What does your patient need to know?

When a patient presents with a history of genital warts (GWs), find out when and where the lesions started; where the lesions are currently located; what new lesions have developed; what treatments have been administered (eg, physician applied, prescription) and which one(s) worked; what side effects to treatments have been experienced and at what dose; does a partner(s) have similar lesions; is there a history of other sexually transmitted diseases or genital cancer; is he/she immunocompromised (eg, human immunodeficiency virus, transplant, medications); and what is his/her sexual orientation.

Once all of the information has been gathered and the entire anogenital region has been examined, a treatment plan can be formulated. If the patient is immunocompromised or is a man who has sex with men, the risk for anogenital malignancy due to human papillomavirus (HPV) is higher, and GWs, which can be coinfected with oncogenic HPV types, should be treated more aggressively. If the patient is still getting new lesions, use of only a destructive method such as cryotherapy will likely lead to suboptimal results.

Any patients with GWs in the anal region but particularly those in high-risk groups such as men who have sex with men and human immunodeficiency virus–infected patients should have an anoscopy to evaluate for lesions on the anal mucosa and in the rectum.

What are your go-to treatments?

Prior treatments need to be taken into account; make sure to understand any side effects and how he/she applied the prior treatment before eliminating it as a viable option. Treatment usually depends on the number of lesions, surface area, anatomic locations involved, and size of the lesions. I start with a 2-pronged approach—a debulking therapy and a patient-applied topical therapy—which allows me to physically remove some of the lesions, typically the larger ones, and then have the patient apply a topical medication at home that will treat the smaller lesions as well as help to clear or decrease the burden of HPV virus on the skin. I use cryotherapy as a debulking agent, but curettage or podophyllin 25% also can be used in the office. I use imiquimod cream 5% as a first-line topical agent at the recommended dose of 3 times weekly; however, if after the first 2 weeks the patient has little response or too much irritation, I titrate the dose so that the patient has mild inflammation on the skin. The dose ultimately can range from daily to once weekly. Some patients who can only tolerate imiquimod once or twice weekly may require zinc oxide paste for the inguinal folds and scrotum to protect from irritation. Alternate topical medications for GWs include sinecatechins ointment 15% or cidofovir ointment 2%.

How do you keep patients compliant?

Start the visit with open communication about the disease, where it came from, what the risks are if it is not treated, and how we can best treat it to make sure we minimize those risks. I explain all of the treatment options as well as our role in treating these lesions and minimizing the risk for disease progression.

What do you do if they refuse treatment?

Most patients with GWs are motivated to be treated. If pain is a concern, such as with cryotherapy, I recommend topical treatments.

What patient resources do you recommend?

The American Academy of Dermatology (https://www.aad.org/public/diseases/contagious-skin-diseases/genital-warts), Harvard Medical School patient education center (Boston, Massachusetts)(http://www.patienteducationcenter.org/articles/genital-warts/), and American Family Physician (http://www.aafp.org/afp/2004/1215/p2345.html) provide patient materials that I recommend.

What does your patient need to know?

When a patient presents with a history of genital warts (GWs), find out when and where the lesions started; where the lesions are currently located; what new lesions have developed; what treatments have been administered (eg, physician applied, prescription) and which one(s) worked; what side effects to treatments have been experienced and at what dose; does a partner(s) have similar lesions; is there a history of other sexually transmitted diseases or genital cancer; is he/she immunocompromised (eg, human immunodeficiency virus, transplant, medications); and what is his/her sexual orientation.

Once all of the information has been gathered and the entire anogenital region has been examined, a treatment plan can be formulated. If the patient is immunocompromised or is a man who has sex with men, the risk for anogenital malignancy due to human papillomavirus (HPV) is higher, and GWs, which can be coinfected with oncogenic HPV types, should be treated more aggressively. If the patient is still getting new lesions, use of only a destructive method such as cryotherapy will likely lead to suboptimal results.

Any patients with GWs in the anal region but particularly those in high-risk groups such as men who have sex with men and human immunodeficiency virus–infected patients should have an anoscopy to evaluate for lesions on the anal mucosa and in the rectum.

What are your go-to treatments?

Prior treatments need to be taken into account; make sure to understand any side effects and how he/she applied the prior treatment before eliminating it as a viable option. Treatment usually depends on the number of lesions, surface area, anatomic locations involved, and size of the lesions. I start with a 2-pronged approach—a debulking therapy and a patient-applied topical therapy—which allows me to physically remove some of the lesions, typically the larger ones, and then have the patient apply a topical medication at home that will treat the smaller lesions as well as help to clear or decrease the burden of HPV virus on the skin. I use cryotherapy as a debulking agent, but curettage or podophyllin 25% also can be used in the office. I use imiquimod cream 5% as a first-line topical agent at the recommended dose of 3 times weekly; however, if after the first 2 weeks the patient has little response or too much irritation, I titrate the dose so that the patient has mild inflammation on the skin. The dose ultimately can range from daily to once weekly. Some patients who can only tolerate imiquimod once or twice weekly may require zinc oxide paste for the inguinal folds and scrotum to protect from irritation. Alternate topical medications for GWs include sinecatechins ointment 15% or cidofovir ointment 2%.

How do you keep patients compliant?

Start the visit with open communication about the disease, where it came from, what the risks are if it is not treated, and how we can best treat it to make sure we minimize those risks. I explain all of the treatment options as well as our role in treating these lesions and minimizing the risk for disease progression.

What do you do if they refuse treatment?

Most patients with GWs are motivated to be treated. If pain is a concern, such as with cryotherapy, I recommend topical treatments.

What patient resources do you recommend?

The American Academy of Dermatology (https://www.aad.org/public/diseases/contagious-skin-diseases/genital-warts), Harvard Medical School patient education center (Boston, Massachusetts)(http://www.patienteducationcenter.org/articles/genital-warts/), and American Family Physician (http://www.aafp.org/afp/2004/1215/p2345.html) provide patient materials that I recommend.

What does your patient need to know?

When a patient presents with a history of genital warts (GWs), find out when and where the lesions started; where the lesions are currently located; what new lesions have developed; what treatments have been administered (eg, physician applied, prescription) and which one(s) worked; what side effects to treatments have been experienced and at what dose; does a partner(s) have similar lesions; is there a history of other sexually transmitted diseases or genital cancer; is he/she immunocompromised (eg, human immunodeficiency virus, transplant, medications); and what is his/her sexual orientation.

Once all of the information has been gathered and the entire anogenital region has been examined, a treatment plan can be formulated. If the patient is immunocompromised or is a man who has sex with men, the risk for anogenital malignancy due to human papillomavirus (HPV) is higher, and GWs, which can be coinfected with oncogenic HPV types, should be treated more aggressively. If the patient is still getting new lesions, use of only a destructive method such as cryotherapy will likely lead to suboptimal results.

Any patients with GWs in the anal region but particularly those in high-risk groups such as men who have sex with men and human immunodeficiency virus–infected patients should have an anoscopy to evaluate for lesions on the anal mucosa and in the rectum.

What are your go-to treatments?

Prior treatments need to be taken into account; make sure to understand any side effects and how he/she applied the prior treatment before eliminating it as a viable option. Treatment usually depends on the number of lesions, surface area, anatomic locations involved, and size of the lesions. I start with a 2-pronged approach—a debulking therapy and a patient-applied topical therapy—which allows me to physically remove some of the lesions, typically the larger ones, and then have the patient apply a topical medication at home that will treat the smaller lesions as well as help to clear or decrease the burden of HPV virus on the skin. I use cryotherapy as a debulking agent, but curettage or podophyllin 25% also can be used in the office. I use imiquimod cream 5% as a first-line topical agent at the recommended dose of 3 times weekly; however, if after the first 2 weeks the patient has little response or too much irritation, I titrate the dose so that the patient has mild inflammation on the skin. The dose ultimately can range from daily to once weekly. Some patients who can only tolerate imiquimod once or twice weekly may require zinc oxide paste for the inguinal folds and scrotum to protect from irritation. Alternate topical medications for GWs include sinecatechins ointment 15% or cidofovir ointment 2%.

How do you keep patients compliant?

Start the visit with open communication about the disease, where it came from, what the risks are if it is not treated, and how we can best treat it to make sure we minimize those risks. I explain all of the treatment options as well as our role in treating these lesions and minimizing the risk for disease progression.

What do you do if they refuse treatment?

Most patients with GWs are motivated to be treated. If pain is a concern, such as with cryotherapy, I recommend topical treatments.

What patient resources do you recommend?

The American Academy of Dermatology (https://www.aad.org/public/diseases/contagious-skin-diseases/genital-warts), Harvard Medical School patient education center (Boston, Massachusetts)(http://www.patienteducationcenter.org/articles/genital-warts/), and American Family Physician (http://www.aafp.org/afp/2004/1215/p2345.html) provide patient materials that I recommend.

An established patient comes into your office with a painful new lesion on the hand. He thinks it may be a wart. You take a focused history of the lesion, do a physical examination, and confirm the diagnosis of verruca vulgaris. You discuss treatment options, risks, and the benefits of treatment, as well as the pathophysiology of warts. The decision is made to proceed that same day with cryosurgical destruction, which is performed. You feel that billing both an office visit with an appended modifier -25 and the benign destruction code 17110 is warranted, but your biller says only the procedure should be reported. Who is correct?

Modifier -25 use has come under increased scrutiny by insurers and regulators. There is a perception that this modifier is frequently used inappropriately or unnecessarily. In fact, the Office of Inspector General reported that 35% of claims using modifier -25 that Medicare allowed did not meet the requirements. The Office of Inspector General has recommended that the “[Centers for Medicare & Medicaid Services] should work with carriers to reduce the number of claims submitted using modifier -25” and “include modifier -25 reviews in their medical review strategies.”¹ Translation: More chart reviews and audits! In my discussions with insurer medical directors, they point to the single diagnosis modifier -25 as likely abused and feel that its use in this context is almost never appropriate. Their audits have been focused on this aspect of dermatologists’ coding. In addition, some private insurers have started to discount reimbursement for office visits billed with modifier -25 by 50% to account for the level of perceived overuse.²

The Current Procedural Terminology description of modifier -25 is relatively clear: Modifier -25 is used to facilitate billing of evaluation and management (E/M) services on the day of a procedure for which separate payment may be made.³ This modifier indicates that a significant, separately identifiable E/M service was performed by the same physician on the day of a procedure. To appropriately bill both the E/M service and the procedure, the physician must indicate that the patient’s condition required an E/M service “above and beyond the usual pre- and post-operative work of a procedure.”⁴ However, it is largely left up to the physician to decide what constitutes the significant, separately identifiable E/M service.

As dermatologists, we all report modifier -25 appropriately as part of our daily practice. Performance of a medically necessary procedure on the same day as an E/M service generally is done to facilitate a prompt diagnosis or streamline treatment of a complex condition. Providing distinct medically necessary services on the same date allows physicians to provide effective and efficient high-quality care, in many cases saving patients a return visit. The most common scenario for using modifier -25 involves multiple concerns and multiple diagnoses, some of which are not associated with a procedure(s) that is performed on the same date of service. With multiple diagnoses, it is straightforward to demonstrate the separate E/M service associated with the nonprocedure-related diagnosis code(s); however, with one diagnosis for both the office visit and the procedure, clear documentation of the separate and identifiable E/M service is critical and is dependent on understanding what is included in the global surgical package.

Insurer payment for procedures includes local or topical anesthesia, the surgical service/procedure itself, immediate postoperative care including dictating the operative note, meeting/discussing the patient’s procedure with family and other physicians, evaluating the patient in postanesthesia/recovery area, and writing orders for the patient. This group of services is called the global surgical package. For minor procedures (ie, those with either 0- or 10-day global periods), the surgical package also includes same-day E/M associated with the decision to perform surgery. An appropriate history and physical examination, as well as the discussion of differential diagnosis, treatment options, and risk and benefits of treatments, are all included in the payment of a minor procedure itself. Therefore, if an E/M service is performed on the same day as a minor procedure to decide whether to proceed with the minor surgical procedure, this E/M service cannot be separately reported. Moreover, the fact that the patient is new to the physician is not sufficient to allow reporting of an E/M service with such a minor procedure. For major procedures (ie, those with 90-day postoperative periods), the decision for surgery is excluded from the global surgical package.

Therefore, it is clear that the clinical scenario for verruca vulgaris treatment as described at the start of this article does not meet criteria for an office visit billed in addition to the destruction. The E/M services performed prior to the patient’s verruca vulgaris treatment are integral to and necessary for the decision to perform the procedure. Making and confirming the diagnosis of a condition or lesion prior to a procedure either by physical evaluation or by interpretation of a pathology report is part of the evaluation required to make the decision to proceed with a particular procedure.

There are clinical scenarios in which a physician can support additional E/M services beyond that of the procedure with just one diagnosis. If a patient presents with warts on the hand and face with resultant cryosurgical destruction done on the hand and a prescription for imiquimod to be used on the face to induce immunologic clearance of viral infection and decrease the risk of scarring, it is clear that a significant and separately identifiable E/M service exists. The evaluation of the facial warts and the prescription of medication and discussion of the risks, benefits, and therapeutic effects of that prescription is definitely distinct from the procedure. Similarly, if an evaluation of a patient with a rash results in only a diagnostic biopsy with no separate cognitive services other than the decision to perform the biopsy, an office visit charge in addition to the biopsy charge would not be warranted. However, if in addition to the biopsy the rash also is treated with topical or systemic steroids because of pruritus or a more extensive evaluation for systemic complications is required, an office visit charge is appropriate.

The frequent use of modifier -25 is a critical part of a high-quality and cost-effective dermatology practice. Same-day performance of E/M services and minor procedures allows for more rapid and efficient diagnosis and treatment of various conditions as well as minimizing unnecessary office visits. However, modifier -25 use, particularly in the context of the same diagnosis for the office visit and the procedure, is under intense insurer scrutiny. Careful and complete documentation of the additional E/M service provided, including the additional history, physical examination results, and treatment considerations above and beyond those typically required by the minor procedure, will reduce the likelihood of redeterminations from reviews and audits. Understanding Medicare guidelines and National Correct Coding Initiative recommendations will help keep the dermatologist out of hot water.⁵

An established patient comes into your office with a painful new lesion on the hand. He thinks it may be a wart. You take a focused history of the lesion, do a physical examination, and confirm the diagnosis of verruca vulgaris. You discuss treatment options, risks, and the benefits of treatment, as well as the pathophysiology of warts. The decision is made to proceed that same day with cryosurgical destruction, which is performed. You feel that billing both an office visit with an appended modifier -25 and the benign destruction code 17110 is warranted, but your biller says only the procedure should be reported. Who is correct?

Modifier -25 use has come under increased scrutiny by insurers and regulators. There is a perception that this modifier is frequently used inappropriately or unnecessarily. In fact, the Office of Inspector General reported that 35% of claims using modifier -25 that Medicare allowed did not meet the requirements. The Office of Inspector General has recommended that the “[Centers for Medicare & Medicaid Services] should work with carriers to reduce the number of claims submitted using modifier -25” and “include modifier -25 reviews in their medical review strategies.”¹ Translation: More chart reviews and audits! In my discussions with insurer medical directors, they point to the single diagnosis modifier -25 as likely abused and feel that its use in this context is almost never appropriate. Their audits have been focused on this aspect of dermatologists’ coding. In addition, some private insurers have started to discount reimbursement for office visits billed with modifier -25 by 50% to account for the level of perceived overuse.²

The Current Procedural Terminology description of modifier -25 is relatively clear: Modifier -25 is used to facilitate billing of evaluation and management (E/M) services on the day of a procedure for which separate payment may be made.³ This modifier indicates that a significant, separately identifiable E/M service was performed by the same physician on the day of a procedure. To appropriately bill both the E/M service and the procedure, the physician must indicate that the patient’s condition required an E/M service “above and beyond the usual pre- and post-operative work of a procedure.”⁴ However, it is largely left up to the physician to decide what constitutes the significant, separately identifiable E/M service.

As dermatologists, we all report modifier -25 appropriately as part of our daily practice. Performance of a medically necessary procedure on the same day as an E/M service generally is done to facilitate a prompt diagnosis or streamline treatment of a complex condition. Providing distinct medically necessary services on the same date allows physicians to provide effective and efficient high-quality care, in many cases saving patients a return visit. The most common scenario for using modifier -25 involves multiple concerns and multiple diagnoses, some of which are not associated with a procedure(s) that is performed on the same date of service. With multiple diagnoses, it is straightforward to demonstrate the separate E/M service associated with the nonprocedure-related diagnosis code(s); however, with one diagnosis for both the office visit and the procedure, clear documentation of the separate and identifiable E/M service is critical and is dependent on understanding what is included in the global surgical package.

Insurer payment for procedures includes local or topical anesthesia, the surgical service/procedure itself, immediate postoperative care including dictating the operative note, meeting/discussing the patient’s procedure with family and other physicians, evaluating the patient in postanesthesia/recovery area, and writing orders for the patient. This group of services is called the global surgical package. For minor procedures (ie, those with either 0- or 10-day global periods), the surgical package also includes same-day E/M associated with the decision to perform surgery. An appropriate history and physical examination, as well as the discussion of differential diagnosis, treatment options, and risk and benefits of treatments, are all included in the payment of a minor procedure itself. Therefore, if an E/M service is performed on the same day as a minor procedure to decide whether to proceed with the minor surgical procedure, this E/M service cannot be separately reported. Moreover, the fact that the patient is new to the physician is not sufficient to allow reporting of an E/M service with such a minor procedure. For major procedures (ie, those with 90-day postoperative periods), the decision for surgery is excluded from the global surgical package.

Therefore, it is clear that the clinical scenario for verruca vulgaris treatment as described at the start of this article does not meet criteria for an office visit billed in addition to the destruction. The E/M services performed prior to the patient’s verruca vulgaris treatment are integral to and necessary for the decision to perform the procedure. Making and confirming the diagnosis of a condition or lesion prior to a procedure either by physical evaluation or by interpretation of a pathology report is part of the evaluation required to make the decision to proceed with a particular procedure.

There are clinical scenarios in which a physician can support additional E/M services beyond that of the procedure with just one diagnosis. If a patient presents with warts on the hand and face with resultant cryosurgical destruction done on the hand and a prescription for imiquimod to be used on the face to induce immunologic clearance of viral infection and decrease the risk of scarring, it is clear that a significant and separately identifiable E/M service exists. The evaluation of the facial warts and the prescription of medication and discussion of the risks, benefits, and therapeutic effects of that prescription is definitely distinct from the procedure. Similarly, if an evaluation of a patient with a rash results in only a diagnostic biopsy with no separate cognitive services other than the decision to perform the biopsy, an office visit charge in addition to the biopsy charge would not be warranted. However, if in addition to the biopsy the rash also is treated with topical or systemic steroids because of pruritus or a more extensive evaluation for systemic complications is required, an office visit charge is appropriate.

