoretical molecular and biological processes that have been identified in other research and could explain underlying mechanisms of schizophrenia. For example, research has suggested that NRG3, which is near the region associated with abstraction and mental flexibility, and affects certain cellular signaling pathways, could be a locus of susceptibility; in particular, some variants of NRG3 have been associated with cognitive and psychotic symptom severity in previous research.

“Although shared genetic substrates appear likely, this is not a study of schizophrenia but rather a study of neurophysiological and neurocognitive deficits that occur in the general population but are more pronounced in the context of schizophrenia and have implications for treatment,” wrote Dr. Greenwood, of the University of California, San Diego, and associates.

One limitation of the study is that, as investigators have demonstrated elsewhere, P values can prove highly variable over the course of replication studies, so the significance thresholds shown in this study still could run the risk of hiding false positives or negatives.

“As many of the 11 endophenotypes have been endorsed as targets for the development of novel treatments for schizophrenia, a better understanding of the corresponding cellular and molecular processes may pave the way for precision-based medicine in schizophrenia and perhaps other psychiatric illnesses with a shared genetic liability,” Dr. Greenwood and associates concluded.

[email protected]

SOURCE: Greenwood TA et al. JAMA Psychiatry. 2019 Oct 9. doi: 10.1001/jamapsychiatry.2019.2850.

The second phase of a genetic study of patients with schizophrenia identified seven genome-wide significant regions of interest that need additional exploration and validation, new research shows.

Lead author Tiffany A. Greenwood, PhD, noted that clinically diverse schizophrenia patients often are grouped together to get large sample sizes for genome-wide association studies, which can miss specific features of the heterogeneous disorder. Instead, in phase 2 of the Consortium on the Genetics of Schizophrenia (COGS) study, Dr. Greenwood and associates sought to connect features clustered into endophenotypes with certain genetic regions of interest.

“As stable biomarkers of the underlying brain dysfunctions, endophenotypes hold promise for parsing clinical heterogeneity of schizophrenia and refining the genetic signal,” they wrote. The study was published in JAMA Psychiatry.

Using this approach among 1,533 participants, they found seven regions exceeding the conventional genome-wide association significance of P less than 5 x 10^–8, including regions associated with the endophenotypes of face memory (chromosome 3p21; effect size, –0.72; P = 4.2 x 10^–8), antisaccade task (chromosome 9q31; effect size, –0.24; P = 3.5 x 10^–8), and abstraction and mental flexibility (chromosome 10q23; effect size, –0.56; P = 1.5 x 10^–8).

Those endophenotypes and genes intersect theoretical molecular and biological processes that have been identified in other research and could explain underlying mechanisms of schizophrenia. For example, research has suggested that NRG3, which is near the region associated with abstraction and mental flexibility, and affects certain cellular signaling pathways, could be a locus of susceptibility; in particular, some variants of NRG3 have been associated with cognitive and psychotic symptom severity in previous research.

“Although shared genetic substrates appear likely, this is not a study of schizophrenia but rather a study of neurophysiological and neurocognitive deficits that occur in the general population but are more pronounced in the context of schizophrenia and have implications for treatment,” wrote Dr. Greenwood, of the University of California, San Diego, and associates.

One limitation of the study is that, as investigators have demonstrated elsewhere, P values can prove highly variable over the course of replication studies, so the significance thresholds shown in this study still could run the risk of hiding false positives or negatives.

“As many of the 11 endophenotypes have been endorsed as targets for the development of novel treatments for schizophrenia, a better understanding of the corresponding cellular and molecular processes may pave the way for precision-based medicine in schizophrenia and perhaps other psychiatric illnesses with a shared genetic liability,” Dr. Greenwood and associates concluded.

[email protected]

SOURCE: Greenwood TA et al. JAMA Psychiatry. 2019 Oct 9. doi: 10.1001/jamapsychiatry.2019.2850.

We are 20 minutes into the visit. My patient is 77 years old, a retired school administrator. She was sent to the oncology clinic for a new diagnosis of lung cancer with metastases to the liver and bones.

I was asking my usual questions – how did this all begin? – and I was hearing the usual answers. The cough that didn’t get better with antibiotics. The unintentional weight loss. The chest x-ray that looked “fuzzy.”

I continue: How many packs of cigarettes a day, and for how many years? Any family history of cancer?

These were my standard questions. They were met by hers: “How did I get this?”

I recently hosted a podcast on common, difficult questions we hear in hematology and oncology. How long do I have to live? What would you do if this were your family member?

This was another. There are variations to be sure. How, why, why me, what did I do, what didn’t I do, did my doctor miss it, if I had this or that test would they have caught it sooner?

When I was an internist, I talked about prevention. Meeting a new patient meant sizing them up for risk factors. In their habits I saw opportunities for healthier choices. In their family histories I gathered warning signs.

Now, I ask the same probing questions, but the purpose is not the same. Smoking, alcohol, family history, I ask these of everyone, I reassure them. It’s no longer about assessing risk. It’s not to place blame. But they read into the fact that I am asking, because they have asked themselves the same.

They ask why.

I try not to overdo the pity. I say that I’m sorry this is happening, but I don’t dwell. What I want to convey is the opposite – it’s normalcy. What I want to convey is: I’ve seen this a million times. This is where we are, and here is where we go. We don’t dwell or regret or wonder what if. My patients don’t want sympathy – at least, not from their doctor. They want a plan.

They ask: How did I get this?

It’s bad luck, I say. It’s a genetic mutation causing a cell to replicate.

My answers do not always satisfy their questions. Because it’s not a question seeking an informational answer. The truth is, medically and existentially, I don’t know. None of us do. The question is an existential itch no medical jargon can scratch.

I have a modern Hippocratic oath tacked to a wall in my room. “I will prevent disease whenever I can, because prevention is preferable to cure,” it says. True, but that offers little solace to those who already have the illness. Yes, we need prevention. And we need a path forward when tragedy has already struck.

I am humbled when I meet a new cancer patient because the visit is a metaphor for a nonjudgmental life. There’s something beautiful about meeting someone exactly where they are, where decisions made in the past are as irrelevant to me now as they were to the cancer.

When they inevitably ask “how did I get this?” and I answer, what I’m really saying is this: I don’t care what you did, or didn’t do, or how we got here. But we are here, and so I am here with you, and from now on the only place we care about is here and now, the only direction forward.

Dr. Yurkiewicz is a fellow in hematology and oncology at Stanford (Calif.) University. Follow her on Twitter @ilanayurkiewicz and listen to her each week on the Blood & Cancer podcast.

We are 20 minutes into the visit. My patient is 77 years old, a retired school administrator. She was sent to the oncology clinic for a new diagnosis of lung cancer with metastases to the liver and bones.

I was asking my usual questions – how did this all begin? – and I was hearing the usual answers. The cough that didn’t get better with antibiotics. The unintentional weight loss. The chest x-ray that looked “fuzzy.”

I continue: How many packs of cigarettes a day, and for how many years? Any family history of cancer?

These were my standard questions. They were met by hers: “How did I get this?”

I recently hosted a podcast on common, difficult questions we hear in hematology and oncology. How long do I have to live? What would you do if this were your family member?

This was another. There are variations to be sure. How, why, why me, what did I do, what didn’t I do, did my doctor miss it, if I had this or that test would they have caught it sooner?

When I was an internist, I talked about prevention. Meeting a new patient meant sizing them up for risk factors. In their habits I saw opportunities for healthier choices. In their family histories I gathered warning signs.

Now, I ask the same probing questions, but the purpose is not the same. Smoking, alcohol, family history, I ask these of everyone, I reassure them. It’s no longer about assessing risk. It’s not to place blame. But they read into the fact that I am asking, because they have asked themselves the same.

They ask why.

I try not to overdo the pity. I say that I’m sorry this is happening, but I don’t dwell. What I want to convey is the opposite – it’s normalcy. What I want to convey is: I’ve seen this a million times. This is where we are, and here is where we go. We don’t dwell or regret or wonder what if. My patients don’t want sympathy – at least, not from their doctor. They want a plan.

They ask: How did I get this?

It’s bad luck, I say. It’s a genetic mutation causing a cell to replicate.

My answers do not always satisfy their questions. Because it’s not a question seeking an informational answer. The truth is, medically and existentially, I don’t know. None of us do. The question is an existential itch no medical jargon can scratch.

I have a modern Hippocratic oath tacked to a wall in my room. “I will prevent disease whenever I can, because prevention is preferable to cure,” it says. True, but that offers little solace to those who already have the illness. Yes, we need prevention. And we need a path forward when tragedy has already struck.

I am humbled when I meet a new cancer patient because the visit is a metaphor for a nonjudgmental life. There’s something beautiful about meeting someone exactly where they are, where decisions made in the past are as irrelevant to me now as they were to the cancer.

When they inevitably ask “how did I get this?” and I answer, what I’m really saying is this: I don’t care what you did, or didn’t do, or how we got here. But we are here, and so I am here with you, and from now on the only place we care about is here and now, the only direction forward.

Dr. Yurkiewicz is a fellow in hematology and oncology at Stanford (Calif.) University. Follow her on Twitter @ilanayurkiewicz and listen to her each week on the Blood & Cancer podcast.

We are 20 minutes into the visit. My patient is 77 years old, a retired school administrator. She was sent to the oncology clinic for a new diagnosis of lung cancer with metastases to the liver and bones.

I was asking my usual questions – how did this all begin? – and I was hearing the usual answers. The cough that didn’t get better with antibiotics. The unintentional weight loss. The chest x-ray that looked “fuzzy.”

I continue: How many packs of cigarettes a day, and for how many years? Any family history of cancer?

These were my standard questions. They were met by hers: “How did I get this?”

I recently hosted a podcast on common, difficult questions we hear in hematology and oncology. How long do I have to live? What would you do if this were your family member?

This was another. There are variations to be sure. How, why, why me, what did I do, what didn’t I do, did my doctor miss it, if I had this or that test would they have caught it sooner?

When I was an internist, I talked about prevention. Meeting a new patient meant sizing them up for risk factors. In their habits I saw opportunities for healthier choices. In their family histories I gathered warning signs.

Now, I ask the same probing questions, but the purpose is not the same. Smoking, alcohol, family history, I ask these of everyone, I reassure them. It’s no longer about assessing risk. It’s not to place blame. But they read into the fact that I am asking, because they have asked themselves the same.

They ask why.

I try not to overdo the pity. I say that I’m sorry this is happening, but I don’t dwell. What I want to convey is the opposite – it’s normalcy. What I want to convey is: I’ve seen this a million times. This is where we are, and here is where we go. We don’t dwell or regret or wonder what if. My patients don’t want sympathy – at least, not from their doctor. They want a plan.

They ask: How did I get this?

It’s bad luck, I say. It’s a genetic mutation causing a cell to replicate.

My answers do not always satisfy their questions. Because it’s not a question seeking an informational answer. The truth is, medically and existentially, I don’t know. None of us do. The question is an existential itch no medical jargon can scratch.

I have a modern Hippocratic oath tacked to a wall in my room. “I will prevent disease whenever I can, because prevention is preferable to cure,” it says. True, but that offers little solace to those who already have the illness. Yes, we need prevention. And we need a path forward when tragedy has already struck.

I am humbled when I meet a new cancer patient because the visit is a metaphor for a nonjudgmental life. There’s something beautiful about meeting someone exactly where they are, where decisions made in the past are as irrelevant to me now as they were to the cancer.

When they inevitably ask “how did I get this?” and I answer, what I’m really saying is this: I don’t care what you did, or didn’t do, or how we got here. But we are here, and so I am here with you, and from now on the only place we care about is here and now, the only direction forward.

Dr. Yurkiewicz is a fellow in hematology and oncology at Stanford (Calif.) University. Follow her on Twitter @ilanayurkiewicz and listen to her each week on the Blood & Cancer podcast.

Sexually transmitted infections (STIs) such as gonorrhea, chlamydia, and syphilis are still increasing in incidence and probably will continue to do so in the near future. Moreover, drug-resistant strains of Neisseria gonorrhoeae are emerging, as are less-known organisms such as Mycoplasma genitalium.

Now the good news: new tests for STIs are available or are coming! Based on nucleic acid amplification, these tests can be performed at the point of care, so that patients can leave the clinic with an accurate diagnosis and proper treatment for themselves and their sexual partners. Also, the tests can be run on samples collected by the patients themselves, either swabs or urine collections, eliminating the need for invasive sampling and making doctor-shy patients more likely to come in to be treated.¹ We hope that by using these sensitive and accurate tests we can begin to bend the upward curve of STIs and be better antimicrobial stewards.²

This article reviews current issues surrounding STI control, and provides detailed guidance on recognizing, testing for, and treating gonorrhea, chlamydia, trichomoniasis, and M genitalium infection.

STI RATES ARE HIGH AND RISING

STIs are among the most common acute infectious diseases worldwide, with an estimated 1 million new curable cases every day.³ Further, STIs have major impacts on sexual, reproductive, and psychological health.

In the United States, rates of reportable STIs (chlamydia, gonorrhea, and syphilis) are rising.⁴ In addition, more-sensitive tests for trichomoniasis, which is not a reportable infection in any state, have revealed it to be more prevalent than previously thought.⁵

BARRIERS AND CHALLENGES TO DIAGNOSIS

The medical system does not fully meet the needs of some populations, including young people and men who have sex with men, regarding their sexual and reproductive health. 

Ongoing barriers among young people include reluctance to use available health services, limited access to STI testing, worries about confidentiality, and the shame and stigma associated with STIs.⁶

Men who have sex with men have a higher incidence of STIs than other groups. Since STIs are associated with a higher risk of human immunodeficiency virus (HIV) infection, it is important to detect, diagnose, and manage STIs in this group—and in all high-risk groups. Rectal STIs are an independent risk factor for incident HIV infection.⁷ In addition, many men who have sex with men face challenges navigating the emotional, physical, and cognitive aspects of adolescence, a voyage further complicated by mental health issues, unprotected sexual encounters, and substance abuse in many, especially among minority youth.⁸ These same factors also impair their ability to access resources for preventing and treating HIV and other STIs.

STI diagnosis is often missed

Most people who have STIs feel no symptoms, which increases the importance of risk-based screening to detect these infections.^9,10 In many other cases, STIs manifest with nonspecific genitourinary symptoms that are mistaken for urinary tract infection. Tomas et al¹¹ found that of 264 women who presented to an emergency department with genitourinary symptoms or were being treated for urinary tract infection, 175 were given a diagnosis of a urinary tract infection. Of these, 100 (57%) were treated without performing a urine culture; 60 (23%) of the 264 women had 1 or more positive STI tests, 22 (37%) of whom did not receive treatment for an STI.

Poor follow-up of patients and partners

Patients with STIs need to be retested 3 months after treatment to make sure the treatment was effective. Another reason for follow-up is that these patients are at higher risk of another infection within a year.¹²

Although treating patients’ partners has been shown to reduce reinfection rates, fewer than one-third of STIs (including HIV infections) were recognized through partner notification between 2010 and 2012 in a Dutch study, in men who have sex with men and in women.¹³ Challenges included partners who could not be identified among men who have sex with men, failure of heterosexual men to notify their partners, and lower rates of partner notification for HIV.  

In the United States, “expedited partner therapy” allows healthcare providers to provide a prescription or medications to partners of patients diagnosed with chlamydia or gonorrhea without examining the partner.¹⁴ While this approach is legal in most states, implementation can be challenging.¹⁵

STI EVALUATION

History and physical examination

A complete sexual history helps in estimating the patient’s risk of an STI and applying appropriate risk-based screening. Factors such as sexual practices, use of barrier protection, and history of STIs should be discussed.

Physical examination is also important. Although some patients may experience discomfort during a genital or pelvic examination, omitting this step may lead to missed diagnoses in women with STIs.¹⁶

Laboratory testing

Laboratory testing for STIs helps ensure accurate diagnosis and treatment. Empiric treatment without testing could give a patient a false sense of health by missing an infection that is not currently causing symptoms but that could later worsen or have lasting complications. Failure to test patients also misses the opportunity for partner notification, linkage to services, and follow-up testing.

Many of the most common STIs, including gonorrhea, chlamydia, and trichomoniasis, can be detected using vaginal, cervical, or urethral swabs or first-catch urine (from the initial urine stream). In studies that compared various sampling methods,¹⁷ self-collected urine samples for gonorrhea in men were nearly as good as clinician-collected swabs of the urethra. In women, self-collected vaginal swabs for gonorrhea and chlamydia were nearly as good as clinician-collected vaginal swabs. While urine specimens are acceptable for chlamydia testing in women, their sensitivity may be slightly lower than with vaginal and endocervical swab specimens.^18,19

A major advantage of urine specimens for STI testing is that collection is noninvasive and is therefore more likely to be acceptable to patients. Urine testing can also be conducted in a variety of nonclinical settings such as health fairs, pharmacy-based screening programs, and express STI testing sites, thus increasing availability.

To prevent further transmission and morbidity and to aid in public health efforts, it is critical to recognize the cause of infectious cervicitis and urethritis and to screen for STIs according to guidelines.¹² Table 1 summarizes current screening and laboratory testing recommendations.

Gonorrhea and chlamydia are the 2 most frequently reported STIs in the United States, with more than 550,000 cases of gonorrhea and 1.7 million cases of chlamydia reported in 2017.⁴

Both infections present similarly: cervicitis or urethritis characterized by discharge (mucopurulent discharge with gonorrhea) and dysuria. Untreated, they can lead to pelvic inflammatory disease, inflammation, and infertility.

Extragenital infections can be asymptomatic or cause exudative pharyngitis or proctitis. Most people in whom chlamydia is detected from pharyngeal specimens are asymptomatic. When pharyngeal symptoms exist secondary to gonorrheal infection, they typically include sore throat and pharyngeal exudates. However, Komaroff et al,²⁰ in a study of 192 men and women who presented with sore throat, found that only 2 (1%) tested positive for N gonorrhoeae.

Screening for gonorrhea and chlamydia

Best practices include screening for gonorrhea and chlamydia as follows^21–23:

• Every year in sexually active women through age 25 (including during pregnancy) and in older women who have risk factors for infection¹²
• At least every year in men who have sex with men, at all sites of sexual contact (urethra, pharynx, rectum), along with testing for HIV and syphilis
• Every 3 to 6 months in men who have sex with men who have multiple or anonymous partners, who are sexually active and use illicit drugs, or who have partners who use illicit drugs
• Possibly every year in young men who live in high-prevalence areas or who are seen in certain clinical settings, such as STI and adolescent clinics.

Specimens. A vaginal swab is preferred for screening in women. Several studies have shown that self-collected swabs have clinical sensitivity and specificity comparable to that of provider-collected samples.^17,24 First-catch urine or endocervical swabs have similar performance characteristics and are also acceptable. In men, urethral swabs or first-catch urine samples are appropriate for screening for urogenital infections.

Testing methods. Testing for both pathogens should be done simultaneously with a nucleic acid amplification test (NAAT). Commercially available NAATs are more sensitive than culture and antigen testing for detecting gonorrhea and chlamydia.^25–27

Most assays are approved by the US Food and Drug Administration (FDA) for testing vaginal, urethral, cervical, and urine specimens. Until recently, no commercial assay was cleared for testing extragenital sites, but recommendations for screening extragenital sites prompted many clinical laboratories to validate throat and rectal swabs for use with NAATs, which are more sensitive than culture at these sites.^25,28 The recent FDA approval of extragenital specimen types for 2 commercially available assays may increase the availability of testing for these sites.

Data on the utility of NAATs for detecting chlamydia and gonorrhea in children are limited, and many clinical laboratories have not validated molecular methods for testing in children. Current guidelines specific to this population should be followed regarding test methods and preferred specimen types.^12,29,30

Although gonococcal infection is usually diagnosed with culture-independent molecular methods, antimicrobial resistance is emerging. Thus, failure of the combination of ceftriaxone and azithromycin should prompt culture-based follow-up testing to determine antimicrobial susceptibility.

Strategies for treatment and control

Historically, people treated for gonorrhea have been treated for chlamydia at the same time, as these diseases tend to go together. This can be with a single intramuscular dose of ceftriaxone for the gonorrhea plus a single oral dose of azithromycin for the chlamydia.¹² For patients who have only gonorrhea, this double regimen may help prevent the development of resistant gonorrhea strains.

All the patient’s sexual partners in the previous 60 days should be tested and treated, and expedited partner therapy should be offered if possible. Patients should be advised to have no sexual contact until they complete the treatment, or 7 days after single-dose treatment. Testing should be repeated 3 months after treatment.

A member of the Mycoplasmataceae family, M genitalium was originally identified as a pathogen in the early 1980s but has only recently emerged as an important cause of STI. Studies indicate that it is responsible for 10% to 20% of cases of nongonococcal urethritis and 10% to 30% of cases of cervicitis.^31–33 Additionally, 2% to 22% of cases of pelvic inflammatory disease have evidence of M genitalium.^34,35

However, data on M genitalium prevalence are suspect because the organism is hard to identify—lacking a cell wall, it is undetectable by Gram stain.³⁶ Although it has been isolated in respiratory and synovial fluids, it has so far been recognized to be clinically important only in the urogenital tract. It can persist for years in infected patients by exploiting specialized cell-surface structures to invade cells.³⁶ Once inside a cell, it triggers secretion of mycoplasmal toxins and destructive metabolites such as hydrogen peroxide, evading the host immune system as it does so.³⁷

Testing guidelines for M genitalium

Current guidelines do not recommend routine screening for M genitalium, and no commercial test was available until recently.¹² Although evidence suggests that M genitalium is independently associated with preterm birth and miscarriages,³⁸ routine screening of pregnant women is not recommended.¹²

Testing for M genitalium should be considered in cases of persistent or recurrent nongonococcal urethritis in patients who test negative for gonorrhea and chlamydia or for whom treatment has failed.¹² Many isolates exhibit genotypic resistance to macrolide antibiotics, which are often the first-line therapy for nongonococcal urethritis.³⁹

Further study is needed to evaluate the potential impact of routine screening for M genitalium on the reproductive and sexual health of at-risk populations.

Diagnostic tests for M genitalium

Awareness of M genitalium as a cause of nongonococcal urethritis has been hampered by a dearth of diagnostic tests.⁴⁰ The organism’s fastidious requirements and extremely slow growth preclude culture as a practical method of diagnosis.⁴¹ Serologic assays are dogged by cross-reactivity and poor sensitivity.^42,43 Thus, molecular assays for detecting M genitalium and associated resistance markers are preferred for diagnosis.¹²

Several molecular tests are approved, available, and in use in Europe for diagnosing M genitalium infection,⁴⁰ and in January 2019 the FDA approved a molecular test that can detect M genitalium in urine specimens and vaginal, endocervical, urethral, and penile meatal swabs. Although vaginal swabs are preferred for this assay because they have higher sensitivity (92% for provider-collected and 99% for patient-collected swabs), urine specimens are acceptable, with a sensitivity of 78%.⁴⁴

At least 1 company is seeking FDA clearance for another molecular diagnostic assay for detecting M genitalium and markers of macrolide resistance in urine and genital swab specimens. Such assays may facilitate appropriate treatment.

Clinicians should stay abreast of diagnostic testing options, which are likely to become more readily available soon.

A high rate of macrolide resistance

Because M genitalium lacks a cell wall, antibiotics such as beta-lactams that target cell wall synthesis are ineffective.

Regimens for treating M genitalium are outlined in Table 2.¹² Azithromycin is more effective than doxycycline. However, as many as 50% of strains were macrolide-resistant in a cohort of US female patients.⁴⁵ Given the high incidence of treatment failure with azithromycin 1 g, it is thought that this regimen might select for resistance. For cases in which symptoms persist, a 1- to 2-week course of moxifloxacin is recommended.¹² However, this has not been validated by clinical trials, and failures of the 7-day regimen have been reported.⁴⁶

Partners of patients who test positive for M genitalium should also be tested and undergo clinically applicable screening for nongonococcal urethritis, cervicitis, and pelvic inflammatory disease.¹²

TRICHOMONIASIS

Trichomoniasis, caused by the parasite Trichomonas vaginalis, is the most prevalent nonviral STI in the United States. It disproportionately affects black women, in whom the prevalence is 13%, compared with 1% in non-Hispanic white women.⁴⁷ It is also present in 26% of women with symptoms who are seen in STI clinics and is highly prevalent in incarcerated populations. It is uncommon in men who have sex with men.⁴⁸

In men, trichomoniasis manifests as urethritis, epididymitis, or prostatitis. While most infected women have no symptoms, they may experience vaginitis with discharge that is diffuse, frothy, pruritic, malodorous, or yellow-green. Vaginal and cervical erythema (“strawberry cervix”) can also occur.

Screening for trichomoniasis

Current guidelines of the US Centers for Disease Control and Prevention (CDC) recommend testing for T vaginalis in women who have symptoms and routinely screening in women who are HIV-positive, regardless of symptoms. There is no evidence to support routine screening of pregnant women without symptoms, and pregnant women who do have symptoms should be evaluated according to the same guidelines as for nonpregnant women.¹² Testing can be considered in patients who have no symptoms but who engage in high-risk behaviors and in areas of high prevalence.

A lack of studies using sensitive methods for T vaginalis detection has hampered a true estimation of disease burden and at-risk populations. Screening recommendations may evolve in upcoming clinical guidelines as the field advances.

As infection can recur, women should be retested 3 months after initial diagnosis.¹²

NAAT is the preferred test for trichomoniasis

Commercially available diagnostic tests for trichomoniasis include culture, antigen testing, and NAAT.⁴⁹ While many clinicians do their own wet-mount microscopy for a rapid result, this method has low sensitivity.⁵⁰ Similarly, antigen testing and culture perform poorly compared with NAATs, which are the gold standard for detection.^51,52 A major advantage of NAATs for T vaginalis detection is that they combine high sensitivity and fast results, facilitating diagnosis and appropriate treatment of patients and their partners.

In spite of these benefits, adoption of molecular diagnostic testing for T vaginalis has lagged behind that for chlamydia and gonorrhea.⁵³ FDA-cleared NAATs are available for testing vaginal, cervical, or urine specimens from women, but until recently, there were no approved assays for testing in men. The Cepheid Xpert TV assay, which is valid for male urine specimens to diagnose other sexually transmitted diseases, has demonstrated excellent diagnostic sensitivity for T vaginalis in men and women.⁵⁴ Interestingly, a large proportion of male patients in this study had no symptoms, suggesting that screening of men in high-risk groups may be warranted.

7-day metronidazole treatment beats single-dose treatment

The first-line treatment for trichomoniasis has been a single dose of metronidazole 2 g by mouth, but in a recent randomized controlled trial,⁵⁵ a course of 500 mg by mouth twice a day for 7 days was 45% more effective at 4 weeks than a single dose, and it should now be the preferred regimen.

In clinical trials,⁵⁶ a single dose of tinidazole 2 g orally was equivalent or superior to metronidazole 2 g and had fewer gastrointestinal side effects, but it is more expensive.

Sexually transmitted infections (STIs) such as gonorrhea, chlamydia, and syphilis are still increasing in incidence and probably will continue to do so in the near future. Moreover, drug-resistant strains of Neisseria gonorrhoeae are emerging, as are less-known organisms such as Mycoplasma genitalium.

Now the good news: new tests for STIs are available or are coming! Based on nucleic acid amplification, these tests can be performed at the point of care, so that patients can leave the clinic with an accurate diagnosis and proper treatment for themselves and their sexual partners. Also, the tests can be run on samples collected by the patients themselves, either swabs or urine collections, eliminating the need for invasive sampling and making doctor-shy patients more likely to come in to be treated.¹ We hope that by using these sensitive and accurate tests we can begin to bend the upward curve of STIs and be better antimicrobial stewards.²

This article reviews current issues surrounding STI control, and provides detailed guidance on recognizing, testing for, and treating gonorrhea, chlamydia, trichomoniasis, and M genitalium infection.

STI RATES ARE HIGH AND RISING

STIs are among the most common acute infectious diseases worldwide, with an estimated 1 million new curable cases every day.³ Further, STIs have major impacts on sexual, reproductive, and psychological health.

In the United States, rates of reportable STIs (chlamydia, gonorrhea, and syphilis) are rising.⁴ In addition, more-sensitive tests for trichomoniasis, which is not a reportable infection in any state, have revealed it to be more prevalent than previously thought.⁵

BARRIERS AND CHALLENGES TO DIAGNOSIS

The medical system does not fully meet the needs of some populations, including young people and men who have sex with men, regarding their sexual and reproductive health. 

Ongoing barriers among young people include reluctance to use available health services, limited access to STI testing, worries about confidentiality, and the shame and stigma associated with STIs.⁶

Men who have sex with men have a higher incidence of STIs than other groups. Since STIs are associated with a higher risk of human immunodeficiency virus (HIV) infection, it is important to detect, diagnose, and manage STIs in this group—and in all high-risk groups. Rectal STIs are an independent risk factor for incident HIV infection.⁷ In addition, many men who have sex with men face challenges navigating the emotional, physical, and cognitive aspects of adolescence, a voyage further complicated by mental health issues, unprotected sexual encounters, and substance abuse in many, especially among minority youth.⁸ These same factors also impair their ability to access resources for preventing and treating HIV and other STIs.