The frequent use of modifier -25 is a critical part of a high-quality and cost-effective dermatology practice. Same-day performance of E/M services and minor procedures allows for more rapid and efficient diagnosis and treatment of various conditions as well as minimizing unnecessary office visits. However, modifier -25 use, particularly in the context of the same diagnosis for the office visit and the procedure, is under intense insurer scrutiny. Careful and complete documentation of the additional E/M service provided, including the additional history, physical examination results, and treatment considerations above and beyond those typically required by the minor procedure, will reduce the likelihood of redeterminations from reviews and audits. Understanding Medicare guidelines and National Correct Coding Initiative recommendations will help keep the dermatologist out of hot water.⁵

An established patient comes into your office with a painful new lesion on the hand. He thinks it may be a wart. You take a focused history of the lesion, do a physical examination, and confirm the diagnosis of verruca vulgaris. You discuss treatment options, risks, and the benefits of treatment, as well as the pathophysiology of warts. The decision is made to proceed that same day with cryosurgical destruction, which is performed. You feel that billing both an office visit with an appended modifier -25 and the benign destruction code 17110 is warranted, but your biller says only the procedure should be reported. Who is correct?

Modifier -25 use has come under increased scrutiny by insurers and regulators. There is a perception that this modifier is frequently used inappropriately or unnecessarily. In fact, the Office of Inspector General reported that 35% of claims using modifier -25 that Medicare allowed did not meet the requirements. The Office of Inspector General has recommended that the “[Centers for Medicare & Medicaid Services] should work with carriers to reduce the number of claims submitted using modifier -25” and “include modifier -25 reviews in their medical review strategies.”¹ Translation: More chart reviews and audits! In my discussions with insurer medical directors, they point to the single diagnosis modifier -25 as likely abused and feel that its use in this context is almost never appropriate. Their audits have been focused on this aspect of dermatologists’ coding. In addition, some private insurers have started to discount reimbursement for office visits billed with modifier -25 by 50% to account for the level of perceived overuse.²

The Current Procedural Terminology description of modifier -25 is relatively clear: Modifier -25 is used to facilitate billing of evaluation and management (E/M) services on the day of a procedure for which separate payment may be made.³ This modifier indicates that a significant, separately identifiable E/M service was performed by the same physician on the day of a procedure. To appropriately bill both the E/M service and the procedure, the physician must indicate that the patient’s condition required an E/M service “above and beyond the usual pre- and post-operative work of a procedure.”⁴ However, it is largely left up to the physician to decide what constitutes the significant, separately identifiable E/M service.

As dermatologists, we all report modifier -25 appropriately as part of our daily practice. Performance of a medically necessary procedure on the same day as an E/M service generally is done to facilitate a prompt diagnosis or streamline treatment of a complex condition. Providing distinct medically necessary services on the same date allows physicians to provide effective and efficient high-quality care, in many cases saving patients a return visit. The most common scenario for using modifier -25 involves multiple concerns and multiple diagnoses, some of which are not associated with a procedure(s) that is performed on the same date of service. With multiple diagnoses, it is straightforward to demonstrate the separate E/M service associated with the nonprocedure-related diagnosis code(s); however, with one diagnosis for both the office visit and the procedure, clear documentation of the separate and identifiable E/M service is critical and is dependent on understanding what is included in the global surgical package.

Insurer payment for procedures includes local or topical anesthesia, the surgical service/procedure itself, immediate postoperative care including dictating the operative note, meeting/discussing the patient’s procedure with family and other physicians, evaluating the patient in postanesthesia/recovery area, and writing orders for the patient. This group of services is called the global surgical package. For minor procedures (ie, those with either 0- or 10-day global periods), the surgical package also includes same-day E/M associated with the decision to perform surgery. An appropriate history and physical examination, as well as the discussion of differential diagnosis, treatment options, and risk and benefits of treatments, are all included in the payment of a minor procedure itself. Therefore, if an E/M service is performed on the same day as a minor procedure to decide whether to proceed with the minor surgical procedure, this E/M service cannot be separately reported. Moreover, the fact that the patient is new to the physician is not sufficient to allow reporting of an E/M service with such a minor procedure. For major procedures (ie, those with 90-day postoperative periods), the decision for surgery is excluded from the global surgical package.

Therefore, it is clear that the clinical scenario for verruca vulgaris treatment as described at the start of this article does not meet criteria for an office visit billed in addition to the destruction. The E/M services performed prior to the patient’s verruca vulgaris treatment are integral to and necessary for the decision to perform the procedure. Making and confirming the diagnosis of a condition or lesion prior to a procedure either by physical evaluation or by interpretation of a pathology report is part of the evaluation required to make the decision to proceed with a particular procedure.

There are clinical scenarios in which a physician can support additional E/M services beyond that of the procedure with just one diagnosis. If a patient presents with warts on the hand and face with resultant cryosurgical destruction done on the hand and a prescription for imiquimod to be used on the face to induce immunologic clearance of viral infection and decrease the risk of scarring, it is clear that a significant and separately identifiable E/M service exists. The evaluation of the facial warts and the prescription of medication and discussion of the risks, benefits, and therapeutic effects of that prescription is definitely distinct from the procedure. Similarly, if an evaluation of a patient with a rash results in only a diagnostic biopsy with no separate cognitive services other than the decision to perform the biopsy, an office visit charge in addition to the biopsy charge would not be warranted. However, if in addition to the biopsy the rash also is treated with topical or systemic steroids because of pruritus or a more extensive evaluation for systemic complications is required, an office visit charge is appropriate.

The frequent use of modifier -25 is a critical part of a high-quality and cost-effective dermatology practice. Same-day performance of E/M services and minor procedures allows for more rapid and efficient diagnosis and treatment of various conditions as well as minimizing unnecessary office visits. However, modifier -25 use, particularly in the context of the same diagnosis for the office visit and the procedure, is under intense insurer scrutiny. Careful and complete documentation of the additional E/M service provided, including the additional history, physical examination results, and treatment considerations above and beyond those typically required by the minor procedure, will reduce the likelihood of redeterminations from reviews and audits. Understanding Medicare guidelines and National Correct Coding Initiative recommendations will help keep the dermatologist out of hot water.⁵

From the University of Colorado Denver, Aurora, CO.

Abstract

• Objective: To provide evidence-based guidance on strategies that are likely or unlikely to be successful in navigating HPV vaccine conversations with patients and parents.
• Methods: Nonsystematic review of the literature.
• Results: This review highlights some of the most recent innovations in provider HPV vaccine communication and describes provider communication strategies that have been found to improve adolescent vaccination rates in rigorous scientific studies. Promising strategies for which additional research is needed and strategies that probably do not work are also described. 
• Conclusion: By understanding what works, what may work, and what not to do when it comes to communicating with families about HPV vaccines, providers can be better prepared for maximizing the impact they can have on adolescent HPV vaccination rates.

Key words: human papillomavirus; vaccine hesitancy; health communication; parents; immunization.

In the United States, more than 14 million people newly acquire genital human papillomavirus (HPV) annually, and 75 million Americans are infected at any given time [1]. As the most common sexually transmitted disease, more than 80% of sexually active U.S. adults are estimated to be infected with HPV by the age of 50 [1,2]. Although the majority of infections are “silent” and resolve without clinical sequelae, a proportion of infected individuals will go on to develop HPV-related diseases. In women, these include cervical cancer and pre-cancer (ie, abnormal Pap smears); cancers of the vagina, vulva, anus, and oropharynx; and genital warts [3]. Males also bear a high burden of HPV-related disease in the form of penile, anal, and oropharyngeal cancers, as well as genital warts [3]. While once thought of as primarily a “woman’s disease” [4], recent research demonstrates men are also significantly impacted by HPV—particularly in the form of oropharyngeal cancers, which are 2 to 3 times more common in men than in women [5]. In fact, it is estimated by the year 2020 more men will die of HPV-related oropharyngeal cancer than women will die of cervical cancer [6,7]. The combined cost of HPV-associated cancers and other conditions is estimated to be $8 billion per year in the United States [8–11].

HPV Vaccines

Effective HPV vaccines have been available for females aged 9 to 26 years since 2006 (bivalent and quadrivalent vaccines) and for males aged 9 to 26 since 2010 (quadrivalent vaccine only) [12]. These vaccines have been shown in clinical trials to be highly efficacious in preventing HPV infection, cervical pre-cancer, and anal, vaginal, penile, and vulvar cancers caused by the HPV types covered in the vaccine [2]. Although their effectiveness against head and neck cancer has not been studied in clinical trials, most experts believe that these vaccines will also provide protection against at least a proportion of these cancers [13,14]. In 2015 the U.S. Food and Drug Administration approved licensure of a 9-valent HPV vaccine that will soon replace the quadrivalent vaccine in the U.S. market [15]. The 9-valent vaccine is licensed for both males and females aged 9 to 26 and is expected to prevent an even higher proportion of HPV-related cancers than earlier HPV vaccines due to the protection against 5 additional oncogenic HPV types [15].

Despite the potential of HPV vaccines to drastically reduce the incidence of HPV-related cancers and other diseases, these vaccines are not being as widely used in the United States as was hoped. The most recent national data from 2015 demonstrates that only 41.9% of girls and 28.1% of boys have received all 3 doses recommended in the HPV vaccine series [16]. This level of vaccine utilization is significantly lower than the Healthy People 2020 goal of 80% coverage [17], and also significantly lower than that of other developed countries such as Australia and the United Kingdom, which have achieved vaccination levels of ~70% among their target adolescent populations [18,19]. In the future, these low vaccination levels will likely be mitigated somewhat by the recent approval from the FDA and recent recommendation from the Advisory Committee on Immunization Practices (ACIP) for only 2 doses of the 9-valent HPV vaccine (spaced 6 to 12 months apart) for adolescents less than 15 years of age [20,21]. Three doses are still recommended for those aged 15 to 26 years.

Provider Communication About HPV Vaccines

How providers communicate with parents and patients about HPV vaccines is a key influential factor driving current U.S. adolescent HPV vaccination levels [22,23]. Numerous studies demonstrate that a provider’s recommendation generally has the largest impact on whether or not an adolescent receives the vaccine, even above that of parent factors such as attitudes and beliefs about the vaccine and patient characteristics such as age and insurance status [23–31]. Moreover, parents consistently cite their adolescent’s provider as one of the most trusted and impactful resources for obtaining vaccine information [22,32].

Unfortunately, research also shows that providers often fail to adequately recommend the HPV vaccine for their patients, especially for 11 to 12 year olds for whom the vaccine is preferentially recommended [33,34]. For example, in a national study of parents done in 2013, not being recommended by a provider was one of the top 5 reasons parents of males and of females aged 11 to 17 gave for not getting their adolescent vaccinated against HPV [35]. Supporting this also is a 2014 study of 776 pediatricians and family medicine providers nationally, in which Gilkey and colleagues found that more than 1 out of 4 providers did not highly endorse the HPV vaccine for 11 to 12 year olds despite this having been the recommended practice from ACIP for the prior 8 years for girls and 4 years for boys. This is in comparison to the other adolescents vaccines that were reported in the same study as being endorsed highly by these providers > 95% of the time [36].

Recognizing that providers’ HPV vaccine recommendations are often suboptimal, researchers have begun to define what components comprise “high-quality” HPV vaccine recommendations. This has been operationalized by one research group as (1) timeliness—routinely recommending the vaccine starting when the patient is ≤ 12 years; (2) consistency—recommending the vaccine for all eligible adolescents as opposed to an approach based on providers’ perception of their patients’ risk for HPV infection; (3) urgency—recommending that the vaccine be given on the same day the vaccine is being discussed, rather than offering the option of getting it at a future visit; and (4) strength—using language that clearly conveys that the provider believes the vaccine is very important for the adolescent to receive. A national study of primary care providers done in 2014 examined how frequently these quality components were implemented [37]. The results were startling and discouraging. Nearly half of providers (49%) reported they usually recommended that 11 to 12 year olds get the vaccine at a later visit, 41% used a risk-based approach for deciding when to recommend the vaccine, 27% did not tell the parents the vaccine was “very or extremely important,” and a large proportion did not start routinely recommending the vaccine before the age of 13 (39% for male patients and 25% for females) [37].

Much research has now accumulated to explain the underlying reasons why providers may not give consistent and high-quality HPV vaccine recommendations to all eligible adolescents [22]. Issues such as providers’ own knowledge about HPV-related diseases, personal beliefs about the vaccine’s safety and necessity, concern that a discussion about the vaccine will necessitate a discussion about adolescent sexuality with the parent, belief that parents will not want their child vaccinated if asked, perceptions that a provider can adequately select those patients most “in need” of HPV vaccination, and concern that raising the vaccine discussion with vaccine-hesitant parents will result in prolonged discussions have been shown to impact whether and how providers communicate about HPV vaccination during clinical visits [22,36–45]. Now that these barriers have been defined and described, there is a great need to use this knowledge to develop and evaluate interventions that help to mitigate these barriers and improve providers’ vaccine communication abilities. Such interventions are needed not only for HPV, but for all vaccines [46,47].

Possible Strategies for Helping Providers Communicate About HPV Vaccines

Before discussing these interventions, it is worth noting that several of the passive and active strategies have been shown in clinical trials to improve adolescent HPV vaccination rates. Although these are beyond the scope of this article, inclusion of these strategies should certainly be considered by any practice as a mechanism to increase vaccination levels, especially given that the most successful interventions to improve vaccination levels consist of multiple components [48]. Also useful is a recently described “taxonomy of vaccine communication interventions” that provides additional perspective on the scope and complexity of interventions to improve vaccine delivery [49]. There are several well-written review articles that describe interventions that focus on passive and active strategies at the practice or community level [50–52].

Interpersonal Communication Strategies Shown to Increase HPV Vaccination

Presumptive Communication

One of the first studies to examine the specific “way” in which providers communicate about vaccines focused not on HPV but rather on young childhood vaccines. In 2013 Opel and colleagues performed a study in which they taped clinical encounters between a pediatrician and a parent of a child aged 1 to 19 months [53]. Of the 111 encounters recorded, 50% of parents were classified as vaccine hesitant. Parents were aware they were being taped but not aware that the overall purpose of the study was to examine providers’ communication related to vaccination. The researchers found that providers generally used one of 2 communication styles to introduce the vaccine discussion. The first, called the “presumptive” style, assumed that parents would agree to vaccination and presented the vaccines as routine (ie, “We have to do some shots today”). The second style, called “participatory,” was more parent-oriented and used language suggesting shared decision-making (ie, “So what do you want to do about shots today?”). The study showed that the odds of resisting the provider’s vaccine recommendations were significantly higher when providers used a participatory approach than a presumptive one, suggesting that even small changes in language can have a major impact on the likelihood of vaccination. However, given the study design, causality between providers’ recommendation style and parents vaccination decisions could not be delineated.

In 2015 Moss and colleagues performed a study that examined the use of these 2 communication styles with regard to HPV vaccination [54]. This study used data from the 2010 National Immunization Survey–Teen, a national survey on childhood vaccination that includes provider verification of vaccines given [16]. Researchers categorized provider vaccine communication styles into “provider-driven,” which was similar to the presumptive style described Opel, and “patient-driven,” which was similar to Opel’s permissive style. Parents who received a more provider-driven style of HPV vaccine recommendation were far more likely to have allowed their adolescent to be vaccinated than those receiving patient-driven recommendations [54]. Further supporting this communication approach are results from a qualitative study done by Hughes and colleagues in which triads of mothers, adolescents, and providers were interviewed after a preventive care visit to assess the communication that occurred regarding HPV vaccination [39]. Providers’ communication style was categorized into 1 of 3 groups: paternalistic (clinician makes the vaccination decision and communicates this to the family); informed (patient and family gathers information from the clinician and other sources to reach a vaccination decision); and shared (medical and personal information is exchanged between the provider and family and then a decision is reached jointly). Providers who typically adopted the paternalistic approach perceived that they had the highest success in convincing parents to vaccinate—a perception that was confirmed in quantitative assessments of vaccination status among adolescents in the study sample [39]. Our own research demonstrates that learning and implementing a presumptive/paternalistic HPV vaccine recommendation style is easy for primary care providers to do and is perceived as often shortening the time taken during clinical visits to discuss the vaccine [55,56]. Thus, providers should consider opening the HPV vaccine conversation using this approach, and then turning to some of the other communication techniques described below when met with parental resistance or questions.

Motivational Interviewing

A second communication technique that seems effective for promoting HPV vaccination, especially for vaccine hesitant parents, is motivational interviewing. Motivational interviewing describes a communication technique in which the provider leverages a parents’ or patients’ intrinsic motivation to engage in a preferred health behavior [57]. Motivational interviewing was originally developed to combat substance abuse [58,59] but has subsequently been successfully applied to a number of other health issues [60–64]. There is growing interest from public health and medical providers in using this technique for increasing vaccination [39,65–68]. Our research group performed a large, cluster-randomized controlled trial of 16 pediatric and family medicine clinics to examine the impact of a provider communication “toolkit” on adolescent HPV vaccine series initiation and completion [50,69]. The toolkit consisted of motivational interviewing training for providers related to HPV vaccination and training on 3 tangible resources providers could also use with parents—an HPV fact sheet, an HPV vaccine decision aid, and an educational website. Results from the study demonstrated that motivational interviewing was the toolkit component most widely utilized by providers and was also perceived as being the most useful. More importantly, HPV vaccine series initiation levels were significantly higher among adolescents in practices receiving the toolkit than in control practices. There was no impact on HPV vaccine series completion (unpublished results). The usefulness of motivational interviewing for vaccination is further supported by a small study in which community pharmacists receiving motivational interviewing training for adult vaccination reported significantly higher patient readiness to receive vaccines following their interaction with the pharmacist than those who did not receive the training [70]. Finally, Perkins et al performed a cluster randomized controlled trial that evaluated the impact of a provider-focused intervention on adolescent HPV vaccination rates. The intervention included frequent provider support meetings, education on HPV infection and vaccination, feedback on providers’ individual HPV vaccination rates, provider incentives, and “basic motivational interviewing principles with vaccine-hesitant parents.” HPV vaccination series initiation and completion rates were significantly higher in intervention practices than controls, and this effect was sustained for at least 6 months after the active intervention period was over [67]. However, it was unknown how much the motivational interviewing contributed to these results. Based on the above information, and the long history of success of motivational interviewing for improving patient compliance with other recommended health behaviors, this technique appears to have a reasonable evidence base and should be considered for communicating with families that express resistance to HPV vaccination.

Personalized Communication

Parents’ reasons for not having their adolescent vaccinated against HPV are often complex and multifactorial [71,72]. Personalized approaches are needed to account for each parent’s unique informational needs, beliefs, and prior experiences [65]. Unfortunately, given the short amount of time allotted for clinical visits, it is often difficult to provide adequate information to parents during these encounters [73–75]. Indeed, concern about prolonged HPV vaccine discussions has been identified as an important barrier for providers that cause some to forgo recommending the vaccine [36,75].

One potential solution to this issue is to leverage technology in the form of web-based interventions that use software to tailor materials to each individual’s unique informational needs. Feasibility for this idea comes from the knowledge that many parents already use the web to research health issues related to their children [76], and that doctors’ offices are increasingly using patient portals and other web-based resources to help parents prepare for upcoming visits, especially those focused on health maintenance [77,78]. Tailored messaging interventions have been shown across populations and health issues to generally result in superior adherence with health behaviors when compared to untailored controls [79–82]. Several researchers have thus begun exploring whether such a personalized communication strategy may be similarly effective for adolescent HPV vaccination [50,83–85]. As an example, Maertens and colleagues used community-based participatory research techniques to develop a web-based tailored messaging intervention for Latinos regarding HPV vaccination [86]. A subsequent randomized controlled trial of the intervention in over 1200 parents of adolescents and young adults demonstrated that the intervention improved participants’ intentions to vaccinate compared to usual care [87], and among adolescents, higher HPV vaccine series initiation levels (unpublished data). Although additional work is needed to understand the feasibility of implementing such an intervention more broadly, additional support for the usefulness of a tailored messaging approach comes from a study of female university students that demonstrated higher HPV vaccination intentions after exposure to tailored information compared to untailored information. However, the impact on actual HPV vaccine utilization was not measured in the study [84]. Contrasting results were found in a different study of university students where researchers failed to find an impact of message tailoring on HPV vaccination utilization. However, this study was limited by a low response rate (~50%) to the follow up survey where vaccination status was assessed, and also by overall low levels of HPV vaccine initiation among the entire study sample (8%) [85]. Given the low number of studies in this area, and some conflicting data, additional research is needed to better understand the impact of personalized communication on HPV vaccination levels. However, results from these studies suggest that a modest benefit may be achieved with this approach, especially if coupled with other, evidence-based, clinic-level interventions to promote vaccination (eg, vaccine reminders, extended office hours), as is suggested by the Task Force on Community Preventive Services [48].

Focusing Communication on Cancer Prevention

HPV vaccines are unique in that they are only 1 of 2 vaccines for cancer prevention (the other being hepatitis B). Provider and parent surveys suggest that while most providers do mention cancer prevention when discussing HPV vaccines [40,88,89], this may be more commonly done with female patients than males [22]. Focusing on cancer prevention rather than sexual transmissibility is a communication technique suggested by the Centers for Disease Control and Prevention (CDC) as many parents cite this aspect of the vaccine as one of the most compelling reasons for vaccinating [45,90]. CDC’s “You are the Key” program [91] uses cancer prevention as a central theme in their physician and patient communication materials, based on significant prior market research on the acceptability and impact of such messages among parents and providers. In 2016 Malo and colleagues tested the potential impact of brief messages related to HPV vaccination, including cancer prevention messages, among a national sample of 776 medical providers and 1504 parents of adolescents [92]. In addition to their potential to motivate parents to vaccination, associations between parental endorsement of each message and their adolescent’s vaccination status were also examined. The cancer prevention messages were among those most highly endorsed by both parents and providers as being motivating for parents to get their adolescent vaccinated. More importantly, among parents these endorsements were associated with a significantly higher likelihood of the adolescent having been vaccinated against HPV. Interestingly, one of the briefest messages in the study, “I [the physician] strongly believe in the importance of this cancer preventing vaccine for [child’s name],” was perceived as the most persuasive message by parents.

Further support for the positive impact of framing HPV vaccines primarily as cancer prevention comes from another national study of 1495 parents of 11 to 17 year olds that examined 3 measures of quality of their adolescent provider’s HPV vaccine recommendation, and the relationship between recommendation quality and likelihood of adolescent HPV vaccination [40]. The 3 quality indicators assessed included providing information about cancer prevention, encouraging the vaccine “strongly,” and recommending it be given on the same day as it was being discussed. While 49% of parents reported receiving no HPV vaccine recommendation from their adolescents’ provider, of those that did, 86% received a cancer prevention message. Parents who had been given high quality recommendations that included either 2 or 3 of the quality indicator measures had over 9 times the odds of vaccine series initiation and 3 times the odds of vaccine series follow through than those who had not received any recommendation, and also significantly higher odds of vaccination than parents who had received low quality recommendations (ie, included only 1 indicator). Taken together, these results suggest that focusing discussions about HPV vaccines on their ability to prevent cancer is likely to be persuasive for some parents.

Strategies That Are Promising But Not Thoroughly Tested

Helping Parents Create Vaccination Plans

A recent commentary suggested that instead of focusing on changing beliefs or “educating” parents and patients about the need for a given vaccine, perhaps a better way to craft interventions for increasing vaccination is to focus on structuring the environment to make vaccination “easy” [93,94]. Examples of this include strategies such as extended office hours and making the vaccine available in other locations such as schools and pharmacies, both of which have been shown in some populations and settings to improve vaccine utilization [48,95]. One aspect of structuring a vaccine-conducive environment that relates to provider communication is helping parents create “implementation intentions” for future vaccination visits. In its most obvious form, this would mean providers provide office resources that facilitate making an appointment for the next dose in the HPV vaccine series during a clinic visit where the first dose was provided. But such an approach could also potentially extend to parents who are on the fence about the vaccine—to make an appointment before the parent leaves the office with an unvaccinated child to either re-discuss the vaccine in the future or to actually start the vaccine series. Support for such a strategy comes primarily from the social sciences, which suggest that implementation intentions work by increasing attention to specific cues to action, making it more likely that that the cue will be acted upon [96–98]. Creating implementation intentions has been shown to be helpful for improving adherence with a variety of health behaviors [99–105], and there is a growing evidence base related to how implementation intentions may facilitate vaccination specifically. For example Vet and colleagues performed a randomized controlled trial among 616 men who have sex with men with either strong or weak intentions to receive the hepatitis B vaccine [106]. Half of the participants were asked to create an implementation intention plan where they described when, where and how they would obtain the vaccine. Those in the control arm were not given this prompt. Regardless of whether their initial vaccination intention was weak or strong, those who had been asked to create an implementation plan had more than double the likelihood of actually getting the vaccine than participants who did not receive the implementation plan prompt. Similarly, a study of influenza vaccination rates among corporate employees found that those who were asked to write down the day and time they planned to go to employee health to get the free vaccine were somewhat more likely (4% higher) to be vaccinated than those who did not receive this prompt [107]. In addition, a study of elderly individuals found that influenza vaccination rates were significantly higher among those who had received “action instructions” on how, when and where to get the vaccine than those who did not [108]. These studies suggest that helping parents craft a definitive follow-up plan regarding vaccination could have a significant impact on vaccination rates—particularly for vaccines like HPV that require multiple doses.

Treating all Adolescent Vaccines the Same

Prior research has demonstrated that providers often communicate differently about HPV vaccines than other adolescent vaccines such as the tetanus-diphtheria-pertussis (Tdap) and meningococcal (MCV) vaccines [22,36]. Providers often tend to discuss the HPV vaccine last among these 3 vaccines, provide weaker endorsements of the vaccine, and pre-emptively give much more detail about the HPV compared to the other vaccines, even in the absence of a parent’s request for additional information [36,39,41]. The CDC and the American Academy of Pediatrics now suggest putting HPV at the beginning or middle of the list of vaccines recommended to the adolescent (ie, “HPV, Tdap and MCV”), and treating all recommended vaccines equivalently in terms of the level of detail provided to parents in the absence of a parent’s request for more information [109,110]. Through these suggestions have face validity, their specific impact on HPV vaccination rates, and on patient and provider satisfaction with the visit have yet to be evaluated.

Strategies that Probably Don’t Work

Presenting Myths and Facts

Research related to promoting other vaccines provides insight into communication activities that probably would not work well for promoting HPV vaccination. A 2012 study by Nyhan and colleagues examined the impact of 2 different messages related to influenza vaccines on participants’ beliefs about the vaccine’s safety and intentions to get vaccinated [111]. One group received information to correct the commonly held belief that influenza vaccine can cause the flu while the other received information about the risks associated with contracting an influenza infection. While the correction of myths did improve participants’ perceptions of the vaccine’s safety, information about influenza dangers did not. Neither message impacted intentions to vaccinate in the study subjects overall. However, in sub-analyses the correction of myths actually decreased intentions to vaccinate among those with high baseline levels of concern about the vaccine’s side effects—that is, among those most concerned that the flu vaccine can give someone the flu, correcting this myth actually decreased the likelihood that they would receive the vaccine. Similar findings have been reported in other studies related to vaccination [112–114], and suggest that the “threat” generated by providing information opposing a person’s beliefs may actually entrench these beliefs further as part of the threat response—a phenomenon known as attitude polarization [115]. These results also are consistent with the concept of negativity bias, which posits that negative information influences people’s risk perceptions more than positive information, and that the more strongly a risk is attempted to be negated, the lower the effectiveness and perceived trust of the information [116].

Using Fear Appeals

One tactic that has been suggested by some as a way to promote vaccination is to provide graphic depictions of the possible sequelae of vaccine-preventable diseases. The thought behind this idea is that because vaccination is so successful, most parents will have never experienced significant impacts from vaccine preventable diseases that, in the past, had been a major motivator for parents to vaccinate. Thus, in order to counter beliefs about “controversial” issues like vaccination, highly emotionally compelling and engaging information may be especially useful. This is a common tactic used by anti-vaccination groups to spread their own messages [117]. However, several studies suggested that using “fear appeals” (aka scare tactics) such as this to promote vaccination can actually have a negative effect on vaccination intentions. For example, in a 2011 study of a nationally representative sample of parents of children < 18 years, 4 different message formats were tested for their impact on parental intentions to vaccinate a future child with the measles-mumps-rubella vaccine (MMR) [113]. Message formats included correcting the misinformation that MMR causes autism, presenting information on MMR-related disease risk, providing a dramatic narrative about a child endangered by measles, and showing pictures of infants affected by these diseases. Counter to the study’s hypotheses, the dramatic narrative message actually increased parents’ perceptions that MMR vaccines had serious side effects, and the pictures increased parents’ belief that the MMR vaccine could cause autism. These counter-intuitive results are consistent with other studies that have examined the impact of message framing on adults’ vaccination intentions for HPV and influenza [108,118,119]. Taken together, fear appeals seem unlikely to sway many hesitant parents towards HPV vaccination.

Looking Into the Future

Moving forward, additional interventions to improve providers’ ability to communicate with families about HPV vaccination will undoubtedly be developed. A major area of interest in this regard is leveraging the power of technology and the internet, including using social media, mobile technologies, and online interventions to augment the provider/parent interaction that occurs during the clinical visit [50,120]. Web-based approaches have the benefit of generally being low cost and easy to disseminate to large populations. Such interventions have already been developed for a number of other health issues, some of which have proven effective [121,122]. However, use of the internet to promote healthy behaviors in general, and vaccination specifically, is still in its infancy. There is still much to be learned about how to create effective web-based tools, how to engage patients with them, and how to assess their impact on health outcomes [123].

Another interesting area for future research is identifying psychological “levers” to motivate parents’ vaccination intentions [94]. One example is focusing on using parents’ values (ie, protecting my child from harm) as an intervention target rather than beliefs or attitudes. This is because values tend to be inherent and static over time, compared to beliefs and attitudes, which are subject to change depending on the context [124]. Prior research has shown that interventions that leverage values rather than facts can be an effective way to overcome beliefs that are highly emotional or controversial, and that individuals are more likely to trust sources and individuals with shared values than those without [125], suggesting that this may be a useful way to motivate parents toward vaccinating their children. Self-affirmation is another example of a psychological lever that has a significant evidence base from the social science literature as a helpful tool for moving patients towards a desired health behavior [126,127], but it has not been extensively applied to the field of vaccination. Researchers in the field of vaccine delivery are increasingly recognizing the potential value of these unique intervention approaches [101,128–134], and it may be fruitful in the future to more closely examine the efficacy of interventions that target things like values, self-affirmation or other psychological levers to change parents’ HPV vaccination behaviors.

A final notable area for intervention research related to HPV vaccination is the use of video games. Although not likely to be used directly during patient visits, this strategy could be conceptualized as a potential way to augment the information conveyed to a parent by a provider directly during a clinical encounter. A meta-analysis from 2016 identified 16 different “serious” video games that were used to train and educate users about specific vaccine preventable diseases (usually influenza, none for HPV) and the need for vaccination [135]. In many of them, the objective of the game was to protect a virtual community from a vaccine preventable disease and/or manage outbreaks. Only 2 of the games evaluated outcomes in the short term (ie, at the time the game was being played). None have evaluated longer-term impacts such as vaccination intention or utilization. In the era of “plugged in” parents and adolescents, video games represent a unique but understudied mechanism for helping providers “communicate,” albeit indirectly, with families about the need for vaccination. Imagine providing a prescription to an HPV-vaccine hesitant family to “go play Zombie Wars HPV!” One would expect the curiosity factor alone would result in significant engagement with this intervention tool.

Conclusion

With persistently lagging HPV vaccination rates among U.S. adolescents, there is a growing need for effective interventions to improve adolescent HPV vaccine utilization. How providers communicate with families is one of the most influential factors in parents’ vaccination decisions. Emerging research is beginning to delineate potentially effective communication techniques such as presumptive approaches to making the vaccine recommendation, framing the vaccine as cancer preventing, and using motivational interviewing and personalized messaging when met with parental vaccine resistance. Moving forward the list of evidence-based interventions to improve providers’ HPV vaccine communication is likely to grow, and to increasingly leverage technology based solutions. However, given the complexities of the vaccination decision [136] and the ever growing spread of vaccine hesitancy [137], it is unlikely that a single intervention approach will be effective for getting adolescent HPV vaccine levels up to the national goal of 80% coverage. As has been recognized in the past, the most effective interventions for HPV vaccination in the future are likely to be multicomponent, including not only provider communication strategies but also clinic-, community-, and parent-level interventions [48].

Corresponding author: Amanda Dempsey, MD, PhD, MPH, 13199 East Montview Blvd, Suite 300, Aurora, CO 80045, [email protected].

Financial disclosures: None reported.

From the University of Colorado Denver, Aurora, CO.

Abstract

• Objective: To provide evidence-based guidance on strategies that are likely or unlikely to be successful in navigating HPV vaccine conversations with patients and parents.
• Methods: Nonsystematic review of the literature.
• Results: This review highlights some of the most recent innovations in provider HPV vaccine communication and describes provider communication strategies that have been found to improve adolescent vaccination rates in rigorous scientific studies. Promising strategies for which additional research is needed and strategies that probably do not work are also described. 
• Conclusion: By understanding what works, what may work, and what not to do when it comes to communicating with families about HPV vaccines, providers can be better prepared for maximizing the impact they can have on adolescent HPV vaccination rates.

Key words: human papillomavirus; vaccine hesitancy; health communication; parents; immunization.

In the United States, more than 14 million people newly acquire genital human papillomavirus (HPV) annually, and 75 million Americans are infected at any given time [1]. As the most common sexually transmitted disease, more than 80% of sexually active U.S. adults are estimated to be infected with HPV by the age of 50 [1,2]. Although the majority of infections are “silent” and resolve without clinical sequelae, a proportion of infected individuals will go on to develop HPV-related diseases. In women, these include cervical cancer and pre-cancer (ie, abnormal Pap smears); cancers of the vagina, vulva, anus, and oropharynx; and genital warts [3]. Males also bear a high burden of HPV-related disease in the form of penile, anal, and oropharyngeal cancers, as well as genital warts [3]. While once thought of as primarily a “woman’s disease” [4], recent research demonstrates men are also significantly impacted by HPV—particularly in the form of oropharyngeal cancers, which are 2 to 3 times more common in men than in women [5]. In fact, it is estimated by the year 2020 more men will die of HPV-related oropharyngeal cancer than women will die of cervical cancer [6,7]. The combined cost of HPV-associated cancers and other conditions is estimated to be $8 billion per year in the United States [8–11].

HPV Vaccines

Effective HPV vaccines have been available for females aged 9 to 26 years since 2006 (bivalent and quadrivalent vaccines) and for males aged 9 to 26 since 2010 (quadrivalent vaccine only) [12]. These vaccines have been shown in clinical trials to be highly efficacious in preventing HPV infection, cervical pre-cancer, and anal, vaginal, penile, and vulvar cancers caused by the HPV types covered in the vaccine [2]. Although their effectiveness against head and neck cancer has not been studied in clinical trials, most experts believe that these vaccines will also provide protection against at least a proportion of these cancers [13,14]. In 2015 the U.S. Food and Drug Administration approved licensure of a 9-valent HPV vaccine that will soon replace the quadrivalent vaccine in the U.S. market [15]. The 9-valent vaccine is licensed for both males and females aged 9 to 26 and is expected to prevent an even higher proportion of HPV-related cancers than earlier HPV vaccines due to the protection against 5 additional oncogenic HPV types [15].

Despite the potential of HPV vaccines to drastically reduce the incidence of HPV-related cancers and other diseases, these vaccines are not being as widely used in the United States as was hoped. The most recent national data from 2015 demonstrates that only 41.9% of girls and 28.1% of boys have received all 3 doses recommended in the HPV vaccine series [16]. This level of vaccine utilization is significantly lower than the Healthy People 2020 goal of 80% coverage [17], and also significantly lower than that of other developed countries such as Australia and the United Kingdom, which have achieved vaccination levels of ~70% among their target adolescent populations [18,19]. In the future, these low vaccination levels will likely be mitigated somewhat by the recent approval from the FDA and recent recommendation from the Advisory Committee on Immunization Practices (ACIP) for only 2 doses of the 9-valent HPV vaccine (spaced 6 to 12 months apart) for adolescents less than 15 years of age [20,21]. Three doses are still recommended for those aged 15 to 26 years.