STI diagnosis is often missed

Most people who have STIs feel no symptoms, which increases the importance of risk-based screening to detect these infections.^9,10 In many other cases, STIs manifest with nonspecific genitourinary symptoms that are mistaken for urinary tract infection. Tomas et al¹¹ found that of 264 women who presented to an emergency department with genitourinary symptoms or were being treated for urinary tract infection, 175 were given a diagnosis of a urinary tract infection. Of these, 100 (57%) were treated without performing a urine culture; 60 (23%) of the 264 women had 1 or more positive STI tests, 22 (37%) of whom did not receive treatment for an STI.

Poor follow-up of patients and partners

Patients with STIs need to be retested 3 months after treatment to make sure the treatment was effective. Another reason for follow-up is that these patients are at higher risk of another infection within a year.¹²

Although treating patients’ partners has been shown to reduce reinfection rates, fewer than one-third of STIs (including HIV infections) were recognized through partner notification between 2010 and 2012 in a Dutch study, in men who have sex with men and in women.¹³ Challenges included partners who could not be identified among men who have sex with men, failure of heterosexual men to notify their partners, and lower rates of partner notification for HIV.  

In the United States, “expedited partner therapy” allows healthcare providers to provide a prescription or medications to partners of patients diagnosed with chlamydia or gonorrhea without examining the partner.¹⁴ While this approach is legal in most states, implementation can be challenging.¹⁵

STI EVALUATION

History and physical examination

A complete sexual history helps in estimating the patient’s risk of an STI and applying appropriate risk-based screening. Factors such as sexual practices, use of barrier protection, and history of STIs should be discussed.

Physical examination is also important. Although some patients may experience discomfort during a genital or pelvic examination, omitting this step may lead to missed diagnoses in women with STIs.¹⁶

Laboratory testing

Laboratory testing for STIs helps ensure accurate diagnosis and treatment. Empiric treatment without testing could give a patient a false sense of health by missing an infection that is not currently causing symptoms but that could later worsen or have lasting complications. Failure to test patients also misses the opportunity for partner notification, linkage to services, and follow-up testing.

Many of the most common STIs, including gonorrhea, chlamydia, and trichomoniasis, can be detected using vaginal, cervical, or urethral swabs or first-catch urine (from the initial urine stream). In studies that compared various sampling methods,¹⁷ self-collected urine samples for gonorrhea in men were nearly as good as clinician-collected swabs of the urethra. In women, self-collected vaginal swabs for gonorrhea and chlamydia were nearly as good as clinician-collected vaginal swabs. While urine specimens are acceptable for chlamydia testing in women, their sensitivity may be slightly lower than with vaginal and endocervical swab specimens.^18,19

A major advantage of urine specimens for STI testing is that collection is noninvasive and is therefore more likely to be acceptable to patients. Urine testing can also be conducted in a variety of nonclinical settings such as health fairs, pharmacy-based screening programs, and express STI testing sites, thus increasing availability.

To prevent further transmission and morbidity and to aid in public health efforts, it is critical to recognize the cause of infectious cervicitis and urethritis and to screen for STIs according to guidelines.¹² Table 1 summarizes current screening and laboratory testing recommendations.

Gonorrhea and chlamydia are the 2 most frequently reported STIs in the United States, with more than 550,000 cases of gonorrhea and 1.7 million cases of chlamydia reported in 2017.⁴

Both infections present similarly: cervicitis or urethritis characterized by discharge (mucopurulent discharge with gonorrhea) and dysuria. Untreated, they can lead to pelvic inflammatory disease, inflammation, and infertility.

Extragenital infections can be asymptomatic or cause exudative pharyngitis or proctitis. Most people in whom chlamydia is detected from pharyngeal specimens are asymptomatic. When pharyngeal symptoms exist secondary to gonorrheal infection, they typically include sore throat and pharyngeal exudates. However, Komaroff et al,²⁰ in a study of 192 men and women who presented with sore throat, found that only 2 (1%) tested positive for N gonorrhoeae.

Screening for gonorrhea and chlamydia

Best practices include screening for gonorrhea and chlamydia as follows^21–23:

• Every year in sexually active women through age 25 (including during pregnancy) and in older women who have risk factors for infection¹²
• At least every year in men who have sex with men, at all sites of sexual contact (urethra, pharynx, rectum), along with testing for HIV and syphilis
• Every 3 to 6 months in men who have sex with men who have multiple or anonymous partners, who are sexually active and use illicit drugs, or who have partners who use illicit drugs
• Possibly every year in young men who live in high-prevalence areas or who are seen in certain clinical settings, such as STI and adolescent clinics.

Specimens. A vaginal swab is preferred for screening in women. Several studies have shown that self-collected swabs have clinical sensitivity and specificity comparable to that of provider-collected samples.^17,24 First-catch urine or endocervical swabs have similar performance characteristics and are also acceptable. In men, urethral swabs or first-catch urine samples are appropriate for screening for urogenital infections.

Testing methods. Testing for both pathogens should be done simultaneously with a nucleic acid amplification test (NAAT). Commercially available NAATs are more sensitive than culture and antigen testing for detecting gonorrhea and chlamydia.^25–27

Most assays are approved by the US Food and Drug Administration (FDA) for testing vaginal, urethral, cervical, and urine specimens. Until recently, no commercial assay was cleared for testing extragenital sites, but recommendations for screening extragenital sites prompted many clinical laboratories to validate throat and rectal swabs for use with NAATs, which are more sensitive than culture at these sites.^25,28 The recent FDA approval of extragenital specimen types for 2 commercially available assays may increase the availability of testing for these sites.

Data on the utility of NAATs for detecting chlamydia and gonorrhea in children are limited, and many clinical laboratories have not validated molecular methods for testing in children. Current guidelines specific to this population should be followed regarding test methods and preferred specimen types.^12,29,30

Although gonococcal infection is usually diagnosed with culture-independent molecular methods, antimicrobial resistance is emerging. Thus, failure of the combination of ceftriaxone and azithromycin should prompt culture-based follow-up testing to determine antimicrobial susceptibility.

Strategies for treatment and control

Historically, people treated for gonorrhea have been treated for chlamydia at the same time, as these diseases tend to go together. This can be with a single intramuscular dose of ceftriaxone for the gonorrhea plus a single oral dose of azithromycin for the chlamydia.¹² For patients who have only gonorrhea, this double regimen may help prevent the development of resistant gonorrhea strains.

All the patient’s sexual partners in the previous 60 days should be tested and treated, and expedited partner therapy should be offered if possible. Patients should be advised to have no sexual contact until they complete the treatment, or 7 days after single-dose treatment. Testing should be repeated 3 months after treatment.

A member of the Mycoplasmataceae family, M genitalium was originally identified as a pathogen in the early 1980s but has only recently emerged as an important cause of STI. Studies indicate that it is responsible for 10% to 20% of cases of nongonococcal urethritis and 10% to 30% of cases of cervicitis.^31–33 Additionally, 2% to 22% of cases of pelvic inflammatory disease have evidence of M genitalium.^34,35

However, data on M genitalium prevalence are suspect because the organism is hard to identify—lacking a cell wall, it is undetectable by Gram stain.³⁶ Although it has been isolated in respiratory and synovial fluids, it has so far been recognized to be clinically important only in the urogenital tract. It can persist for years in infected patients by exploiting specialized cell-surface structures to invade cells.³⁶ Once inside a cell, it triggers secretion of mycoplasmal toxins and destructive metabolites such as hydrogen peroxide, evading the host immune system as it does so.³⁷

Testing guidelines for M genitalium

Current guidelines do not recommend routine screening for M genitalium, and no commercial test was available until recently.¹² Although evidence suggests that M genitalium is independently associated with preterm birth and miscarriages,³⁸ routine screening of pregnant women is not recommended.¹²

Testing for M genitalium should be considered in cases of persistent or recurrent nongonococcal urethritis in patients who test negative for gonorrhea and chlamydia or for whom treatment has failed.¹² Many isolates exhibit genotypic resistance to macrolide antibiotics, which are often the first-line therapy for nongonococcal urethritis.³⁹

Further study is needed to evaluate the potential impact of routine screening for M genitalium on the reproductive and sexual health of at-risk populations.

Diagnostic tests for M genitalium

Awareness of M genitalium as a cause of nongonococcal urethritis has been hampered by a dearth of diagnostic tests.⁴⁰ The organism’s fastidious requirements and extremely slow growth preclude culture as a practical method of diagnosis.⁴¹ Serologic assays are dogged by cross-reactivity and poor sensitivity.^42,43 Thus, molecular assays for detecting M genitalium and associated resistance markers are preferred for diagnosis.¹²

Several molecular tests are approved, available, and in use in Europe for diagnosing M genitalium infection,⁴⁰ and in January 2019 the FDA approved a molecular test that can detect M genitalium in urine specimens and vaginal, endocervical, urethral, and penile meatal swabs. Although vaginal swabs are preferred for this assay because they have higher sensitivity (92% for provider-collected and 99% for patient-collected swabs), urine specimens are acceptable, with a sensitivity of 78%.⁴⁴

At least 1 company is seeking FDA clearance for another molecular diagnostic assay for detecting M genitalium and markers of macrolide resistance in urine and genital swab specimens. Such assays may facilitate appropriate treatment.

Clinicians should stay abreast of diagnostic testing options, which are likely to become more readily available soon.

A high rate of macrolide resistance

Because M genitalium lacks a cell wall, antibiotics such as beta-lactams that target cell wall synthesis are ineffective.

Regimens for treating M genitalium are outlined in Table 2.¹² Azithromycin is more effective than doxycycline. However, as many as 50% of strains were macrolide-resistant in a cohort of US female patients.⁴⁵ Given the high incidence of treatment failure with azithromycin 1 g, it is thought that this regimen might select for resistance. For cases in which symptoms persist, a 1- to 2-week course of moxifloxacin is recommended.¹² However, this has not been validated by clinical trials, and failures of the 7-day regimen have been reported.⁴⁶

Partners of patients who test positive for M genitalium should also be tested and undergo clinically applicable screening for nongonococcal urethritis, cervicitis, and pelvic inflammatory disease.¹²

TRICHOMONIASIS

Trichomoniasis, caused by the parasite Trichomonas vaginalis, is the most prevalent nonviral STI in the United States. It disproportionately affects black women, in whom the prevalence is 13%, compared with 1% in non-Hispanic white women.⁴⁷ It is also present in 26% of women with symptoms who are seen in STI clinics and is highly prevalent in incarcerated populations. It is uncommon in men who have sex with men.⁴⁸

In men, trichomoniasis manifests as urethritis, epididymitis, or prostatitis. While most infected women have no symptoms, they may experience vaginitis with discharge that is diffuse, frothy, pruritic, malodorous, or yellow-green. Vaginal and cervical erythema (“strawberry cervix”) can also occur.

Screening for trichomoniasis

Current guidelines of the US Centers for Disease Control and Prevention (CDC) recommend testing for T vaginalis in women who have symptoms and routinely screening in women who are HIV-positive, regardless of symptoms. There is no evidence to support routine screening of pregnant women without symptoms, and pregnant women who do have symptoms should be evaluated according to the same guidelines as for nonpregnant women.¹² Testing can be considered in patients who have no symptoms but who engage in high-risk behaviors and in areas of high prevalence.

A lack of studies using sensitive methods for T vaginalis detection has hampered a true estimation of disease burden and at-risk populations. Screening recommendations may evolve in upcoming clinical guidelines as the field advances.

As infection can recur, women should be retested 3 months after initial diagnosis.¹²

NAAT is the preferred test for trichomoniasis

Commercially available diagnostic tests for trichomoniasis include culture, antigen testing, and NAAT.⁴⁹ While many clinicians do their own wet-mount microscopy for a rapid result, this method has low sensitivity.⁵⁰ Similarly, antigen testing and culture perform poorly compared with NAATs, which are the gold standard for detection.^51,52 A major advantage of NAATs for T vaginalis detection is that they combine high sensitivity and fast results, facilitating diagnosis and appropriate treatment of patients and their partners.

In spite of these benefits, adoption of molecular diagnostic testing for T vaginalis has lagged behind that for chlamydia and gonorrhea.⁵³ FDA-cleared NAATs are available for testing vaginal, cervical, or urine specimens from women, but until recently, there were no approved assays for testing in men. The Cepheid Xpert TV assay, which is valid for male urine specimens to diagnose other sexually transmitted diseases, has demonstrated excellent diagnostic sensitivity for T vaginalis in men and women.⁵⁴ Interestingly, a large proportion of male patients in this study had no symptoms, suggesting that screening of men in high-risk groups may be warranted.

7-day metronidazole treatment beats single-dose treatment

The first-line treatment for trichomoniasis has been a single dose of metronidazole 2 g by mouth, but in a recent randomized controlled trial,⁵⁵ a course of 500 mg by mouth twice a day for 7 days was 45% more effective at 4 weeks than a single dose, and it should now be the preferred regimen.

In clinical trials,⁵⁶ a single dose of tinidazole 2 g orally was equivalent or superior to metronidazole 2 g and had fewer gastrointestinal side effects, but it is more expensive.

Sexually transmitted infections (STIs) such as gonorrhea, chlamydia, and syphilis are still increasing in incidence and probably will continue to do so in the near future. Moreover, drug-resistant strains of Neisseria gonorrhoeae are emerging, as are less-known organisms such as Mycoplasma genitalium.

Now the good news: new tests for STIs are available or are coming! Based on nucleic acid amplification, these tests can be performed at the point of care, so that patients can leave the clinic with an accurate diagnosis and proper treatment for themselves and their sexual partners. Also, the tests can be run on samples collected by the patients themselves, either swabs or urine collections, eliminating the need for invasive sampling and making doctor-shy patients more likely to come in to be treated.¹ We hope that by using these sensitive and accurate tests we can begin to bend the upward curve of STIs and be better antimicrobial stewards.²

This article reviews current issues surrounding STI control, and provides detailed guidance on recognizing, testing for, and treating gonorrhea, chlamydia, trichomoniasis, and M genitalium infection.

STI RATES ARE HIGH AND RISING

STIs are among the most common acute infectious diseases worldwide, with an estimated 1 million new curable cases every day.³ Further, STIs have major impacts on sexual, reproductive, and psychological health.

In the United States, rates of reportable STIs (chlamydia, gonorrhea, and syphilis) are rising.⁴ In addition, more-sensitive tests for trichomoniasis, which is not a reportable infection in any state, have revealed it to be more prevalent than previously thought.⁵

BARRIERS AND CHALLENGES TO DIAGNOSIS

The medical system does not fully meet the needs of some populations, including young people and men who have sex with men, regarding their sexual and reproductive health. 

Ongoing barriers among young people include reluctance to use available health services, limited access to STI testing, worries about confidentiality, and the shame and stigma associated with STIs.⁶

Men who have sex with men have a higher incidence of STIs than other groups. Since STIs are associated with a higher risk of human immunodeficiency virus (HIV) infection, it is important to detect, diagnose, and manage STIs in this group—and in all high-risk groups. Rectal STIs are an independent risk factor for incident HIV infection.⁷ In addition, many men who have sex with men face challenges navigating the emotional, physical, and cognitive aspects of adolescence, a voyage further complicated by mental health issues, unprotected sexual encounters, and substance abuse in many, especially among minority youth.⁸ These same factors also impair their ability to access resources for preventing and treating HIV and other STIs.

STI diagnosis is often missed

Most people who have STIs feel no symptoms, which increases the importance of risk-based screening to detect these infections.^9,10 In many other cases, STIs manifest with nonspecific genitourinary symptoms that are mistaken for urinary tract infection. Tomas et al¹¹ found that of 264 women who presented to an emergency department with genitourinary symptoms or were being treated for urinary tract infection, 175 were given a diagnosis of a urinary tract infection. Of these, 100 (57%) were treated without performing a urine culture; 60 (23%) of the 264 women had 1 or more positive STI tests, 22 (37%) of whom did not receive treatment for an STI.

Poor follow-up of patients and partners

Patients with STIs need to be retested 3 months after treatment to make sure the treatment was effective. Another reason for follow-up is that these patients are at higher risk of another infection within a year.¹²

Although treating patients’ partners has been shown to reduce reinfection rates, fewer than one-third of STIs (including HIV infections) were recognized through partner notification between 2010 and 2012 in a Dutch study, in men who have sex with men and in women.¹³ Challenges included partners who could not be identified among men who have sex with men, failure of heterosexual men to notify their partners, and lower rates of partner notification for HIV.  

In the United States, “expedited partner therapy” allows healthcare providers to provide a prescription or medications to partners of patients diagnosed with chlamydia or gonorrhea without examining the partner.¹⁴ While this approach is legal in most states, implementation can be challenging.¹⁵

STI EVALUATION

History and physical examination

A complete sexual history helps in estimating the patient’s risk of an STI and applying appropriate risk-based screening. Factors such as sexual practices, use of barrier protection, and history of STIs should be discussed.

Physical examination is also important. Although some patients may experience discomfort during a genital or pelvic examination, omitting this step may lead to missed diagnoses in women with STIs.¹⁶

Laboratory testing

Laboratory testing for STIs helps ensure accurate diagnosis and treatment. Empiric treatment without testing could give a patient a false sense of health by missing an infection that is not currently causing symptoms but that could later worsen or have lasting complications. Failure to test patients also misses the opportunity for partner notification, linkage to services, and follow-up testing.

Many of the most common STIs, including gonorrhea, chlamydia, and trichomoniasis, can be detected using vaginal, cervical, or urethral swabs or first-catch urine (from the initial urine stream). In studies that compared various sampling methods,¹⁷ self-collected urine samples for gonorrhea in men were nearly as good as clinician-collected swabs of the urethra. In women, self-collected vaginal swabs for gonorrhea and chlamydia were nearly as good as clinician-collected vaginal swabs. While urine specimens are acceptable for chlamydia testing in women, their sensitivity may be slightly lower than with vaginal and endocervical swab specimens.^18,19

A major advantage of urine specimens for STI testing is that collection is noninvasive and is therefore more likely to be acceptable to patients. Urine testing can also be conducted in a variety of nonclinical settings such as health fairs, pharmacy-based screening programs, and express STI testing sites, thus increasing availability.

To prevent further transmission and morbidity and to aid in public health efforts, it is critical to recognize the cause of infectious cervicitis and urethritis and to screen for STIs according to guidelines.¹² Table 1 summarizes current screening and laboratory testing recommendations.

Gonorrhea and chlamydia are the 2 most frequently reported STIs in the United States, with more than 550,000 cases of gonorrhea and 1.7 million cases of chlamydia reported in 2017.⁴

Both infections present similarly: cervicitis or urethritis characterized by discharge (mucopurulent discharge with gonorrhea) and dysuria. Untreated, they can lead to pelvic inflammatory disease, inflammation, and infertility.

Extragenital infections can be asymptomatic or cause exudative pharyngitis or proctitis. Most people in whom chlamydia is detected from pharyngeal specimens are asymptomatic. When pharyngeal symptoms exist secondary to gonorrheal infection, they typically include sore throat and pharyngeal exudates. However, Komaroff et al,²⁰ in a study of 192 men and women who presented with sore throat, found that only 2 (1%) tested positive for N gonorrhoeae.

Screening for gonorrhea and chlamydia

Best practices include screening for gonorrhea and chlamydia as follows^21–23:

• Every year in sexually active women through age 25 (including during pregnancy) and in older women who have risk factors for infection¹²
• At least every year in men who have sex with men, at all sites of sexual contact (urethra, pharynx, rectum), along with testing for HIV and syphilis
• Every 3 to 6 months in men who have sex with men who have multiple or anonymous partners, who are sexually active and use illicit drugs, or who have partners who use illicit drugs
• Possibly every year in young men who live in high-prevalence areas or who are seen in certain clinical settings, such as STI and adolescent clinics.

Specimens. A vaginal swab is preferred for screening in women. Several studies have shown that self-collected swabs have clinical sensitivity and specificity comparable to that of provider-collected samples.^17,24 First-catch urine or endocervical swabs have similar performance characteristics and are also acceptable. In men, urethral swabs or first-catch urine samples are appropriate for screening for urogenital infections.

Testing methods. Testing for both pathogens should be done simultaneously with a nucleic acid amplification test (NAAT). Commercially available NAATs are more sensitive than culture and antigen testing for detecting gonorrhea and chlamydia.^25–27

Most assays are approved by the US Food and Drug Administration (FDA) for testing vaginal, urethral, cervical, and urine specimens. Until recently, no commercial assay was cleared for testing extragenital sites, but recommendations for screening extragenital sites prompted many clinical laboratories to validate throat and rectal swabs for use with NAATs, which are more sensitive than culture at these sites.^25,28 The recent FDA approval of extragenital specimen types for 2 commercially available assays may increase the availability of testing for these sites.

Data on the utility of NAATs for detecting chlamydia and gonorrhea in children are limited, and many clinical laboratories have not validated molecular methods for testing in children. Current guidelines specific to this population should be followed regarding test methods and preferred specimen types.^12,29,30

Although gonococcal infection is usually diagnosed with culture-independent molecular methods, antimicrobial resistance is emerging. Thus, failure of the combination of ceftriaxone and azithromycin should prompt culture-based follow-up testing to determine antimicrobial susceptibility.

Strategies for treatment and control

Historically, people treated for gonorrhea have been treated for chlamydia at the same time, as these diseases tend to go together. This can be with a single intramuscular dose of ceftriaxone for the gonorrhea plus a single oral dose of azithromycin for the chlamydia.¹² For patients who have only gonorrhea, this double regimen may help prevent the development of resistant gonorrhea strains.

All the patient’s sexual partners in the previous 60 days should be tested and treated, and expedited partner therapy should be offered if possible. Patients should be advised to have no sexual contact until they complete the treatment, or 7 days after single-dose treatment. Testing should be repeated 3 months after treatment.

A member of the Mycoplasmataceae family, M genitalium was originally identified as a pathogen in the early 1980s but has only recently emerged as an important cause of STI. Studies indicate that it is responsible for 10% to 20% of cases of nongonococcal urethritis and 10% to 30% of cases of cervicitis.^31–33 Additionally, 2% to 22% of cases of pelvic inflammatory disease have evidence of M genitalium.^34,35

However, data on M genitalium prevalence are suspect because the organism is hard to identify—lacking a cell wall, it is undetectable by Gram stain.³⁶ Although it has been isolated in respiratory and synovial fluids, it has so far been recognized to be clinically important only in the urogenital tract. It can persist for years in infected patients by exploiting specialized cell-surface structures to invade cells.³⁶ Once inside a cell, it triggers secretion of mycoplasmal toxins and destructive metabolites such as hydrogen peroxide, evading the host immune system as it does so.³⁷

Testing guidelines for M genitalium

Current guidelines do not recommend routine screening for M genitalium, and no commercial test was available until recently.¹² Although evidence suggests that M genitalium is independently associated with preterm birth and miscarriages,³⁸ routine screening of pregnant women is not recommended.¹²

Testing for M genitalium should be considered in cases of persistent or recurrent nongonococcal urethritis in patients who test negative for gonorrhea and chlamydia or for whom treatment has failed.¹² Many isolates exhibit genotypic resistance to macrolide antibiotics, which are often the first-line therapy for nongonococcal urethritis.³⁹

Further study is needed to evaluate the potential impact of routine screening for M genitalium on the reproductive and sexual health of at-risk populations.

Diagnostic tests for M genitalium

Awareness of M genitalium as a cause of nongonococcal urethritis has been hampered by a dearth of diagnostic tests.⁴⁰ The organism’s fastidious requirements and extremely slow growth preclude culture as a practical method of diagnosis.⁴¹ Serologic assays are dogged by cross-reactivity and poor sensitivity.^42,43 Thus, molecular assays for detecting M genitalium and associated resistance markers are preferred for diagnosis.¹²

Several molecular tests are approved, available, and in use in Europe for diagnosing M genitalium infection,⁴⁰ and in January 2019 the FDA approved a molecular test that can detect M genitalium in urine specimens and vaginal, endocervical, urethral, and penile meatal swabs. Although vaginal swabs are preferred for this assay because they have higher sensitivity (92% for provider-collected and 99% for patient-collected swabs), urine specimens are acceptable, with a sensitivity of 78%.⁴⁴

At least 1 company is seeking FDA clearance for another molecular diagnostic assay for detecting M genitalium and markers of macrolide resistance in urine and genital swab specimens. Such assays may facilitate appropriate treatment.

Clinicians should stay abreast of diagnostic testing options, which are likely to become more readily available soon.

A high rate of macrolide resistance

Because M genitalium lacks a cell wall, antibiotics such as beta-lactams that target cell wall synthesis are ineffective.

Regimens for treating M genitalium are outlined in Table 2.¹² Azithromycin is more effective than doxycycline. However, as many as 50% of strains were macrolide-resistant in a cohort of US female patients.⁴⁵ Given the high incidence of treatment failure with azithromycin 1 g, it is thought that this regimen might select for resistance. For cases in which symptoms persist, a 1- to 2-week course of moxifloxacin is recommended.¹² However, this has not been validated by clinical trials, and failures of the 7-day regimen have been reported.⁴⁶

Partners of patients who test positive for M genitalium should also be tested and undergo clinically applicable screening for nongonococcal urethritis, cervicitis, and pelvic inflammatory disease.¹²

TRICHOMONIASIS

Trichomoniasis, caused by the parasite Trichomonas vaginalis, is the most prevalent nonviral STI in the United States. It disproportionately affects black women, in whom the prevalence is 13%, compared with 1% in non-Hispanic white women.⁴⁷ It is also present in 26% of women with symptoms who are seen in STI clinics and is highly prevalent in incarcerated populations. It is uncommon in men who have sex with men.⁴⁸

In men, trichomoniasis manifests as urethritis, epididymitis, or prostatitis. While most infected women have no symptoms, they may experience vaginitis with discharge that is diffuse, frothy, pruritic, malodorous, or yellow-green. Vaginal and cervical erythema (“strawberry cervix”) can also occur.

Screening for trichomoniasis

Current guidelines of the US Centers for Disease Control and Prevention (CDC) recommend testing for T vaginalis in women who have symptoms and routinely screening in women who are HIV-positive, regardless of symptoms. There is no evidence to support routine screening of pregnant women without symptoms, and pregnant women who do have symptoms should be evaluated according to the same guidelines as for nonpregnant women.¹² Testing can be considered in patients who have no symptoms but who engage in high-risk behaviors and in areas of high prevalence.

A lack of studies using sensitive methods for T vaginalis detection has hampered a true estimation of disease burden and at-risk populations. Screening recommendations may evolve in upcoming clinical guidelines as the field advances.

As infection can recur, women should be retested 3 months after initial diagnosis.¹²

NAAT is the preferred test for trichomoniasis

Commercially available diagnostic tests for trichomoniasis include culture, antigen testing, and NAAT.⁴⁹ While many clinicians do their own wet-mount microscopy for a rapid result, this method has low sensitivity.⁵⁰ Similarly, antigen testing and culture perform poorly compared with NAATs, which are the gold standard for detection.^51,52 A major advantage of NAATs for T vaginalis detection is that they combine high sensitivity and fast results, facilitating diagnosis and appropriate treatment of patients and their partners.

In spite of these benefits, adoption of molecular diagnostic testing for T vaginalis has lagged behind that for chlamydia and gonorrhea.⁵³ FDA-cleared NAATs are available for testing vaginal, cervical, or urine specimens from women, but until recently, there were no approved assays for testing in men. The Cepheid Xpert TV assay, which is valid for male urine specimens to diagnose other sexually transmitted diseases, has demonstrated excellent diagnostic sensitivity for T vaginalis in men and women.⁵⁴ Interestingly, a large proportion of male patients in this study had no symptoms, suggesting that screening of men in high-risk groups may be warranted.

7-day metronidazole treatment beats single-dose treatment

The first-line treatment for trichomoniasis has been a single dose of metronidazole 2 g by mouth, but in a recent randomized controlled trial,⁵⁵ a course of 500 mg by mouth twice a day for 7 days was 45% more effective at 4 weeks than a single dose, and it should now be the preferred regimen.

In clinical trials,⁵⁶ a single dose of tinidazole 2 g orally was equivalent or superior to metronidazole 2 g and had fewer gastrointestinal side effects, but it is more expensive.

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

All cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus.¹ But most cases of Barrett esophagus go undiagnosed. And Barrett esophagus is seen in 5% to 15% of patients with gastroesophageal reflux disease.² These facts clearly emphasize the need for screening. Here, we review the rationale and recommendations for screening and surveillance, as well as the range of treatment options.

SCOPE OF THE PROBLEM

The American Cancer Society estimated there were 17,290 new cases of esophageal cancer and 15,850 deaths from it in the United States in 2018.³ Of the 2 main histologic types of esophageal cancer, adenocarcinoma and squamous cell cancer, adenocarcinoma is more common in the United States.