Provider Communication About HPV Vaccines

How providers communicate with parents and patients about HPV vaccines is a key influential factor driving current U.S. adolescent HPV vaccination levels [22,23]. Numerous studies demonstrate that a provider’s recommendation generally has the largest impact on whether or not an adolescent receives the vaccine, even above that of parent factors such as attitudes and beliefs about the vaccine and patient characteristics such as age and insurance status [23–31]. Moreover, parents consistently cite their adolescent’s provider as one of the most trusted and impactful resources for obtaining vaccine information [22,32].

Unfortunately, research also shows that providers often fail to adequately recommend the HPV vaccine for their patients, especially for 11 to 12 year olds for whom the vaccine is preferentially recommended [33,34]. For example, in a national study of parents done in 2013, not being recommended by a provider was one of the top 5 reasons parents of males and of females aged 11 to 17 gave for not getting their adolescent vaccinated against HPV [35]. Supporting this also is a 2014 study of 776 pediatricians and family medicine providers nationally, in which Gilkey and colleagues found that more than 1 out of 4 providers did not highly endorse the HPV vaccine for 11 to 12 year olds despite this having been the recommended practice from ACIP for the prior 8 years for girls and 4 years for boys. This is in comparison to the other adolescents vaccines that were reported in the same study as being endorsed highly by these providers > 95% of the time [36].

Recognizing that providers’ HPV vaccine recommendations are often suboptimal, researchers have begun to define what components comprise “high-quality” HPV vaccine recommendations. This has been operationalized by one research group as (1) timeliness—routinely recommending the vaccine starting when the patient is ≤ 12 years; (2) consistency—recommending the vaccine for all eligible adolescents as opposed to an approach based on providers’ perception of their patients’ risk for HPV infection; (3) urgency—recommending that the vaccine be given on the same day the vaccine is being discussed, rather than offering the option of getting it at a future visit; and (4) strength—using language that clearly conveys that the provider believes the vaccine is very important for the adolescent to receive. A national study of primary care providers done in 2014 examined how frequently these quality components were implemented [37]. The results were startling and discouraging. Nearly half of providers (49%) reported they usually recommended that 11 to 12 year olds get the vaccine at a later visit, 41% used a risk-based approach for deciding when to recommend the vaccine, 27% did not tell the parents the vaccine was “very or extremely important,” and a large proportion did not start routinely recommending the vaccine before the age of 13 (39% for male patients and 25% for females) [37].

Much research has now accumulated to explain the underlying reasons why providers may not give consistent and high-quality HPV vaccine recommendations to all eligible adolescents [22]. Issues such as providers’ own knowledge about HPV-related diseases, personal beliefs about the vaccine’s safety and necessity, concern that a discussion about the vaccine will necessitate a discussion about adolescent sexuality with the parent, belief that parents will not want their child vaccinated if asked, perceptions that a provider can adequately select those patients most “in need” of HPV vaccination, and concern that raising the vaccine discussion with vaccine-hesitant parents will result in prolonged discussions have been shown to impact whether and how providers communicate about HPV vaccination during clinical visits [22,36–45]. Now that these barriers have been defined and described, there is a great need to use this knowledge to develop and evaluate interventions that help to mitigate these barriers and improve providers’ vaccine communication abilities. Such interventions are needed not only for HPV, but for all vaccines [46,47].

Possible Strategies for Helping Providers Communicate About HPV Vaccines

Before discussing these interventions, it is worth noting that several of the passive and active strategies have been shown in clinical trials to improve adolescent HPV vaccination rates. Although these are beyond the scope of this article, inclusion of these strategies should certainly be considered by any practice as a mechanism to increase vaccination levels, especially given that the most successful interventions to improve vaccination levels consist of multiple components [48]. Also useful is a recently described “taxonomy of vaccine communication interventions” that provides additional perspective on the scope and complexity of interventions to improve vaccine delivery [49]. There are several well-written review articles that describe interventions that focus on passive and active strategies at the practice or community level [50–52].

Interpersonal Communication Strategies Shown to Increase HPV Vaccination

Presumptive Communication

One of the first studies to examine the specific “way” in which providers communicate about vaccines focused not on HPV but rather on young childhood vaccines. In 2013 Opel and colleagues performed a study in which they taped clinical encounters between a pediatrician and a parent of a child aged 1 to 19 months [53]. Of the 111 encounters recorded, 50% of parents were classified as vaccine hesitant. Parents were aware they were being taped but not aware that the overall purpose of the study was to examine providers’ communication related to vaccination. The researchers found that providers generally used one of 2 communication styles to introduce the vaccine discussion. The first, called the “presumptive” style, assumed that parents would agree to vaccination and presented the vaccines as routine (ie, “We have to do some shots today”). The second style, called “participatory,” was more parent-oriented and used language suggesting shared decision-making (ie, “So what do you want to do about shots today?”). The study showed that the odds of resisting the provider’s vaccine recommendations were significantly higher when providers used a participatory approach than a presumptive one, suggesting that even small changes in language can have a major impact on the likelihood of vaccination. However, given the study design, causality between providers’ recommendation style and parents vaccination decisions could not be delineated.

In 2015 Moss and colleagues performed a study that examined the use of these 2 communication styles with regard to HPV vaccination [54]. This study used data from the 2010 National Immunization Survey–Teen, a national survey on childhood vaccination that includes provider verification of vaccines given [16]. Researchers categorized provider vaccine communication styles into “provider-driven,” which was similar to the presumptive style described Opel, and “patient-driven,” which was similar to Opel’s permissive style. Parents who received a more provider-driven style of HPV vaccine recommendation were far more likely to have allowed their adolescent to be vaccinated than those receiving patient-driven recommendations [54]. Further supporting this communication approach are results from a qualitative study done by Hughes and colleagues in which triads of mothers, adolescents, and providers were interviewed after a preventive care visit to assess the communication that occurred regarding HPV vaccination [39]. Providers’ communication style was categorized into 1 of 3 groups: paternalistic (clinician makes the vaccination decision and communicates this to the family); informed (patient and family gathers information from the clinician and other sources to reach a vaccination decision); and shared (medical and personal information is exchanged between the provider and family and then a decision is reached jointly). Providers who typically adopted the paternalistic approach perceived that they had the highest success in convincing parents to vaccinate—a perception that was confirmed in quantitative assessments of vaccination status among adolescents in the study sample [39]. Our own research demonstrates that learning and implementing a presumptive/paternalistic HPV vaccine recommendation style is easy for primary care providers to do and is perceived as often shortening the time taken during clinical visits to discuss the vaccine [55,56]. Thus, providers should consider opening the HPV vaccine conversation using this approach, and then turning to some of the other communication techniques described below when met with parental resistance or questions.

Motivational Interviewing

A second communication technique that seems effective for promoting HPV vaccination, especially for vaccine hesitant parents, is motivational interviewing. Motivational interviewing describes a communication technique in which the provider leverages a parents’ or patients’ intrinsic motivation to engage in a preferred health behavior [57]. Motivational interviewing was originally developed to combat substance abuse [58,59] but has subsequently been successfully applied to a number of other health issues [60–64]. There is growing interest from public health and medical providers in using this technique for increasing vaccination [39,65–68]. Our research group performed a large, cluster-randomized controlled trial of 16 pediatric and family medicine clinics to examine the impact of a provider communication “toolkit” on adolescent HPV vaccine series initiation and completion [50,69]. The toolkit consisted of motivational interviewing training for providers related to HPV vaccination and training on 3 tangible resources providers could also use with parents—an HPV fact sheet, an HPV vaccine decision aid, and an educational website. Results from the study demonstrated that motivational interviewing was the toolkit component most widely utilized by providers and was also perceived as being the most useful. More importantly, HPV vaccine series initiation levels were significantly higher among adolescents in practices receiving the toolkit than in control practices. There was no impact on HPV vaccine series completion (unpublished results). The usefulness of motivational interviewing for vaccination is further supported by a small study in which community pharmacists receiving motivational interviewing training for adult vaccination reported significantly higher patient readiness to receive vaccines following their interaction with the pharmacist than those who did not receive the training [70]. Finally, Perkins et al performed a cluster randomized controlled trial that evaluated the impact of a provider-focused intervention on adolescent HPV vaccination rates. The intervention included frequent provider support meetings, education on HPV infection and vaccination, feedback on providers’ individual HPV vaccination rates, provider incentives, and “basic motivational interviewing principles with vaccine-hesitant parents.” HPV vaccination series initiation and completion rates were significantly higher in intervention practices than controls, and this effect was sustained for at least 6 months after the active intervention period was over [67]. However, it was unknown how much the motivational interviewing contributed to these results. Based on the above information, and the long history of success of motivational interviewing for improving patient compliance with other recommended health behaviors, this technique appears to have a reasonable evidence base and should be considered for communicating with families that express resistance to HPV vaccination.

Personalized Communication

Parents’ reasons for not having their adolescent vaccinated against HPV are often complex and multifactorial [71,72]. Personalized approaches are needed to account for each parent’s unique informational needs, beliefs, and prior experiences [65]. Unfortunately, given the short amount of time allotted for clinical visits, it is often difficult to provide adequate information to parents during these encounters [73–75]. Indeed, concern about prolonged HPV vaccine discussions has been identified as an important barrier for providers that cause some to forgo recommending the vaccine [36,75].

One potential solution to this issue is to leverage technology in the form of web-based interventions that use software to tailor materials to each individual’s unique informational needs. Feasibility for this idea comes from the knowledge that many parents already use the web to research health issues related to their children [76], and that doctors’ offices are increasingly using patient portals and other web-based resources to help parents prepare for upcoming visits, especially those focused on health maintenance [77,78]. Tailored messaging interventions have been shown across populations and health issues to generally result in superior adherence with health behaviors when compared to untailored controls [79–82]. Several researchers have thus begun exploring whether such a personalized communication strategy may be similarly effective for adolescent HPV vaccination [50,83–85]. As an example, Maertens and colleagues used community-based participatory research techniques to develop a web-based tailored messaging intervention for Latinos regarding HPV vaccination [86]. A subsequent randomized controlled trial of the intervention in over 1200 parents of adolescents and young adults demonstrated that the intervention improved participants’ intentions to vaccinate compared to usual care [87], and among adolescents, higher HPV vaccine series initiation levels (unpublished data). Although additional work is needed to understand the feasibility of implementing such an intervention more broadly, additional support for the usefulness of a tailored messaging approach comes from a study of female university students that demonstrated higher HPV vaccination intentions after exposure to tailored information compared to untailored information. However, the impact on actual HPV vaccine utilization was not measured in the study [84]. Contrasting results were found in a different study of university students where researchers failed to find an impact of message tailoring on HPV vaccination utilization. However, this study was limited by a low response rate (~50%) to the follow up survey where vaccination status was assessed, and also by overall low levels of HPV vaccine initiation among the entire study sample (8%) [85]. Given the low number of studies in this area, and some conflicting data, additional research is needed to better understand the impact of personalized communication on HPV vaccination levels. However, results from these studies suggest that a modest benefit may be achieved with this approach, especially if coupled with other, evidence-based, clinic-level interventions to promote vaccination (eg, vaccine reminders, extended office hours), as is suggested by the Task Force on Community Preventive Services [48].

Focusing Communication on Cancer Prevention

HPV vaccines are unique in that they are only 1 of 2 vaccines for cancer prevention (the other being hepatitis B). Provider and parent surveys suggest that while most providers do mention cancer prevention when discussing HPV vaccines [40,88,89], this may be more commonly done with female patients than males [22]. Focusing on cancer prevention rather than sexual transmissibility is a communication technique suggested by the Centers for Disease Control and Prevention (CDC) as many parents cite this aspect of the vaccine as one of the most compelling reasons for vaccinating [45,90]. CDC’s “You are the Key” program [91] uses cancer prevention as a central theme in their physician and patient communication materials, based on significant prior market research on the acceptability and impact of such messages among parents and providers. In 2016 Malo and colleagues tested the potential impact of brief messages related to HPV vaccination, including cancer prevention messages, among a national sample of 776 medical providers and 1504 parents of adolescents [92]. In addition to their potential to motivate parents to vaccination, associations between parental endorsement of each message and their adolescent’s vaccination status were also examined. The cancer prevention messages were among those most highly endorsed by both parents and providers as being motivating for parents to get their adolescent vaccinated. More importantly, among parents these endorsements were associated with a significantly higher likelihood of the adolescent having been vaccinated against HPV. Interestingly, one of the briefest messages in the study, “I [the physician] strongly believe in the importance of this cancer preventing vaccine for [child’s name],” was perceived as the most persuasive message by parents.

Further support for the positive impact of framing HPV vaccines primarily as cancer prevention comes from another national study of 1495 parents of 11 to 17 year olds that examined 3 measures of quality of their adolescent provider’s HPV vaccine recommendation, and the relationship between recommendation quality and likelihood of adolescent HPV vaccination [40]. The 3 quality indicators assessed included providing information about cancer prevention, encouraging the vaccine “strongly,” and recommending it be given on the same day as it was being discussed. While 49% of parents reported receiving no HPV vaccine recommendation from their adolescents’ provider, of those that did, 86% received a cancer prevention message. Parents who had been given high quality recommendations that included either 2 or 3 of the quality indicator measures had over 9 times the odds of vaccine series initiation and 3 times the odds of vaccine series follow through than those who had not received any recommendation, and also significantly higher odds of vaccination than parents who had received low quality recommendations (ie, included only 1 indicator). Taken together, these results suggest that focusing discussions about HPV vaccines on their ability to prevent cancer is likely to be persuasive for some parents.

Strategies That Are Promising But Not Thoroughly Tested

Helping Parents Create Vaccination Plans

A recent commentary suggested that instead of focusing on changing beliefs or “educating” parents and patients about the need for a given vaccine, perhaps a better way to craft interventions for increasing vaccination is to focus on structuring the environment to make vaccination “easy” [93,94]. Examples of this include strategies such as extended office hours and making the vaccine available in other locations such as schools and pharmacies, both of which have been shown in some populations and settings to improve vaccine utilization [48,95]. One aspect of structuring a vaccine-conducive environment that relates to provider communication is helping parents create “implementation intentions” for future vaccination visits. In its most obvious form, this would mean providers provide office resources that facilitate making an appointment for the next dose in the HPV vaccine series during a clinic visit where the first dose was provided. But such an approach could also potentially extend to parents who are on the fence about the vaccine—to make an appointment before the parent leaves the office with an unvaccinated child to either re-discuss the vaccine in the future or to actually start the vaccine series. Support for such a strategy comes primarily from the social sciences, which suggest that implementation intentions work by increasing attention to specific cues to action, making it more likely that that the cue will be acted upon [96–98]. Creating implementation intentions has been shown to be helpful for improving adherence with a variety of health behaviors [99–105], and there is a growing evidence base related to how implementation intentions may facilitate vaccination specifically. For example Vet and colleagues performed a randomized controlled trial among 616 men who have sex with men with either strong or weak intentions to receive the hepatitis B vaccine [106]. Half of the participants were asked to create an implementation intention plan where they described when, where and how they would obtain the vaccine. Those in the control arm were not given this prompt. Regardless of whether their initial vaccination intention was weak or strong, those who had been asked to create an implementation plan had more than double the likelihood of actually getting the vaccine than participants who did not receive the implementation plan prompt. Similarly, a study of influenza vaccination rates among corporate employees found that those who were asked to write down the day and time they planned to go to employee health to get the free vaccine were somewhat more likely (4% higher) to be vaccinated than those who did not receive this prompt [107]. In addition, a study of elderly individuals found that influenza vaccination rates were significantly higher among those who had received “action instructions” on how, when and where to get the vaccine than those who did not [108]. These studies suggest that helping parents craft a definitive follow-up plan regarding vaccination could have a significant impact on vaccination rates—particularly for vaccines like HPV that require multiple doses.

Treating all Adolescent Vaccines the Same

Prior research has demonstrated that providers often communicate differently about HPV vaccines than other adolescent vaccines such as the tetanus-diphtheria-pertussis (Tdap) and meningococcal (MCV) vaccines [22,36]. Providers often tend to discuss the HPV vaccine last among these 3 vaccines, provide weaker endorsements of the vaccine, and pre-emptively give much more detail about the HPV compared to the other vaccines, even in the absence of a parent’s request for additional information [36,39,41]. The CDC and the American Academy of Pediatrics now suggest putting HPV at the beginning or middle of the list of vaccines recommended to the adolescent (ie, “HPV, Tdap and MCV”), and treating all recommended vaccines equivalently in terms of the level of detail provided to parents in the absence of a parent’s request for more information [109,110]. Through these suggestions have face validity, their specific impact on HPV vaccination rates, and on patient and provider satisfaction with the visit have yet to be evaluated.

Strategies that Probably Don’t Work

Presenting Myths and Facts

Research related to promoting other vaccines provides insight into communication activities that probably would not work well for promoting HPV vaccination. A 2012 study by Nyhan and colleagues examined the impact of 2 different messages related to influenza vaccines on participants’ beliefs about the vaccine’s safety and intentions to get vaccinated [111]. One group received information to correct the commonly held belief that influenza vaccine can cause the flu while the other received information about the risks associated with contracting an influenza infection. While the correction of myths did improve participants’ perceptions of the vaccine’s safety, information about influenza dangers did not. Neither message impacted intentions to vaccinate in the study subjects overall. However, in sub-analyses the correction of myths actually decreased intentions to vaccinate among those with high baseline levels of concern about the vaccine’s side effects—that is, among those most concerned that the flu vaccine can give someone the flu, correcting this myth actually decreased the likelihood that they would receive the vaccine. Similar findings have been reported in other studies related to vaccination [112–114], and suggest that the “threat” generated by providing information opposing a person’s beliefs may actually entrench these beliefs further as part of the threat response—a phenomenon known as attitude polarization [115]. These results also are consistent with the concept of negativity bias, which posits that negative information influences people’s risk perceptions more than positive information, and that the more strongly a risk is attempted to be negated, the lower the effectiveness and perceived trust of the information [116].

Using Fear Appeals

One tactic that has been suggested by some as a way to promote vaccination is to provide graphic depictions of the possible sequelae of vaccine-preventable diseases. The thought behind this idea is that because vaccination is so successful, most parents will have never experienced significant impacts from vaccine preventable diseases that, in the past, had been a major motivator for parents to vaccinate. Thus, in order to counter beliefs about “controversial” issues like vaccination, highly emotionally compelling and engaging information may be especially useful. This is a common tactic used by anti-vaccination groups to spread their own messages [117]. However, several studies suggested that using “fear appeals” (aka scare tactics) such as this to promote vaccination can actually have a negative effect on vaccination intentions. For example, in a 2011 study of a nationally representative sample of parents of children < 18 years, 4 different message formats were tested for their impact on parental intentions to vaccinate a future child with the measles-mumps-rubella vaccine (MMR) [113]. Message formats included correcting the misinformation that MMR causes autism, presenting information on MMR-related disease risk, providing a dramatic narrative about a child endangered by measles, and showing pictures of infants affected by these diseases. Counter to the study’s hypotheses, the dramatic narrative message actually increased parents’ perceptions that MMR vaccines had serious side effects, and the pictures increased parents’ belief that the MMR vaccine could cause autism. These counter-intuitive results are consistent with other studies that have examined the impact of message framing on adults’ vaccination intentions for HPV and influenza [108,118,119]. Taken together, fear appeals seem unlikely to sway many hesitant parents towards HPV vaccination.

Looking Into the Future

Moving forward, additional interventions to improve providers’ ability to communicate with families about HPV vaccination will undoubtedly be developed. A major area of interest in this regard is leveraging the power of technology and the internet, including using social media, mobile technologies, and online interventions to augment the provider/parent interaction that occurs during the clinical visit [50,120]. Web-based approaches have the benefit of generally being low cost and easy to disseminate to large populations. Such interventions have already been developed for a number of other health issues, some of which have proven effective [121,122]. However, use of the internet to promote healthy behaviors in general, and vaccination specifically, is still in its infancy. There is still much to be learned about how to create effective web-based tools, how to engage patients with them, and how to assess their impact on health outcomes [123].

Another interesting area for future research is identifying psychological “levers” to motivate parents’ vaccination intentions [94]. One example is focusing on using parents’ values (ie, protecting my child from harm) as an intervention target rather than beliefs or attitudes. This is because values tend to be inherent and static over time, compared to beliefs and attitudes, which are subject to change depending on the context [124]. Prior research has shown that interventions that leverage values rather than facts can be an effective way to overcome beliefs that are highly emotional or controversial, and that individuals are more likely to trust sources and individuals with shared values than those without [125], suggesting that this may be a useful way to motivate parents toward vaccinating their children. Self-affirmation is another example of a psychological lever that has a significant evidence base from the social science literature as a helpful tool for moving patients towards a desired health behavior [126,127], but it has not been extensively applied to the field of vaccination. Researchers in the field of vaccine delivery are increasingly recognizing the potential value of these unique intervention approaches [101,128–134], and it may be fruitful in the future to more closely examine the efficacy of interventions that target things like values, self-affirmation or other psychological levers to change parents’ HPV vaccination behaviors.

A final notable area for intervention research related to HPV vaccination is the use of video games. Although not likely to be used directly during patient visits, this strategy could be conceptualized as a potential way to augment the information conveyed to a parent by a provider directly during a clinical encounter. A meta-analysis from 2016 identified 16 different “serious” video games that were used to train and educate users about specific vaccine preventable diseases (usually influenza, none for HPV) and the need for vaccination [135]. In many of them, the objective of the game was to protect a virtual community from a vaccine preventable disease and/or manage outbreaks. Only 2 of the games evaluated outcomes in the short term (ie, at the time the game was being played). None have evaluated longer-term impacts such as vaccination intention or utilization. In the era of “plugged in” parents and adolescents, video games represent a unique but understudied mechanism for helping providers “communicate,” albeit indirectly, with families about the need for vaccination. Imagine providing a prescription to an HPV-vaccine hesitant family to “go play Zombie Wars HPV!” One would expect the curiosity factor alone would result in significant engagement with this intervention tool.

Conclusion

With persistently lagging HPV vaccination rates among U.S. adolescents, there is a growing need for effective interventions to improve adolescent HPV vaccine utilization. How providers communicate with families is one of the most influential factors in parents’ vaccination decisions. Emerging research is beginning to delineate potentially effective communication techniques such as presumptive approaches to making the vaccine recommendation, framing the vaccine as cancer preventing, and using motivational interviewing and personalized messaging when met with parental vaccine resistance. Moving forward the list of evidence-based interventions to improve providers’ HPV vaccine communication is likely to grow, and to increasingly leverage technology based solutions. However, given the complexities of the vaccination decision [136] and the ever growing spread of vaccine hesitancy [137], it is unlikely that a single intervention approach will be effective for getting adolescent HPV vaccine levels up to the national goal of 80% coverage. As has been recognized in the past, the most effective interventions for HPV vaccination in the future are likely to be multicomponent, including not only provider communication strategies but also clinic-, community-, and parent-level interventions [48].

Corresponding author: Amanda Dempsey, MD, PhD, MPH, 13199 East Montview Blvd, Suite 300, Aurora, CO 80045, [email protected].

Financial disclosures: None reported.

From the University of Colorado Denver, Aurora, CO.

Abstract

• Objective: To provide evidence-based guidance on strategies that are likely or unlikely to be successful in navigating HPV vaccine conversations with patients and parents.
• Methods: Nonsystematic review of the literature.
• Results: This review highlights some of the most recent innovations in provider HPV vaccine communication and describes provider communication strategies that have been found to improve adolescent vaccination rates in rigorous scientific studies. Promising strategies for which additional research is needed and strategies that probably do not work are also described. 
• Conclusion: By understanding what works, what may work, and what not to do when it comes to communicating with families about HPV vaccines, providers can be better prepared for maximizing the impact they can have on adolescent HPV vaccination rates.

Key words: human papillomavirus; vaccine hesitancy; health communication; parents; immunization.

In the United States, more than 14 million people newly acquire genital human papillomavirus (HPV) annually, and 75 million Americans are infected at any given time [1]. As the most common sexually transmitted disease, more than 80% of sexually active U.S. adults are estimated to be infected with HPV by the age of 50 [1,2]. Although the majority of infections are “silent” and resolve without clinical sequelae, a proportion of infected individuals will go on to develop HPV-related diseases. In women, these include cervical cancer and pre-cancer (ie, abnormal Pap smears); cancers of the vagina, vulva, anus, and oropharynx; and genital warts [3]. Males also bear a high burden of HPV-related disease in the form of penile, anal, and oropharyngeal cancers, as well as genital warts [3]. While once thought of as primarily a “woman’s disease” [4], recent research demonstrates men are also significantly impacted by HPV—particularly in the form of oropharyngeal cancers, which are 2 to 3 times more common in men than in women [5]. In fact, it is estimated by the year 2020 more men will die of HPV-related oropharyngeal cancer than women will die of cervical cancer [6,7]. The combined cost of HPV-associated cancers and other conditions is estimated to be $8 billion per year in the United States [8–11].

HPV Vaccines

Effective HPV vaccines have been available for females aged 9 to 26 years since 2006 (bivalent and quadrivalent vaccines) and for males aged 9 to 26 since 2010 (quadrivalent vaccine only) [12]. These vaccines have been shown in clinical trials to be highly efficacious in preventing HPV infection, cervical pre-cancer, and anal, vaginal, penile, and vulvar cancers caused by the HPV types covered in the vaccine [2]. Although their effectiveness against head and neck cancer has not been studied in clinical trials, most experts believe that these vaccines will also provide protection against at least a proportion of these cancers [13,14]. In 2015 the U.S. Food and Drug Administration approved licensure of a 9-valent HPV vaccine that will soon replace the quadrivalent vaccine in the U.S. market [15]. The 9-valent vaccine is licensed for both males and females aged 9 to 26 and is expected to prevent an even higher proportion of HPV-related cancers than earlier HPV vaccines due to the protection against 5 additional oncogenic HPV types [15].

Despite the potential of HPV vaccines to drastically reduce the incidence of HPV-related cancers and other diseases, these vaccines are not being as widely used in the United States as was hoped. The most recent national data from 2015 demonstrates that only 41.9% of girls and 28.1% of boys have received all 3 doses recommended in the HPV vaccine series [16]. This level of vaccine utilization is significantly lower than the Healthy People 2020 goal of 80% coverage [17], and also significantly lower than that of other developed countries such as Australia and the United Kingdom, which have achieved vaccination levels of ~70% among their target adolescent populations [18,19]. In the future, these low vaccination levels will likely be mitigated somewhat by the recent approval from the FDA and recent recommendation from the Advisory Committee on Immunization Practices (ACIP) for only 2 doses of the 9-valent HPV vaccine (spaced 6 to 12 months apart) for adolescents less than 15 years of age [20,21]. Three doses are still recommended for those aged 15 to 26 years.

Provider Communication About HPV Vaccines

How providers communicate with parents and patients about HPV vaccines is a key influential factor driving current U.S. adolescent HPV vaccination levels [22,23]. Numerous studies demonstrate that a provider’s recommendation generally has the largest impact on whether or not an adolescent receives the vaccine, even above that of parent factors such as attitudes and beliefs about the vaccine and patient characteristics such as age and insurance status [23–31]. Moreover, parents consistently cite their adolescent’s provider as one of the most trusted and impactful resources for obtaining vaccine information [22,32].

Unfortunately, research also shows that providers often fail to adequately recommend the HPV vaccine for their patients, especially for 11 to 12 year olds for whom the vaccine is preferentially recommended [33,34]. For example, in a national study of parents done in 2013, not being recommended by a provider was one of the top 5 reasons parents of males and of females aged 11 to 17 gave for not getting their adolescent vaccinated against HPV [35]. Supporting this also is a 2014 study of 776 pediatricians and family medicine providers nationally, in which Gilkey and colleagues found that more than 1 out of 4 providers did not highly endorse the HPV vaccine for 11 to 12 year olds despite this having been the recommended practice from ACIP for the prior 8 years for girls and 4 years for boys. This is in comparison to the other adolescents vaccines that were reported in the same study as being endorsed highly by these providers > 95% of the time [36].

Recognizing that providers’ HPV vaccine recommendations are often suboptimal, researchers have begun to define what components comprise “high-quality” HPV vaccine recommendations. This has been operationalized by one research group as (1) timeliness—routinely recommending the vaccine starting when the patient is ≤ 12 years; (2) consistency—recommending the vaccine for all eligible adolescents as opposed to an approach based on providers’ perception of their patients’ risk for HPV infection; (3) urgency—recommending that the vaccine be given on the same day the vaccine is being discussed, rather than offering the option of getting it at a future visit; and (4) strength—using language that clearly conveys that the provider believes the vaccine is very important for the adolescent to receive. A national study of primary care providers done in 2014 examined how frequently these quality components were implemented [37]. The results were startling and discouraging. Nearly half of providers (49%) reported they usually recommended that 11 to 12 year olds get the vaccine at a later visit, 41% used a risk-based approach for deciding when to recommend the vaccine, 27% did not tell the parents the vaccine was “very or extremely important,” and a large proportion did not start routinely recommending the vaccine before the age of 13 (39% for male patients and 25% for females) [37].

Much research has now accumulated to explain the underlying reasons why providers may not give consistent and high-quality HPV vaccine recommendations to all eligible adolescents [22]. Issues such as providers’ own knowledge about HPV-related diseases, personal beliefs about the vaccine’s safety and necessity, concern that a discussion about the vaccine will necessitate a discussion about adolescent sexuality with the parent, belief that parents will not want their child vaccinated if asked, perceptions that a provider can adequately select those patients most “in need” of HPV vaccination, and concern that raising the vaccine discussion with vaccine-hesitant parents will result in prolonged discussions have been shown to impact whether and how providers communicate about HPV vaccination during clinical visits [22,36–45]. Now that these barriers have been defined and described, there is a great need to use this knowledge to develop and evaluate interventions that help to mitigate these barriers and improve providers’ vaccine communication abilities. Such interventions are needed not only for HPV, but for all vaccines [46,47].

Possible Strategies for Helping Providers Communicate About HPV Vaccines

Before discussing these interventions, it is worth noting that several of the passive and active strategies have been shown in clinical trials to improve adolescent HPV vaccination rates. Although these are beyond the scope of this article, inclusion of these strategies should certainly be considered by any practice as a mechanism to increase vaccination levels, especially given that the most successful interventions to improve vaccination levels consist of multiple components [48]. Also useful is a recently described “taxonomy of vaccine communication interventions” that provides additional perspective on the scope and complexity of interventions to improve vaccine delivery [49]. There are several well-written review articles that describe interventions that focus on passive and active strategies at the practice or community level [50–52].

Interpersonal Communication Strategies Shown to Increase HPV Vaccination

Presumptive Communication

One of the first studies to examine the specific “way” in which providers communicate about vaccines focused not on HPV but rather on young childhood vaccines. In 2013 Opel and colleagues performed a study in which they taped clinical encounters between a pediatrician and a parent of a child aged 1 to 19 months [53]. Of the 111 encounters recorded, 50% of parents were classified as vaccine hesitant. Parents were aware they were being taped but not aware that the overall purpose of the study was to examine providers’ communication related to vaccination. The researchers found that providers generally used one of 2 communication styles to introduce the vaccine discussion. The first, called the “presumptive” style, assumed that parents would agree to vaccination and presented the vaccines as routine (ie, “We have to do some shots today”). The second style, called “participatory,” was more parent-oriented and used language suggesting shared decision-making (ie, “So what do you want to do about shots today?”). The study showed that the odds of resisting the provider’s vaccine recommendations were significantly higher when providers used a participatory approach than a presumptive one, suggesting that even small changes in language can have a major impact on the likelihood of vaccination. However, given the study design, causality between providers’ recommendation style and parents vaccination decisions could not be delineated.

In 2015 Moss and colleagues performed a study that examined the use of these 2 communication styles with regard to HPV vaccination [54]. This study used data from the 2010 National Immunization Survey–Teen, a national survey on childhood vaccination that includes provider verification of vaccines given [16]. Researchers categorized provider vaccine communication styles into “provider-driven,” which was similar to the presumptive style described Opel, and “patient-driven,” which was similar to Opel’s permissive style. Parents who received a more provider-driven style of HPV vaccine recommendation were far more likely to have allowed their adolescent to be vaccinated than those receiving patient-driven recommendations [54]. Further supporting this communication approach are results from a qualitative study done by Hughes and colleagues in which triads of mothers, adolescents, and providers were interviewed after a preventive care visit to assess the communication that occurred regarding HPV vaccination [39]. Providers’ communication style was categorized into 1 of 3 groups: paternalistic (clinician makes the vaccination decision and communicates this to the family); informed (patient and family gathers information from the clinician and other sources to reach a vaccination decision); and shared (medical and personal information is exchanged between the provider and family and then a decision is reached jointly). Providers who typically adopted the paternalistic approach perceived that they had the highest success in convincing parents to vaccinate—a perception that was confirmed in quantitative assessments of vaccination status among adolescents in the study sample [39]. Our own research demonstrates that learning and implementing a presumptive/paternalistic HPV vaccine recommendation style is easy for primary care providers to do and is perceived as often shortening the time taken during clinical visits to discuss the vaccine [55,56]. Thus, providers should consider opening the HPV vaccine conversation using this approach, and then turning to some of the other communication techniques described below when met with parental resistance or questions.

Motivational Interviewing

A second communication technique that seems effective for promoting HPV vaccination, especially for vaccine hesitant parents, is motivational interviewing. Motivational interviewing describes a communication technique in which the provider leverages a parents’ or patients’ intrinsic motivation to engage in a preferred health behavior [57]. Motivational interviewing was originally developed to combat substance abuse [58,59] but has subsequently been successfully applied to a number of other health issues [60–64]. There is growing interest from public health and medical providers in using this technique for increasing vaccination [39,65–68]. Our research group performed a large, cluster-randomized controlled trial of 16 pediatric and family medicine clinics to examine the impact of a provider communication “toolkit” on adolescent HPV vaccine series initiation and completion [50,69]. The toolkit consisted of motivational interviewing training for providers related to HPV vaccination and training on 3 tangible resources providers could also use with parents—an HPV fact sheet, an HPV vaccine decision aid, and an educational website. Results from the study demonstrated that motivational interviewing was the toolkit component most widely utilized by providers and was also perceived as being the most useful. More importantly, HPV vaccine series initiation levels were significantly higher among adolescents in practices receiving the toolkit than in control practices. There was no impact on HPV vaccine series completion (unpublished results). The usefulness of motivational interviewing for vaccination is further supported by a small study in which community pharmacists receiving motivational interviewing training for adult vaccination reported significantly higher patient readiness to receive vaccines following their interaction with the pharmacist than those who did not receive the training [70]. Finally, Perkins et al performed a cluster randomized controlled trial that evaluated the impact of a provider-focused intervention on adolescent HPV vaccination rates. The intervention included frequent provider support meetings, education on HPV infection and vaccination, feedback on providers’ individual HPV vaccination rates, provider incentives, and “basic motivational interviewing principles with vaccine-hesitant parents.” HPV vaccination series initiation and completion rates were significantly higher in intervention practices than controls, and this effect was sustained for at least 6 months after the active intervention period was over [67]. However, it was unknown how much the motivational interviewing contributed to these results. Based on the above information, and the long history of success of motivational interviewing for improving patient compliance with other recommended health behaviors, this technique appears to have a reasonable evidence base and should be considered for communicating with families that express resistance to HPV vaccination.

Personalized Communication

Parents’ reasons for not having their adolescent vaccinated against HPV are often complex and multifactorial [71,72]. Personalized approaches are needed to account for each parent’s unique informational needs, beliefs, and prior experiences [65]. Unfortunately, given the short amount of time allotted for clinical visits, it is often difficult to provide adequate information to parents during these encounters [73–75]. Indeed, concern about prolonged HPV vaccine discussions has been identified as an important barrier for providers that cause some to forgo recommending the vaccine [36,75].

One potential solution to this issue is to leverage technology in the form of web-based interventions that use software to tailor materials to each individual’s unique informational needs. Feasibility for this idea comes from the knowledge that many parents already use the web to research health issues related to their children [76], and that doctors’ offices are increasingly using patient portals and other web-based resources to help parents prepare for upcoming visits, especially those focused on health maintenance [77,78]. Tailored messaging interventions have been shown across populations and health issues to generally result in superior adherence with health behaviors when compared to untailored controls [79–82]. Several researchers have thus begun exploring whether such a personalized communication strategy may be similarly effective for adolescent HPV vaccination [50,83–85]. As an example, Maertens and colleagues used community-based participatory research techniques to develop a web-based tailored messaging intervention for Latinos regarding HPV vaccination [86]. A subsequent randomized controlled trial of the intervention in over 1200 parents of adolescents and young adults demonstrated that the intervention improved participants’ intentions to vaccinate compared to usual care [87], and among adolescents, higher HPV vaccine series initiation levels (unpublished data). Although additional work is needed to understand the feasibility of implementing such an intervention more broadly, additional support for the usefulness of a tailored messaging approach comes from a study of female university students that demonstrated higher HPV vaccination intentions after exposure to tailored information compared to untailored information. However, the impact on actual HPV vaccine utilization was not measured in the study [84]. Contrasting results were found in a different study of university students where researchers failed to find an impact of message tailoring on HPV vaccination utilization. However, this study was limited by a low response rate (~50%) to the follow up survey where vaccination status was assessed, and also by overall low levels of HPV vaccine initiation among the entire study sample (8%) [85]. Given the low number of studies in this area, and some conflicting data, additional research is needed to better understand the impact of personalized communication on HPV vaccination levels. However, results from these studies suggest that a modest benefit may be achieved with this approach, especially if coupled with other, evidence-based, clinic-level interventions to promote vaccination (eg, vaccine reminders, extended office hours), as is suggested by the Task Force on Community Preventive Services [48].