The precursor lesion is Barrett esophagus, defined as an extension of salmon-colored mucosa at least 1 cm into the tubular esophagus proximal to the gastroesophageal junction, with biopsy confirmation of intestinal metaplasia.⁴

The natural course of progression to dysplasia and cancer in Barrett esophagus is unknown but is thought to be stepwise, from no dysplasia to low-grade dysplasia to high-grade dysplasia and cancer, and the cancer risk depends on the degree of dysplasia: the annual risk is 0.33% if there is no dysplasia, 0.54% with low-grade dysplasia, and 7% with high-grade dysplasia.⁴

Although all cases of esophageal adenocarcinoma are thought to arise from Barrett esophagus,¹ more than 90% of patients with newly diagnosed esophageal adenocarcinoma do not have a prior diagnosis of Barrett esophagus.⁵ Therefore, there is a substantial unmet need to expand screening for Barrett esophagus in people at risk.

GASTROESOPHAGEAL REFLUX DISEASE IS A RISK FACTOR FOR CANCER

The rationale behind screening is that detecting Barrett esophagus early and intervening in a timely manner in patients at higher risk of developing adenocarcinoma will decrease mortality.

Chronic gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma (odds ratio [OR] 7.7, 95% confidence interval [CI] 5.3–11.4), and the risk increases when symptoms are long-standing (> 20 years) or severe (OR 43.5, 95% CI 18.3–103.5) or occur daily (OR 5.5, 95% CI 3.2–9.3).⁶

Reflux symptoms are scored as follows:

• Heartburn only, 1 point
• Regurgitation only, 1 point
• Heartburn with regurgitation, 1.5 points
• Nightly symptoms (2 points if yes, 0 if no)
• Symptoms once a week, 0 points; 2 to 6 times a week, 1 point; 7 to 15 times a week, 2 points; more than 15 times a week, 3 points.⁶

A score of 4.5 or higher indicates severe reflux disease. However, it is worth noting that the annual incidence of esophageal adenocarcinoma in patients with long-term gastroesophageal reflux disease is less than 0.001%.⁷

RISK FACTORS FOR BARRETT ESOPHAGUS

Risk factors for Barrett esophagus include:

Male sex. Barrett esophagus is more prevalent in men than in women, at a ratio of 2 to 1; but in individuals under age 50, the ratio is 4 to 1.⁸

Age 50 or older. Barrett esophagus is usually diagnosed in the sixth to seventh decade of life, and the prevalence increases from 2.1% in the third decade to 9.3% in the sixth decade.⁹

White race. It is more prevalent in whites than in blacks (5.0% vs 1.5%, P < .0001).¹⁰

Central obesity. Waist circumference is an independent risk factor: every 5-cm increase carries an OR of 1.14 (95% CI 1.03–1.27, P = .02).¹¹

Cigarette smoking increases the risk of Barrett esophagus (OR 1.42; 95% CI 1.15–1.76).¹²

A family history of Barrett esophagus or esophageal adenocarcinoma is a strong risk factor (OR 12, 95% CI 3.3–44.8). In 1 study, the risk in first- and second-degree relatives of patients with Barrett esophagus was 24%, compared with 5% in a control population (P < .005).¹³

SCREENING GUIDELINES AND DRAWBACKS

The standard screening method is esophagogastroduodenoscopy with sedation, with careful visual inspection and 4-quadrant biopsies every 2 cm using the Seattle protocol, ie, including biopsy of any mucosal irregularities in salmon-colored mucosa above the gastroesophageal junction (Figure 1).⁴

Endoscopic screening is cost-effective, costing $10,440 per quality-adjusted life-year saved, which is well below the accepted threshold of less than $100,000.¹⁴ However, it is still expensive, invasive, and not ideal for screening large populations.

Less-invasive methods under study

Less-invasive, less-expensive methods being tested for mass screening include:

Unsedated transnasal endoscopy. Done with only topical anesthesia, it has high diagnostic accuracy and is quicker and more cost-effective than standard esophagogastroduodenoscopy, with fewer adverse effects. However, the procedure has not yet gained widespread acceptance for regular use by gastroenterologists.¹⁵

A swallowable sponge. Another promising test is cell collection using the Cytosponge Cell Collection Device (Medtronic, Minneapolis, MN). An encapsulated compressed sponge with a string attached is swallowed; in the stomach, the capsule dissolves, and the sponge expands and is then withdrawn using the attached string. The obtained cytology sample from the lower esophagus is then tested for trefoil factor 3, a protein biomarker for Barrett esophagus.¹⁶

A retractable balloon. The EsoCheck Cell Collection Device is a retractable balloon attached to a string. When swallowed, it gathers distal esophageal cells for detecting methylated DNA markers for Barrett esophagus.¹⁷

Esophageal capsule endoscopy uses a camera to visualize the esophagus, but lacks the ability to obtain biopsy samples.

Other screening methods are being tested, although data are limited. Liquid biopsy uses a blood sample to detect microRNAs that are dysregulated in cancer. The “electronic nose” is a device that detects exhaled volatile organic compounds altered in Barrett esophagus. Another test involves taking an oral wash sample to study the oral microbiome for a pattern associated with adenocarcinoma.^18–21

Surveillance in Barrett esophagus aims to detect premalignant changes or early-stage adenocarcinoma to provide longer survival and lower cancer-related mortality. Recent evidence suggests that patients with esophageal adenocarcinoma that is diagnosed in a Barrett esophagus surveillance program have an earlier stage of disease and therefore a survival benefit.²²

Patient education is essential

Before enrolling a patient in a surveillance program, the clinician should explain the risks, benefits, and limitations, the importance of periodic endoscopy, and the possible eventual need for endoscopic therapy or surgery.

The endoscopic procedure

Surveillance involves examination by high-definition white-light endoscopy, with random 4-quadrant biopsies every 2 cm (or every 1 cm in patients with a history of dysplasia) and biopsy of any mucosal irregularity (nodule, ulcer, or other visible lesion). The degree of dysplasia determines the frequency of follow-up surveillance intervals and the need for endoscopic eradication therapy, as presented in professional society guidelines (Table 1).^4,23,24

Advanced methods for detecting dysplasia

Newer endoscopic surveillance techniques include dye-based chromoendoscopy, narrow-band imaging, confocal laser endomicroscopy, volumetric laser endomicroscopy, and wide-area transepithelial sampling with computer-assisted 3-dimensional analysis. All these techniques are used to increase the detection of dysplasia. Of these, dye-based chromoendoscopy, narrow-band imaging, and confocal laser endomicroscopy meet current criteria of the American Society for Gastrointestinal Endoscopy for preservation and incorporation of valuable endoscopic innovations.²³

MANAGEMENT OF NONDYSPLASTIC BARRETT ESOPHAGUS

A proton pump inhibitor (PPI) is recommended to control reflux symptoms in patients with nondysplastic Barrett esophagus. But it is important to counsel patients on additional ways to protect against esophageal adenocarcinoma, such as:

• Low to moderate alcohol consumption
• Regular physical activity
• Increased dietary intake of fruits, vegetables, folate, fiber, beta-carotene, and vitamin C
• Weight control
• Smoking cessation.²⁵

Surveillance endoscopy with 4-quadrant biopsies at 2-cm intervals is recommended every 3 to 5 years (Table 1).

DOES CHEMOPREVENTION HAVE A ROLE?

Chemoprevention is an exciting area of research in preventing progression to adenocarcinoma in patients with Barrett esophagus. Various drugs such as aspirin, other nonsteroidal anti-inflammatory drugs (NSAIDs), PPIs, metformin, and statins have been studied.

Aspirin

Aspirin has been shown to prevent development of Barrett esophagus in patients with reflux disease,²⁶ but more studies are needed to validate those findings.

PPIs

Gastroesophageal reflux disease is a primary risk factor for esophageal adenocarcinoma, and gastric acid suppression with PPIs reduces cancer risk. PPI therapy is associated with a 71% decrease in the risk of high-grade dysplasia and adenocarcinoma in patients with Barrett esophagus (OR 0.29, 95% CI 0.12–0.79).²⁷ Long-term therapy (> 2 to 3 years) has a higher protective effect (adjusted OR 0.45, 95% CI 0.19–1.06) than short-term therapy (< 2 to 3 years) (adjusted OR 1.09, 95% CI 0.47–2.56).²⁷

NSAIDs

NSAIDs, including aspirin, have been associated with decreased risk of colon, stomach, lung, breast, and esophageal cancer due to their potential to inhibit cyclooxygenase 2 (COX-2) enzymes.

A meta-analysis demonstrated that aspirin and NSAIDs led to a 32% reduction in the risk of adenocarcinoma (OR 0.68, 95% CI 0.56–0.83). The benefit was even greater if the drug was taken daily or more frequently (OR 0.56, 95% CI 0.43–0.73, P < .001) or was taken for 10 or more years (OR 0.63, 95% CI 0.45–0.90, P = .04).²⁸

PPI plus aspirin

The best evidence for the role of PPIs and aspirin in reducing the risk of dysplasia comes from the Aspirin and Esomeprazole Chemoprevention in Barrett’s Metaplasia Trial.²⁹ This randomized, controlled trial compared 4 regimens consisting of esomeprazole (a PPI) in either a high dose (40 mg twice daily) or a low dose (20 mg once daily) plus either aspirin (300 or 320 mg per day) or no aspirin in 2,557 patients with Barrett esophagus. The composite end point was the time to all-cause mortality, adenocarcinoma, or high-grade dysplasia.

At a median follow-up of 8.9 years, the combination of high-dose esomeprazole plus aspirin had the strongest effect compared with low-dose esomeprazole without aspirin (time ratio 1.59, 95% CI 1.14–2.23, P = .0068). The number needed to treat was 34 for esomeprazole and 43 for aspirin.²⁹

Based on these data, we can conclude that aspirin and PPIs can prevent dysplasia and all-cause mortality in Barrett esophagus.

Metformin: No evidence of benefit

Metformin was studied as a protective agent against obesity-associated cancers including esophageal adenocarcinoma, as it reduces insulin levels.

In a randomized controlled trial³⁰ in 74 patients with Barrett esophagus, metformin (starting at 500 mg daily, increasing to 2,000 mg/day by week 4) was compared with placebo. At 12 weeks, the percent change in esophageal levels of the biomarker pS6K1—an intracellular mediator of insulin and insulin-like growth factor activation in Barrett epithelium—did not differ significantly between the 2 groups (1.4% with metformin vs −14.7% with placebo; 1-sided P = .80). This suggested that metformin did not significantly alter proliferation or apoptosis in Barrett epithelium, despite reducing serum insulin levels and insulin resistance. Thus, metformin did not demonstrate a chemoprotective effect in preventing the progression of Barrett esophagus to adenocarcinoma.

Vitamin D affects genes regulating proliferation, apoptosis, and differentiation, and has therefore been studied as a potential antineoplastic agent. Its deficiency has also been associated with increased risk of esophageal adenocarcinoma. However, its efficacy in chemoprevention is unclear.³¹

One study found no association between serum 25-hydroxyvitamin D levels and prevalence of dysplasia in Barrett esophagus (P = .90). An increase in vitamin D levels had no effect on progression to dysplasia or cancer (for every 5-nmol/L increase from baseline, hazard ratio 0.98, P = .62).³²

In another study, supplementation with vitamin D₃ (cholecalciferol 50,000 IU weekly) plus a PPI for 12 weeks significantly improved the serum 25-hydroxyvitamin D levels without significant changes in gene expression from Barrett epithelium.³³ These findings were confirmed in a meta-analysis that showed no consistent association between vitamin D exposure and risk of esophageal neoplasm.³⁴

Thus, there is currently no evidence to support vitamin D for chemoprevention in Barrett esophagus or esophageal adenocarcinoma.

Statins

In addition to lowering cholesterol, statins have antiproliferative, pro-apoptotic, anti-angiogenic, and immunomodulatory effects that prevent cancer, leading to a 41% reduction in the risk of adenocarcinoma in patients with Barrett esophagus in one study (adjusted OR 0.59, 95% CI 0.45–0.78); the number needed to treat with statins to prevent 1 case of adenocarcinoma was 389.³⁵

A meta-analysis also showed that statin use was associated with a lower risk of progression of Barrett esophagus (OR 0.48, 95% CI 0.31–0.73).³⁶

In general, statins appear promising for chemoprevention, but more study is needed.

When is chemoprevention appropriate?

Chemoprevention is not recommended for all patients with Barrett esophagus, given that the condition affects 1% to 2% of the US adult population, and very few patients have progression to esophageal adenocarcinoma. Rather, chemoprevention may be considered in patients with Barrett esophagus and multiple risk factors for adenocarcinoma.

INDEFINITE DYSPLASIA

In Barrett esophagus with indefinite dysplasia, either the epithelial abnormalities are insufficient for a diagnosis of dysplasia, or the nature of the epithelial abnormalities is uncertain due to inflammation or technical difficulties with specimen processing. The risk of high-grade dysplasia or cancer within 1 year of the diagnosis of indefinite dysplasia varies between 1.9% and 15%.³⁷ The recommendation for management is to optimize acid-suppressive therapy for 3 to 6 months and then to repeat esophagogastroduodenoscopy. If indefinite dysplasia is noted again, repeat endoscopy in 12 months is recommended.²

ENDOSCOPIC ERADICATION: AN OVERVIEW

Because dysplasia in Barrett esophagus carries a high risk of progression to cancer, the standard of care is endoscopic mucosal resection of visible lesions, followed by ablation of the flat mucosa, with the aim of achieving complete eradication of intestinal metaplasia.^4,38 The initial endoscopic treatment is followed by outpatient sessions every 8 to 10 weeks until the dysplasia is eradicated. A key part of treatment during this time is maximal acid suppression with a PPI twice daily and a histamine-2 blocker at night. In rare cases, fundoplication is required to control reflux refractory to medical therapy.

After eradication is confirmed, continued surveillance is necessary, as recurrences have been reported at a rate of 4.8% per year for intestinal metaplasia, and 2% per year for dysplasia.³⁹

Current endoscopic resection techniques

Endoscopic resection techniques include mucosal resection, submucosal dissection, radio­frequency ablation, cryotherapy, argon plasma coagulation, and photodynamic therapy (Figure 2).

In mucosal resection, the lesion is either suctioned into a band ligator, after which a band is placed around the lesion, or suctioned into a cap fitted at the end of the endoscope, after which the lesion is removed using a snare.

In submucosal dissection, a liquid is injected into the submucosa to lift the lesion, making it easier to remove. The procedure is technically complex and requires additional training.

In radiofrequency ablation, a special catheter is passed through the endoscope to ablate the affected epithelium by thermal injury. Argon plasma coagulation works in a similar way, but uses ionized argon gas to induce thermal coagulation of metaplastic epithelium.

Cryotherapy produces cellular injury by rapid freezing and thawing of tissue using a cryogen such as liquid nitrogen or nitrous oxide.

In photodynamic therapy, a photosensitizer (porfimer sodium) is administered and taken up preferentially by metaplastic epithelium; it is then activated by transmission of red light using the endoscope, leading to destruction of the metaplastic epithelium.

Of the different techniques, radiofrequency ablation has the most evidence for efficacy and hence is the most commonly used.

All of these procedures are generally well tolerated and have favorable side-effect profiles. After radiofrequency ablation with or without mucosal resection, esophageal strictures are noted in 5.6% of patients, and bleeding and perforation occur rarely (1% and 0.6% of patients, respectively).⁴⁰ Submucosal dissection is associated with a higher rate of complications such as stricture formation (11% of patients) and bleeding or perforation (1.5% of patients).⁴¹

Most patients with low-grade dysplasia (73%) are down-staged to nondysplastic Barrett esophagus or to indefinite for dysplasia after review by expert pathologists.⁴² Patients with confirmed and persistent low-grade dysplasia are at higher risk of progression.⁴³

Once low-grade dysplasia is confirmed by a second gastrointestinal pathologist, the patient should undergo endoscopic ablation. A landmark study by Shaheen et al⁴⁴ demonstrated the benefit of radiofrequency ablation in achieving complete eradication of dysplasia (90.5% vs 22.7% for a sham procedure) and complete eradication of intestinal metaplasia (77.4% vs 2.3% for a sham procedure). In another trial of 136 patients with low-grade dysplasia followed for 3 years, Phoa et al⁴⁵ demonstrated that radiofrequency ablation reduced the rate of progression to high-grade dysplasia by 25% and to adenocarcinoma by 7.4% compared with endoscopic surveillance.

Patients with confirmed low-grade dysplasia who do not undergo eradication therapy should have surveillance endoscopy every 6 to 12 months (Table 1).

HIGH-GRADE DYSPLASIA: RECOMMENDED MANAGEMENT

As with low-grade dysplasia, the diagnosis of high-grade dysplasia needs to be confirmed by a second pathologist with gastrointestinal expertise. In the past, the treatment was esophagectomy, but due to lower morbidity and equivalent efficacy of radiofrequency ablation,⁴⁶ the current treatment of choice is endoscopic mucosal resection of raised lesions, followed by radiofrequency ablation of the entire affected segment.

In the study by Shaheen et al,⁴⁴ 42 patients with high-grade dysplasia were randomized to radiofrequency ablation and 21 to a sham procedure, and 81% of ablation patients achieved complete eradication of dysplasia vs 19% with the sham procedure. Eradication of intestinal metaplasia was achieved in 77% of ablation patients vs 2% of patients with the sham therapy. Results of 3-year follow-up from the same cohort showed complete eradication of dysplasia in 98% and of intestinal metaplasia in 91%.⁴⁷

Endoscopic eradication therapy is recommended for all patients with Barrett esophagus and high-grade dysplasia without a life-limiting comorbidity. Alternatively, surveillance every 3 months is an option if the patient does not wish to undergo eradication therapy. Radiofrequency ablation is more cost-effective than esophagectomy or endoscopic surveillance followed by treatment once patients develop adenocarcinoma.^48,49

EARLY ESOPHAGEAL ADENOCARCINOMA: RECOMMENDED MANAGEMENT

Adenocarcinoma limited to the mucosa and without evidence of nodal involvement can be resected endoscopically. In patients with localized cancer, mucosal resection is done not only for therapeutic purposes but also for staging. Ideal management is multidisciplinary, including a gastroenterologist, thoracic surgeon, oncologist, pathologist, and radiation oncologist.

If lesions have features suggesting submucosal invasion or are greater than 1.5 cm in size, or if it is difficult to separate (ie, lift) the mucosa from the submucosal layer with injection of saline, then submucosal dissection is recommended.⁵⁰ Because of the risk of metachronous lesions, ablation of the remaining Barrett esophagus mucosa is recommended after resection of cancer.

Endoscopic eradication is highly effective and durable for the treatment of intramucosal esophageal adenocarcinoma. In a study of 1,000 patients, 963 patients (96.3%) had achieved a complete response; 12 patients (3.7%) underwent surgery after eradication failed during a follow-up of almost 5 years.⁵¹ Metachronous lesions or recurrence of cancer developed during the follow-up period in 140 patients (14.5%) but were successfully treated endoscopically in 115, resulting in a long-term complete remission rate of 93.8%.

POSTABLATION MANAGEMENT

Because of the risk of recurrence of dysplasia after ablation, long-term PPI therapy and surveillance are recommended.

Surveillance endoscopy involves 4-quadrant biopsies taken every 1 cm from the entire length of segment where Barrett esophagus had been seen before ablation.

The timing of surveillance intervals depends on the preablation grade of dysplasia. For low-grade dysplasia, the recommendation is every 6 months for the first year after ablation and, if there is no recurrence of dysplasia, annually after that.² After treatment of high-grade dysplasia or intramucosal adenocarcinoma, the recommendation is every 3 months for the first year, every 6 months in the second year, and then annually.²

Migraine is the second leading cause of years of life lived with a disability globally.¹ It affects people of all ages, but particularly during the years associated with the highest productivity in terms of work and family life.

Migraine is a genetic neurologic disease that can be influenced or triggered by environmental factors. However, triggers do not cause migraine. For example, stress does not cause migraine, but it can exacerbate it.

Primary care physicians can help patients reduce the likelihood of a migraine attack, the severity of symptoms, or both by offering lifestyle counseling centered around the mnemonic SEEDS: sleep, exercise, eat, diary, and stress. In this article, each factor is discussed individually for its current support in the literature along with best-practice recommendations.

S IS FOR SLEEP

Before optimizing sleep hygiene, screen for sleep apnea, especially in those who have chronic daily headache upon awakening. An excellent tool is the STOP-Bang screening questionnaire⁵ (www.stopbang.ca/osa/screening.php). Patients respond “yes” or “no” to the following questions:

• Snoring: Do you snore loudly (louder than talking or loud enough to be heard through closed doors)?
• Tired: Do you often feel tired, fatigued, or sleepy during the daytime?
• Observed: Has anyone observed you stop breathing during your sleep?
• Pressure: Do you have or are you being treated for high blood pressure?
• Body mass index greater than 35 kg/m²?
• Age over 50?
• Neck circumference larger than 40 cm (females) or  42 cm (males)?
• Gender—male?

Each “yes” answer is scored as 1 point. A score less than 3 indicates low risk of obstructive sleep apnea; 3 to 4 indicates moderate risk; and 5 or more indicates high risk. Optimization of sleep apnea with continuous positive airway pressure therapy can improve sleep apnea headache.⁶ The improved sleep from reduced arousals may also mitigate migraine symptoms.

Behavioral modification for sleep hygiene can convert chronic migraine to episodic migraine.⁷ One such program is stimulus control therapy, which focuses on using cues to initiate sleep (Table 1). Patients are encouraged to keep the bedroom quiet, dark, and cool, and to go to sleep at the same time every night. Importantly, the bed should be associated only with sleep. If patients are unable to fall asleep within 20 to 30 minutes, they should leave the room so they do not associate the bed with frustration and anxiety. Use of phones, tablets, and television in the bedroom is discouraged as these devices may make it more difficult to fall asleep.⁸

The next option is sleep restriction, which is useful for comorbid insomnia. Patients keep a sleep diary to better understand their sleep-wake cycle. The goal is 90% sleep efficiency, meaning that 90% of the time in bed (TIB) is spent asleep. For example, if the patient is in bed 8 hours but asleep only 4 hours, sleep efficiency is 50%. The goal is to reduce TIB to match the time asleep and to agree on a prescribed daily wake-up time. When the patient is consistently sleeping 90% of the TIB, add 30-minute increments until he or she is appropriately sleeping 7 to 8 hours at night.⁹ Naps are not recommended.

Let patients know that their migraine may worsen until a new routine sleep pattern emerges. This method is not recommended for patients with untreated sleep apnea.

E IS FOR EXERCISE

Exercise is broadly recommended for a healthy lifestyle; some evidence suggests that it can also be useful in the management of migraine.¹⁰ Low levels of physical activity and a sedentary lifestyle are associated with migraine.¹¹ It is unclear if patients with migraine are less likely to exercise because they want to avoid triggering a migraine or if a sedentary lifestyle increases their risk.

Exercise has been studied for its prophylactic benefits in migraine, and one hypothesis relates to beta-endorphins. Levels of beta-endorphins are reduced in the cerebrospinal fluid of patients with migraine.¹² Exercise programs may increase levels while reducing headache frequency and duration.¹³ One study showed that pain thresholds do not change with exercise programs, suggesting that it is avoidance behavior that is positively altered rather than the underlying pain pathways.¹⁴

A systematic review and meta-analysis based on 5 randomized controlled trials and 1 nonrandomized controlled clinical trial showed that exercise reduced monthly migraine days by only 0.6 (± 0.3) days, but the data also suggested that as the exercise intensity increased, so did the positive effects.¹⁰

Some data suggest that exercise may also reduce migraine duration and severity as well as the need for abortive medication.¹⁰ Two studies in this systematic review^15,16 showed that exercise benefits were equivalent to those of migraine preventives such as amitriptyline and topiramate; the combination of amitriptyline and exercise was more beneficial than exercise alone. Multiple types of exercise were beneficial, including walking, jogging, cross-training, and cycling when done for least 6 weeks and for 30 to 50 minutes 3 to 5 times a week.

These findings are in line with the current recommendations for general health from the American College of Sports Medicine, ie, moderate to vigorous cardio­respiratory exercise for 30 to 60 minutes 3 to 5 times a week (or 150 minutes per week). The daily exercise can be continuous or done in intervals of less than 20 minutes. For those with a sedentary lifestyle, as is seen in a significant proportion of the migraine population, light to moderate exercise for less than 20 minutes is still beneficial.¹⁷

Based on this evidence, the best current recommendation for patients with migraine is to engage in graded moderate cardiorespiratory exercise, although any exercise is better than none. If a patient is sedentary or has poor exercise tolerance, or both, exercising once a week for shorter time periods may be a manageable place to start.

Some patients may identify exercise as a trigger or exacerbating factor in migraine. These patients may need appropriate prophylactic and abortive therapies before starting an exercise regimen.

THE SECOND E IS FOR EAT (FOOD AND DRINK)

Many patients believe that some foods trigger migraine attacks, but further study is needed. The most consistent food triggers appear to be red wine and caffeine (withdrawal).^18,19 Interestingly, patients with migraine report low levels of alcohol consumption,²⁰ but it is unclear if that is because alcohol has a protective effect or if patients avoid it.

Some patients may crave certain foods in the prodromal phase of an attack, eat the food, experience the attack, and falsely conclude that the food caused the attack.²¹ Premonitory symptoms include fatigue, cognitive changes, homeostatic changes, sensory hyperresponsiveness, and food cravings.²¹ It is difficult to distinguish between premonitory phase food cravings and true triggers because premonitory symptoms can precede headache by 48 to 72 hours, and the timing for a trigger to be considered causal is not known.²²

Chocolate is often thought to be a migraine trigger, but the evidence argues against this and even suggests that sweet cravings are a part of the premonitory phase.²³ Monosodium glutamate is often identified as a trigger as well, but the literature is inconsistent and does not support a causal relationship.²⁴ Identifying true food triggers in migraine is difficult, and patients with migraine may have poor quality diets, with some foods acting as true triggers for certain patients.²⁵ These possibilities have led to the development of many “migraine diets,” including elimination diets.

Elimination diets

Elimination diets involve avoiding specific food items over a period of time and then adding them back in one at a time to gauge whether they cause a reaction in the body. A number of these diets have been studied for their effects on headache and migraine:

Gluten-free diets restrict foods that contain wheat, rye, and barley. A systematic review of gluten-free diets in patients with celiac disease found that headache or migraine frequency decreased by 51.6% to 100% based on multiple cohort studies (N = 42,388).²⁶ There are no studies on the use of a gluten-free diet for migraine in patients without celiac disease.

Immunoglobulin G-elimination diets restrict foods that serve as antigens for IgG. However, data supporting these diets are inconsistent. Two small randomized controlled trials found that the diets improved migraine symptoms, but a larger study found no improvement in the number of migraine days at 12 weeks, although there was an initially significant effect at 4 weeks.^27–29

Antihistamine diets restrict foods that have high levels of histamines, including fermented dairy, vegetables, soy products,  wine, beer, alcohol, and those that cause histamine release regardless of IgE testing results. A prospective single-arm study of antihistamine diets in patients with chronic headache reported symptom improvement, which could be applied to certain comorbidities such as  mast cell activation syndrome.³⁰ Another prospective nonrandomized controlled study eliminated foods based on positive IgE skin-prick testing for allergy in patients with recurrent migraine and found that it reduced headache frequency.³¹

Tyramine-free diets are often recommended due to the presumption that tyramine-containing foods (eg, aged cheese, cured or smoked meats and fish, and beer) are triggers. However, multiple studies have reviewed this theory with inconsistent results,³² and the only study of a tyramine-free diet was negative.³³ In addition, commonly purported high-tyramine foods have lower tyramine levels than previously thought.³⁴

Low-fat diets in migraine are supported by 2 small randomized controlled trials and a prospective study showing a decrease in symptom severity; the results for frequency are inconsistent.^35–37

Low-glycemic index diets are supported in migraine by 1 randomized controlled trial that showed improvement in migraine frequency in a diet group and in a control group of patients who took a standard migraine-preventive medication to manage their symptoms.³⁸

Other migraine diets

Diets high in certain foods or ingredient ratios, as opposed to elimination diets, have also been studied in patients with migraine. One promising diet containing high levels of omega-3 fatty acids and low levels of omega-6 fatty acids was shown in a systematic review to reduce the duration of migraine but not the frequency or severity.³⁹ A more recent randomized controlled trial of this diet in chronic migraine also showed that it decreased migraine frequency.⁴⁰

The ketogenic diet (high fat, low carbohydrate) had promising results in a randomized controlled trial in overweight women with migraine and in a prospective study.^41,42 However, a prospective study of the Atkins diet in teenagers with chronic daily headaches showed no benefit.⁴³ The ketogenic diet is difficult to follow and may work in part due to weight loss alone, although ketogenesis itself may also play a role.^41,44

Sodium levels have been shown to be higher in the cerebrospinal fluid of patients with migraine than in controls, particularly during an attack.⁴⁵ For a prehypertensive population or an elderly population, a low-sodium diet may be beneficial based on 2 prospective trials.^46,47 However, a younger female population without hypertension and low-to-normal body mass index had a reduced probability of migraine while consuming a high-sodium diet.⁴⁸

Counseling about sodium intake should be tailored to specific patient populations. For example, a diet low in sodium may be appropriate for patients with vascular risk factors such as hypertension, whereas a high-sodium diet may be appropriate in patients with comorbidities like postural tachycardia syndrome or in those with a propensity for low blood pressure or low body mass index.