Focusing Communication on Cancer Prevention

HPV vaccines are unique in that they are only 1 of 2 vaccines for cancer prevention (the other being hepatitis B). Provider and parent surveys suggest that while most providers do mention cancer prevention when discussing HPV vaccines [40,88,89], this may be more commonly done with female patients than males [22]. Focusing on cancer prevention rather than sexual transmissibility is a communication technique suggested by the Centers for Disease Control and Prevention (CDC) as many parents cite this aspect of the vaccine as one of the most compelling reasons for vaccinating [45,90]. CDC’s “You are the Key” program [91] uses cancer prevention as a central theme in their physician and patient communication materials, based on significant prior market research on the acceptability and impact of such messages among parents and providers. In 2016 Malo and colleagues tested the potential impact of brief messages related to HPV vaccination, including cancer prevention messages, among a national sample of 776 medical providers and 1504 parents of adolescents [92]. In addition to their potential to motivate parents to vaccination, associations between parental endorsement of each message and their adolescent’s vaccination status were also examined. The cancer prevention messages were among those most highly endorsed by both parents and providers as being motivating for parents to get their adolescent vaccinated. More importantly, among parents these endorsements were associated with a significantly higher likelihood of the adolescent having been vaccinated against HPV. Interestingly, one of the briefest messages in the study, “I [the physician] strongly believe in the importance of this cancer preventing vaccine for [child’s name],” was perceived as the most persuasive message by parents.

Further support for the positive impact of framing HPV vaccines primarily as cancer prevention comes from another national study of 1495 parents of 11 to 17 year olds that examined 3 measures of quality of their adolescent provider’s HPV vaccine recommendation, and the relationship between recommendation quality and likelihood of adolescent HPV vaccination [40]. The 3 quality indicators assessed included providing information about cancer prevention, encouraging the vaccine “strongly,” and recommending it be given on the same day as it was being discussed. While 49% of parents reported receiving no HPV vaccine recommendation from their adolescents’ provider, of those that did, 86% received a cancer prevention message. Parents who had been given high quality recommendations that included either 2 or 3 of the quality indicator measures had over 9 times the odds of vaccine series initiation and 3 times the odds of vaccine series follow through than those who had not received any recommendation, and also significantly higher odds of vaccination than parents who had received low quality recommendations (ie, included only 1 indicator). Taken together, these results suggest that focusing discussions about HPV vaccines on their ability to prevent cancer is likely to be persuasive for some parents.

Strategies That Are Promising But Not Thoroughly Tested

Helping Parents Create Vaccination Plans

A recent commentary suggested that instead of focusing on changing beliefs or “educating” parents and patients about the need for a given vaccine, perhaps a better way to craft interventions for increasing vaccination is to focus on structuring the environment to make vaccination “easy” [93,94]. Examples of this include strategies such as extended office hours and making the vaccine available in other locations such as schools and pharmacies, both of which have been shown in some populations and settings to improve vaccine utilization [48,95]. One aspect of structuring a vaccine-conducive environment that relates to provider communication is helping parents create “implementation intentions” for future vaccination visits. In its most obvious form, this would mean providers provide office resources that facilitate making an appointment for the next dose in the HPV vaccine series during a clinic visit where the first dose was provided. But such an approach could also potentially extend to parents who are on the fence about the vaccine—to make an appointment before the parent leaves the office with an unvaccinated child to either re-discuss the vaccine in the future or to actually start the vaccine series. Support for such a strategy comes primarily from the social sciences, which suggest that implementation intentions work by increasing attention to specific cues to action, making it more likely that that the cue will be acted upon [96–98]. Creating implementation intentions has been shown to be helpful for improving adherence with a variety of health behaviors [99–105], and there is a growing evidence base related to how implementation intentions may facilitate vaccination specifically. For example Vet and colleagues performed a randomized controlled trial among 616 men who have sex with men with either strong or weak intentions to receive the hepatitis B vaccine [106]. Half of the participants were asked to create an implementation intention plan where they described when, where and how they would obtain the vaccine. Those in the control arm were not given this prompt. Regardless of whether their initial vaccination intention was weak or strong, those who had been asked to create an implementation plan had more than double the likelihood of actually getting the vaccine than participants who did not receive the implementation plan prompt. Similarly, a study of influenza vaccination rates among corporate employees found that those who were asked to write down the day and time they planned to go to employee health to get the free vaccine were somewhat more likely (4% higher) to be vaccinated than those who did not receive this prompt [107]. In addition, a study of elderly individuals found that influenza vaccination rates were significantly higher among those who had received “action instructions” on how, when and where to get the vaccine than those who did not [108]. These studies suggest that helping parents craft a definitive follow-up plan regarding vaccination could have a significant impact on vaccination rates—particularly for vaccines like HPV that require multiple doses.

Treating all Adolescent Vaccines the Same

Prior research has demonstrated that providers often communicate differently about HPV vaccines than other adolescent vaccines such as the tetanus-diphtheria-pertussis (Tdap) and meningococcal (MCV) vaccines [22,36]. Providers often tend to discuss the HPV vaccine last among these 3 vaccines, provide weaker endorsements of the vaccine, and pre-emptively give much more detail about the HPV compared to the other vaccines, even in the absence of a parent’s request for additional information [36,39,41]. The CDC and the American Academy of Pediatrics now suggest putting HPV at the beginning or middle of the list of vaccines recommended to the adolescent (ie, “HPV, Tdap and MCV”), and treating all recommended vaccines equivalently in terms of the level of detail provided to parents in the absence of a parent’s request for more information [109,110]. Through these suggestions have face validity, their specific impact on HPV vaccination rates, and on patient and provider satisfaction with the visit have yet to be evaluated.

Strategies that Probably Don’t Work

Presenting Myths and Facts

Research related to promoting other vaccines provides insight into communication activities that probably would not work well for promoting HPV vaccination. A 2012 study by Nyhan and colleagues examined the impact of 2 different messages related to influenza vaccines on participants’ beliefs about the vaccine’s safety and intentions to get vaccinated [111]. One group received information to correct the commonly held belief that influenza vaccine can cause the flu while the other received information about the risks associated with contracting an influenza infection. While the correction of myths did improve participants’ perceptions of the vaccine’s safety, information about influenza dangers did not. Neither message impacted intentions to vaccinate in the study subjects overall. However, in sub-analyses the correction of myths actually decreased intentions to vaccinate among those with high baseline levels of concern about the vaccine’s side effects—that is, among those most concerned that the flu vaccine can give someone the flu, correcting this myth actually decreased the likelihood that they would receive the vaccine. Similar findings have been reported in other studies related to vaccination [112–114], and suggest that the “threat” generated by providing information opposing a person’s beliefs may actually entrench these beliefs further as part of the threat response—a phenomenon known as attitude polarization [115]. These results also are consistent with the concept of negativity bias, which posits that negative information influences people’s risk perceptions more than positive information, and that the more strongly a risk is attempted to be negated, the lower the effectiveness and perceived trust of the information [116].

Using Fear Appeals

One tactic that has been suggested by some as a way to promote vaccination is to provide graphic depictions of the possible sequelae of vaccine-preventable diseases. The thought behind this idea is that because vaccination is so successful, most parents will have never experienced significant impacts from vaccine preventable diseases that, in the past, had been a major motivator for parents to vaccinate. Thus, in order to counter beliefs about “controversial” issues like vaccination, highly emotionally compelling and engaging information may be especially useful. This is a common tactic used by anti-vaccination groups to spread their own messages [117]. However, several studies suggested that using “fear appeals” (aka scare tactics) such as this to promote vaccination can actually have a negative effect on vaccination intentions. For example, in a 2011 study of a nationally representative sample of parents of children < 18 years, 4 different message formats were tested for their impact on parental intentions to vaccinate a future child with the measles-mumps-rubella vaccine (MMR) [113]. Message formats included correcting the misinformation that MMR causes autism, presenting information on MMR-related disease risk, providing a dramatic narrative about a child endangered by measles, and showing pictures of infants affected by these diseases. Counter to the study’s hypotheses, the dramatic narrative message actually increased parents’ perceptions that MMR vaccines had serious side effects, and the pictures increased parents’ belief that the MMR vaccine could cause autism. These counter-intuitive results are consistent with other studies that have examined the impact of message framing on adults’ vaccination intentions for HPV and influenza [108,118,119]. Taken together, fear appeals seem unlikely to sway many hesitant parents towards HPV vaccination.

Looking Into the Future

Moving forward, additional interventions to improve providers’ ability to communicate with families about HPV vaccination will undoubtedly be developed. A major area of interest in this regard is leveraging the power of technology and the internet, including using social media, mobile technologies, and online interventions to augment the provider/parent interaction that occurs during the clinical visit [50,120]. Web-based approaches have the benefit of generally being low cost and easy to disseminate to large populations. Such interventions have already been developed for a number of other health issues, some of which have proven effective [121,122]. However, use of the internet to promote healthy behaviors in general, and vaccination specifically, is still in its infancy. There is still much to be learned about how to create effective web-based tools, how to engage patients with them, and how to assess their impact on health outcomes [123].

Another interesting area for future research is identifying psychological “levers” to motivate parents’ vaccination intentions [94]. One example is focusing on using parents’ values (ie, protecting my child from harm) as an intervention target rather than beliefs or attitudes. This is because values tend to be inherent and static over time, compared to beliefs and attitudes, which are subject to change depending on the context [124]. Prior research has shown that interventions that leverage values rather than facts can be an effective way to overcome beliefs that are highly emotional or controversial, and that individuals are more likely to trust sources and individuals with shared values than those without [125], suggesting that this may be a useful way to motivate parents toward vaccinating their children. Self-affirmation is another example of a psychological lever that has a significant evidence base from the social science literature as a helpful tool for moving patients towards a desired health behavior [126,127], but it has not been extensively applied to the field of vaccination. Researchers in the field of vaccine delivery are increasingly recognizing the potential value of these unique intervention approaches [101,128–134], and it may be fruitful in the future to more closely examine the efficacy of interventions that target things like values, self-affirmation or other psychological levers to change parents’ HPV vaccination behaviors.

A final notable area for intervention research related to HPV vaccination is the use of video games. Although not likely to be used directly during patient visits, this strategy could be conceptualized as a potential way to augment the information conveyed to a parent by a provider directly during a clinical encounter. A meta-analysis from 2016 identified 16 different “serious” video games that were used to train and educate users about specific vaccine preventable diseases (usually influenza, none for HPV) and the need for vaccination [135]. In many of them, the objective of the game was to protect a virtual community from a vaccine preventable disease and/or manage outbreaks. Only 2 of the games evaluated outcomes in the short term (ie, at the time the game was being played). None have evaluated longer-term impacts such as vaccination intention or utilization. In the era of “plugged in” parents and adolescents, video games represent a unique but understudied mechanism for helping providers “communicate,” albeit indirectly, with families about the need for vaccination. Imagine providing a prescription to an HPV-vaccine hesitant family to “go play Zombie Wars HPV!” One would expect the curiosity factor alone would result in significant engagement with this intervention tool.

Conclusion

With persistently lagging HPV vaccination rates among U.S. adolescents, there is a growing need for effective interventions to improve adolescent HPV vaccine utilization. How providers communicate with families is one of the most influential factors in parents’ vaccination decisions. Emerging research is beginning to delineate potentially effective communication techniques such as presumptive approaches to making the vaccine recommendation, framing the vaccine as cancer preventing, and using motivational interviewing and personalized messaging when met with parental vaccine resistance. Moving forward the list of evidence-based interventions to improve providers’ HPV vaccine communication is likely to grow, and to increasingly leverage technology based solutions. However, given the complexities of the vaccination decision [136] and the ever growing spread of vaccine hesitancy [137], it is unlikely that a single intervention approach will be effective for getting adolescent HPV vaccine levels up to the national goal of 80% coverage. As has been recognized in the past, the most effective interventions for HPV vaccination in the future are likely to be multicomponent, including not only provider communication strategies but also clinic-, community-, and parent-level interventions [48].

Corresponding author: Amanda Dempsey, MD, PhD, MPH, 13199 East Montview Blvd, Suite 300, Aurora, CO 80045, [email protected].

Financial disclosures: None reported.

Study Overview

Objective. To evaluate the impact on overall survival of adding antiandrogen (bicalutamide) therapy to radiation in patients with prostate cancer who have an elevated prostate-specific antigen (PSA) after radical prostatectomy (either as persistence or as a relapse) and no evidence of metastatic disease.

Design. Phase III, randomized, double-blind, placebo-controlled trial.

Setting and participants. The trial was designed by NRG Oncology (Philadelphia, PA), sponsored by the National Cancer Institute, and conducted at NRG Oncology member sites, which included community-based hospitals. Eligible patients had undergone radical prostatectomy and had disease that was originally assessed, on the basis of pathological testing, as tumor stage T2 (confined to the prostate but with a positive surgical margin) or T3 (with histologic extension of tumor beyond the prostatic capsule) without nodal involvement. Patients also had to have a detectable PSA level between 0.2 and 4.0 ng/mL at least 8 weeks after surgery.  All the patients underwent abdominal and pelvic computed tomography (CT) and bone scans to rule out metastatic disease. Patients who received prior chemotherapy or radiation therapy for prostate cancer were excluded. Most patients had not received prior hormonal therapy for prostate cancer.

Intervention. All eligible patients received salvage radiation therapy within 12 weeks after randomization. Radiation was directed to the original prostatic site, the tumor resection bed, and the membranous urethra at a total dose of 64.8 Gy given in 36 daily fractions. In addition to the radiation therapy, patients were randomly assigned to receive either 150 mg of bicalutamide or 1 placebo tablet daily, beginning at the initiation of radiation therapy and continuing for 24 months.  Tablets were administered in a double blind fashion. Follow-up evaluations occurred every 3 months for 2 years, then every 6 months for 3 years, and then yearly. Bone and CT scans were performed either at biochemical recurrence or as indicated clinically. If metastatic disease was detected on follow up or if the serum PSA level rose to more than 4.0 ng/mL, maximum androgen blockade was recommended.

Main outcome measure. The main outcome was overall survival rate at 12 years. Secondary end points were disease-specific death, distant metastases, local disease progression, non–disease-specific death, any prostate cancer progression including a second biochemical recurrence, and adverse events.

Main results. 840 patients were randomized between March 1998 and March 2003, with 760 patients eligible for evaluation (384 patients in bicalutamide group and 376 in placebo group). Demographic and tumor-related characteristics of the 2 groups were similar. In both groups the majority of  patients were white (89.6% in bicalutamide arm; 86.2 in placebo), had Karnofsky performance status score of 100% (77.1 % in bicalutamide arm; 74.5% in placebo), and had positive surgical margin (75% in bicalutamide arm; 74.7% in placebo). Median age was 65 years, and median PSA level at trial entry was 0.6 ng/mL.  The median follow-up among the surviving patients was 13 years.

A total of 21 patients in the bicalutamide group and 46 patients in the placebo group died from prostate cancer. The actuarial rate of overall survival at 12 years was 76.3% in the bicalutamide group and 71.3% in the placebo group (hazard ratio [HR] for death 0.77 [95% confidence interval {CI} 0.59 to 0.99; P = 0.04), resulting in a 23% relative reduction in the risk of death in patients who received bicalutamide. The 12-year incidence of death from prostate cancer was 5.8% in the bicalutamide group versus 13.4% in the placebo group (HR 0.49 [95% CI 0.32 to 0.74]; P < 0.001), resulting in a 51% lower rate of death from prostate cancer in bicalutamide patients. Post hoc subgroup analyses showed that the greatest overall survival benefit was seen in subgroups of patients with more aggressive prostate cancer, such as those with high PSA level at trial entry (1.5 to 4.0 ng/mL) or Gleason score of 7.  There were too few patients with Gleason scores of 8, 9, or 10 to draw meaningful conclusions about this subgroup. There appeared to be a larger benefit in patients with positive surgical margins than in those with negative surgical margins.

Adherence to radiation therapy was similar between the 2 trial groups, and addition of bicalutamide to radiation therapy did not result in an increase in adverse events associated with radiation therapy, such as cystitis, colitis, or sexual dysfunction. The rates of hot flashes and cardiovascular deaths were not significantly higher in the bicalutamide group than in the placebo group. However, gynecomastia was reported significantly more frequently in the bicalutamide group (69.7%) than  in the placebo group (10.9%, P < 0.001).

Conclusion. The addition of 24 months of antiandrogen therapy with daily bicalutamide to salvage radiation therapy resulted in significantly higher rates of long-term overall survival and lower incidence of death from prostate cancer as compared to the addition of placebo. This benefit appeared to be without a significant cost in terms of toxicity.

Commentary

Prostate cancer is the second most common cancer in men worldwide, with an estimated 1,618,000 cases and 366,000 deaths in 2015 [1]. The current lifetime risk of developing prostate cancer for men living in the United States is estimated to be 1 in 6 [2]. Most prostate cancers are diagnosed in the localized stage, which is often treated with radical prostatectomy. All prostate tissue is removed during a successful radical prostatectomy. Postoperatively, detectable serum PSA is indicative of residual prostatic tissue, which presumably represents disease recurrence. This elevation in PSA after surgery in the absence of systemic metastatic disease is termed bio-chemical recurrence. The current standard of care for patients who develop biochemical recurrence is salvage radiation therapy. The prognosis for these patients is related to initial tumor characteristics—grade, volume and local stage. However, approximately 50% of the patients who are treated with salvage radiation therapy for biochemical recurrence will have further disease progression and may ultimately die from prostate cancer [3,4]. This is especially true when aggressive disease features are present. Radiation therapy combined with androgen-deprivation therapy using GnRH agonists or antiandrogen therapy (bicalutamide, flutamide) prolongs survival among some men with an intact prostate. This combination represents a rationale approach to prolong survival among men who develop non-metastatic biochemical relapse after radical prostatectomy.