Encourage routine meals and hydration

The standard advice for patients with migraine is to consume regular meals. Headaches have been associated with fasting, and those with migraine are predisposed to attacks in the setting of fasting.^49,50 Migraine is more common when meals are skipped, particularly breakfast.⁵¹

It is unclear how fasting lowers the migraine threshold. Nutritional studies show that skipping meals, particularly breakfast, increases low-grade inflammation and impairs  glucose metabolism by affecting insulin and fat oxidation metabolism.⁵² However, hypoglycemia itself is not a consistent cause of headache or migraine attacks.⁵³ As described above, a randomized controlled trial of a low-glycemic index diet actually decreased migraine frequency and severity.³⁸ Skipping meals also reduces energy and is associated with reduced physical activity, perhaps leading to multiple compounding triggers that further lower the migraine threshold.^54,55

When counseling patients about the need to eat breakfast, consider what they normally consume (eg, is breakfast just a cup of coffee?). Replacing simple carbohydrates with protein, fats, and fiber may be beneficial for general health, but the effects on migraine are not known, nor is the optimal composition of breakfast foods.⁵⁵

The optimal timing of breakfast relative to awakening is also unclear, but in general, it should be eaten within 30 to 60 minutes of rising. Also consider patients’ work hours—delayed-phase or shift workers have altered sleep cycles.

Recommendations vary in regard to hydration. Headache is associated with fluid restriction and dehydration,^56,57 but only a few studies suggest that rehydration and increased hydration status can improve migraine.⁵⁸ In fact, a single post hoc analysis of a metoclopramide study showed that intravenous fluid alone for patients with migraine in the emergency room did not improve pain outcomes.⁵⁹

The amount of water patients should drink daily in the setting of migraine is also unknown, but a study showed benefit with 4 L, which equates to a daily intake of 16 eight-ounce glasses.⁶⁰ One review on general health that could be extrapolated given the low risk of the intervention indicated that 1.8 L daily (7 to 8 eight-ounce glasses) promoted a euhydration status in most people, although many factors contribute to hydration status.⁶¹

Caffeine intake is also a major consideration. Caffeine is a nonspecific adenosine receptor antagonist that modulates adenosine receptors like the pronociceptive 2A receptor, leading to changes integral to the neuropathophysiology of migraine.⁶² Caffeine has analgesic properties at doses greater than 65 to 200 mg and augments the effects of analgesics such as acetaminophen and aspirin. Chronic caffeine use can lead to withdrawal symptoms when intake is stopped abruptly; this is thought to be due to upregulation of adenosine receptors, but the effect varies based on genetic predisposition.¹⁹

The risk of chronic daily headache may relate to high use of caffeine preceding the onset of chronification, and caffeine abstinence may improve response to acute migraine treatment.^19,63 There is a dose-dependent risk of headache.^64,65 Current recommendations suggest limiting caffeine consumption to less than 200 mg per day or stopping caffeine consumption altogether based on the quantity required for caffeine-withdrawal headache.⁶⁶ Varying  the caffeine dose from day to day may also trigger headache due to the high sensitivity to caffeine withdrawal.

While many diets have shown potential benefit in patients with migraine, more studies are needed before any one “migraine diet” can be recommended. Caution should be taken, as there is risk of adverse effects from nutrient deficiencies or excess levels, especially if the patient is not under the care of a healthcare professional who is familiar with the diet.

Whether it is beneficial to avoid specific food triggers at this time is unclear and still controversial even within the migraine community because some of these foods may be misattributed as triggers instead of premonitory cravings driven by the hypothalamus. It is important to counsel patients with migraine to eat a healthy diet with consistent meals, to maintain adequate hydration, and to keep their caffeine intake low or at least consistent, although these teachings are predominantly based on limited studies with extrapolation from nutrition research.

D IS FOR DIARY

A headache diary is a recommended part of headache management and may enhance the accuracy of diagnosis and assist in treatment modifications. Paper and electronic diaries have been used. Electronic diaries may be more accurate for real-time use, but patients may be more likely to complete a paper one.⁶⁷ Patients prefer electronic diaries over long paper forms,⁶⁸ but a practical issue to consider is easy electronic access.

Patients can start keeping a headache diary before the initial consultation to assist with diagnosis, or early in their management. A first-appointment diary mailed with instructions is a feasible option.⁶⁹ These types of diaries ask detailed questions to help diagnose all major primary headache types including menstrual migraine and to identify concomitant medication-overuse headache. Physicians and patients generally report improved communication with use of a diary.⁷⁰

Some providers distinguish between a headache diary and a calendar. In standard practice, a headache diary is the general term referring to both, but the literature differentiates between the two. Both should at least include headache frequency, with possible inclusion of other factors such as headache duration, headache intensity, analgesic use, headache impact on function, and absenteeism. Potential triggers including menses can also be tracked. The calendar version can fit on a single page and can be used for simple tracking of headache frequency and analgesia use.

One of the simplest calendars to use is the “stoplight” calendar. Red days are when a patient is completely debilitated in bed. On a yellow day, function at work, school, or daily activities is significantly reduced by migraine, but the patient is not bedbound. A green day is when headache is present but function is not affected. No color is placed if the patient is 100% headache-free.

Acute treatment use can be written in or, to improve compliance, a checkmark can be placed on days of treatment. Patients who are tracking menses circle the days of menstruation. The calendar-diary should be brought to every appointment to track treatment response and medication use.

THE SECOND S IS FOR STRESS

Behavioral management such as cognitive behavioral therapy in migraine has been shown to decrease catastrophizing, migraine disability, and headache severity and frequency.⁷⁴ Both depression and anxiety can improve along with migraine.⁷⁵ Cognitive behavioral therapy can be provided in individualized sessions or group sessions, either in person or online.^74,76,77 The effects become more prominent about 5 weeks into treatment.⁷⁸

Biofeedback, which uses behavioral techniques paired with physiologic autonomic measures, has been extensively studied, and shows benefit in migraine, including in meta-analysis.⁷⁹ The types of biofeedback measurements used include electromyography, electroencephalography, temperature, sweat sensors, heart rate, blood volume pulse feedback, and respiration bands. While biofeedback is generally done under the guidance of a therapist, it can still be useful with minimal therapist contact and supplemental audio.⁸⁰

Mindfulness, or the awareness of thoughts, feelings, and sensations in the present moment without judgment, is a behavioral technique that can be done alone or paired with another technique. It is often taught through a mindfulness-based stress-reduction  program, which relies on a standardized approach. A meta-analysis showed that mindfulness improves pain intensity, headache frequency, disability, self-efficacy, and quality of life.⁸¹ It may work by encouraging pain acceptance.⁸²

Relaxation techniques are also employed in migraine management, either alone or in conjunction with techniques mentioned  above, such as mindfulness. They include progressive muscle relaxation and deep breathing. Relaxation has been shown to be effective when done by professional trainers as well as lay trainers in both individual and group settings.^83,84

In patients with intractable headache, more-intensive inpatient and outpatient programs have been tried. Inpatient admissions with multidisciplinary programs that include a focus on behavioral techniques often paired with lifestyle education and sometimes pharmacologic management can be beneficial.^85,86 These programs have also been successfully conducted as multiple outpatient sessions.^86–88

Stress management is an important aspect of migraine management. These treatments often involve homework and require active participation.

LIFESTYLE FOR ALL

All patients with migraine should initiate lifestyle modifications (see Advice to patients with migraine: SEEDS for success). Modifications with the highest level of evidence, specifically behavioral techniques, have had the most reproducible results. A headache diary is an essential tool to identify patterns and needs for optimization of acute or preventive treatment regimens. The strongest evidence is for the behavioral management techniques for stress reduction.

Migraine is the second leading cause of years of life lived with a disability globally.¹ It affects people of all ages, but particularly during the years associated with the highest productivity in terms of work and family life.

Migraine is a genetic neurologic disease that can be influenced or triggered by environmental factors. However, triggers do not cause migraine. For example, stress does not cause migraine, but it can exacerbate it.

Primary care physicians can help patients reduce the likelihood of a migraine attack, the severity of symptoms, or both by offering lifestyle counseling centered around the mnemonic SEEDS: sleep, exercise, eat, diary, and stress. In this article, each factor is discussed individually for its current support in the literature along with best-practice recommendations.

S IS FOR SLEEP

Before optimizing sleep hygiene, screen for sleep apnea, especially in those who have chronic daily headache upon awakening. An excellent tool is the STOP-Bang screening questionnaire⁵ (www.stopbang.ca/osa/screening.php). Patients respond “yes” or “no” to the following questions:

• Snoring: Do you snore loudly (louder than talking or loud enough to be heard through closed doors)?
• Tired: Do you often feel tired, fatigued, or sleepy during the daytime?
• Observed: Has anyone observed you stop breathing during your sleep?
• Pressure: Do you have or are you being treated for high blood pressure?
• Body mass index greater than 35 kg/m²?
• Age over 50?
• Neck circumference larger than 40 cm (females) or  42 cm (males)?
• Gender—male?

Each “yes” answer is scored as 1 point. A score less than 3 indicates low risk of obstructive sleep apnea; 3 to 4 indicates moderate risk; and 5 or more indicates high risk. Optimization of sleep apnea with continuous positive airway pressure therapy can improve sleep apnea headache.⁶ The improved sleep from reduced arousals may also mitigate migraine symptoms.

Behavioral modification for sleep hygiene can convert chronic migraine to episodic migraine.⁷ One such program is stimulus control therapy, which focuses on using cues to initiate sleep (Table 1). Patients are encouraged to keep the bedroom quiet, dark, and cool, and to go to sleep at the same time every night. Importantly, the bed should be associated only with sleep. If patients are unable to fall asleep within 20 to 30 minutes, they should leave the room so they do not associate the bed with frustration and anxiety. Use of phones, tablets, and television in the bedroom is discouraged as these devices may make it more difficult to fall asleep.⁸

The next option is sleep restriction, which is useful for comorbid insomnia. Patients keep a sleep diary to better understand their sleep-wake cycle. The goal is 90% sleep efficiency, meaning that 90% of the time in bed (TIB) is spent asleep. For example, if the patient is in bed 8 hours but asleep only 4 hours, sleep efficiency is 50%. The goal is to reduce TIB to match the time asleep and to agree on a prescribed daily wake-up time. When the patient is consistently sleeping 90% of the TIB, add 30-minute increments until he or she is appropriately sleeping 7 to 8 hours at night.⁹ Naps are not recommended.

Let patients know that their migraine may worsen until a new routine sleep pattern emerges. This method is not recommended for patients with untreated sleep apnea.

E IS FOR EXERCISE

Exercise is broadly recommended for a healthy lifestyle; some evidence suggests that it can also be useful in the management of migraine.¹⁰ Low levels of physical activity and a sedentary lifestyle are associated with migraine.¹¹ It is unclear if patients with migraine are less likely to exercise because they want to avoid triggering a migraine or if a sedentary lifestyle increases their risk.

Exercise has been studied for its prophylactic benefits in migraine, and one hypothesis relates to beta-endorphins. Levels of beta-endorphins are reduced in the cerebrospinal fluid of patients with migraine.¹² Exercise programs may increase levels while reducing headache frequency and duration.¹³ One study showed that pain thresholds do not change with exercise programs, suggesting that it is avoidance behavior that is positively altered rather than the underlying pain pathways.¹⁴

A systematic review and meta-analysis based on 5 randomized controlled trials and 1 nonrandomized controlled clinical trial showed that exercise reduced monthly migraine days by only 0.6 (± 0.3) days, but the data also suggested that as the exercise intensity increased, so did the positive effects.¹⁰

Some data suggest that exercise may also reduce migraine duration and severity as well as the need for abortive medication.¹⁰ Two studies in this systematic review^15,16 showed that exercise benefits were equivalent to those of migraine preventives such as amitriptyline and topiramate; the combination of amitriptyline and exercise was more beneficial than exercise alone. Multiple types of exercise were beneficial, including walking, jogging, cross-training, and cycling when done for least 6 weeks and for 30 to 50 minutes 3 to 5 times a week.

These findings are in line with the current recommendations for general health from the American College of Sports Medicine, ie, moderate to vigorous cardio­respiratory exercise for 30 to 60 minutes 3 to 5 times a week (or 150 minutes per week). The daily exercise can be continuous or done in intervals of less than 20 minutes. For those with a sedentary lifestyle, as is seen in a significant proportion of the migraine population, light to moderate exercise for less than 20 minutes is still beneficial.¹⁷

Based on this evidence, the best current recommendation for patients with migraine is to engage in graded moderate cardiorespiratory exercise, although any exercise is better than none. If a patient is sedentary or has poor exercise tolerance, or both, exercising once a week for shorter time periods may be a manageable place to start.

Some patients may identify exercise as a trigger or exacerbating factor in migraine. These patients may need appropriate prophylactic and abortive therapies before starting an exercise regimen.

THE SECOND E IS FOR EAT (FOOD AND DRINK)

Many patients believe that some foods trigger migraine attacks, but further study is needed. The most consistent food triggers appear to be red wine and caffeine (withdrawal).^18,19 Interestingly, patients with migraine report low levels of alcohol consumption,²⁰ but it is unclear if that is because alcohol has a protective effect or if patients avoid it.

Some patients may crave certain foods in the prodromal phase of an attack, eat the food, experience the attack, and falsely conclude that the food caused the attack.²¹ Premonitory symptoms include fatigue, cognitive changes, homeostatic changes, sensory hyperresponsiveness, and food cravings.²¹ It is difficult to distinguish between premonitory phase food cravings and true triggers because premonitory symptoms can precede headache by 48 to 72 hours, and the timing for a trigger to be considered causal is not known.²²

Chocolate is often thought to be a migraine trigger, but the evidence argues against this and even suggests that sweet cravings are a part of the premonitory phase.²³ Monosodium glutamate is often identified as a trigger as well, but the literature is inconsistent and does not support a causal relationship.²⁴ Identifying true food triggers in migraine is difficult, and patients with migraine may have poor quality diets, with some foods acting as true triggers for certain patients.²⁵ These possibilities have led to the development of many “migraine diets,” including elimination diets.

Elimination diets

Elimination diets involve avoiding specific food items over a period of time and then adding them back in one at a time to gauge whether they cause a reaction in the body. A number of these diets have been studied for their effects on headache and migraine:

Gluten-free diets restrict foods that contain wheat, rye, and barley. A systematic review of gluten-free diets in patients with celiac disease found that headache or migraine frequency decreased by 51.6% to 100% based on multiple cohort studies (N = 42,388).²⁶ There are no studies on the use of a gluten-free diet for migraine in patients without celiac disease.

Immunoglobulin G-elimination diets restrict foods that serve as antigens for IgG. However, data supporting these diets are inconsistent. Two small randomized controlled trials found that the diets improved migraine symptoms, but a larger study found no improvement in the number of migraine days at 12 weeks, although there was an initially significant effect at 4 weeks.^27–29

Antihistamine diets restrict foods that have high levels of histamines, including fermented dairy, vegetables, soy products,  wine, beer, alcohol, and those that cause histamine release regardless of IgE testing results. A prospective single-arm study of antihistamine diets in patients with chronic headache reported symptom improvement, which could be applied to certain comorbidities such as  mast cell activation syndrome.³⁰ Another prospective nonrandomized controlled study eliminated foods based on positive IgE skin-prick testing for allergy in patients with recurrent migraine and found that it reduced headache frequency.³¹

Tyramine-free diets are often recommended due to the presumption that tyramine-containing foods (eg, aged cheese, cured or smoked meats and fish, and beer) are triggers. However, multiple studies have reviewed this theory with inconsistent results,³² and the only study of a tyramine-free diet was negative.³³ In addition, commonly purported high-tyramine foods have lower tyramine levels than previously thought.³⁴

Low-fat diets in migraine are supported by 2 small randomized controlled trials and a prospective study showing a decrease in symptom severity; the results for frequency are inconsistent.^35–37

Low-glycemic index diets are supported in migraine by 1 randomized controlled trial that showed improvement in migraine frequency in a diet group and in a control group of patients who took a standard migraine-preventive medication to manage their symptoms.³⁸

Other migraine diets

Diets high in certain foods or ingredient ratios, as opposed to elimination diets, have also been studied in patients with migraine. One promising diet containing high levels of omega-3 fatty acids and low levels of omega-6 fatty acids was shown in a systematic review to reduce the duration of migraine but not the frequency or severity.³⁹ A more recent randomized controlled trial of this diet in chronic migraine also showed that it decreased migraine frequency.⁴⁰

The ketogenic diet (high fat, low carbohydrate) had promising results in a randomized controlled trial in overweight women with migraine and in a prospective study.^41,42 However, a prospective study of the Atkins diet in teenagers with chronic daily headaches showed no benefit.⁴³ The ketogenic diet is difficult to follow and may work in part due to weight loss alone, although ketogenesis itself may also play a role.^41,44

Sodium levels have been shown to be higher in the cerebrospinal fluid of patients with migraine than in controls, particularly during an attack.⁴⁵ For a prehypertensive population or an elderly population, a low-sodium diet may be beneficial based on 2 prospective trials.^46,47 However, a younger female population without hypertension and low-to-normal body mass index had a reduced probability of migraine while consuming a high-sodium diet.⁴⁸

Counseling about sodium intake should be tailored to specific patient populations. For example, a diet low in sodium may be appropriate for patients with vascular risk factors such as hypertension, whereas a high-sodium diet may be appropriate in patients with comorbidities like postural tachycardia syndrome or in those with a propensity for low blood pressure or low body mass index.

Encourage routine meals and hydration

The standard advice for patients with migraine is to consume regular meals. Headaches have been associated with fasting, and those with migraine are predisposed to attacks in the setting of fasting.^49,50 Migraine is more common when meals are skipped, particularly breakfast.⁵¹

It is unclear how fasting lowers the migraine threshold. Nutritional studies show that skipping meals, particularly breakfast, increases low-grade inflammation and impairs  glucose metabolism by affecting insulin and fat oxidation metabolism.⁵² However, hypoglycemia itself is not a consistent cause of headache or migraine attacks.⁵³ As described above, a randomized controlled trial of a low-glycemic index diet actually decreased migraine frequency and severity.³⁸ Skipping meals also reduces energy and is associated with reduced physical activity, perhaps leading to multiple compounding triggers that further lower the migraine threshold.^54,55

When counseling patients about the need to eat breakfast, consider what they normally consume (eg, is breakfast just a cup of coffee?). Replacing simple carbohydrates with protein, fats, and fiber may be beneficial for general health, but the effects on migraine are not known, nor is the optimal composition of breakfast foods.⁵⁵

The optimal timing of breakfast relative to awakening is also unclear, but in general, it should be eaten within 30 to 60 minutes of rising. Also consider patients’ work hours—delayed-phase or shift workers have altered sleep cycles.

Recommendations vary in regard to hydration. Headache is associated with fluid restriction and dehydration,^56,57 but only a few studies suggest that rehydration and increased hydration status can improve migraine.⁵⁸ In fact, a single post hoc analysis of a metoclopramide study showed that intravenous fluid alone for patients with migraine in the emergency room did not improve pain outcomes.⁵⁹

The amount of water patients should drink daily in the setting of migraine is also unknown, but a study showed benefit with 4 L, which equates to a daily intake of 16 eight-ounce glasses.⁶⁰ One review on general health that could be extrapolated given the low risk of the intervention indicated that 1.8 L daily (7 to 8 eight-ounce glasses) promoted a euhydration status in most people, although many factors contribute to hydration status.⁶¹

Caffeine intake is also a major consideration. Caffeine is a nonspecific adenosine receptor antagonist that modulates adenosine receptors like the pronociceptive 2A receptor, leading to changes integral to the neuropathophysiology of migraine.⁶² Caffeine has analgesic properties at doses greater than 65 to 200 mg and augments the effects of analgesics such as acetaminophen and aspirin. Chronic caffeine use can lead to withdrawal symptoms when intake is stopped abruptly; this is thought to be due to upregulation of adenosine receptors, but the effect varies based on genetic predisposition.¹⁹

The risk of chronic daily headache may relate to high use of caffeine preceding the onset of chronification, and caffeine abstinence may improve response to acute migraine treatment.^19,63 There is a dose-dependent risk of headache.^64,65 Current recommendations suggest limiting caffeine consumption to less than 200 mg per day or stopping caffeine consumption altogether based on the quantity required for caffeine-withdrawal headache.⁶⁶ Varying  the caffeine dose from day to day may also trigger headache due to the high sensitivity to caffeine withdrawal.

While many diets have shown potential benefit in patients with migraine, more studies are needed before any one “migraine diet” can be recommended. Caution should be taken, as there is risk of adverse effects from nutrient deficiencies or excess levels, especially if the patient is not under the care of a healthcare professional who is familiar with the diet.

Whether it is beneficial to avoid specific food triggers at this time is unclear and still controversial even within the migraine community because some of these foods may be misattributed as triggers instead of premonitory cravings driven by the hypothalamus. It is important to counsel patients with migraine to eat a healthy diet with consistent meals, to maintain adequate hydration, and to keep their caffeine intake low or at least consistent, although these teachings are predominantly based on limited studies with extrapolation from nutrition research.

D IS FOR DIARY

A headache diary is a recommended part of headache management and may enhance the accuracy of diagnosis and assist in treatment modifications. Paper and electronic diaries have been used. Electronic diaries may be more accurate for real-time use, but patients may be more likely to complete a paper one.⁶⁷ Patients prefer electronic diaries over long paper forms,⁶⁸ but a practical issue to consider is easy electronic access.

Patients can start keeping a headache diary before the initial consultation to assist with diagnosis, or early in their management. A first-appointment diary mailed with instructions is a feasible option.⁶⁹ These types of diaries ask detailed questions to help diagnose all major primary headache types including menstrual migraine and to identify concomitant medication-overuse headache. Physicians and patients generally report improved communication with use of a diary.⁷⁰

Some providers distinguish between a headache diary and a calendar. In standard practice, a headache diary is the general term referring to both, but the literature differentiates between the two. Both should at least include headache frequency, with possible inclusion of other factors such as headache duration, headache intensity, analgesic use, headache impact on function, and absenteeism. Potential triggers including menses can also be tracked. The calendar version can fit on a single page and can be used for simple tracking of headache frequency and analgesia use.

One of the simplest calendars to use is the “stoplight” calendar. Red days are when a patient is completely debilitated in bed. On a yellow day, function at work, school, or daily activities is significantly reduced by migraine, but the patient is not bedbound. A green day is when headache is present but function is not affected. No color is placed if the patient is 100% headache-free.

Acute treatment use can be written in or, to improve compliance, a checkmark can be placed on days of treatment. Patients who are tracking menses circle the days of menstruation. The calendar-diary should be brought to every appointment to track treatment response and medication use.

THE SECOND S IS FOR STRESS

Behavioral management such as cognitive behavioral therapy in migraine has been shown to decrease catastrophizing, migraine disability, and headache severity and frequency.⁷⁴ Both depression and anxiety can improve along with migraine.⁷⁵ Cognitive behavioral therapy can be provided in individualized sessions or group sessions, either in person or online.^74,76,77 The effects become more prominent about 5 weeks into treatment.⁷⁸

Biofeedback, which uses behavioral techniques paired with physiologic autonomic measures, has been extensively studied, and shows benefit in migraine, including in meta-analysis.⁷⁹ The types of biofeedback measurements used include electromyography, electroencephalography, temperature, sweat sensors, heart rate, blood volume pulse feedback, and respiration bands. While biofeedback is generally done under the guidance of a therapist, it can still be useful with minimal therapist contact and supplemental audio.⁸⁰

Mindfulness, or the awareness of thoughts, feelings, and sensations in the present moment without judgment, is a behavioral technique that can be done alone or paired with another technique. It is often taught through a mindfulness-based stress-reduction  program, which relies on a standardized approach. A meta-analysis showed that mindfulness improves pain intensity, headache frequency, disability, self-efficacy, and quality of life.⁸¹ It may work by encouraging pain acceptance.⁸²

Relaxation techniques are also employed in migraine management, either alone or in conjunction with techniques mentioned  above, such as mindfulness. They include progressive muscle relaxation and deep breathing. Relaxation has been shown to be effective when done by professional trainers as well as lay trainers in both individual and group settings.^83,84

In patients with intractable headache, more-intensive inpatient and outpatient programs have been tried. Inpatient admissions with multidisciplinary programs that include a focus on behavioral techniques often paired with lifestyle education and sometimes pharmacologic management can be beneficial.^85,86 These programs have also been successfully conducted as multiple outpatient sessions.^86–88

Stress management is an important aspect of migraine management. These treatments often involve homework and require active participation.

LIFESTYLE FOR ALL

All patients with migraine should initiate lifestyle modifications (see Advice to patients with migraine: SEEDS for success). Modifications with the highest level of evidence, specifically behavioral techniques, have had the most reproducible results. A headache diary is an essential tool to identify patterns and needs for optimization of acute or preventive treatment regimens. The strongest evidence is for the behavioral management techniques for stress reduction.

Migraine is the second leading cause of years of life lived with a disability globally.¹ It affects people of all ages, but particularly during the years associated with the highest productivity in terms of work and family life.

Migraine is a genetic neurologic disease that can be influenced or triggered by environmental factors. However, triggers do not cause migraine. For example, stress does not cause migraine, but it can exacerbate it.

Primary care physicians can help patients reduce the likelihood of a migraine attack, the severity of symptoms, or both by offering lifestyle counseling centered around the mnemonic SEEDS: sleep, exercise, eat, diary, and stress. In this article, each factor is discussed individually for its current support in the literature along with best-practice recommendations.

S IS FOR SLEEP

Before optimizing sleep hygiene, screen for sleep apnea, especially in those who have chronic daily headache upon awakening. An excellent tool is the STOP-Bang screening questionnaire⁵ (www.stopbang.ca/osa/screening.php). Patients respond “yes” or “no” to the following questions:

• Snoring: Do you snore loudly (louder than talking or loud enough to be heard through closed doors)?
• Tired: Do you often feel tired, fatigued, or sleepy during the daytime?
• Observed: Has anyone observed you stop breathing during your sleep?
• Pressure: Do you have or are you being treated for high blood pressure?
• Body mass index greater than 35 kg/m²?
• Age over 50?
• Neck circumference larger than 40 cm (females) or  42 cm (males)?
• Gender—male?

Each “yes” answer is scored as 1 point. A score less than 3 indicates low risk of obstructive sleep apnea; 3 to 4 indicates moderate risk; and 5 or more indicates high risk. Optimization of sleep apnea with continuous positive airway pressure therapy can improve sleep apnea headache.⁶ The improved sleep from reduced arousals may also mitigate migraine symptoms.

Behavioral modification for sleep hygiene can convert chronic migraine to episodic migraine.⁷ One such program is stimulus control therapy, which focuses on using cues to initiate sleep (Table 1). Patients are encouraged to keep the bedroom quiet, dark, and cool, and to go to sleep at the same time every night. Importantly, the bed should be associated only with sleep. If patients are unable to fall asleep within 20 to 30 minutes, they should leave the room so they do not associate the bed with frustration and anxiety. Use of phones, tablets, and television in the bedroom is discouraged as these devices may make it more difficult to fall asleep.⁸

The next option is sleep restriction, which is useful for comorbid insomnia. Patients keep a sleep diary to better understand their sleep-wake cycle. The goal is 90% sleep efficiency, meaning that 90% of the time in bed (TIB) is spent asleep. For example, if the patient is in bed 8 hours but asleep only 4 hours, sleep efficiency is 50%. The goal is to reduce TIB to match the time asleep and to agree on a prescribed daily wake-up time. When the patient is consistently sleeping 90% of the TIB, add 30-minute increments until he or she is appropriately sleeping 7 to 8 hours at night.⁹ Naps are not recommended.

Let patients know that their migraine may worsen until a new routine sleep pattern emerges. This method is not recommended for patients with untreated sleep apnea.

E IS FOR EXERCISE

Exercise is broadly recommended for a healthy lifestyle; some evidence suggests that it can also be useful in the management of migraine.¹⁰ Low levels of physical activity and a sedentary lifestyle are associated with migraine.¹¹ It is unclear if patients with migraine are less likely to exercise because they want to avoid triggering a migraine or if a sedentary lifestyle increases their risk.

Exercise has been studied for its prophylactic benefits in migraine, and one hypothesis relates to beta-endorphins. Levels of beta-endorphins are reduced in the cerebrospinal fluid of patients with migraine.¹² Exercise programs may increase levels while reducing headache frequency and duration.¹³ One study showed that pain thresholds do not change with exercise programs, suggesting that it is avoidance behavior that is positively altered rather than the underlying pain pathways.¹⁴

A systematic review and meta-analysis based on 5 randomized controlled trials and 1 nonrandomized controlled clinical trial showed that exercise reduced monthly migraine days by only 0.6 (± 0.3) days, but the data also suggested that as the exercise intensity increased, so did the positive effects.¹⁰

Some data suggest that exercise may also reduce migraine duration and severity as well as the need for abortive medication.¹⁰ Two studies in this systematic review^15,16 showed that exercise benefits were equivalent to those of migraine preventives such as amitriptyline and topiramate; the combination of amitriptyline and exercise was more beneficial than exercise alone. Multiple types of exercise were beneficial, including walking, jogging, cross-training, and cycling when done for least 6 weeks and for 30 to 50 minutes 3 to 5 times a week.

These findings are in line with the current recommendations for general health from the American College of Sports Medicine, ie, moderate to vigorous cardio­respiratory exercise for 30 to 60 minutes 3 to 5 times a week (or 150 minutes per week). The daily exercise can be continuous or done in intervals of less than 20 minutes. For those with a sedentary lifestyle, as is seen in a significant proportion of the migraine population, light to moderate exercise for less than 20 minutes is still beneficial.¹⁷

Based on this evidence, the best current recommendation for patients with migraine is to engage in graded moderate cardiorespiratory exercise, although any exercise is better than none. If a patient is sedentary or has poor exercise tolerance, or both, exercising once a week for shorter time periods may be a manageable place to start.

Some patients may identify exercise as a trigger or exacerbating factor in migraine. These patients may need appropriate prophylactic and abortive therapies before starting an exercise regimen.

THE SECOND E IS FOR EAT (FOOD AND DRINK)

Many patients believe that some foods trigger migraine attacks, but further study is needed. The most consistent food triggers appear to be red wine and caffeine (withdrawal).^18,19 Interestingly, patients with migraine report low levels of alcohol consumption,²⁰ but it is unclear if that is because alcohol has a protective effect or if patients avoid it.

Some patients may crave certain foods in the prodromal phase of an attack, eat the food, experience the attack, and falsely conclude that the food caused the attack.²¹ Premonitory symptoms include fatigue, cognitive changes, homeostatic changes, sensory hyperresponsiveness, and food cravings.²¹ It is difficult to distinguish between premonitory phase food cravings and true triggers because premonitory symptoms can precede headache by 48 to 72 hours, and the timing for a trigger to be considered causal is not known.²²

Chocolate is often thought to be a migraine trigger, but the evidence argues against this and even suggests that sweet cravings are a part of the premonitory phase.²³ Monosodium glutamate is often identified as a trigger as well, but the literature is inconsistent and does not support a causal relationship.²⁴ Identifying true food triggers in migraine is difficult, and patients with migraine may have poor quality diets, with some foods acting as true triggers for certain patients.²⁵ These possibilities have led to the development of many “migraine diets,” including elimination diets.

Elimination diets

Elimination diets involve avoiding specific food items over a period of time and then adding them back in one at a time to gauge whether they cause a reaction in the body. A number of these diets have been studied for their effects on headache and migraine:

Gluten-free diets restrict foods that contain wheat, rye, and barley. A systematic review of gluten-free diets in patients with celiac disease found that headache or migraine frequency decreased by 51.6% to 100% based on multiple cohort studies (N = 42,388).²⁶ There are no studies on the use of a gluten-free diet for migraine in patients without celiac disease.

Immunoglobulin G-elimination diets restrict foods that serve as antigens for IgG. However, data supporting these diets are inconsistent. Two small randomized controlled trials found that the diets improved migraine symptoms, but a larger study found no improvement in the number of migraine days at 12 weeks, although there was an initially significant effect at 4 weeks.^27–29

Antihistamine diets restrict foods that have high levels of histamines, including fermented dairy, vegetables, soy products,  wine, beer, alcohol, and those that cause histamine release regardless of IgE testing results. A prospective single-arm study of antihistamine diets in patients with chronic headache reported symptom improvement, which could be applied to certain comorbidities such as  mast cell activation syndrome.³⁰ Another prospective nonrandomized controlled study eliminated foods based on positive IgE skin-prick testing for allergy in patients with recurrent migraine and found that it reduced headache frequency.³¹

Tyramine-free diets are often recommended due to the presumption that tyramine-containing foods (eg, aged cheese, cured or smoked meats and fish, and beer) are triggers. However, multiple studies have reviewed this theory with inconsistent results,³² and the only study of a tyramine-free diet was negative.³³ In addition, commonly purported high-tyramine foods have lower tyramine levels than previously thought.³⁴

Low-fat diets in migraine are supported by 2 small randomized controlled trials and a prospective study showing a decrease in symptom severity; the results for frequency are inconsistent.^35–37

Low-glycemic index diets are supported in migraine by 1 randomized controlled trial that showed improvement in migraine frequency in a diet group and in a control group of patients who took a standard migraine-preventive medication to manage their symptoms.³⁸

Other migraine diets

Diets high in certain foods or ingredient ratios, as opposed to elimination diets, have also been studied in patients with migraine. One promising diet containing high levels of omega-3 fatty acids and low levels of omega-6 fatty acids was shown in a systematic review to reduce the duration of migraine but not the frequency or severity.³⁹ A more recent randomized controlled trial of this diet in chronic migraine also showed that it decreased migraine frequency.⁴⁰

The ketogenic diet (high fat, low carbohydrate) had promising results in a randomized controlled trial in overweight women with migraine and in a prospective study.^41,42 However, a prospective study of the Atkins diet in teenagers with chronic daily headaches showed no benefit.⁴³ The ketogenic diet is difficult to follow and may work in part due to weight loss alone, although ketogenesis itself may also play a role.^41,44

Sodium levels have been shown to be higher in the cerebrospinal fluid of patients with migraine than in controls, particularly during an attack.⁴⁵ For a prehypertensive population or an elderly population, a low-sodium diet may be beneficial based on 2 prospective trials.^46,47 However, a younger female population without hypertension and low-to-normal body mass index had a reduced probability of migraine while consuming a high-sodium diet.⁴⁸

Counseling about sodium intake should be tailored to specific patient populations. For example, a diet low in sodium may be appropriate for patients with vascular risk factors such as hypertension, whereas a high-sodium diet may be appropriate in patients with comorbidities like postural tachycardia syndrome or in those with a propensity for low blood pressure or low body mass index.

Encourage routine meals and hydration

The standard advice for patients with migraine is to consume regular meals. Headaches have been associated with fasting, and those with migraine are predisposed to attacks in the setting of fasting.^49,50 Migraine is more common when meals are skipped, particularly breakfast.⁵¹

It is unclear how fasting lowers the migraine threshold. Nutritional studies show that skipping meals, particularly breakfast, increases low-grade inflammation and impairs  glucose metabolism by affecting insulin and fat oxidation metabolism.⁵² However, hypoglycemia itself is not a consistent cause of headache or migraine attacks.⁵³ As described above, a randomized controlled trial of a low-glycemic index diet actually decreased migraine frequency and severity.³⁸ Skipping meals also reduces energy and is associated with reduced physical activity, perhaps leading to multiple compounding triggers that further lower the migraine threshold.^54,55

When counseling patients about the need to eat breakfast, consider what they normally consume (eg, is breakfast just a cup of coffee?). Replacing simple carbohydrates with protein, fats, and fiber may be beneficial for general health, but the effects on migraine are not known, nor is the optimal composition of breakfast foods.⁵⁵

The optimal timing of breakfast relative to awakening is also unclear, but in general, it should be eaten within 30 to 60 minutes of rising. Also consider patients’ work hours—delayed-phase or shift workers have altered sleep cycles.

Recommendations vary in regard to hydration. Headache is associated with fluid restriction and dehydration,^56,57 but only a few studies suggest that rehydration and increased hydration status can improve migraine.⁵⁸ In fact, a single post hoc analysis of a metoclopramide study showed that intravenous fluid alone for patients with migraine in the emergency room did not improve pain outcomes.⁵⁹

The amount of water patients should drink daily in the setting of migraine is also unknown, but a study showed benefit with 4 L, which equates to a daily intake of 16 eight-ounce glasses.⁶⁰ One review on general health that could be extrapolated given the low risk of the intervention indicated that 1.8 L daily (7 to 8 eight-ounce glasses) promoted a euhydration status in most people, although many factors contribute to hydration status.⁶¹

Caffeine intake is also a major consideration. Caffeine is a nonspecific adenosine receptor antagonist that modulates adenosine receptors like the pronociceptive 2A receptor, leading to changes integral to the neuropathophysiology of migraine.⁶² Caffeine has analgesic properties at doses greater than 65 to 200 mg and augments the effects of analgesics such as acetaminophen and aspirin. Chronic caffeine use can lead to withdrawal symptoms when intake is stopped abruptly; this is thought to be due to upregulation of adenosine receptors, but the effect varies based on genetic predisposition.¹⁹

The risk of chronic daily headache may relate to high use of caffeine preceding the onset of chronification, and caffeine abstinence may improve response to acute migraine treatment.^19,63 There is a dose-dependent risk of headache.^64,65 Current recommendations suggest limiting caffeine consumption to less than 200 mg per day or stopping caffeine consumption altogether based on the quantity required for caffeine-withdrawal headache.⁶⁶ Varying  the caffeine dose from day to day may also trigger headache due to the high sensitivity to caffeine withdrawal.

While many diets have shown potential benefit in patients with migraine, more studies are needed before any one “migraine diet” can be recommended. Caution should be taken, as there is risk of adverse effects from nutrient deficiencies or excess levels, especially if the patient is not under the care of a healthcare professional who is familiar with the diet.

Whether it is beneficial to avoid specific food triggers at this time is unclear and still controversial even within the migraine community because some of these foods may be misattributed as triggers instead of premonitory cravings driven by the hypothalamus. It is important to counsel patients with migraine to eat a healthy diet with consistent meals, to maintain adequate hydration, and to keep their caffeine intake low or at least consistent, although these teachings are predominantly based on limited studies with extrapolation from nutrition research.

D IS FOR DIARY

A headache diary is a recommended part of headache management and may enhance the accuracy of diagnosis and assist in treatment modifications. Paper and electronic diaries have been used. Electronic diaries may be more accurate for real-time use, but patients may be more likely to complete a paper one.⁶⁷ Patients prefer electronic diaries over long paper forms,⁶⁸ but a practical issue to consider is easy electronic access.

Patients can start keeping a headache diary before the initial consultation to assist with diagnosis, or early in their management. A first-appointment diary mailed with instructions is a feasible option.⁶⁹ These types of diaries ask detailed questions to help diagnose all major primary headache types including menstrual migraine and to identify concomitant medication-overuse headache. Physicians and patients generally report improved communication with use of a diary.⁷⁰

Some providers distinguish between a headache diary and a calendar. In standard practice, a headache diary is the general term referring to both, but the literature differentiates between the two. Both should at least include headache frequency, with possible inclusion of other factors such as headache duration, headache intensity, analgesic use, headache impact on function, and absenteeism. Potential triggers including menses can also be tracked. The calendar version can fit on a single page and can be used for simple tracking of headache frequency and analgesia use.

One of the simplest calendars to use is the “stoplight” calendar. Red days are when a patient is completely debilitated in bed. On a yellow day, function at work, school, or daily activities is significantly reduced by migraine, but the patient is not bedbound. A green day is when headache is present but function is not affected. No color is placed if the patient is 100% headache-free.

Acute treatment use can be written in or, to improve compliance, a checkmark can be placed on days of treatment. Patients who are tracking menses circle the days of menstruation. The calendar-diary should be brought to every appointment to track treatment response and medication use.

THE SECOND S IS FOR STRESS

Behavioral management such as cognitive behavioral therapy in migraine has been shown to decrease catastrophizing, migraine disability, and headache severity and frequency.⁷⁴ Both depression and anxiety can improve along with migraine.⁷⁵ Cognitive behavioral therapy can be provided in individualized sessions or group sessions, either in person or online.^74,76,77 The effects become more prominent about 5 weeks into treatment.⁷⁸

Biofeedback, which uses behavioral techniques paired with physiologic autonomic measures, has been extensively studied, and shows benefit in migraine, including in meta-analysis.⁷⁹ The types of biofeedback measurements used include electromyography, electroencephalography, temperature, sweat sensors, heart rate, blood volume pulse feedback, and respiration bands. While biofeedback is generally done under the guidance of a therapist, it can still be useful with minimal therapist contact and supplemental audio.⁸⁰

Mindfulness, or the awareness of thoughts, feelings, and sensations in the present moment without judgment, is a behavioral technique that can be done alone or paired with another technique. It is often taught through a mindfulness-based stress-reduction  program, which relies on a standardized approach. A meta-analysis showed that mindfulness improves pain intensity, headache frequency, disability, self-efficacy, and quality of life.⁸¹ It may work by encouraging pain acceptance.⁸²

Relaxation techniques are also employed in migraine management, either alone or in conjunction with techniques mentioned  above, such as mindfulness. They include progressive muscle relaxation and deep breathing. Relaxation has been shown to be effective when done by professional trainers as well as lay trainers in both individual and group settings.^83,84

In patients with intractable headache, more-intensive inpatient and outpatient programs have been tried. Inpatient admissions with multidisciplinary programs that include a focus on behavioral techniques often paired with lifestyle education and sometimes pharmacologic management can be beneficial.^85,86 These programs have also been successfully conducted as multiple outpatient sessions.^86–88

Stress management is an important aspect of migraine management. These treatments often involve homework and require active participation.

LIFESTYLE FOR ALL

All patients with migraine should initiate lifestyle modifications (see Advice to patients with migraine: SEEDS for success). Modifications with the highest level of evidence, specifically behavioral techniques, have had the most reproducible results. A headache diary is an essential tool to identify patterns and needs for optimization of acute or preventive treatment regimens. The strongest evidence is for the behavioral management techniques for stress reduction.

Lyme disease is a complex multisystem bacterial infection affecting the skin, joints, heart, and nervous system. The full spectrum of disease was first recognized and the disease was named in the 1970s during an outbreak of arthritis in children in the town of Lyme, Connecticut.¹

This review describes the epidemiology and pathogenesis of Lyme disease, the advantages and disadvantages of current diagnostic methods, and diagnostic algorithms.

THE MOST COMMON TICK-BORNE INFECTION IN NORTH AMERICA

Lyme disease is the most common tick-borne infection in North America.^2,3 In the United States, more than 30,000 cases are reported annually. In fact, in 2017, the number of cases was about 42,000, a 16% increase from the previous year, according to the US Centers for Disease Control and Prevention (CDC).

From Sigal LH. Myths and facts about Lyme disease. Cleve Clin J Med 1997; 64(4):203–209.

The infection is caused by Borrelia burgdorferi, a particularly arthritogenic spirochete transmitted by Ixodes scapularis (the black-legged deer tick, (Figure 1) and Ixodes pacificus (the Western black-legged tick). Although the infection can occur at any time of the year, its peak incidence is in May to late September, coinciding with increased outdoor recreational activity in areas where ticks live.^3,4 The typical tick habitat consists of deciduous woodland with sufficient humidity provided by a good layer of decaying vegetation. However, people can contract Lyme disease in their own backyard.³

Most cases of Lyme disease are seen in the northeastern United States, mainly in suburban and rural areas.^2,3 Other areas affected include the midwestern states of Minnesota, Wisconsin, and Michigan, as well as northern California.⁴ Fourteen states and the District of Columbia report a high average incidence (> 10 cases per 100,000 persons) (Table 1).²

FIRST COMES IgM, THEN IgG

The pathogenesis and the different stages of infection should inform laboratory testing in Lyme disease.

It is estimated that only 5% of infected ticks that bite people actually transmit their spirochetes to the human host.⁵ However, once infected, the patient’s innate immune system mounts a response that results in the classic erythema migrans rash at the bite site. A rash develops in only about 85% of patients who are infected and can appear at any time between 3 and 30 days, but most commonly after 7 days. Hence, a rash occurring within the first few hours of tick contact is not erythema migrans and does not indicate infection, but rather an early reaction to tick salivary antigens.⁵

Antibody levels remain below the detection limits of currently available serologic tests in the first 7 days after exposure. Immunoglobulin M (IgM) antibody titers peak between 8 and 14 days after tick contact, but IgM antibodies may never develop if the patient is started on early appropriate antimicrobial therapy.⁵

If the infection is not treated, the spirochete may disseminate through the blood from the bite site to different tissues.³ Both cell-mediated and antibody-mediated immunity swing into action to kill the spirochetes at this stage. The IgM antibody response occurs in 1 to 2 weeks, followed by a robust IgG response in 2 to 4 weeks.⁶

Because IgM can also cross-react with antigens other than those associated with B burgdorferi, the IgM test is less specific than the IgG test for Lyme disease.

Once a patient is exposed and mounts an antibody-mediated response to the spirochete, the antibody profile may persist for months to years, even after successful antibiotic treatment and cure of the disease.⁵

Despite the immune system’s robust series of defenses, untreated B burgdorferi infection can persist, as the organism has a bag of tricks to evade destruction. It can decrease its expression of specific immunogenic surface-exposed proteins, change its antigenic properties through recombination, and bind to the patient’s extracellular matrix proteins to facilitate further dissemination.³

Certain host-genetic factors also play a role in the pathogenesis of Lyme disease, such as the HLA-DR4 allele, which has been associated with antibiotic-refractory Lyme-related arthritis.³

LYME DISEASE EVOLVES THROUGH STAGES

Lyme disease evolves through stages broadly classified as early and late infection, with significant variability in its presentation.⁷

Early infection

Early disease is further subdivided into “localized” infection (stage 1), characterized by a single erythema migrans lesion and local lymphadenopathy, and “disseminated” infection (stage 2), associated with multiple erythema migrans lesions distant from the bite site, facial nerve palsy, radiculoneuritis, meningitis, carditis, or migratory arthritis or arthralgia.⁸

Highly specific physical findings include erythema migrans, cranial nerve palsy, high-grade or progressive conduction block, and recurrent migratory polyarthritis. Less specific symptoms and signs of Lyme disease include arthralgia, myalgia, neck stiffness, palpitations, and myocarditis.⁵

Erythema migrans lesions are evident in at least 85% of patients with early disease.⁹ If they are not apparent on physical examination, they may be located at hidden sites and may be atypical in appearance or transient.⁵

If treatment is not started in the initial stage of the disease, 60% of infected patients may develop disseminated infection.⁵ Progressive, untreated infection can manifest with Lyme arthritis and neuroborreliosis.⁷

Noncutaneous manifestations are less common now than in the past due to increased awareness of the disease and early initiation of treatment.¹⁰

Late infection

Manifestations of late (stage 3) infection include oligoarthritis (affecting any joint but often the knee) and neuroborreliosis. Clinical signs and symptoms of Lyme disease may take months to resolve even after appropriate antimicrobial therapy is completed. This should not be interpreted as ongoing, persistent infection, but as related to host immune-mediated activity.⁵

The usefulness of a laboratory test depends on the individual patient’s pretest probability of infection, which in turn depends on the patient’s epidemiologic risk of exposure and clinical features of Lyme disease. Patients with a high pretest probability—eg, a history of a tick bite followed by the classic erythema migrans rash—do not need testing and can start antimicrobial therapy right away.¹¹

Serologic tests are the gold standard

Prompt diagnosis is important, as early Lyme disease is easily treatable without any future sequelae.¹¹

Tests for Lyme disease can be divided into direct methods, which detect the spirochete itself by culture or by polymerase chain reaction (PCR), and indirect methods, which detect antibodies (Table 2). Direct tests lack sensitivity for Lyme disease; hence, serologic tests remain the gold standard. Currently recommended is a standard 2-tier testing strategy using an enzyme-linked immunosorbent assay (ELISA) followed by Western blot for confirmation.

DIRECT METHODS

Culture lacks sensitivity

A number of factors limit the sensitivity of direct culture for diagnosing Lyme disease. B burgdorferi does not grow easily in culture, requiring special media, low temperatures, and long periods of incubation. Only a relatively few spirochetes are present in human tissues and body fluids to begin with, and bacterial counts are further reduced with duration and dissemination of infection.⁵ All of these limit the possibility of detecting this organism.

Polymerase chain reaction may help in some situations

Molecular assays are not part of the standard evaluation and should be used only in conjunction with serologic testing.⁷ These tests have high specificity but lack consistent sensitivity.

That said, PCR testing may be useful:

• In early infection, before antibody responses develop
• In reinfection, when serologic tests are not reliable because the antibodies persist for many years after an infection in many patients
• In endemic areas where serologic testing has high false-positive rates due to high baseline population seropositivity for anti-Borrelia antibodies caused by subclinical infection.³

PCR assays that target plasmid-borne genes encoding outer surface proteins A and C (OspA and OspC) and VisE (variable major protein-like sequence, expressed) are more sensitive than those that detect chromosomal 16s ribosomal ribonucleic acid (rRNA) genes, as plasmid-rich “blebs” are shed in larger concentrations than chromosomal DNA during active infection.⁷ However, these plasmid-contained genes persist in body tissues and fluids even after the infection is cleared, and their detection may not necessarily correlate with ongoing disease.⁸ Detection of chromosomal 16s rRNA genes is a better predictor of true organism viability.

The sensitivity of PCR for borrelial DNA depends on the type of sample. If a skin biopsy sample is taken of the leading edge of an erythema migrans lesion, the sensitivity is 69% and the specificity is 100%. In patients with Lyme arthritis, PCR of the synovial fluid has a sensitivity of up to 80%. However, the sensitivity of PCR of the cerebrospinal fluid of patients with neurologic manifestations of Lyme disease is only 19%.⁷ PCR of other clinical samples, including blood and urine, is not recommended, as spirochetes are primarily confined to tissues, and very few are present in these body fluids.^3,12

The disadvantage of PCR is that a positive result does not always mean active infection, as the DNA of the dead microbe persists for several months even after successful treatment.⁸

INDIRECT METHODS

Enzyme-linked immunosorbent assay

ELISAs detect anti-Borrelia antibodies. Early-generation ELISAs, still used in many laboratories, use whole-cell extracts of B burgdorferi. Examples are the Vidas Lyme screen (Biomérieux, biomerieux-usa.com) and the Wampole B burgdorferi IgG/M EIA II assay (Alere, www.alere.com). Newer ELISAs use recombinant proteins.¹³

Three major targets for ELISA antibodies are flagellin (Fla), outer surface protein C (OspC), and VisE, especially the invariable region 6 (IR6). Among these, VisE-IR6 is the most conserved region in B burgdorferi.

Early-generation assays have a sensitivity of 89% and specificity of 72%.¹¹ However, the patient’s serum may have antibodies that cross-react with unrelated bacterial antigens, leading to false-positive results (Table 3). Whole-cell sonicate assays are not recommended as an independent test and must be confirmed with Western blot testing when assay results are indeterminate or positive.¹¹

Newer-generation ELISAs detect antibodies targeting recombinant proteins of VisE, especially a synthetic peptide C6, within IR6.¹³ VisE-IR6 is the most conserved region of the B burgdorferi complex, and its detection is a highly specific finding, supporting the diagnosis of Lyme disease. Antibodies against VisE-IR6 antigen are the earliest to develop.⁵ An example of a newer-generation serologic test is the VisE C6 Lyme EIA kit, approved as a first-tier test by the US Food and Drug Administration in 2001. This test has a specificity of 99%,^14,15 and its specificity is further increased when used in conjunction with Western blot (99.5%).¹⁵ The advantage of the C6 antibody test is that it is more sensitive than 2-tier testing during early infection (sensitivity 29%–74% vs 17%–40% in early localized infection, and 56%–90% vs 27%–78% in early disseminated infection).⁶

During early infection, older and newer ELISAs are less sensitive because of the limited number of antigens expressed at this stage.¹³ All patients suspected of having early Lyme disease who are seronegative at initial testing should have follow-up testing to look for seroconversion.¹³

Western blot

Western blot (immunoblot) testing identifies IgM and IgG antibodies against specific B burgdorferi antigens. It is considered positive if it detects at least 2 of a possible 3 specific IgM bands in the first 4 weeks of disease or at least 5 of 10 specific IgG bands after 4 weeks of disease (Table 4 and Figure 2).¹⁶

The nature of the bands indicates the duration of infection: Western blot bands against 23-kD OspC and 41-kD FlaB are seen in early localized infection, whereas bands against all 3 B burgdorferi proteins will be seen after several weeks of disease.¹⁷ The IgM result should be interpreted carefully, as only 2 bands are required for the test to be positive, and IgM binds to antigen less specifically than IgG.¹²

If clinical symptoms and signs of Lyme disease have been present for more than 1 month, IgM reactivity alone should not be used to support the diagnosis, in view of the likelihood of a false-positive test result in this situation.¹⁸ This is called the “1-month rule” in the diagnosis of Lyme disease.¹³

In early localized infection, Western blot is only half as sensitive as ELISA testing. Since the overall sensitivity of a 2-step algorithm is equal to that of its least sensitive component, 2-tiered testing is not useful in early disease.¹³

Although currently considered the most specific test for confirmation of Lyme disease, Western blot has limitations. It is technically and interpretively complex and is thus not universally available.¹³ The blots are scored by visual examination, compromising the reproducibility of the test, although densitometric blot analysis techniques and automated scanning and scoring attempt to address some of these limitations.¹³ Like the ELISA, Western blot can have false-positive results in healthy individuals without tick exposure, as nonspecific IgM immunoblots develop faint bands. This is because of cross-reaction between B burgdorferi antigens and antigens from other microorganisms. Around 50% of healthy adults show low-level serum IgG reactivity against the FlaB antigen, leading to false-positive results as well. In cases in which the Western blot result is indeterminate, other etiologies must be considered.

False-positive IgM Western blots are a significant problem. In a 5-year retrospective study done at 63 US Air Force healthcare facilities, 113 (53.3%) of 212 IgM Western blots were falsely positive.¹⁹ A false-positive test was defined as one that failed to meet seropositivity (a first-tier test omitted or negative, > 30 days of symptoms with negative IgG blot), lack of exposure including residing in areas without documented tick habitats, patients having atypical or no symptoms, and negative serology within 30 days of a positive test.

In a similar study done in a highly endemic area, 50 (27.5%) of 182 patients had a false-positive test.²⁰ Physicians need to be careful when interpreting IgM Western blots. It is always important to consider locale, epidemiology, and symptoms when interpreting the test.

Limitations of serologic tests for Lyme disease

Currently available serologic tests have inherent limitations:

• Antibodies against B burgdorferi take at least 1 week to develop
• The background rate of seropositivity in endemic areas can be up to 4%, affecting the utility of a positive test result
• Serologic tests cannot be used as tests of cure because antibodies can persist for months to years even after appropriate antimicrobial therapy and cure of disease; thus, a positive serologic result could represent active infection or remote exposure²¹
• Antibodies can cross-react with related bacteria, including other borrelial or treponemal spirochetes
• False-positive serologic test results can also occur in association with other medical conditions such as polyclonal gammopathies and systemic lupus erythematosus.¹²

RECOMMENDATIONS FOR TESTING

Standard 2-tier testing

The CDC released recommendations for diagnosing Lyme disease after a second national conference of serologic diagnosis of Lyme disease in October 1994.¹⁸ The 2-tiered testing method, involving a sensitive ELISA followed by the Western blot to confirm positive and indeterminate ELISA results, was suggested as the gold standard for diagnosis (Figure 3). Of note, negative ELISA results do not require further testing.¹¹

The sensitivity of 2-tiered testing depends on the stage of the disease. Unfortunately, this method has a wide range of sensitivity (17% to 78%) in stage 1 disease. In the same stage, the sensitivity increases from 14.1% in patients with a single erythema migrans lesion and early localized infection to 65.4% in those with multiple lesions. The algorithm has excellent sensitivity in late stage 3 infection (96% to 100%).⁵

A 2-step ELISA algorithm

A 2-step ELISA algorithm (without the Western blot) that includes the whole-cell sonicate assay followed by the VisE C6 peptide assay actually showed higher sensitivity and comparable specificity compared with 2-tiered testing in early localized disease (sensitivity 61%–74% vs 29%–48%, respectively; specificity 99.5% for both methods).²² This higher sensitivity was even more pronounced in early disseminated infection (sensitivity 100% vs 40%, respectively). By late infection, the sensitivities of both testing strategies reached 100%. Compared with the Western blot, the 2-step ELISA algorithm was simpler to execute in a reproducible fashion.⁵

The Infectious Diseases Society of America is revising its current guidelines, with an update expected late this year, which may shift the recommendation from 2-tiered testing to the 2-step ELISA algorithm.

Multiplex testing

To overcome the intrinsic problems of protein-based assays, a multiplexed, array-based assay for the diagnosis of tick-borne infections called Tick-Borne Disease Serochip (TBD-Serochip) was established using recombinant antigens that identify key immunodominant epitopes.⁸ More studies are needed to establish the validity and usefulness of these tests in clinical practice.