The study by Shipley and colleagues reports the long-term outcomes of a randomized trial comparing salvage radiation plus 2 years of antiandrogen therapy to salvage radiation and placebo. Starting daily bicalutamide 150 mg orally with salvage radiation and continuing for 2 years was associated with a 23% improvement in overall survival and a 51% lower rate of death from prostate cancer, as compared to the placebo group. The number needed to treat (NNT) with bicalutamide to prevent one death from prostate cancer over 12 years was 20. By comparison, standard treatment of prostate cancer with surgery or radiation has an NNT of 27, which demonstrates the magnitude of the benefit of addition of antiandrogen therapy to salvage radiation. The benefit appears to be greatest in patients with poor prognostic factors such as higher Gleason scores (8 to 10), a higher PSA level at entry (0.7 to 4.0 ng/mL), or positive surgical margins. In contrast, patients with lower Gleason score or negative margins seemed to benefit less from the addition of antiandrogen therapy to salvage radiation. Two years of bicalutamide was not associated with increased incidence of radiation-related toxicities or cardiovascular death. As expected, the primary adverse effect of bicalutamide was gynecomastia, which was seen in 70% of the men treated. This adverse effect can be distressing but can be mitigated by prophylactic radiation of the breast or by the administration of tamoxifen, which were not done as preventative measures in this trial.

While the addition of bicalutamide to radiation did show a clear benefit to overall survival, questions remain about whether bicalutamide is the best drug to use. As the authors note, at present GnRH agonists such as leuprolide are considered first-line hormonal therapy with radiation for most patients with prostate cancer, and bicalutamide at the dose used in this study (150 mg) is not approved. We do not know how GnRH agonists will perform either as a single agent or in combination with antiandrogen for patients who develop biochemical relapse, as the use of GnRH agonists with radiation therapy has not been evaluated in patients who develop biochemical relapse in randomized clinical trials. Two trials exploring the use of androgen deprivation therapy with salvage radiation therapy in patients with biochemical recurrence (Radiotherapy and Androgen Deprivation in Combination After Local Surgery–Hormone Duration [RADICALS-HD] and the Groupe d’Etude des Tumeurs Uro-Génitales [GETUG]-16 trial) have finished enrollment; however, we will have to wait until the overall survival data matures before drawing any meaningful conclusions from them [5,6]. We know that patients with certain aggressive disease features based on tumor stage, grade, and volume are more likely to develop biochemical recurrence. As such, it is logical to consider evaluating the role of androgen deprivation therapy with adjuvant radiation therapy in patients who are at high risk of biochemical relapse with the goal of prolonging survival and reducing the risk of metastases. The RADICALS (Radiation Therapy and Androgen Deprivation Therapy in Treating Patients Who Have Undergone Surgery for Prostate Cancer) trial, which is evaluating the role of androgen deprivation therapy after adjuvant radiation therapy in patients who are at high risk of developing biochemical relapse, will help to address this issue.

Applications for Clinical Practice

Adding an antiandrogen agent (bicalutamide) to salvage radiation therapy in this randomized, double-blind, placebo-controlled trial resulted in higher rates of overall survival, disease-specific survival, and metastasis-free interval than radiation therapy alone for patients who developed biochemical relapse after radical prostatectomy for pathological T2/T3 and node-negative prostate cancer. We eagerly await the results of clinical trials evaluating the role of GnRH agonists in combination with salvage radiation therapy in patients in this setting. Given the long natural history of prostate cancer and the relatively low event rate, such studies can take over a decade to show differences in overall survival. Thus, until such data is available, 24 months of bicalutamide in combination with salvage radiation should be considered the new standard of care for patients (especially those at high risk) who develop non-metastatic biochemical relapse after prostatectomy.

—Devalkumar Rajyaguru, MD, and Lori Rosenstein, MD,
Gundersen Health System, La Crosse, WI

Study Overview

Objective. To evaluate the impact on overall survival of adding antiandrogen (bicalutamide) therapy to radiation in patients with prostate cancer who have an elevated prostate-specific antigen (PSA) after radical prostatectomy (either as persistence or as a relapse) and no evidence of metastatic disease.

Design. Phase III, randomized, double-blind, placebo-controlled trial.

Setting and participants. The trial was designed by NRG Oncology (Philadelphia, PA), sponsored by the National Cancer Institute, and conducted at NRG Oncology member sites, which included community-based hospitals. Eligible patients had undergone radical prostatectomy and had disease that was originally assessed, on the basis of pathological testing, as tumor stage T2 (confined to the prostate but with a positive surgical margin) or T3 (with histologic extension of tumor beyond the prostatic capsule) without nodal involvement. Patients also had to have a detectable PSA level between 0.2 and 4.0 ng/mL at least 8 weeks after surgery.  All the patients underwent abdominal and pelvic computed tomography (CT) and bone scans to rule out metastatic disease. Patients who received prior chemotherapy or radiation therapy for prostate cancer were excluded. Most patients had not received prior hormonal therapy for prostate cancer.

Intervention. All eligible patients received salvage radiation therapy within 12 weeks after randomization. Radiation was directed to the original prostatic site, the tumor resection bed, and the membranous urethra at a total dose of 64.8 Gy given in 36 daily fractions. In addition to the radiation therapy, patients were randomly assigned to receive either 150 mg of bicalutamide or 1 placebo tablet daily, beginning at the initiation of radiation therapy and continuing for 24 months.  Tablets were administered in a double blind fashion. Follow-up evaluations occurred every 3 months for 2 years, then every 6 months for 3 years, and then yearly. Bone and CT scans were performed either at biochemical recurrence or as indicated clinically. If metastatic disease was detected on follow up or if the serum PSA level rose to more than 4.0 ng/mL, maximum androgen blockade was recommended.

Main outcome measure. The main outcome was overall survival rate at 12 years. Secondary end points were disease-specific death, distant metastases, local disease progression, non–disease-specific death, any prostate cancer progression including a second biochemical recurrence, and adverse events.

Main results. 840 patients were randomized between March 1998 and March 2003, with 760 patients eligible for evaluation (384 patients in bicalutamide group and 376 in placebo group). Demographic and tumor-related characteristics of the 2 groups were similar. In both groups the majority of  patients were white (89.6% in bicalutamide arm; 86.2 in placebo), had Karnofsky performance status score of 100% (77.1 % in bicalutamide arm; 74.5% in placebo), and had positive surgical margin (75% in bicalutamide arm; 74.7% in placebo). Median age was 65 years, and median PSA level at trial entry was 0.6 ng/mL.  The median follow-up among the surviving patients was 13 years.

A total of 21 patients in the bicalutamide group and 46 patients in the placebo group died from prostate cancer. The actuarial rate of overall survival at 12 years was 76.3% in the bicalutamide group and 71.3% in the placebo group (hazard ratio [HR] for death 0.77 [95% confidence interval {CI} 0.59 to 0.99; P = 0.04), resulting in a 23% relative reduction in the risk of death in patients who received bicalutamide. The 12-year incidence of death from prostate cancer was 5.8% in the bicalutamide group versus 13.4% in the placebo group (HR 0.49 [95% CI 0.32 to 0.74]; P < 0.001), resulting in a 51% lower rate of death from prostate cancer in bicalutamide patients. Post hoc subgroup analyses showed that the greatest overall survival benefit was seen in subgroups of patients with more aggressive prostate cancer, such as those with high PSA level at trial entry (1.5 to 4.0 ng/mL) or Gleason score of 7.  There were too few patients with Gleason scores of 8, 9, or 10 to draw meaningful conclusions about this subgroup. There appeared to be a larger benefit in patients with positive surgical margins than in those with negative surgical margins.

Adherence to radiation therapy was similar between the 2 trial groups, and addition of bicalutamide to radiation therapy did not result in an increase in adverse events associated with radiation therapy, such as cystitis, colitis, or sexual dysfunction. The rates of hot flashes and cardiovascular deaths were not significantly higher in the bicalutamide group than in the placebo group. However, gynecomastia was reported significantly more frequently in the bicalutamide group (69.7%) than  in the placebo group (10.9%, P < 0.001).

Conclusion. The addition of 24 months of antiandrogen therapy with daily bicalutamide to salvage radiation therapy resulted in significantly higher rates of long-term overall survival and lower incidence of death from prostate cancer as compared to the addition of placebo. This benefit appeared to be without a significant cost in terms of toxicity.

Commentary

Prostate cancer is the second most common cancer in men worldwide, with an estimated 1,618,000 cases and 366,000 deaths in 2015 [1]. The current lifetime risk of developing prostate cancer for men living in the United States is estimated to be 1 in 6 [2]. Most prostate cancers are diagnosed in the localized stage, which is often treated with radical prostatectomy. All prostate tissue is removed during a successful radical prostatectomy. Postoperatively, detectable serum PSA is indicative of residual prostatic tissue, which presumably represents disease recurrence. This elevation in PSA after surgery in the absence of systemic metastatic disease is termed bio-chemical recurrence. The current standard of care for patients who develop biochemical recurrence is salvage radiation therapy. The prognosis for these patients is related to initial tumor characteristics—grade, volume and local stage. However, approximately 50% of the patients who are treated with salvage radiation therapy for biochemical recurrence will have further disease progression and may ultimately die from prostate cancer [3,4]. This is especially true when aggressive disease features are present. Radiation therapy combined with androgen-deprivation therapy using GnRH agonists or antiandrogen therapy (bicalutamide, flutamide) prolongs survival among some men with an intact prostate. This combination represents a rationale approach to prolong survival among men who develop non-metastatic biochemical relapse after radical prostatectomy.

The study by Shipley and colleagues reports the long-term outcomes of a randomized trial comparing salvage radiation plus 2 years of antiandrogen therapy to salvage radiation and placebo. Starting daily bicalutamide 150 mg orally with salvage radiation and continuing for 2 years was associated with a 23% improvement in overall survival and a 51% lower rate of death from prostate cancer, as compared to the placebo group. The number needed to treat (NNT) with bicalutamide to prevent one death from prostate cancer over 12 years was 20. By comparison, standard treatment of prostate cancer with surgery or radiation has an NNT of 27, which demonstrates the magnitude of the benefit of addition of antiandrogen therapy to salvage radiation. The benefit appears to be greatest in patients with poor prognostic factors such as higher Gleason scores (8 to 10), a higher PSA level at entry (0.7 to 4.0 ng/mL), or positive surgical margins. In contrast, patients with lower Gleason score or negative margins seemed to benefit less from the addition of antiandrogen therapy to salvage radiation. Two years of bicalutamide was not associated with increased incidence of radiation-related toxicities or cardiovascular death. As expected, the primary adverse effect of bicalutamide was gynecomastia, which was seen in 70% of the men treated. This adverse effect can be distressing but can be mitigated by prophylactic radiation of the breast or by the administration of tamoxifen, which were not done as preventative measures in this trial.

While the addition of bicalutamide to radiation did show a clear benefit to overall survival, questions remain about whether bicalutamide is the best drug to use. As the authors note, at present GnRH agonists such as leuprolide are considered first-line hormonal therapy with radiation for most patients with prostate cancer, and bicalutamide at the dose used in this study (150 mg) is not approved. We do not know how GnRH agonists will perform either as a single agent or in combination with antiandrogen for patients who develop biochemical relapse, as the use of GnRH agonists with radiation therapy has not been evaluated in patients who develop biochemical relapse in randomized clinical trials. Two trials exploring the use of androgen deprivation therapy with salvage radiation therapy in patients with biochemical recurrence (Radiotherapy and Androgen Deprivation in Combination After Local Surgery–Hormone Duration [RADICALS-HD] and the Groupe d’Etude des Tumeurs Uro-Génitales [GETUG]-16 trial) have finished enrollment; however, we will have to wait until the overall survival data matures before drawing any meaningful conclusions from them [5,6]. We know that patients with certain aggressive disease features based on tumor stage, grade, and volume are more likely to develop biochemical recurrence. As such, it is logical to consider evaluating the role of androgen deprivation therapy with adjuvant radiation therapy in patients who are at high risk of biochemical relapse with the goal of prolonging survival and reducing the risk of metastases. The RADICALS (Radiation Therapy and Androgen Deprivation Therapy in Treating Patients Who Have Undergone Surgery for Prostate Cancer) trial, which is evaluating the role of androgen deprivation therapy after adjuvant radiation therapy in patients who are at high risk of developing biochemical relapse, will help to address this issue.

Applications for Clinical Practice

Adding an antiandrogen agent (bicalutamide) to salvage radiation therapy in this randomized, double-blind, placebo-controlled trial resulted in higher rates of overall survival, disease-specific survival, and metastasis-free interval than radiation therapy alone for patients who developed biochemical relapse after radical prostatectomy for pathological T2/T3 and node-negative prostate cancer. We eagerly await the results of clinical trials evaluating the role of GnRH agonists in combination with salvage radiation therapy in patients in this setting. Given the long natural history of prostate cancer and the relatively low event rate, such studies can take over a decade to show differences in overall survival. Thus, until such data is available, 24 months of bicalutamide in combination with salvage radiation should be considered the new standard of care for patients (especially those at high risk) who develop non-metastatic biochemical relapse after prostatectomy.

—Devalkumar Rajyaguru, MD, and Lori Rosenstein, MD,
Gundersen Health System, La Crosse, WI

Study Overview

Objective. To evaluate the impact on overall survival of adding antiandrogen (bicalutamide) therapy to radiation in patients with prostate cancer who have an elevated prostate-specific antigen (PSA) after radical prostatectomy (either as persistence or as a relapse) and no evidence of metastatic disease.

Design. Phase III, randomized, double-blind, placebo-controlled trial.

Setting and participants. The trial was designed by NRG Oncology (Philadelphia, PA), sponsored by the National Cancer Institute, and conducted at NRG Oncology member sites, which included community-based hospitals. Eligible patients had undergone radical prostatectomy and had disease that was originally assessed, on the basis of pathological testing, as tumor stage T2 (confined to the prostate but with a positive surgical margin) or T3 (with histologic extension of tumor beyond the prostatic capsule) without nodal involvement. Patients also had to have a detectable PSA level between 0.2 and 4.0 ng/mL at least 8 weeks after surgery.  All the patients underwent abdominal and pelvic computed tomography (CT) and bone scans to rule out metastatic disease. Patients who received prior chemotherapy or radiation therapy for prostate cancer were excluded. Most patients had not received prior hormonal therapy for prostate cancer.

Intervention. All eligible patients received salvage radiation therapy within 12 weeks after randomization. Radiation was directed to the original prostatic site, the tumor resection bed, and the membranous urethra at a total dose of 64.8 Gy given in 36 daily fractions. In addition to the radiation therapy, patients were randomly assigned to receive either 150 mg of bicalutamide or 1 placebo tablet daily, beginning at the initiation of radiation therapy and continuing for 24 months.  Tablets were administered in a double blind fashion. Follow-up evaluations occurred every 3 months for 2 years, then every 6 months for 3 years, and then yearly. Bone and CT scans were performed either at biochemical recurrence or as indicated clinically. If metastatic disease was detected on follow up or if the serum PSA level rose to more than 4.0 ng/mL, maximum androgen blockade was recommended.

Main outcome measure. The main outcome was overall survival rate at 12 years. Secondary end points were disease-specific death, distant metastases, local disease progression, non–disease-specific death, any prostate cancer progression including a second biochemical recurrence, and adverse events.

Main results. 840 patients were randomized between March 1998 and March 2003, with 760 patients eligible for evaluation (384 patients in bicalutamide group and 376 in placebo group). Demographic and tumor-related characteristics of the 2 groups were similar. In both groups the majority of  patients were white (89.6% in bicalutamide arm; 86.2 in placebo), had Karnofsky performance status score of 100% (77.1 % in bicalutamide arm; 74.5% in placebo), and had positive surgical margin (75% in bicalutamide arm; 74.7% in placebo). Median age was 65 years, and median PSA level at trial entry was 0.6 ng/mL.  The median follow-up among the surviving patients was 13 years.

A total of 21 patients in the bicalutamide group and 46 patients in the placebo group died from prostate cancer. The actuarial rate of overall survival at 12 years was 76.3% in the bicalutamide group and 71.3% in the placebo group (hazard ratio [HR] for death 0.77 [95% confidence interval {CI} 0.59 to 0.99; P = 0.04), resulting in a 23% relative reduction in the risk of death in patients who received bicalutamide. The 12-year incidence of death from prostate cancer was 5.8% in the bicalutamide group versus 13.4% in the placebo group (HR 0.49 [95% CI 0.32 to 0.74]; P < 0.001), resulting in a 51% lower rate of death from prostate cancer in bicalutamide patients. Post hoc subgroup analyses showed that the greatest overall survival benefit was seen in subgroups of patients with more aggressive prostate cancer, such as those with high PSA level at trial entry (1.5 to 4.0 ng/mL) or Gleason score of 7.  There were too few patients with Gleason scores of 8, 9, or 10 to draw meaningful conclusions about this subgroup. There appeared to be a larger benefit in patients with positive surgical margins than in those with negative surgical margins.

Adherence to radiation therapy was similar between the 2 trial groups, and addition of bicalutamide to radiation therapy did not result in an increase in adverse events associated with radiation therapy, such as cystitis, colitis, or sexual dysfunction. The rates of hot flashes and cardiovascular deaths were not significantly higher in the bicalutamide group than in the placebo group. However, gynecomastia was reported significantly more frequently in the bicalutamide group (69.7%) than  in the placebo group (10.9%, P < 0.001).

Conclusion. The addition of 24 months of antiandrogen therapy with daily bicalutamide to salvage radiation therapy resulted in significantly higher rates of long-term overall survival and lower incidence of death from prostate cancer as compared to the addition of placebo. This benefit appeared to be without a significant cost in terms of toxicity.

Commentary

Prostate cancer is the second most common cancer in men worldwide, with an estimated 1,618,000 cases and 366,000 deaths in 2015 [1]. The current lifetime risk of developing prostate cancer for men living in the United States is estimated to be 1 in 6 [2]. Most prostate cancers are diagnosed in the localized stage, which is often treated with radical prostatectomy. All prostate tissue is removed during a successful radical prostatectomy. Postoperatively, detectable serum PSA is indicative of residual prostatic tissue, which presumably represents disease recurrence. This elevation in PSA after surgery in the absence of systemic metastatic disease is termed bio-chemical recurrence. The current standard of care for patients who develop biochemical recurrence is salvage radiation therapy. The prognosis for these patients is related to initial tumor characteristics—grade, volume and local stage. However, approximately 50% of the patients who are treated with salvage radiation therapy for biochemical recurrence will have further disease progression and may ultimately die from prostate cancer [3,4]. This is especially true when aggressive disease features are present. Radiation therapy combined with androgen-deprivation therapy using GnRH agonists or antiandrogen therapy (bicalutamide, flutamide) prolongs survival among some men with an intact prostate. This combination represents a rationale approach to prolong survival among men who develop non-metastatic biochemical relapse after radical prostatectomy.