Who should not be tested?

The American College of Physicians⁶ recommends against testing in patients:

• Presenting with nonspecific symptoms (eg, headache, myalgia, fatigue, arthralgia) without objective signs of Lyme disease
• With low pretest probability of infection based on epidemiologic exposures and clinical features
• Living in Lyme-endemic areas with no history of tick exposure⁶
• Presenting less than 1 week after tick exposure⁵
• Seeking a test of cure for treated Lyme disease.

DIAGNOSIS IN SPECIAL SITUATIONS

Early Lyme disease

The classic erythema migrans lesion on physical examination of a patient with suspected Lyme disease is diagnostic and does not require laboratory confirmation.¹⁰

In ambiguous cases, 2-tiered testing of a serum sample during the acute presentation and again 4 to 6 weeks later can be useful. In patients who remain seronegative on paired serum samples despite symptoms lasting longer than 6 weeks and no antibiotic treatment in the interim, the diagnosis of Lyme disease is unlikely, and another diagnosis should be sought.³

Antimicrobial therapy may block the serologic response; hence, negative serologic testing in patients started on empiric antibiotics should not rule out Lyme disease.⁶

PCR or bacterial culture testing is not recommended in the evaluation of suspected early Lyme disease.

Central nervous system Lyme disease

Central nervous system Lyme disease is diagnosed by 2-tiered testing using peripheral blood samples because all patients with this infectious manifestation should have mounted an adequate IgG response in the blood.¹¹

B cells migrate to and proliferate inside the central nervous system, leading to intrathecal production of anti-Borrelia antibodies. An index of cerebrospinal fluid to serum antibody greater than 1 is thus also indicative of neuroborreliosis.¹² Thus, performing lumbar puncture to detect intrathecal production of antibodies may support the diagnosis of central nervous system Lyme disease; however, it is not necessary.¹¹

Antibodies persist in the central nervous system for many years after appropriate antimicrobial treatment.

Lyme arthritis

Articular involvement in Lyme disease is characterized by a robust humoral response such that a negative IgG serologic test virtually rules out Lyme arthritis.²³ PCR testing of synovial fluid for borrelial DNA has a sensitivity of 80% but may become falsely negative after 1 to 2 months of antibiotic treatment.^24,25 In an algorithm suggested by Puius et al,²³ PCR testing of synovial fluid should be done in patients who have minimal to no response after 2 months of appropriate oral antimicrobial therapy to determine whether intravenous antibiotics are merited.

Table 5 summarizes the tests of choice in different clinical stages of infection.

Acknowledgment: The authors would like to acknowledge Anita Modi, MD, and Ceena N. Jacob, MD, for reviewing the manuscript and providing valuable suggestions, and Belinda Yen-Lieberman, PhD, for contributing pictures of the Western blot test results.

Lyme disease is a complex multisystem bacterial infection affecting the skin, joints, heart, and nervous system. The full spectrum of disease was first recognized and the disease was named in the 1970s during an outbreak of arthritis in children in the town of Lyme, Connecticut.¹

This review describes the epidemiology and pathogenesis of Lyme disease, the advantages and disadvantages of current diagnostic methods, and diagnostic algorithms.

THE MOST COMMON TICK-BORNE INFECTION IN NORTH AMERICA

Lyme disease is the most common tick-borne infection in North America.^2,3 In the United States, more than 30,000 cases are reported annually. In fact, in 2017, the number of cases was about 42,000, a 16% increase from the previous year, according to the US Centers for Disease Control and Prevention (CDC).

From Sigal LH. Myths and facts about Lyme disease. Cleve Clin J Med 1997; 64(4):203–209.

The infection is caused by Borrelia burgdorferi, a particularly arthritogenic spirochete transmitted by Ixodes scapularis (the black-legged deer tick, (Figure 1) and Ixodes pacificus (the Western black-legged tick). Although the infection can occur at any time of the year, its peak incidence is in May to late September, coinciding with increased outdoor recreational activity in areas where ticks live.^3,4 The typical tick habitat consists of deciduous woodland with sufficient humidity provided by a good layer of decaying vegetation. However, people can contract Lyme disease in their own backyard.³

Most cases of Lyme disease are seen in the northeastern United States, mainly in suburban and rural areas.^2,3 Other areas affected include the midwestern states of Minnesota, Wisconsin, and Michigan, as well as northern California.⁴ Fourteen states and the District of Columbia report a high average incidence (> 10 cases per 100,000 persons) (Table 1).²

FIRST COMES IgM, THEN IgG

The pathogenesis and the different stages of infection should inform laboratory testing in Lyme disease.

It is estimated that only 5% of infected ticks that bite people actually transmit their spirochetes to the human host.⁵ However, once infected, the patient’s innate immune system mounts a response that results in the classic erythema migrans rash at the bite site. A rash develops in only about 85% of patients who are infected and can appear at any time between 3 and 30 days, but most commonly after 7 days. Hence, a rash occurring within the first few hours of tick contact is not erythema migrans and does not indicate infection, but rather an early reaction to tick salivary antigens.⁵

Antibody levels remain below the detection limits of currently available serologic tests in the first 7 days after exposure. Immunoglobulin M (IgM) antibody titers peak between 8 and 14 days after tick contact, but IgM antibodies may never develop if the patient is started on early appropriate antimicrobial therapy.⁵

If the infection is not treated, the spirochete may disseminate through the blood from the bite site to different tissues.³ Both cell-mediated and antibody-mediated immunity swing into action to kill the spirochetes at this stage. The IgM antibody response occurs in 1 to 2 weeks, followed by a robust IgG response in 2 to 4 weeks.⁶

Because IgM can also cross-react with antigens other than those associated with B burgdorferi, the IgM test is less specific than the IgG test for Lyme disease.

Once a patient is exposed and mounts an antibody-mediated response to the spirochete, the antibody profile may persist for months to years, even after successful antibiotic treatment and cure of the disease.⁵

Despite the immune system’s robust series of defenses, untreated B burgdorferi infection can persist, as the organism has a bag of tricks to evade destruction. It can decrease its expression of specific immunogenic surface-exposed proteins, change its antigenic properties through recombination, and bind to the patient’s extracellular matrix proteins to facilitate further dissemination.³

Certain host-genetic factors also play a role in the pathogenesis of Lyme disease, such as the HLA-DR4 allele, which has been associated with antibiotic-refractory Lyme-related arthritis.³

LYME DISEASE EVOLVES THROUGH STAGES

Lyme disease evolves through stages broadly classified as early and late infection, with significant variability in its presentation.⁷

Early infection

Early disease is further subdivided into “localized” infection (stage 1), characterized by a single erythema migrans lesion and local lymphadenopathy, and “disseminated” infection (stage 2), associated with multiple erythema migrans lesions distant from the bite site, facial nerve palsy, radiculoneuritis, meningitis, carditis, or migratory arthritis or arthralgia.⁸

Highly specific physical findings include erythema migrans, cranial nerve palsy, high-grade or progressive conduction block, and recurrent migratory polyarthritis. Less specific symptoms and signs of Lyme disease include arthralgia, myalgia, neck stiffness, palpitations, and myocarditis.⁵

Erythema migrans lesions are evident in at least 85% of patients with early disease.⁹ If they are not apparent on physical examination, they may be located at hidden sites and may be atypical in appearance or transient.⁵

If treatment is not started in the initial stage of the disease, 60% of infected patients may develop disseminated infection.⁵ Progressive, untreated infection can manifest with Lyme arthritis and neuroborreliosis.⁷

Noncutaneous manifestations are less common now than in the past due to increased awareness of the disease and early initiation of treatment.¹⁰

Late infection

Manifestations of late (stage 3) infection include oligoarthritis (affecting any joint but often the knee) and neuroborreliosis. Clinical signs and symptoms of Lyme disease may take months to resolve even after appropriate antimicrobial therapy is completed. This should not be interpreted as ongoing, persistent infection, but as related to host immune-mediated activity.⁵

The usefulness of a laboratory test depends on the individual patient’s pretest probability of infection, which in turn depends on the patient’s epidemiologic risk of exposure and clinical features of Lyme disease. Patients with a high pretest probability—eg, a history of a tick bite followed by the classic erythema migrans rash—do not need testing and can start antimicrobial therapy right away.¹¹

Serologic tests are the gold standard

Prompt diagnosis is important, as early Lyme disease is easily treatable without any future sequelae.¹¹

Tests for Lyme disease can be divided into direct methods, which detect the spirochete itself by culture or by polymerase chain reaction (PCR), and indirect methods, which detect antibodies (Table 2). Direct tests lack sensitivity for Lyme disease; hence, serologic tests remain the gold standard. Currently recommended is a standard 2-tier testing strategy using an enzyme-linked immunosorbent assay (ELISA) followed by Western blot for confirmation.

DIRECT METHODS

Culture lacks sensitivity

A number of factors limit the sensitivity of direct culture for diagnosing Lyme disease. B burgdorferi does not grow easily in culture, requiring special media, low temperatures, and long periods of incubation. Only a relatively few spirochetes are present in human tissues and body fluids to begin with, and bacterial counts are further reduced with duration and dissemination of infection.⁵ All of these limit the possibility of detecting this organism.

Polymerase chain reaction may help in some situations

Molecular assays are not part of the standard evaluation and should be used only in conjunction with serologic testing.⁷ These tests have high specificity but lack consistent sensitivity.

That said, PCR testing may be useful:

• In early infection, before antibody responses develop
• In reinfection, when serologic tests are not reliable because the antibodies persist for many years after an infection in many patients
• In endemic areas where serologic testing has high false-positive rates due to high baseline population seropositivity for anti-Borrelia antibodies caused by subclinical infection.³

PCR assays that target plasmid-borne genes encoding outer surface proteins A and C (OspA and OspC) and VisE (variable major protein-like sequence, expressed) are more sensitive than those that detect chromosomal 16s ribosomal ribonucleic acid (rRNA) genes, as plasmid-rich “blebs” are shed in larger concentrations than chromosomal DNA during active infection.⁷ However, these plasmid-contained genes persist in body tissues and fluids even after the infection is cleared, and their detection may not necessarily correlate with ongoing disease.⁸ Detection of chromosomal 16s rRNA genes is a better predictor of true organism viability.

The sensitivity of PCR for borrelial DNA depends on the type of sample. If a skin biopsy sample is taken of the leading edge of an erythema migrans lesion, the sensitivity is 69% and the specificity is 100%. In patients with Lyme arthritis, PCR of the synovial fluid has a sensitivity of up to 80%. However, the sensitivity of PCR of the cerebrospinal fluid of patients with neurologic manifestations of Lyme disease is only 19%.⁷ PCR of other clinical samples, including blood and urine, is not recommended, as spirochetes are primarily confined to tissues, and very few are present in these body fluids.^3,12

The disadvantage of PCR is that a positive result does not always mean active infection, as the DNA of the dead microbe persists for several months even after successful treatment.⁸

INDIRECT METHODS

Enzyme-linked immunosorbent assay

ELISAs detect anti-Borrelia antibodies. Early-generation ELISAs, still used in many laboratories, use whole-cell extracts of B burgdorferi. Examples are the Vidas Lyme screen (Biomérieux, biomerieux-usa.com) and the Wampole B burgdorferi IgG/M EIA II assay (Alere, www.alere.com). Newer ELISAs use recombinant proteins.¹³

Three major targets for ELISA antibodies are flagellin (Fla), outer surface protein C (OspC), and VisE, especially the invariable region 6 (IR6). Among these, VisE-IR6 is the most conserved region in B burgdorferi.

Early-generation assays have a sensitivity of 89% and specificity of 72%.¹¹ However, the patient’s serum may have antibodies that cross-react with unrelated bacterial antigens, leading to false-positive results (Table 3). Whole-cell sonicate assays are not recommended as an independent test and must be confirmed with Western blot testing when assay results are indeterminate or positive.¹¹

Newer-generation ELISAs detect antibodies targeting recombinant proteins of VisE, especially a synthetic peptide C6, within IR6.¹³ VisE-IR6 is the most conserved region of the B burgdorferi complex, and its detection is a highly specific finding, supporting the diagnosis of Lyme disease. Antibodies against VisE-IR6 antigen are the earliest to develop.⁵ An example of a newer-generation serologic test is the VisE C6 Lyme EIA kit, approved as a first-tier test by the US Food and Drug Administration in 2001. This test has a specificity of 99%,^14,15 and its specificity is further increased when used in conjunction with Western blot (99.5%).¹⁵ The advantage of the C6 antibody test is that it is more sensitive than 2-tier testing during early infection (sensitivity 29%–74% vs 17%–40% in early localized infection, and 56%–90% vs 27%–78% in early disseminated infection).⁶

During early infection, older and newer ELISAs are less sensitive because of the limited number of antigens expressed at this stage.¹³ All patients suspected of having early Lyme disease who are seronegative at initial testing should have follow-up testing to look for seroconversion.¹³

Western blot

Western blot (immunoblot) testing identifies IgM and IgG antibodies against specific B burgdorferi antigens. It is considered positive if it detects at least 2 of a possible 3 specific IgM bands in the first 4 weeks of disease or at least 5 of 10 specific IgG bands after 4 weeks of disease (Table 4 and Figure 2).¹⁶

The nature of the bands indicates the duration of infection: Western blot bands against 23-kD OspC and 41-kD FlaB are seen in early localized infection, whereas bands against all 3 B burgdorferi proteins will be seen after several weeks of disease.¹⁷ The IgM result should be interpreted carefully, as only 2 bands are required for the test to be positive, and IgM binds to antigen less specifically than IgG.¹²

If clinical symptoms and signs of Lyme disease have been present for more than 1 month, IgM reactivity alone should not be used to support the diagnosis, in view of the likelihood of a false-positive test result in this situation.¹⁸ This is called the “1-month rule” in the diagnosis of Lyme disease.¹³

In early localized infection, Western blot is only half as sensitive as ELISA testing. Since the overall sensitivity of a 2-step algorithm is equal to that of its least sensitive component, 2-tiered testing is not useful in early disease.¹³

Although currently considered the most specific test for confirmation of Lyme disease, Western blot has limitations. It is technically and interpretively complex and is thus not universally available.¹³ The blots are scored by visual examination, compromising the reproducibility of the test, although densitometric blot analysis techniques and automated scanning and scoring attempt to address some of these limitations.¹³ Like the ELISA, Western blot can have false-positive results in healthy individuals without tick exposure, as nonspecific IgM immunoblots develop faint bands. This is because of cross-reaction between B burgdorferi antigens and antigens from other microorganisms. Around 50% of healthy adults show low-level serum IgG reactivity against the FlaB antigen, leading to false-positive results as well. In cases in which the Western blot result is indeterminate, other etiologies must be considered.

False-positive IgM Western blots are a significant problem. In a 5-year retrospective study done at 63 US Air Force healthcare facilities, 113 (53.3%) of 212 IgM Western blots were falsely positive.¹⁹ A false-positive test was defined as one that failed to meet seropositivity (a first-tier test omitted or negative, > 30 days of symptoms with negative IgG blot), lack of exposure including residing in areas without documented tick habitats, patients having atypical or no symptoms, and negative serology within 30 days of a positive test.

In a similar study done in a highly endemic area, 50 (27.5%) of 182 patients had a false-positive test.²⁰ Physicians need to be careful when interpreting IgM Western blots. It is always important to consider locale, epidemiology, and symptoms when interpreting the test.

Limitations of serologic tests for Lyme disease

Currently available serologic tests have inherent limitations:

• Antibodies against B burgdorferi take at least 1 week to develop
• The background rate of seropositivity in endemic areas can be up to 4%, affecting the utility of a positive test result
• Serologic tests cannot be used as tests of cure because antibodies can persist for months to years even after appropriate antimicrobial therapy and cure of disease; thus, a positive serologic result could represent active infection or remote exposure²¹
• Antibodies can cross-react with related bacteria, including other borrelial or treponemal spirochetes
• False-positive serologic test results can also occur in association with other medical conditions such as polyclonal gammopathies and systemic lupus erythematosus.¹²

RECOMMENDATIONS FOR TESTING

Standard 2-tier testing

The CDC released recommendations for diagnosing Lyme disease after a second national conference of serologic diagnosis of Lyme disease in October 1994.¹⁸ The 2-tiered testing method, involving a sensitive ELISA followed by the Western blot to confirm positive and indeterminate ELISA results, was suggested as the gold standard for diagnosis (Figure 3). Of note, negative ELISA results do not require further testing.¹¹

The sensitivity of 2-tiered testing depends on the stage of the disease. Unfortunately, this method has a wide range of sensitivity (17% to 78%) in stage 1 disease. In the same stage, the sensitivity increases from 14.1% in patients with a single erythema migrans lesion and early localized infection to 65.4% in those with multiple lesions. The algorithm has excellent sensitivity in late stage 3 infection (96% to 100%).⁵

A 2-step ELISA algorithm

A 2-step ELISA algorithm (without the Western blot) that includes the whole-cell sonicate assay followed by the VisE C6 peptide assay actually showed higher sensitivity and comparable specificity compared with 2-tiered testing in early localized disease (sensitivity 61%–74% vs 29%–48%, respectively; specificity 99.5% for both methods).²² This higher sensitivity was even more pronounced in early disseminated infection (sensitivity 100% vs 40%, respectively). By late infection, the sensitivities of both testing strategies reached 100%. Compared with the Western blot, the 2-step ELISA algorithm was simpler to execute in a reproducible fashion.⁵

The Infectious Diseases Society of America is revising its current guidelines, with an update expected late this year, which may shift the recommendation from 2-tiered testing to the 2-step ELISA algorithm.

Multiplex testing

To overcome the intrinsic problems of protein-based assays, a multiplexed, array-based assay for the diagnosis of tick-borne infections called Tick-Borne Disease Serochip (TBD-Serochip) was established using recombinant antigens that identify key immunodominant epitopes.⁸ More studies are needed to establish the validity and usefulness of these tests in clinical practice.

Who should not be tested?

The American College of Physicians⁶ recommends against testing in patients:

• Presenting with nonspecific symptoms (eg, headache, myalgia, fatigue, arthralgia) without objective signs of Lyme disease
• With low pretest probability of infection based on epidemiologic exposures and clinical features
• Living in Lyme-endemic areas with no history of tick exposure⁶
• Presenting less than 1 week after tick exposure⁵
• Seeking a test of cure for treated Lyme disease.

DIAGNOSIS IN SPECIAL SITUATIONS

Early Lyme disease

The classic erythema migrans lesion on physical examination of a patient with suspected Lyme disease is diagnostic and does not require laboratory confirmation.¹⁰

In ambiguous cases, 2-tiered testing of a serum sample during the acute presentation and again 4 to 6 weeks later can be useful. In patients who remain seronegative on paired serum samples despite symptoms lasting longer than 6 weeks and no antibiotic treatment in the interim, the diagnosis of Lyme disease is unlikely, and another diagnosis should be sought.³

Antimicrobial therapy may block the serologic response; hence, negative serologic testing in patients started on empiric antibiotics should not rule out Lyme disease.⁶

PCR or bacterial culture testing is not recommended in the evaluation of suspected early Lyme disease.

Central nervous system Lyme disease

Central nervous system Lyme disease is diagnosed by 2-tiered testing using peripheral blood samples because all patients with this infectious manifestation should have mounted an adequate IgG response in the blood.¹¹

B cells migrate to and proliferate inside the central nervous system, leading to intrathecal production of anti-Borrelia antibodies. An index of cerebrospinal fluid to serum antibody greater than 1 is thus also indicative of neuroborreliosis.¹² Thus, performing lumbar puncture to detect intrathecal production of antibodies may support the diagnosis of central nervous system Lyme disease; however, it is not necessary.¹¹

Antibodies persist in the central nervous system for many years after appropriate antimicrobial treatment.

Lyme arthritis

Articular involvement in Lyme disease is characterized by a robust humoral response such that a negative IgG serologic test virtually rules out Lyme arthritis.²³ PCR testing of synovial fluid for borrelial DNA has a sensitivity of 80% but may become falsely negative after 1 to 2 months of antibiotic treatment.^24,25 In an algorithm suggested by Puius et al,²³ PCR testing of synovial fluid should be done in patients who have minimal to no response after 2 months of appropriate oral antimicrobial therapy to determine whether intravenous antibiotics are merited.

Table 5 summarizes the tests of choice in different clinical stages of infection.

Acknowledgment: The authors would like to acknowledge Anita Modi, MD, and Ceena N. Jacob, MD, for reviewing the manuscript and providing valuable suggestions, and Belinda Yen-Lieberman, PhD, for contributing pictures of the Western blot test results.

Lyme disease is a complex multisystem bacterial infection affecting the skin, joints, heart, and nervous system. The full spectrum of disease was first recognized and the disease was named in the 1970s during an outbreak of arthritis in children in the town of Lyme, Connecticut.¹

This review describes the epidemiology and pathogenesis of Lyme disease, the advantages and disadvantages of current diagnostic methods, and diagnostic algorithms.

THE MOST COMMON TICK-BORNE INFECTION IN NORTH AMERICA

Lyme disease is the most common tick-borne infection in North America.^2,3 In the United States, more than 30,000 cases are reported annually. In fact, in 2017, the number of cases was about 42,000, a 16% increase from the previous year, according to the US Centers for Disease Control and Prevention (CDC).

From Sigal LH. Myths and facts about Lyme disease. Cleve Clin J Med 1997; 64(4):203–209.

The infection is caused by Borrelia burgdorferi, a particularly arthritogenic spirochete transmitted by Ixodes scapularis (the black-legged deer tick, (Figure 1) and Ixodes pacificus (the Western black-legged tick). Although the infection can occur at any time of the year, its peak incidence is in May to late September, coinciding with increased outdoor recreational activity in areas where ticks live.^3,4 The typical tick habitat consists of deciduous woodland with sufficient humidity provided by a good layer of decaying vegetation. However, people can contract Lyme disease in their own backyard.³

Most cases of Lyme disease are seen in the northeastern United States, mainly in suburban and rural areas.^2,3 Other areas affected include the midwestern states of Minnesota, Wisconsin, and Michigan, as well as northern California.⁴ Fourteen states and the District of Columbia report a high average incidence (> 10 cases per 100,000 persons) (Table 1).²

FIRST COMES IgM, THEN IgG

The pathogenesis and the different stages of infection should inform laboratory testing in Lyme disease.

It is estimated that only 5% of infected ticks that bite people actually transmit their spirochetes to the human host.⁵ However, once infected, the patient’s innate immune system mounts a response that results in the classic erythema migrans rash at the bite site. A rash develops in only about 85% of patients who are infected and can appear at any time between 3 and 30 days, but most commonly after 7 days. Hence, a rash occurring within the first few hours of tick contact is not erythema migrans and does not indicate infection, but rather an early reaction to tick salivary antigens.⁵

Antibody levels remain below the detection limits of currently available serologic tests in the first 7 days after exposure. Immunoglobulin M (IgM) antibody titers peak between 8 and 14 days after tick contact, but IgM antibodies may never develop if the patient is started on early appropriate antimicrobial therapy.⁵

If the infection is not treated, the spirochete may disseminate through the blood from the bite site to different tissues.³ Both cell-mediated and antibody-mediated immunity swing into action to kill the spirochetes at this stage. The IgM antibody response occurs in 1 to 2 weeks, followed by a robust IgG response in 2 to 4 weeks.⁶

Because IgM can also cross-react with antigens other than those associated with B burgdorferi, the IgM test is less specific than the IgG test for Lyme disease.

Once a patient is exposed and mounts an antibody-mediated response to the spirochete, the antibody profile may persist for months to years, even after successful antibiotic treatment and cure of the disease.⁵

Despite the immune system’s robust series of defenses, untreated B burgdorferi infection can persist, as the organism has a bag of tricks to evade destruction. It can decrease its expression of specific immunogenic surface-exposed proteins, change its antigenic properties through recombination, and bind to the patient’s extracellular matrix proteins to facilitate further dissemination.³

Certain host-genetic factors also play a role in the pathogenesis of Lyme disease, such as the HLA-DR4 allele, which has been associated with antibiotic-refractory Lyme-related arthritis.³

LYME DISEASE EVOLVES THROUGH STAGES

Lyme disease evolves through stages broadly classified as early and late infection, with significant variability in its presentation.⁷

Early infection

Early disease is further subdivided into “localized” infection (stage 1), characterized by a single erythema migrans lesion and local lymphadenopathy, and “disseminated” infection (stage 2), associated with multiple erythema migrans lesions distant from the bite site, facial nerve palsy, radiculoneuritis, meningitis, carditis, or migratory arthritis or arthralgia.⁸

Highly specific physical findings include erythema migrans, cranial nerve palsy, high-grade or progressive conduction block, and recurrent migratory polyarthritis. Less specific symptoms and signs of Lyme disease include arthralgia, myalgia, neck stiffness, palpitations, and myocarditis.⁵

Erythema migrans lesions are evident in at least 85% of patients with early disease.⁹ If they are not apparent on physical examination, they may be located at hidden sites and may be atypical in appearance or transient.⁵

If treatment is not started in the initial stage of the disease, 60% of infected patients may develop disseminated infection.⁵ Progressive, untreated infection can manifest with Lyme arthritis and neuroborreliosis.⁷

Noncutaneous manifestations are less common now than in the past due to increased awareness of the disease and early initiation of treatment.¹⁰

Late infection

Manifestations of late (stage 3) infection include oligoarthritis (affecting any joint but often the knee) and neuroborreliosis. Clinical signs and symptoms of Lyme disease may take months to resolve even after appropriate antimicrobial therapy is completed. This should not be interpreted as ongoing, persistent infection, but as related to host immune-mediated activity.⁵

The usefulness of a laboratory test depends on the individual patient’s pretest probability of infection, which in turn depends on the patient’s epidemiologic risk of exposure and clinical features of Lyme disease. Patients with a high pretest probability—eg, a history of a tick bite followed by the classic erythema migrans rash—do not need testing and can start antimicrobial therapy right away.¹¹

Serologic tests are the gold standard

Prompt diagnosis is important, as early Lyme disease is easily treatable without any future sequelae.¹¹

Tests for Lyme disease can be divided into direct methods, which detect the spirochete itself by culture or by polymerase chain reaction (PCR), and indirect methods, which detect antibodies (Table 2). Direct tests lack sensitivity for Lyme disease; hence, serologic tests remain the gold standard. Currently recommended is a standard 2-tier testing strategy using an enzyme-linked immunosorbent assay (ELISA) followed by Western blot for confirmation.

DIRECT METHODS

Culture lacks sensitivity

A number of factors limit the sensitivity of direct culture for diagnosing Lyme disease. B burgdorferi does not grow easily in culture, requiring special media, low temperatures, and long periods of incubation. Only a relatively few spirochetes are present in human tissues and body fluids to begin with, and bacterial counts are further reduced with duration and dissemination of infection.⁵ All of these limit the possibility of detecting this organism.

Polymerase chain reaction may help in some situations

Molecular assays are not part of the standard evaluation and should be used only in conjunction with serologic testing.⁷ These tests have high specificity but lack consistent sensitivity.

That said, PCR testing may be useful:

• In early infection, before antibody responses develop
• In reinfection, when serologic tests are not reliable because the antibodies persist for many years after an infection in many patients
• In endemic areas where serologic testing has high false-positive rates due to high baseline population seropositivity for anti-Borrelia antibodies caused by subclinical infection.³

PCR assays that target plasmid-borne genes encoding outer surface proteins A and C (OspA and OspC) and VisE (variable major protein-like sequence, expressed) are more sensitive than those that detect chromosomal 16s ribosomal ribonucleic acid (rRNA) genes, as plasmid-rich “blebs” are shed in larger concentrations than chromosomal DNA during active infection.⁷ However, these plasmid-contained genes persist in body tissues and fluids even after the infection is cleared, and their detection may not necessarily correlate with ongoing disease.⁸ Detection of chromosomal 16s rRNA genes is a better predictor of true organism viability.

The sensitivity of PCR for borrelial DNA depends on the type of sample. If a skin biopsy sample is taken of the leading edge of an erythema migrans lesion, the sensitivity is 69% and the specificity is 100%. In patients with Lyme arthritis, PCR of the synovial fluid has a sensitivity of up to 80%. However, the sensitivity of PCR of the cerebrospinal fluid of patients with neurologic manifestations of Lyme disease is only 19%.⁷ PCR of other clinical samples, including blood and urine, is not recommended, as spirochetes are primarily confined to tissues, and very few are present in these body fluids.^3,12

The disadvantage of PCR is that a positive result does not always mean active infection, as the DNA of the dead microbe persists for several months even after successful treatment.⁸

INDIRECT METHODS

Enzyme-linked immunosorbent assay

ELISAs detect anti-Borrelia antibodies. Early-generation ELISAs, still used in many laboratories, use whole-cell extracts of B burgdorferi. Examples are the Vidas Lyme screen (Biomérieux, biomerieux-usa.com) and the Wampole B burgdorferi IgG/M EIA II assay (Alere, www.alere.com). Newer ELISAs use recombinant proteins.¹³

Three major targets for ELISA antibodies are flagellin (Fla), outer surface protein C (OspC), and VisE, especially the invariable region 6 (IR6). Among these, VisE-IR6 is the most conserved region in B burgdorferi.

Early-generation assays have a sensitivity of 89% and specificity of 72%.¹¹ However, the patient’s serum may have antibodies that cross-react with unrelated bacterial antigens, leading to false-positive results (Table 3). Whole-cell sonicate assays are not recommended as an independent test and must be confirmed with Western blot testing when assay results are indeterminate or positive.¹¹

Newer-generation ELISAs detect antibodies targeting recombinant proteins of VisE, especially a synthetic peptide C6, within IR6.¹³ VisE-IR6 is the most conserved region of the B burgdorferi complex, and its detection is a highly specific finding, supporting the diagnosis of Lyme disease. Antibodies against VisE-IR6 antigen are the earliest to develop.⁵ An example of a newer-generation serologic test is the VisE C6 Lyme EIA kit, approved as a first-tier test by the US Food and Drug Administration in 2001. This test has a specificity of 99%,^14,15 and its specificity is further increased when used in conjunction with Western blot (99.5%).¹⁵ The advantage of the C6 antibody test is that it is more sensitive than 2-tier testing during early infection (sensitivity 29%–74% vs 17%–40% in early localized infection, and 56%–90% vs 27%–78% in early disseminated infection).⁶

During early infection, older and newer ELISAs are less sensitive because of the limited number of antigens expressed at this stage.¹³ All patients suspected of having early Lyme disease who are seronegative at initial testing should have follow-up testing to look for seroconversion.¹³

Western blot

Western blot (immunoblot) testing identifies IgM and IgG antibodies against specific B burgdorferi antigens. It is considered positive if it detects at least 2 of a possible 3 specific IgM bands in the first 4 weeks of disease or at least 5 of 10 specific IgG bands after 4 weeks of disease (Table 4 and Figure 2).¹⁶

The nature of the bands indicates the duration of infection: Western blot bands against 23-kD OspC and 41-kD FlaB are seen in early localized infection, whereas bands against all 3 B burgdorferi proteins will be seen after several weeks of disease.¹⁷ The IgM result should be interpreted carefully, as only 2 bands are required for the test to be positive, and IgM binds to antigen less specifically than IgG.¹²

If clinical symptoms and signs of Lyme disease have been present for more than 1 month, IgM reactivity alone should not be used to support the diagnosis, in view of the likelihood of a false-positive test result in this situation.¹⁸ This is called the “1-month rule” in the diagnosis of Lyme disease.¹³

In early localized infection, Western blot is only half as sensitive as ELISA testing. Since the overall sensitivity of a 2-step algorithm is equal to that of its least sensitive component, 2-tiered testing is not useful in early disease.¹³

Although currently considered the most specific test for confirmation of Lyme disease, Western blot has limitations. It is technically and interpretively complex and is thus not universally available.¹³ The blots are scored by visual examination, compromising the reproducibility of the test, although densitometric blot analysis techniques and automated scanning and scoring attempt to address some of these limitations.¹³ Like the ELISA, Western blot can have false-positive results in healthy individuals without tick exposure, as nonspecific IgM immunoblots develop faint bands. This is because of cross-reaction between B burgdorferi antigens and antigens from other microorganisms. Around 50% of healthy adults show low-level serum IgG reactivity against the FlaB antigen, leading to false-positive results as well. In cases in which the Western blot result is indeterminate, other etiologies must be considered.

False-positive IgM Western blots are a significant problem. In a 5-year retrospective study done at 63 US Air Force healthcare facilities, 113 (53.3%) of 212 IgM Western blots were falsely positive.¹⁹ A false-positive test was defined as one that failed to meet seropositivity (a first-tier test omitted or negative, > 30 days of symptoms with negative IgG blot), lack of exposure including residing in areas without documented tick habitats, patients having atypical or no symptoms, and negative serology within 30 days of a positive test.

In a similar study done in a highly endemic area, 50 (27.5%) of 182 patients had a false-positive test.²⁰ Physicians need to be careful when interpreting IgM Western blots. It is always important to consider locale, epidemiology, and symptoms when interpreting the test.

Limitations of serologic tests for Lyme disease

Currently available serologic tests have inherent limitations:

• Antibodies against B burgdorferi take at least 1 week to develop
• The background rate of seropositivity in endemic areas can be up to 4%, affecting the utility of a positive test result
• Serologic tests cannot be used as tests of cure because antibodies can persist for months to years even after appropriate antimicrobial therapy and cure of disease; thus, a positive serologic result could represent active infection or remote exposure²¹
• Antibodies can cross-react with related bacteria, including other borrelial or treponemal spirochetes
• False-positive serologic test results can also occur in association with other medical conditions such as polyclonal gammopathies and systemic lupus erythematosus.¹²

RECOMMENDATIONS FOR TESTING

Standard 2-tier testing

The CDC released recommendations for diagnosing Lyme disease after a second national conference of serologic diagnosis of Lyme disease in October 1994.¹⁸ The 2-tiered testing method, involving a sensitive ELISA followed by the Western blot to confirm positive and indeterminate ELISA results, was suggested as the gold standard for diagnosis (Figure 3). Of note, negative ELISA results do not require further testing.¹¹

The sensitivity of 2-tiered testing depends on the stage of the disease. Unfortunately, this method has a wide range of sensitivity (17% to 78%) in stage 1 disease. In the same stage, the sensitivity increases from 14.1% in patients with a single erythema migrans lesion and early localized infection to 65.4% in those with multiple lesions. The algorithm has excellent sensitivity in late stage 3 infection (96% to 100%).⁵

A 2-step ELISA algorithm

A 2-step ELISA algorithm (without the Western blot) that includes the whole-cell sonicate assay followed by the VisE C6 peptide assay actually showed higher sensitivity and comparable specificity compared with 2-tiered testing in early localized disease (sensitivity 61%–74% vs 29%–48%, respectively; specificity 99.5% for both methods).²² This higher sensitivity was even more pronounced in early disseminated infection (sensitivity 100% vs 40%, respectively). By late infection, the sensitivities of both testing strategies reached 100%. Compared with the Western blot, the 2-step ELISA algorithm was simpler to execute in a reproducible fashion.⁵

The Infectious Diseases Society of America is revising its current guidelines, with an update expected late this year, which may shift the recommendation from 2-tiered testing to the 2-step ELISA algorithm.

Multiplex testing

To overcome the intrinsic problems of protein-based assays, a multiplexed, array-based assay for the diagnosis of tick-borne infections called Tick-Borne Disease Serochip (TBD-Serochip) was established using recombinant antigens that identify key immunodominant epitopes.⁸ More studies are needed to establish the validity and usefulness of these tests in clinical practice.

Who should not be tested?

The American College of Physicians⁶ recommends against testing in patients:

• Presenting with nonspecific symptoms (eg, headache, myalgia, fatigue, arthralgia) without objective signs of Lyme disease
• With low pretest probability of infection based on epidemiologic exposures and clinical features
• Living in Lyme-endemic areas with no history of tick exposure⁶
• Presenting less than 1 week after tick exposure⁵
• Seeking a test of cure for treated Lyme disease.

DIAGNOSIS IN SPECIAL SITUATIONS

Early Lyme disease

The classic erythema migrans lesion on physical examination of a patient with suspected Lyme disease is diagnostic and does not require laboratory confirmation.¹⁰

In ambiguous cases, 2-tiered testing of a serum sample during the acute presentation and again 4 to 6 weeks later can be useful. In patients who remain seronegative on paired serum samples despite symptoms lasting longer than 6 weeks and no antibiotic treatment in the interim, the diagnosis of Lyme disease is unlikely, and another diagnosis should be sought.³

Antimicrobial therapy may block the serologic response; hence, negative serologic testing in patients started on empiric antibiotics should not rule out Lyme disease.⁶

PCR or bacterial culture testing is not recommended in the evaluation of suspected early Lyme disease.

Central nervous system Lyme disease

Central nervous system Lyme disease is diagnosed by 2-tiered testing using peripheral blood samples because all patients with this infectious manifestation should have mounted an adequate IgG response in the blood.¹¹

B cells migrate to and proliferate inside the central nervous system, leading to intrathecal production of anti-Borrelia antibodies. An index of cerebrospinal fluid to serum antibody greater than 1 is thus also indicative of neuroborreliosis.¹² Thus, performing lumbar puncture to detect intrathecal production of antibodies may support the diagnosis of central nervous system Lyme disease; however, it is not necessary.¹¹

Antibodies persist in the central nervous system for many years after appropriate antimicrobial treatment.

Lyme arthritis

Articular involvement in Lyme disease is characterized by a robust humoral response such that a negative IgG serologic test virtually rules out Lyme arthritis.²³ PCR testing of synovial fluid for borrelial DNA has a sensitivity of 80% but may become falsely negative after 1 to 2 months of antibiotic treatment.^24,25 In an algorithm suggested by Puius et al,²³ PCR testing of synovial fluid should be done in patients who have minimal to no response after 2 months of appropriate oral antimicrobial therapy to determine whether intravenous antibiotics are merited.

Table 5 summarizes the tests of choice in different clinical stages of infection.

Acknowledgment: The authors would like to acknowledge Anita Modi, MD, and Ceena N. Jacob, MD, for reviewing the manuscript and providing valuable suggestions, and Belinda Yen-Lieberman, PhD, for contributing pictures of the Western blot test results.

Is the relationship between A-fib and sleep apnea more than a coincidence stemming from the number of shared associated comorbidities? Significantly, the treatment of obstructive sleep apnea with continuous positive airway pressure (CPAP) has been shown to decrease the recurrence of A-fib after pharmacologic or electrical conversion and after interventional pulmonary vein interruption.¹ This suggests that at least in some cases, sleep apnea plays an active role in initiating and possibly also maintaining A-fib. The immediate culprit mediators that come to mind are hypoxia and hypercapnea; both are at least partially ameliorated by the successful use of CPAP, and both are reasonable physiologic candidates for induction of A-fib. Hypoxia is supported by clinical observation, and hypercapnea by experimental modeling.²

It is easy for clinicians to conceptualize the organ effects of hypoxia and hypercapnea. We are accustomed to seeing clinical ramifications of these in the emergency department and intensive care unit, particularly those affecting the brain and heart, organs critically dependent on transmembrane ion flow. We may recall from biochemistry classes the effects of hypoxia on intracellular metabolism and the implications on energy stores, mitochondrial function, and ion translocation. Recent work on the cellular effects of hypoxia, including research that resulted in a Nobel prize, has drawn major attention to patterned cellular responses to intermittent and persistent hypoxia. This includes recognition of epigenetic changes resulting in localized cardiac remodeling and fibrosis,³ factors that clearly affect the expression of arrhythmias, including A-fib.

But the interrelationship between A-fib and sleep apnea may be even more convoluted and intriguing. It now seems that most things cardiac are associated with inflammation in some guise, and the A-fib connection with sleep apnea may not be an exception. Almost 20 years ago, it was recognized that A-fib is associated with an elevation in circulating C-reactive protein (CRP),⁴ a biomarker of “inflammation,” although not necessarily an active participant. Recent reviews of this connection have been published,⁵ and successful anti-inflammatory approaches to preventing A-fib using colchicine have been described.⁶ So how does this tie in with sleep apnea?

A number of papers have now demonstrated that sleep apnea is also associated with an elevation in CRP,⁷ perhaps due to increases in tumor necrosis factor (TNF)-alpha in response to the intermittent hypoxia of sleep apnea. TNF can drive the inflammatory response through increased expression of genes regulated by nuclear factor kappa-B.⁸ While it certainly warrants consideration that the elevated biomarkers of inflammation in patients with sleep apnea actually reflect the presence of the frequent comorbidities, including visceral obesity, treating sleep apnea with CPAP (comparable to what I noted above in patients with A-fib) has been shown to reduce circulating CRP levels.⁹

As our understanding of the biologic underpinnings of A-fib and sleep apnea continue to grow, the practical clinical implications of the relationship between them, as described by Ayache et al, may achieve greater clarity. The two conditions commonly coexist, and treating the sleep apnea results in better rhythm-directed outcomes in the A-fib.

Stay tuned, there is certainly more to learn about this.

Is the relationship between A-fib and sleep apnea more than a coincidence stemming from the number of shared associated comorbidities? Significantly, the treatment of obstructive sleep apnea with continuous positive airway pressure (CPAP) has been shown to decrease the recurrence of A-fib after pharmacologic or electrical conversion and after interventional pulmonary vein interruption.¹ This suggests that at least in some cases, sleep apnea plays an active role in initiating and possibly also maintaining A-fib. The immediate culprit mediators that come to mind are hypoxia and hypercapnea; both are at least partially ameliorated by the successful use of CPAP, and both are reasonable physiologic candidates for induction of A-fib. Hypoxia is supported by clinical observation, and hypercapnea by experimental modeling.²

It is easy for clinicians to conceptualize the organ effects of hypoxia and hypercapnea. We are accustomed to seeing clinical ramifications of these in the emergency department and intensive care unit, particularly those affecting the brain and heart, organs critically dependent on transmembrane ion flow. We may recall from biochemistry classes the effects of hypoxia on intracellular metabolism and the implications on energy stores, mitochondrial function, and ion translocation. Recent work on the cellular effects of hypoxia, including research that resulted in a Nobel prize, has drawn major attention to patterned cellular responses to intermittent and persistent hypoxia. This includes recognition of epigenetic changes resulting in localized cardiac remodeling and fibrosis,³ factors that clearly affect the expression of arrhythmias, including A-fib.

But the interrelationship between A-fib and sleep apnea may be even more convoluted and intriguing. It now seems that most things cardiac are associated with inflammation in some guise, and the A-fib connection with sleep apnea may not be an exception. Almost 20 years ago, it was recognized that A-fib is associated with an elevation in circulating C-reactive protein (CRP),⁴ a biomarker of “inflammation,” although not necessarily an active participant. Recent reviews of this connection have been published,⁵ and successful anti-inflammatory approaches to preventing A-fib using colchicine have been described.⁶ So how does this tie in with sleep apnea?

A number of papers have now demonstrated that sleep apnea is also associated with an elevation in CRP,⁷ perhaps due to increases in tumor necrosis factor (TNF)-alpha in response to the intermittent hypoxia of sleep apnea. TNF can drive the inflammatory response through increased expression of genes regulated by nuclear factor kappa-B.⁸ While it certainly warrants consideration that the elevated biomarkers of inflammation in patients with sleep apnea actually reflect the presence of the frequent comorbidities, including visceral obesity, treating sleep apnea with CPAP (comparable to what I noted above in patients with A-fib) has been shown to reduce circulating CRP levels.⁹

As our understanding of the biologic underpinnings of A-fib and sleep apnea continue to grow, the practical clinical implications of the relationship between them, as described by Ayache et al, may achieve greater clarity. The two conditions commonly coexist, and treating the sleep apnea results in better rhythm-directed outcomes in the A-fib.

Stay tuned, there is certainly more to learn about this.

Is the relationship between A-fib and sleep apnea more than a coincidence stemming from the number of shared associated comorbidities? Significantly, the treatment of obstructive sleep apnea with continuous positive airway pressure (CPAP) has been shown to decrease the recurrence of A-fib after pharmacologic or electrical conversion and after interventional pulmonary vein interruption.¹ This suggests that at least in some cases, sleep apnea plays an active role in initiating and possibly also maintaining A-fib. The immediate culprit mediators that come to mind are hypoxia and hypercapnea; both are at least partially ameliorated by the successful use of CPAP, and both are reasonable physiologic candidates for induction of A-fib. Hypoxia is supported by clinical observation, and hypercapnea by experimental modeling.²

It is easy for clinicians to conceptualize the organ effects of hypoxia and hypercapnea. We are accustomed to seeing clinical ramifications of these in the emergency department and intensive care unit, particularly those affecting the brain and heart, organs critically dependent on transmembrane ion flow. We may recall from biochemistry classes the effects of hypoxia on intracellular metabolism and the implications on energy stores, mitochondrial function, and ion translocation. Recent work on the cellular effects of hypoxia, including research that resulted in a Nobel prize, has drawn major attention to patterned cellular responses to intermittent and persistent hypoxia. This includes recognition of epigenetic changes resulting in localized cardiac remodeling and fibrosis,³ factors that clearly affect the expression of arrhythmias, including A-fib.

But the interrelationship between A-fib and sleep apnea may be even more convoluted and intriguing. It now seems that most things cardiac are associated with inflammation in some guise, and the A-fib connection with sleep apnea may not be an exception. Almost 20 years ago, it was recognized that A-fib is associated with an elevation in circulating C-reactive protein (CRP),⁴ a biomarker of “inflammation,” although not necessarily an active participant. Recent reviews of this connection have been published,⁵ and successful anti-inflammatory approaches to preventing A-fib using colchicine have been described.⁶ So how does this tie in with sleep apnea?

A number of papers have now demonstrated that sleep apnea is also associated with an elevation in CRP,⁷ perhaps due to increases in tumor necrosis factor (TNF)-alpha in response to the intermittent hypoxia of sleep apnea. TNF can drive the inflammatory response through increased expression of genes regulated by nuclear factor kappa-B.⁸ While it certainly warrants consideration that the elevated biomarkers of inflammation in patients with sleep apnea actually reflect the presence of the frequent comorbidities, including visceral obesity, treating sleep apnea with CPAP (comparable to what I noted above in patients with A-fib) has been shown to reduce circulating CRP levels.⁹

As our understanding of the biologic underpinnings of A-fib and sleep apnea continue to grow, the practical clinical implications of the relationship between them, as described by Ayache et al, may achieve greater clarity. The two conditions commonly coexist, and treating the sleep apnea results in better rhythm-directed outcomes in the A-fib.

Stay tuned, there is certainly more to learn about this.

A 43-year-old man presented with a 3-week history of halitosis. He was also concerned about the irregular appearance of his tongue, which he had noticed over the past 3 years. He had no history of wearing dentures or of any skin disorder.

A BROAD DIFFERENTIAL DIAGNOSIS

Fissured tongue (scrotal tongue, plicated tongue, lingua plicata) is a common normal variant of the tongue surface with a male preponderance and a reported prevalence of 10% to 20% in the general population, and the incidence increases strikingly with age.¹

The cause is not known, but familial clustering is seen, and a polygenic or autosomal dominant hereditary component is presumed.¹

The condition may be associated with removable dentures, geographic tongue, pernicious anemia, Sjögren syndrome, psoriasis, acromegaly, macroglossia, oral-facial-digital syndrome type I, Pierre Robin syndrome, Down syndrome, and Melkersson Rosenthal syndrome.² It is usually asymptomatic, but if the fissures are deep, food may become lodged in them, resulting in tongue inflammation, burning sensation, and halitosis.¹

Typically, fissures of varying depth extending to the margin are apparent on the dorsal surface of the tongue. The condition is confined to the anterior two-thirds of the tongue, which is of ectodermal origin. Histologically, the epithelium, lamina propria, and musculature are all involved in the formation of the fissures.³ The deeper fissures may lack filliform papillae due to bacterial inflammation.³ The diagnosis is clinical, and treatment includes reassurance, advice on good oral hygiene, and tongue cleansing.¹

A 43-year-old man presented with a 3-week history of halitosis. He was also concerned about the irregular appearance of his tongue, which he had noticed over the past 3 years. He had no history of wearing dentures or of any skin disorder.

A BROAD DIFFERENTIAL DIAGNOSIS

Fissured tongue (scrotal tongue, plicated tongue, lingua plicata) is a common normal variant of the tongue surface with a male preponderance and a reported prevalence of 10% to 20% in the general population, and the incidence increases strikingly with age.¹

The cause is not known, but familial clustering is seen, and a polygenic or autosomal dominant hereditary component is presumed.¹

The condition may be associated with removable dentures, geographic tongue, pernicious anemia, Sjögren syndrome, psoriasis, acromegaly, macroglossia, oral-facial-digital syndrome type I, Pierre Robin syndrome, Down syndrome, and Melkersson Rosenthal syndrome.² It is usually asymptomatic, but if the fissures are deep, food may become lodged in them, resulting in tongue inflammation, burning sensation, and halitosis.¹

Typically, fissures of varying depth extending to the margin are apparent on the dorsal surface of the tongue. The condition is confined to the anterior two-thirds of the tongue, which is of ectodermal origin. Histologically, the epithelium, lamina propria, and musculature are all involved in the formation of the fissures.³ The deeper fissures may lack filliform papillae due to bacterial inflammation.³ The diagnosis is clinical, and treatment includes reassurance, advice on good oral hygiene, and tongue cleansing.¹

A 43-year-old man presented with a 3-week history of halitosis. He was also concerned about the irregular appearance of his tongue, which he had noticed over the past 3 years. He had no history of wearing dentures or of any skin disorder.

A BROAD DIFFERENTIAL DIAGNOSIS

Fissured tongue (scrotal tongue, plicated tongue, lingua plicata) is a common normal variant of the tongue surface with a male preponderance and a reported prevalence of 10% to 20% in the general population, and the incidence increases strikingly with age.¹

The cause is not known, but familial clustering is seen, and a polygenic or autosomal dominant hereditary component is presumed.¹

The condition may be associated with removable dentures, geographic tongue, pernicious anemia, Sjögren syndrome, psoriasis, acromegaly, macroglossia, oral-facial-digital syndrome type I, Pierre Robin syndrome, Down syndrome, and Melkersson Rosenthal syndrome.² It is usually asymptomatic, but if the fissures are deep, food may become lodged in them, resulting in tongue inflammation, burning sensation, and halitosis.¹

Typically, fissures of varying depth extending to the margin are apparent on the dorsal surface of the tongue. The condition is confined to the anterior two-thirds of the tongue, which is of ectodermal origin. Histologically, the epithelium, lamina propria, and musculature are all involved in the formation of the fissures.³ The deeper fissures may lack filliform papillae due to bacterial inflammation.³ The diagnosis is clinical, and treatment includes reassurance, advice on good oral hygiene, and tongue cleansing.¹

A 50-year-old man with acute myeloid leukemia (AML) with a complex karyotype was admitted to the hospital with several days of dull, left-sided abdominal pain. His most recent bone marrow biopsy showed 30% blasts, and immunophenotyping was suggestive of persistent AML (CD13+, CD34+, CD117+, CD33+, CD7+, MPO–). He was on treatment with venetoclax and cytarabine after induction therapy had failed.

On admission, his heart rate was 101 beats per minute and his blood pressure was 122/85 mm Hg. Abdominal examination revealed mild distention, hepatomegaly, and previously known massive splenomegaly, with the splenic tip extending to the umbilicus, and mild tenderness.

Results of laboratory testing revealed persistent pancytopenia:

• Hemoglobin level 6.8 g/dL (reference range 13.0–17.0)
• Total white blood cell count 0.8 × 10⁹/L (4.5–11.0)
• Platelet count 8 × 10⁹/L (150–400).

The next day, he developed severe, acute-onset left-sided abdominal pain. A check of vital signs showed worsening sinus tachycardia at 132 beats per minute and a drop in blood pressure to 90/56 mm Hg. He had worsening diffuse abdominal tenderness with sluggish bowel sounds. His hemoglobin concentration was 6.4 g/dL and platelet count 12 × 10⁹/L.

He received supportive transfusions of blood products. Surgical exploration was deemed risky, given his overall condition and severe thrombocytopenia. Splenic angiography showed no evidence of pseudoaneurysm or focal contrast extravasation. He underwent empiric embolization of the midsplenic artery, after which his hemodynamic status stabilized. He died 4 weeks later of acute respiratory failure from pneumonia.

SPLENIC RUPTURE IN AML

Atraumatic splenic rupture is rare but potentially life-threatening, especially if the diagnosis is delayed. Conditions that can cause splenomegaly and predispose to rupture include infection (infectious mononucleosis, malaria), malignant hematologic disorders (leukemia, lymphoma), other neoplasms, and amyloidosis.¹

The literature includes a few reports of splenic rupture in patients with AML.^2–4 The proposed mechanisms include bleeding from infarction sites or tumor foci, dysregulated hemostasis, and leukostasis.

The classic presentation of splenic rupture is acute-onset left-sided abdominal pain associated with hypotension and decreasing hemoglobin levels. CT of the abdomen is confirmatory, and resuscitation with crystalloids and blood products is a vital initial step in management. Choice of treatment depends on the patient’s surgical risk and hemodynamic status; options include conservative medical management, splenic artery embolization, and exploratory laparotomy.

In patients with AML and splenomegaly presenting with acute abdominal pain, clinicians need to be aware of this potential hematologic emergency.

A 50-year-old man with acute myeloid leukemia (AML) with a complex karyotype was admitted to the hospital with several days of dull, left-sided abdominal pain. His most recent bone marrow biopsy showed 30% blasts, and immunophenotyping was suggestive of persistent AML (CD13+, CD34+, CD117+, CD33+, CD7+, MPO–). He was on treatment with venetoclax and cytarabine after induction therapy had failed.

On admission, his heart rate was 101 beats per minute and his blood pressure was 122/85 mm Hg. Abdominal examination revealed mild distention, hepatomegaly, and previously known massive splenomegaly, with the splenic tip extending to the umbilicus, and mild tenderness.

Results of laboratory testing revealed persistent pancytopenia:

• Hemoglobin level 6.8 g/dL (reference range 13.0–17.0)
• Total white blood cell count 0.8 × 10⁹/L (4.5–11.0)
• Platelet count 8 × 10⁹/L (150–400).

The next day, he developed severe, acute-onset left-sided abdominal pain. A check of vital signs showed worsening sinus tachycardia at 132 beats per minute and a drop in blood pressure to 90/56 mm Hg. He had worsening diffuse abdominal tenderness with sluggish bowel sounds. His hemoglobin concentration was 6.4 g/dL and platelet count 12 × 10⁹/L.

He received supportive transfusions of blood products. Surgical exploration was deemed risky, given his overall condition and severe thrombocytopenia. Splenic angiography showed no evidence of pseudoaneurysm or focal contrast extravasation. He underwent empiric embolization of the midsplenic artery, after which his hemodynamic status stabilized. He died 4 weeks later of acute respiratory failure from pneumonia.

SPLENIC RUPTURE IN AML

Atraumatic splenic rupture is rare but potentially life-threatening, especially if the diagnosis is delayed. Conditions that can cause splenomegaly and predispose to rupture include infection (infectious mononucleosis, malaria), malignant hematologic disorders (leukemia, lymphoma), other neoplasms, and amyloidosis.¹

The literature includes a few reports of splenic rupture in patients with AML.^2–4 The proposed mechanisms include bleeding from infarction sites or tumor foci, dysregulated hemostasis, and leukostasis.

The classic presentation of splenic rupture is acute-onset left-sided abdominal pain associated with hypotension and decreasing hemoglobin levels. CT of the abdomen is confirmatory, and resuscitation with crystalloids and blood products is a vital initial step in management. Choice of treatment depends on the patient’s surgical risk and hemodynamic status; options include conservative medical management, splenic artery embolization, and exploratory laparotomy.

In patients with AML and splenomegaly presenting with acute abdominal pain, clinicians need to be aware of this potential hematologic emergency.

A 50-year-old man with acute myeloid leukemia (AML) with a complex karyotype was admitted to the hospital with several days of dull, left-sided abdominal pain. His most recent bone marrow biopsy showed 30% blasts, and immunophenotyping was suggestive of persistent AML (CD13+, CD34+, CD117+, CD33+, CD7+, MPO–). He was on treatment with venetoclax and cytarabine after induction therapy had failed.

On admission, his heart rate was 101 beats per minute and his blood pressure was 122/85 mm Hg. Abdominal examination revealed mild distention, hepatomegaly, and previously known massive splenomegaly, with the splenic tip extending to the umbilicus, and mild tenderness.

Results of laboratory testing revealed persistent pancytopenia:

• Hemoglobin level 6.8 g/dL (reference range 13.0–17.0)
• Total white blood cell count 0.8 × 10⁹/L (4.5–11.0)
• Platelet count 8 × 10⁹/L (150–400).

The next day, he developed severe, acute-onset left-sided abdominal pain. A check of vital signs showed worsening sinus tachycardia at 132 beats per minute and a drop in blood pressure to 90/56 mm Hg. He had worsening diffuse abdominal tenderness with sluggish bowel sounds. His hemoglobin concentration was 6.4 g/dL and platelet count 12 × 10⁹/L.

He received supportive transfusions of blood products. Surgical exploration was deemed risky, given his overall condition and severe thrombocytopenia. Splenic angiography showed no evidence of pseudoaneurysm or focal contrast extravasation. He underwent empiric embolization of the midsplenic artery, after which his hemodynamic status stabilized. He died 4 weeks later of acute respiratory failure from pneumonia.

SPLENIC RUPTURE IN AML

Atraumatic splenic rupture is rare but potentially life-threatening, especially if the diagnosis is delayed. Conditions that can cause splenomegaly and predispose to rupture include infection (infectious mononucleosis, malaria), malignant hematologic disorders (leukemia, lymphoma), other neoplasms, and amyloidosis.¹

The literature includes a few reports of splenic rupture in patients with AML.^2–4 The proposed mechanisms include bleeding from infarction sites or tumor foci, dysregulated hemostasis, and leukostasis.

The classic presentation of splenic rupture is acute-onset left-sided abdominal pain associated with hypotension and decreasing hemoglobin levels. CT of the abdomen is confirmatory, and resuscitation with crystalloids and blood products is a vital initial step in management. Choice of treatment depends on the patient’s surgical risk and hemodynamic status; options include conservative medical management, splenic artery embolization, and exploratory laparotomy.

In patients with AML and splenomegaly presenting with acute abdominal pain, clinicians need to be aware of this potential hematologic emergency.

Yes. The prevalence of sleep apnea is exceedingly high in patients with atrial fibrillation—50% to 80% compared with 30% to 60% in respective control groups.^1–3 Conversely, atrial fibrillation is more prevalent in those with sleep-disordered breathing than in those without (4.8% vs 0.9%).⁴

Sleep-disordered breathing comprises obstructive sleep apnea and central sleep apnea. Obstructive sleep apnea, characterized by repetitive upper-airway obstruction during sleep, is accompanied by intermittent hypoxia, rises in carbon dioxide, autonomic nervous system fluctuations, and intrathoracic pressure alterations.⁵ Central sleep apnea may be neurally mediated and, in the setting of cardiac disease, is characterized by alterations in chemosensitivity and chemoresponsiveness, leading to a state of high loop gain—ie, a hypersensitive ventilatory control system leading to ventilatory drive oscillations.⁶

Both obstructive and central sleep apnea have been associated with atrial fibrillation. Experimental data implicate obstructive sleep apnea as a trigger of atrial arrhythmogenesis,^7,8 and epidemiologic studies support an association between central sleep apnea, Cheyne-Stokes respiration, and incident atrial fibrillation.⁹

HOW SLEEP APNEA COULD LEAD TO ATRIAL FIBRILLATION

In experiments in animals, intermittent upper-airway obstruction led to forced inspiration, substantial negative intrathoracic pressure, subsequent left atrial distention, and increased susceptibility to atrial fibrillation.¹⁰ The autonomic nervous system may be a mediator of apnea-induced atrial fibrillation, as apnea-induced atrial fibrillation is suppressed with autonomic blockade.¹⁰

Emerging data also support the hypothesis that intermittent hypoxia⁷ and resolution of hypercapnia,⁸ as observed in obstructive sleep apnea, exert atrial electrophysiologic changes that increase vulnerability to atrial arrhythmogenesis.

In a case-crossover study,¹¹ the odds of paroxysmal atrial fibrillation occurring after a respiratory disturbance were 17.9 times higher than after normal breathing (95% confidence interval [CI] 2.2–144.2), though the absolute rate of overall arrhythmia events (including both atrial fibrillation and nonsustained ventricular tachycardia) associated with respiratory disturbances was low (1 excess arrhythmia event per 40,000 respiratory disturbances).

EFFECT OF SLEEP APNEA ON ATRIAL FIBRILLATION MANAGEMENT

Sleep apnea also seems to affect the efficacy of a rhythm-control strategy for atrial fibrillation. For example, patients with obstructive sleep apnea have a higher risk of recurrent atrial fibrillation after cardioversion (82% vs 42% in controls)¹² and up to a 25% greater risk of recurrence after catheter ablation compared with those without obstructive sleep apnea (risk ratio 1.25, 95% CI 1.08–1.45).¹³

Several observational studies showed a higher rate of atrial fibrillation after pulmonary vein isolation in obstructive sleep apnea patients who do not use continuous positive airway pressure (CPAP) than in those who do.^14–17 CPAP therapy appears to exert beneficial effects on cardiac structural remodeling;  cardiac magnetic resonance imaging shows that patients with sleep apnea who received less than 4 hours of CPAP per night had larger left atrial dimensions and increased left ventricular mass compared with those who received more than 4 hours of CPAP at night.¹⁷ However, a need remains for high-quality, large randomized controlled trials to eliminate potential unmeasured biases due to differences that may exist between CPAP users and non-users, such as general adherence to medical therapy and healthcare interventions.

An additional consideration is that the overall utility and value of obtaining a diagnosis of obstructive sleep apnea strictly as it pertains to atrial fibrillation management is affected by whether a rhythm- or rate-control strategy is pursued. In other words, if a patient is deemed to be in permanent atrial fibrillation and a rhythm-control strategy is therefore not pursued, the potential effect of untreated obstructive sleep apnea on atrial fibrillation recurrence could be less important. In this case, however, the other beneficial cardiovascular and systemic effects of diagnosing and treating underlying obstructive sleep apnea would remain.

Epidemiologic and clinic-based studies have supported an association between sleep apnea (mostly central, but also obstructive) and atrial fibrillation.^4,18

Community-based studies such as the Sleep Heart Health Study⁴ and the Outcomes of Sleep Disorders in Older Men Study (MrOS Sleep),¹⁸ involving thousands of participants, have found the strongest cross-sectional associations of both obstructive and central sleep apnea with nocturnal atrial fibrillation. The findings included a 2 to 5 times higher odds of nocturnal atrial fibrillation, particularly in those with a moderate to severe degree of sleep-disordered breathing—even after adjusting for confounding influences (eg, obesity) and self-reported cardiac disease such as heart failure.

In MrOS Sleep, in an older male cohort, both obstructive and central sleep apnea were associated with nocturnal atrial fibrillation, though central sleep apnea and Cheyne-Stokes respirations had a stronger magnitude of association.¹⁸

Further insights can be drawn specifically from patients with heart failure. Sin et al,¹⁹ in a 1999 study, found that in 450 patients with systolic heart failure (85% men), the prevalence of sleep-disordered breathing was 25% to 33% (depending on the apnea-hypopnea index cutoff used) for central sleep apnea, and similarly 27% to 38% for obstructive sleep apnea. The prevalence of atrial fibrillation in this group was 10% in women and 15% in men. Atrial fibrillation was reported as a significant risk factor for central sleep apnea, but not for obstructive sleep apnea (for which only male sex and increasing body mass index were significant risk factors). Directionality was not clearly reported in this retrospective study in terms of timing of sleep studies and other assessments: ie, the report did not clearly state which came first, the atrial fibrillation or the sleep apnea. Therefore, the possibility that central sleep apnea is a predictor of atrial fibrillation cannot be excluded.  

Yumino et al,²⁰ in a study published in 2009, evaluated 218 patients with heart failure (with a left ventricular ejection fraction of ≤ 45%) and reported a prevalence of moderate to severe sleep apnea of 21% for central sleep apnea and 26% for obstructive sleep apnea. In multivariate analysis, atrial fibrillation was independently associated with central sleep apnea but not obstructive sleep apnea.

In recent cohort studies, central sleep apnea was associated with 2 to 3 times higher odds of developing atrial fibrillation, while obstructive sleep apnea was not a predictor of incident atrial fibrillation.^9,21

Although most available studies associate sleep apnea with atrial fibrillation, findings of a case-control study²² did not support a difference in the prevalence of sleep apnea syndrome (defined as apnea index ≥ 5 and apnea-hypopnea index ≥ 15, and the presence of sleep symptoms) in patients with lone atrial fibrillation (no evident cardiovascular disease) compared with controls matched for age, sex, and cardiovascular morbidity.

But observational studies are limited by the potential for residual unmeasured confounding factors and lack of objective cardiac structural data, such as left ventricular ejection fraction and atrial enlargement. Moreover, there can be significant differences in sleep apnea definitions among studies, thus limiting the ability to reach a definitive conclusion about the relationship between sleep apnea and atrial fibrillation.

SCREENING AND DIAGNOSIS

The 2014 joint guidelines of the American Heart Association, American College of Cardiology, and Heart Rhythm Society for the management of atrial fibrillation state that a sleep study may be useful if sleep apnea is suspected.²³ The 2019 focused update of the 2014 guidelines²⁴ state that for overweight and obese patients with atrial fibrillation, weight loss combined with risk-factor modification is recommended (class I recommendation, level of evidence B-R, ie, data derived from 1 or more randomized trials or meta-analysis of such studies). Risk-factor modification in this case includes assessment and treatment of underlying sleep apnea, hypertension, hyperlipidemia, glucose intolerance, and alcohol and tobacco use.

Laboratory polysomnography has long been considered the gold standard for sleep apnea diagnosis. In one study,¹³ obstructive sleep apnea was a greater predictor of atrial fibrillation when diagnosed by polysomnography (risk ratio 1.40, 95% CI 1.16–1.68) compared with identification by screening using the Berlin questionnaire (risk ratio 1.07, 95% CI 0.91–1.27). However, a laboratory sleep study is associated with increased patient burden and limited availability.

Home sleep apnea testing is being increasingly used in the diagnostic evaluation of obstructive sleep apnea and may be a less costly, more available alternative. However, since a home sleep apnea test is less sensitive than polysomnography in detecting obstructive sleep apnea, the American Academy of Sleep Medicine guidelines²⁸ state that if a single home sleep apnea test is negative or inconclusive, polysomnography should be done if there is clinical suspicion of sleep apnea. Moreover, current guidelines from this group recommend that patients with significant cardiorespiratory disease should be tested with polysomnography rather than home sleep apnea testing.²²

Further study is needed to determine the optimal screening method for sleep apnea in patients with atrial fibrillation and to clarify the role of home sleep apnea testing. While keeping in mind the limitations of a screening questionnaire in this population, as a general approach it is reasonable to use a screening questionnaire for sleep apnea. And if the screen is positive, further evaluation with a sleep study is merited, whether by laboratory polysomnography, a home sleep apnea test, or referral to a sleep specialist.

MULTIDISCIPLINARY CARE MAY BE IDEAL

Overall, given the high prevalence of sleep apnea in patients with atrial fibrillation, the deleterious effects of sleep apnea in general, the influence of sleep apnea on atrial fibrillation, and the cardiovascular and other beneficial effects of adequate treatment of sleep apnea, patients with atrial fibrillation should be assessed for sleep apnea.

While the optimal strategy in evaluating for sleep apnea in these patients needs to be further defined, a multidisciplinary approach to care involving a primary care provider, cardiologist, and sleep specialist may be ideal.

Yes. The prevalence of sleep apnea is exceedingly high in patients with atrial fibrillation—50% to 80% compared with 30% to 60% in respective control groups.^1–3 Conversely, atrial fibrillation is more prevalent in those with sleep-disordered breathing than in those without (4.8% vs 0.9%).⁴

Sleep-disordered breathing comprises obstructive sleep apnea and central sleep apnea. Obstructive sleep apnea, characterized by repetitive upper-airway obstruction during sleep, is accompanied by intermittent hypoxia, rises in carbon dioxide, autonomic nervous system fluctuations, and intrathoracic pressure alterations.⁵ Central sleep apnea may be neurally mediated and, in the setting of cardiac disease, is characterized by alterations in chemosensitivity and chemoresponsiveness, leading to a state of high loop gain—ie, a hypersensitive ventilatory control system leading to ventilatory drive oscillations.⁶

Both obstructive and central sleep apnea have been associated with atrial fibrillation. Experimental data implicate obstructive sleep apnea as a trigger of atrial arrhythmogenesis,^7,8 and epidemiologic studies support an association between central sleep apnea, Cheyne-Stokes respiration, and incident atrial fibrillation.⁹

HOW SLEEP APNEA COULD LEAD TO ATRIAL FIBRILLATION

In experiments in animals, intermittent upper-airway obstruction led to forced inspiration, substantial negative intrathoracic pressure, subsequent left atrial distention, and increased susceptibility to atrial fibrillation.¹⁰ The autonomic nervous system may be a mediator of apnea-induced atrial fibrillation, as apnea-induced atrial fibrillation is suppressed with autonomic blockade.¹⁰

Emerging data also support the hypothesis that intermittent hypoxia⁷ and resolution of hypercapnia,⁸ as observed in obstructive sleep apnea, exert atrial electrophysiologic changes that increase vulnerability to atrial arrhythmogenesis.

In a case-crossover study,¹¹ the odds of paroxysmal atrial fibrillation occurring after a respiratory disturbance were 17.9 times higher than after normal breathing (95% confidence interval [CI] 2.2–144.2), though the absolute rate of overall arrhythmia events (including both atrial fibrillation and nonsustained ventricular tachycardia) associated with respiratory disturbances was low (1 excess arrhythmia event per 40,000 respiratory disturbances).

EFFECT OF SLEEP APNEA ON ATRIAL FIBRILLATION MANAGEMENT

Sleep apnea also seems to affect the efficacy of a rhythm-control strategy for atrial fibrillation. For example, patients with obstructive sleep apnea have a higher risk of recurrent atrial fibrillation after cardioversion (82% vs 42% in controls)¹² and up to a 25% greater risk of recurrence after catheter ablation compared with those without obstructive sleep apnea (risk ratio 1.25, 95% CI 1.08–1.45).¹³

Several observational studies showed a higher rate of atrial fibrillation after pulmonary vein isolation in obstructive sleep apnea patients who do not use continuous positive airway pressure (CPAP) than in those who do.^14–17 CPAP therapy appears to exert beneficial effects on cardiac structural remodeling;  cardiac magnetic resonance imaging shows that patients with sleep apnea who received less than 4 hours of CPAP per night had larger left atrial dimensions and increased left ventricular mass compared with those who received more than 4 hours of CPAP at night.¹⁷ However, a need remains for high-quality, large randomized controlled trials to eliminate potential unmeasured biases due to differences that may exist between CPAP users and non-users, such as general adherence to medical therapy and healthcare interventions.

An additional consideration is that the overall utility and value of obtaining a diagnosis of obstructive sleep apnea strictly as it pertains to atrial fibrillation management is affected by whether a rhythm- or rate-control strategy is pursued. In other words, if a patient is deemed to be in permanent atrial fibrillation and a rhythm-control strategy is therefore not pursued, the potential effect of untreated obstructive sleep apnea on atrial fibrillation recurrence could be less important. In this case, however, the other beneficial cardiovascular and systemic effects of diagnosing and treating underlying obstructive sleep apnea would remain.

Epidemiologic and clinic-based studies have supported an association between sleep apnea (mostly central, but also obstructive) and atrial fibrillation.^4,18

Community-based studies such as the Sleep Heart Health Study⁴ and the Outcomes of Sleep Disorders in Older Men Study (MrOS Sleep),¹⁸ involving thousands of participants, have found the strongest cross-sectional associations of both obstructive and central sleep apnea with nocturnal atrial fibrillation. The findings included a 2 to 5 times higher odds of nocturnal atrial fibrillation, particularly in those with a moderate to severe degree of sleep-disordered breathing—even after adjusting for confounding influences (eg, obesity) and self-reported cardiac disease such as heart failure.

In MrOS Sleep, in an older male cohort, both obstructive and central sleep apnea were associated with nocturnal atrial fibrillation, though central sleep apnea and Cheyne-Stokes respirations had a stronger magnitude of association.¹⁸

Further insights can be drawn specifically from patients with heart failure. Sin et al,¹⁹ in a 1999 study, found that in 450 patients with systolic heart failure (85% men), the prevalence of sleep-disordered breathing was 25% to 33% (depending on the apnea-hypopnea index cutoff used) for central sleep apnea, and similarly 27% to 38% for obstructive sleep apnea. The prevalence of atrial fibrillation in this group was 10% in women and 15% in men. Atrial fibrillation was reported as a significant risk factor for central sleep apnea, but not for obstructive sleep apnea (for which only male sex and increasing body mass index were significant risk factors). Directionality was not clearly reported in this retrospective study in terms of timing of sleep studies and other assessments: ie, the report did not clearly state which came first, the atrial fibrillation or the sleep apnea. Therefore, the possibility that central sleep apnea is a predictor of atrial fibrillation cannot be excluded.  

Yumino et al,²⁰ in a study published in 2009, evaluated 218 patients with heart failure (with a left ventricular ejection fraction of ≤ 45%) and reported a prevalence of moderate to severe sleep apnea of 21% for central sleep apnea and 26% for obstructive sleep apnea. In multivariate analysis, atrial fibrillation was independently associated with central sleep apnea but not obstructive sleep apnea.

In recent cohort studies, central sleep apnea was associated with 2 to 3 times higher odds of developing atrial fibrillation, while obstructive sleep apnea was not a predictor of incident atrial fibrillation.^9,21

Although most available studies associate sleep apnea with atrial fibrillation, findings of a case-control study²² did not support a difference in the prevalence of sleep apnea syndrome (defined as apnea index ≥ 5 and apnea-hypopnea index ≥ 15, and the presence of sleep symptoms) in patients with lone atrial fibrillation (no evident cardiovascular disease) compared with controls matched for age, sex, and cardiovascular morbidity.

But observational studies are limited by the potential for residual unmeasured confounding factors and lack of objective cardiac structural data, such as left ventricular ejection fraction and atrial enlargement. Moreover, there can be significant differences in sleep apnea definitions among studies, thus limiting the ability to reach a definitive conclusion about the relationship between sleep apnea and atrial fibrillation.

SCREENING AND DIAGNOSIS

The 2014 joint guidelines of the American Heart Association, American College of Cardiology, and Heart Rhythm Society for the management of atrial fibrillation state that a sleep study may be useful if sleep apnea is suspected.²³ The 2019 focused update of the 2014 guidelines²⁴ state that for overweight and obese patients with atrial fibrillation, weight loss combined with risk-factor modification is recommended (class I recommendation, level of evidence B-R, ie, data derived from 1 or more randomized trials or meta-analysis of such studies). Risk-factor modification in this case includes assessment and treatment of underlying sleep apnea, hypertension, hyperlipidemia, glucose intolerance, and alcohol and tobacco use.

Laboratory polysomnography has long been considered the gold standard for sleep apnea diagnosis. In one study,¹³ obstructive sleep apnea was a greater predictor of atrial fibrillation when diagnosed by polysomnography (risk ratio 1.40, 95% CI 1.16–1.68) compared with identification by screening using the Berlin questionnaire (risk ratio 1.07, 95% CI 0.91–1.27). However, a laboratory sleep study is associated with increased patient burden and limited availability.

Home sleep apnea testing is being increasingly used in the diagnostic evaluation of obstructive sleep apnea and may be a less costly, more available alternative. However, since a home sleep apnea test is less sensitive than polysomnography in detecting obstructive sleep apnea, the American Academy of Sleep Medicine guidelines²⁸ state that if a single home sleep apnea test is negative or inconclusive, polysomnography should be done if there is clinical suspicion of sleep apnea. Moreover, current guidelines from this group recommend that patients with significant cardiorespiratory disease should be tested with polysomnography rather than home sleep apnea testing.²²

Further study is needed to determine the optimal screening method for sleep apnea in patients with atrial fibrillation and to clarify the role of home sleep apnea testing. While keeping in mind the limitations of a screening questionnaire in this population, as a general approach it is reasonable to use a screening questionnaire for sleep apnea. And if the screen is positive, further evaluation with a sleep study is merited, whether by laboratory polysomnography, a home sleep apnea test, or referral to a sleep specialist.

MULTIDISCIPLINARY CARE MAY BE IDEAL

Overall, given the high prevalence of sleep apnea in patients with atrial fibrillation, the deleterious effects of sleep apnea in general, the influence of sleep apnea on atrial fibrillation, and the cardiovascular and other beneficial effects of adequate treatment of sleep apnea, patients with atrial fibrillation should be assessed for sleep apnea.

While the optimal strategy in evaluating for sleep apnea in these patients needs to be further defined, a multidisciplinary approach to care involving a primary care provider, cardiologist, and sleep specialist may be ideal.

Yes. The prevalence of sleep apnea is exceedingly high in patients with atrial fibrillation—50% to 80% compared with 30% to 60% in respective control groups.^1–3 Conversely, atrial fibrillation is more prevalent in those with sleep-disordered breathing than in those without (4.8% vs 0.9%).⁴

Sleep-disordered breathing comprises obstructive sleep apnea and central sleep apnea. Obstructive sleep apnea, characterized by repetitive upper-airway obstruction during sleep, is accompanied by intermittent hypoxia, rises in carbon dioxide, autonomic nervous system fluctuations, and intrathoracic pressure alterations.⁵ Central sleep apnea may be neurally mediated and, in the setting of cardiac disease, is characterized by alterations in chemosensitivity and chemoresponsiveness, leading to a state of high loop gain—ie, a hypersensitive ventilatory control system leading to ventilatory drive oscillations.⁶

Both obstructive and central sleep apnea have been associated with atrial fibrillation. Experimental data implicate obstructive sleep apnea as a trigger of atrial arrhythmogenesis,^7,8 and epidemiologic studies support an association between central sleep apnea, Cheyne-Stokes respiration, and incident atrial fibrillation.⁹

HOW SLEEP APNEA COULD LEAD TO ATRIAL FIBRILLATION

In experiments in animals, intermittent upper-airway obstruction led to forced inspiration, substantial negative intrathoracic pressure, subsequent left atrial distention, and increased susceptibility to atrial fibrillation.¹⁰ The autonomic nervous system may be a mediator of apnea-induced atrial fibrillation, as apnea-induced atrial fibrillation is suppressed with autonomic blockade.¹⁰

Emerging data also support the hypothesis that intermittent hypoxia⁷ and resolution of hypercapnia,⁸ as observed in obstructive sleep apnea, exert atrial electrophysiologic changes that increase vulnerability to atrial arrhythmogenesis.

In a case-crossover study,¹¹ the odds of paroxysmal atrial fibrillation occurring after a respiratory disturbance were 17.9 times higher than after normal breathing (95% confidence interval [CI] 2.2–144.2), though the absolute rate of overall arrhythmia events (including both atrial fibrillation and nonsustained ventricular tachycardia) associated with respiratory disturbances was low (1 excess arrhythmia event per 40,000 respiratory disturbances).

EFFECT OF SLEEP APNEA ON ATRIAL FIBRILLATION MANAGEMENT

Sleep apnea also seems to affect the efficacy of a rhythm-control strategy for atrial fibrillation. For example, patients with obstructive sleep apnea have a higher risk of recurrent atrial fibrillation after cardioversion (82% vs 42% in controls)¹² and up to a 25% greater risk of recurrence after catheter ablation compared with those without obstructive sleep apnea (risk ratio 1.25, 95% CI 1.08–1.45).¹³

Several observational studies showed a higher rate of atrial fibrillation after pulmonary vein isolation in obstructive sleep apnea patients who do not use continuous positive airway pressure (CPAP) than in those who do.^14–17 CPAP therapy appears to exert beneficial effects on cardiac structural remodeling;  cardiac magnetic resonance imaging shows that patients with sleep apnea who received less than 4 hours of CPAP per night had larger left atrial dimensions and increased left ventricular mass compared with those who received more than 4 hours of CPAP at night.¹⁷ However, a need remains for high-quality, large randomized controlled trials to eliminate potential unmeasured biases due to differences that may exist between CPAP users and non-users, such as general adherence to medical therapy and healthcare interventions.

An additional consideration is that the overall utility and value of obtaining a diagnosis of obstructive sleep apnea strictly as it pertains to atrial fibrillation management is affected by whether a rhythm- or rate-control strategy is pursued. In other words, if a patient is deemed to be in permanent atrial fibrillation and a rhythm-control strategy is therefore not pursued, the potential effect of untreated obstructive sleep apnea on atrial fibrillation recurrence could be less important. In this case, however, the other beneficial cardiovascular and systemic effects of diagnosing and treating underlying obstructive sleep apnea would remain.

Epidemiologic and clinic-based studies have supported an association between sleep apnea (mostly central, but also obstructive) and atrial fibrillation.^4,18

Community-based studies such as the Sleep Heart Health Study⁴ and the Outcomes of Sleep Disorders in Older Men Study (MrOS Sleep),¹⁸ involving thousands of participants, have found the strongest cross-sectional associations of both obstructive and central sleep apnea with nocturnal atrial fibrillation. The findings included a 2 to 5 times higher odds of nocturnal atrial fibrillation, particularly in those with a moderate to severe degree of sleep-disordered breathing—even after adjusting for confounding influences (eg, obesity) and self-reported cardiac disease such as heart failure.

In MrOS Sleep, in an older male cohort, both obstructive and central sleep apnea were associated with nocturnal atrial fibrillation, though central sleep apnea and Cheyne-Stokes respirations had a stronger magnitude of association.¹⁸

Further insights can be drawn specifically from patients with heart failure. Sin et al,¹⁹ in a 1999 study, found that in 450 patients with systolic heart failure (85% men), the prevalence of sleep-disordered breathing was 25% to 33% (depending on the apnea-hypopnea index cutoff used) for central sleep apnea, and similarly 27% to 38% for obstructive sleep apnea. The prevalence of atrial fibrillation in this group was 10% in women and 15% in men. Atrial fibrillation was reported as a significant risk factor for central sleep apnea, but not for obstructive sleep apnea (for which only male sex and increasing body mass index were significant risk factors). Directionality was not clearly reported in this retrospective study in terms of timing of sleep studies and other assessments: ie, the report did not clearly state which came first, the atrial fibrillation or the sleep apnea. Therefore, the possibility that central sleep apnea is a predictor of atrial fibrillation cannot be excluded.  

Yumino et al,²⁰ in a study published in 2009, evaluated 218 patients with heart failure (with a left ventricular ejection fraction of ≤ 45%) and reported a prevalence of moderate to severe sleep apnea of 21% for central sleep apnea and 26% for obstructive sleep apnea. In multivariate analysis, atrial fibrillation was independently associated with central sleep apnea but not obstructive sleep apnea.

In recent cohort studies, central sleep apnea was associated with 2 to 3 times higher odds of developing atrial fibrillation, while obstructive sleep apnea was not a predictor of incident atrial fibrillation.^9,21

Although most available studies associate sleep apnea with atrial fibrillation, findings of a case-control study²² did not support a difference in the prevalence of sleep apnea syndrome (defined as apnea index ≥ 5 and apnea-hypopnea index ≥ 15, and the presence of sleep symptoms) in patients with lone atrial fibrillation (no evident cardiovascular disease) compared with controls matched for age, sex, and cardiovascular morbidity.

But observational studies are limited by the potential for residual unmeasured confounding factors and lack of objective cardiac structural data, such as left ventricular ejection fraction and atrial enlargement. Moreover, there can be significant differences in sleep apnea definitions among studies, thus limiting the ability to reach a definitive conclusion about the relationship between sleep apnea and atrial fibrillation.

SCREENING AND DIAGNOSIS

The 2014 joint guidelines of the American Heart Association, American College of Cardiology, and Heart Rhythm Society for the management of atrial fibrillation state that a sleep study may be useful if sleep apnea is suspected.²³ The 2019 focused update of the 2014 guidelines²⁴ state that for overweight and obese patients with atrial fibrillation, weight loss combined with risk-factor modification is recommended (class I recommendation, level of evidence B-R, ie, data derived from 1 or more randomized trials or meta-analysis of such studies). Risk-factor modification in this case includes assessment and treatment of underlying sleep apnea, hypertension, hyperlipidemia, glucose intolerance, and alcohol and tobacco use.

Laboratory polysomnography has long been considered the gold standard for sleep apnea diagnosis. In one study,¹³ obstructive sleep apnea was a greater predictor of atrial fibrillation when diagnosed by polysomnography (risk ratio 1.40, 95% CI 1.16–1.68) compared with identification by screening using the Berlin questionnaire (risk ratio 1.07, 95% CI 0.91–1.27). However, a laboratory sleep study is associated with increased patient burden and limited availability.

Home sleep apnea testing is being increasingly used in the diagnostic evaluation of obstructive sleep apnea and may be a less costly, more available alternative. However, since a home sleep apnea test is less sensitive than polysomnography in detecting obstructive sleep apnea, the American Academy of Sleep Medicine guidelines²⁸ state that if a single home sleep apnea test is negative or inconclusive, polysomnography should be done if there is clinical suspicion of sleep apnea. Moreover, current guidelines from this group recommend that patients with significant cardiorespiratory disease should be tested with polysomnography rather than home sleep apnea testing.²²

Further study is needed to determine the optimal screening method for sleep apnea in patients with atrial fibrillation and to clarify the role of home sleep apnea testing. While keeping in mind the limitations of a screening questionnaire in this population, as a general approach it is reasonable to use a screening questionnaire for sleep apnea. And if the screen is positive, further evaluation with a sleep study is merited, whether by laboratory polysomnography, a home sleep apnea test, or referral to a sleep specialist.

MULTIDISCIPLINARY CARE MAY BE IDEAL

Overall, given the high prevalence of sleep apnea in patients with atrial fibrillation, the deleterious effects of sleep apnea in general, the influence of sleep apnea on atrial fibrillation, and the cardiovascular and other beneficial effects of adequate treatment of sleep apnea, patients with atrial fibrillation should be assessed for sleep apnea.

While the optimal strategy in evaluating for sleep apnea in these patients needs to be further defined, a multidisciplinary approach to care involving a primary care provider, cardiologist, and sleep specialist may be ideal.