The study by Shipley and colleagues reports the long-term outcomes of a randomized trial comparing salvage radiation plus 2 years of antiandrogen therapy to salvage radiation and placebo. Starting daily bicalutamide 150 mg orally with salvage radiation and continuing for 2 years was associated with a 23% improvement in overall survival and a 51% lower rate of death from prostate cancer, as compared to the placebo group. The number needed to treat (NNT) with bicalutamide to prevent one death from prostate cancer over 12 years was 20. By comparison, standard treatment of prostate cancer with surgery or radiation has an NNT of 27, which demonstrates the magnitude of the benefit of addition of antiandrogen therapy to salvage radiation. The benefit appears to be greatest in patients with poor prognostic factors such as higher Gleason scores (8 to 10), a higher PSA level at entry (0.7 to 4.0 ng/mL), or positive surgical margins. In contrast, patients with lower Gleason score or negative margins seemed to benefit less from the addition of antiandrogen therapy to salvage radiation. Two years of bicalutamide was not associated with increased incidence of radiation-related toxicities or cardiovascular death. As expected, the primary adverse effect of bicalutamide was gynecomastia, which was seen in 70% of the men treated. This adverse effect can be distressing but can be mitigated by prophylactic radiation of the breast or by the administration of tamoxifen, which were not done as preventative measures in this trial.

While the addition of bicalutamide to radiation did show a clear benefit to overall survival, questions remain about whether bicalutamide is the best drug to use. As the authors note, at present GnRH agonists such as leuprolide are considered first-line hormonal therapy with radiation for most patients with prostate cancer, and bicalutamide at the dose used in this study (150 mg) is not approved. We do not know how GnRH agonists will perform either as a single agent or in combination with antiandrogen for patients who develop biochemical relapse, as the use of GnRH agonists with radiation therapy has not been evaluated in patients who develop biochemical relapse in randomized clinical trials. Two trials exploring the use of androgen deprivation therapy with salvage radiation therapy in patients with biochemical recurrence (Radiotherapy and Androgen Deprivation in Combination After Local Surgery–Hormone Duration [RADICALS-HD] and the Groupe d’Etude des Tumeurs Uro-Génitales [GETUG]-16 trial) have finished enrollment; however, we will have to wait until the overall survival data matures before drawing any meaningful conclusions from them [5,6]. We know that patients with certain aggressive disease features based on tumor stage, grade, and volume are more likely to develop biochemical recurrence. As such, it is logical to consider evaluating the role of androgen deprivation therapy with adjuvant radiation therapy in patients who are at high risk of biochemical relapse with the goal of prolonging survival and reducing the risk of metastases. The RADICALS (Radiation Therapy and Androgen Deprivation Therapy in Treating Patients Who Have Undergone Surgery for Prostate Cancer) trial, which is evaluating the role of androgen deprivation therapy after adjuvant radiation therapy in patients who are at high risk of developing biochemical relapse, will help to address this issue.

Applications for Clinical Practice

Adding an antiandrogen agent (bicalutamide) to salvage radiation therapy in this randomized, double-blind, placebo-controlled trial resulted in higher rates of overall survival, disease-specific survival, and metastasis-free interval than radiation therapy alone for patients who developed biochemical relapse after radical prostatectomy for pathological T2/T3 and node-negative prostate cancer. We eagerly await the results of clinical trials evaluating the role of GnRH agonists in combination with salvage radiation therapy in patients in this setting. Given the long natural history of prostate cancer and the relatively low event rate, such studies can take over a decade to show differences in overall survival. Thus, until such data is available, 24 months of bicalutamide in combination with salvage radiation should be considered the new standard of care for patients (especially those at high risk) who develop non-metastatic biochemical relapse after prostatectomy.

—Devalkumar Rajyaguru, MD, and Lori Rosenstein, MD,
Gundersen Health System, La Crosse, WI

Primary Cutaneous B-cell Lymphoma

Cutaneous B-cell lymphomas (CBCLs) are a diverse but rare group of cutaneous lympho­proliferative neoplasms that make up approximately 20% of the total number of hematolymphoid neoplasms primary to the skin.¹ These lymphomas are comprised of neoplastic B cells in various stages of differentiation. As a whole, they are rare neoplasms that primarily involve the head, neck, trunk, arms, or legs.¹ Clinically, patients present with nontender, compressible, solitary, red to violaceous papules or nodules. Most CBCLs are considered low-grade malignancies with nonaggressive behavior and excellent prognosis; however, the diffuse large B-cell lymphomas, including but not limited to intravascular and leg type; lymphomatoid granulomatosis; and B-cell lymphoblastic lymphoma can act more aggressively.¹

Histopathologic examination of primary CBCL generally reveals a relatively normal epidermis accompanied by a nodular to diffuse monomorphic lymphocytic cellular infiltrate in the dermis that can occasionally extend into the subcutaneous tissue (quiz image). Although not specific for CBCLs, oftentimes there is an acellular portion of the superficial papillary dermis known as a grenz zone that can serve as a histopathologic clue to the diagnosis of a cutaneous lymphoproliferative disorder. The list of malignant B-cell neoplasms is extensive (eg, cutaneous marginal zone B-cell lymphoma, primary cutaneous follicle center lymphoma, diffuse large B-cell lymphoma, intravascular large B-cell lymphoma), and few are seen in the skin.

The most common type of CBCL is marginal zone B-cell lymphoma, which is considered to be a tumor of mucosa-associated (or skin-associated) lymphoid tissue. It is characterized by a monomorphous population of small mature lymphocytes showing characteristics of the B cells of the marginal zone of the lymph node. Some cells have the features of centrocytes/centroblasts (Figure 1) demonstrated by slightly irregular or indented nuclei and generous amounts of cytoplasm. Larger and more pleomorphic cells such as immunoblasts are similarly noted (Figure 1). The quiz image and Figure 1 demonstrate a cutaneous marginal zone B-cell lymphoma. A histomorphologic clue supporting a diagnosis of marginal zone B-cell lymphoma over reactive lymphoid hyperplasia is a B-cell predominate (B- to T-cell ratio of at least 3 to 1) infiltrate that is comprised of marginal zone-type cells. Immunohistochemistry demonstrating fewer differentiated B cells with light chain restriction may provide additional evidence that supports a clonal and potentially malignant process.

Erythematous to violaceous nodules on the head and neck of older individuals are characteristic of both granuloma faciale and CBCL. Histologically, granuloma faciale is characterized by a dense cellular infiltrate, often with a nodular outline, occupying the mid dermis.² Granuloma faciale typically spares the immediate subepidermis and hair follicles, forming a grenz zone. The cellular infiltrate is polymorphic and consists of eosinophils and neutrophils with scattered plasma cells, mast cells, and lymphocytes in a vasculocentric distribution, eventually with chronic concentric fibrosis (Figure 2).

Leukemia cutis demonstrates a dermal infiltrate that contains atypical mononuclear cells (myeloblasts and myelocytes)(Figure 3).³ These markedly atypical mononuclear cells can have kidney bean-shaped nuclei and percolate through the dermal collagen, resembling single-file cells. They have increased nuclear to cytoplasmic ratios and occasionally have prominent nucleoli. Correlation with immunophenotypic and cytochemical studies is required for specific typing of the leukemic infiltrate.

Similar to primary CBCL, lymphomatoid papulosis (LyP) consists of erythematous papules or nodules that can occur anywhere on the body. In contrast to CBCL, the lesions of LyP classically self-resolve. However, approximately 10% to 20% of patients develop a malignant lymphoma, with mycosis fungoides, Hodgkin disease, and anaplastic large cell lymphoma being the most commonly associated.

Histologic examination of lesions of LyP classically demonstrates a wedge-shaped dermal infiltrate with variable epidermal changes (Figure 4). The wedge-shaped infiltrate is composed of large atypical cells. Three main types of lesions have been delineated: types A, B, and C. Type A is characterized by an increased number of cells with large vesicular nuclei with clumped chromatin, prominent nucleoli, and pronounced cytoplasm. Reed-Sternberg-like cells with an admixture of inflammatory cells including small lymphocytes, macrophages, neutrophils, and eosinophils also are present. Type B neoplastic cells vary in size and feature hyperchromatic, convoluted, or cerebriform nuclei. The infiltrate can be dense and bandlike with fewer cells resembling mycosis fungoides; type B LyP has neoplastic cells, not inflammatory cells. Finally, type C demonstrates solid sheets of large atypical cells resembling anaplastic large cell lymphoma. Immunohistochemically, the atypical cells often are CD4⁺ and CD8^- with variable loss of pan-T-cell antigens. The atypical cells of types A and C express CD30 reactivity.⁴

Merkel cell carcinoma (MCC) is a primary neuroendocrine carcinoma of the skin that usually arises on sun-exposed skin in elderly patients with lesions that histologically and clinically resemble cutaneous lymphoma.⁵ It classically is composed of small, round to oval, basophilic cells with a vesicular nucleus and multiple small nucleoli. Apoptotic cells and mitoses often are present.⁶ One key finding that helps to differentiate MCC from lymphoma is the presence of finely dispersed salt-and-pepper chromatin and molded nuclear contour in MCC (Figure 5).

Immunophenotyping is important in the differentiation of these diagnoses. The atypical cells of LyP are positive for CD3, CD4, and CD30 but are negative for CD8. However, in type B LyP, the large CD30⁺ cells seen in the other types are not commonly seen. In contrast, MCC expresses reactivity with cytokeratins, in particular cytokeratin 20 and CAM5.2, classically in a paranuclear dotlike pattern. In keeping with MCC's neuroendocrine differentiation, the tumor cells will demonstrate reactivity with synaptophysin, chromogranin, and CD56. The immunohistochemistry for leukemia cutis varies depending on the type of leukemia. Acute myelomonocytic leukemia is positive for myeloperoxidase, CD13, CD33, and CD68. The immunophenotype of these marginal zone lymphoma cells is as follows: positive for CD20, CD79a, and Bcl-2; negative for Bcl-6, CD5, CD10, CD23, and cyclin D1 (Bcl-1).⁷ 

Primary Cutaneous B-cell Lymphoma

Cutaneous B-cell lymphomas (CBCLs) are a diverse but rare group of cutaneous lympho­proliferative neoplasms that make up approximately 20% of the total number of hematolymphoid neoplasms primary to the skin.¹ These lymphomas are comprised of neoplastic B cells in various stages of differentiation. As a whole, they are rare neoplasms that primarily involve the head, neck, trunk, arms, or legs.¹ Clinically, patients present with nontender, compressible, solitary, red to violaceous papules or nodules. Most CBCLs are considered low-grade malignancies with nonaggressive behavior and excellent prognosis; however, the diffuse large B-cell lymphomas, including but not limited to intravascular and leg type; lymphomatoid granulomatosis; and B-cell lymphoblastic lymphoma can act more aggressively.¹

Histopathologic examination of primary CBCL generally reveals a relatively normal epidermis accompanied by a nodular to diffuse monomorphic lymphocytic cellular infiltrate in the dermis that can occasionally extend into the subcutaneous tissue (quiz image). Although not specific for CBCLs, oftentimes there is an acellular portion of the superficial papillary dermis known as a grenz zone that can serve as a histopathologic clue to the diagnosis of a cutaneous lymphoproliferative disorder. The list of malignant B-cell neoplasms is extensive (eg, cutaneous marginal zone B-cell lymphoma, primary cutaneous follicle center lymphoma, diffuse large B-cell lymphoma, intravascular large B-cell lymphoma), and few are seen in the skin.

The most common type of CBCL is marginal zone B-cell lymphoma, which is considered to be a tumor of mucosa-associated (or skin-associated) lymphoid tissue. It is characterized by a monomorphous population of small mature lymphocytes showing characteristics of the B cells of the marginal zone of the lymph node. Some cells have the features of centrocytes/centroblasts (Figure 1) demonstrated by slightly irregular or indented nuclei and generous amounts of cytoplasm. Larger and more pleomorphic cells such as immunoblasts are similarly noted (Figure 1). The quiz image and Figure 1 demonstrate a cutaneous marginal zone B-cell lymphoma. A histomorphologic clue supporting a diagnosis of marginal zone B-cell lymphoma over reactive lymphoid hyperplasia is a B-cell predominate (B- to T-cell ratio of at least 3 to 1) infiltrate that is comprised of marginal zone-type cells. Immunohistochemistry demonstrating fewer differentiated B cells with light chain restriction may provide additional evidence that supports a clonal and potentially malignant process.

Erythematous to violaceous nodules on the head and neck of older individuals are characteristic of both granuloma faciale and CBCL. Histologically, granuloma faciale is characterized by a dense cellular infiltrate, often with a nodular outline, occupying the mid dermis.² Granuloma faciale typically spares the immediate subepidermis and hair follicles, forming a grenz zone. The cellular infiltrate is polymorphic and consists of eosinophils and neutrophils with scattered plasma cells, mast cells, and lymphocytes in a vasculocentric distribution, eventually with chronic concentric fibrosis (Figure 2).

Leukemia cutis demonstrates a dermal infiltrate that contains atypical mononuclear cells (myeloblasts and myelocytes)(Figure 3).³ These markedly atypical mononuclear cells can have kidney bean-shaped nuclei and percolate through the dermal collagen, resembling single-file cells. They have increased nuclear to cytoplasmic ratios and occasionally have prominent nucleoli. Correlation with immunophenotypic and cytochemical studies is required for specific typing of the leukemic infiltrate.

Similar to primary CBCL, lymphomatoid papulosis (LyP) consists of erythematous papules or nodules that can occur anywhere on the body. In contrast to CBCL, the lesions of LyP classically self-resolve. However, approximately 10% to 20% of patients develop a malignant lymphoma, with mycosis fungoides, Hodgkin disease, and anaplastic large cell lymphoma being the most commonly associated.

Histologic examination of lesions of LyP classically demonstrates a wedge-shaped dermal infiltrate with variable epidermal changes (Figure 4). The wedge-shaped infiltrate is composed of large atypical cells. Three main types of lesions have been delineated: types A, B, and C. Type A is characterized by an increased number of cells with large vesicular nuclei with clumped chromatin, prominent nucleoli, and pronounced cytoplasm. Reed-Sternberg-like cells with an admixture of inflammatory cells including small lymphocytes, macrophages, neutrophils, and eosinophils also are present. Type B neoplastic cells vary in size and feature hyperchromatic, convoluted, or cerebriform nuclei. The infiltrate can be dense and bandlike with fewer cells resembling mycosis fungoides; type B LyP has neoplastic cells, not inflammatory cells. Finally, type C demonstrates solid sheets of large atypical cells resembling anaplastic large cell lymphoma. Immunohistochemically, the atypical cells often are CD4⁺ and CD8^- with variable loss of pan-T-cell antigens. The atypical cells of types A and C express CD30 reactivity.⁴

Merkel cell carcinoma (MCC) is a primary neuroendocrine carcinoma of the skin that usually arises on sun-exposed skin in elderly patients with lesions that histologically and clinically resemble cutaneous lymphoma.⁵ It classically is composed of small, round to oval, basophilic cells with a vesicular nucleus and multiple small nucleoli. Apoptotic cells and mitoses often are present.⁶ One key finding that helps to differentiate MCC from lymphoma is the presence of finely dispersed salt-and-pepper chromatin and molded nuclear contour in MCC (Figure 5).

Immunophenotyping is important in the differentiation of these diagnoses. The atypical cells of LyP are positive for CD3, CD4, and CD30 but are negative for CD8. However, in type B LyP, the large CD30⁺ cells seen in the other types are not commonly seen. In contrast, MCC expresses reactivity with cytokeratins, in particular cytokeratin 20 and CAM5.2, classically in a paranuclear dotlike pattern. In keeping with MCC's neuroendocrine differentiation, the tumor cells will demonstrate reactivity with synaptophysin, chromogranin, and CD56. The immunohistochemistry for leukemia cutis varies depending on the type of leukemia. Acute myelomonocytic leukemia is positive for myeloperoxidase, CD13, CD33, and CD68. The immunophenotype of these marginal zone lymphoma cells is as follows: positive for CD20, CD79a, and Bcl-2; negative for Bcl-6, CD5, CD10, CD23, and cyclin D1 (Bcl-1).⁷ 

Primary Cutaneous B-cell Lymphoma

Cutaneous B-cell lymphomas (CBCLs) are a diverse but rare group of cutaneous lympho­proliferative neoplasms that make up approximately 20% of the total number of hematolymphoid neoplasms primary to the skin.¹ These lymphomas are comprised of neoplastic B cells in various stages of differentiation. As a whole, they are rare neoplasms that primarily involve the head, neck, trunk, arms, or legs.¹ Clinically, patients present with nontender, compressible, solitary, red to violaceous papules or nodules. Most CBCLs are considered low-grade malignancies with nonaggressive behavior and excellent prognosis; however, the diffuse large B-cell lymphomas, including but not limited to intravascular and leg type; lymphomatoid granulomatosis; and B-cell lymphoblastic lymphoma can act more aggressively.¹

Histopathologic examination of primary CBCL generally reveals a relatively normal epidermis accompanied by a nodular to diffuse monomorphic lymphocytic cellular infiltrate in the dermis that can occasionally extend into the subcutaneous tissue (quiz image). Although not specific for CBCLs, oftentimes there is an acellular portion of the superficial papillary dermis known as a grenz zone that can serve as a histopathologic clue to the diagnosis of a cutaneous lymphoproliferative disorder. The list of malignant B-cell neoplasms is extensive (eg, cutaneous marginal zone B-cell lymphoma, primary cutaneous follicle center lymphoma, diffuse large B-cell lymphoma, intravascular large B-cell lymphoma), and few are seen in the skin.

The most common type of CBCL is marginal zone B-cell lymphoma, which is considered to be a tumor of mucosa-associated (or skin-associated) lymphoid tissue. It is characterized by a monomorphous population of small mature lymphocytes showing characteristics of the B cells of the marginal zone of the lymph node. Some cells have the features of centrocytes/centroblasts (Figure 1) demonstrated by slightly irregular or indented nuclei and generous amounts of cytoplasm. Larger and more pleomorphic cells such as immunoblasts are similarly noted (Figure 1). The quiz image and Figure 1 demonstrate a cutaneous marginal zone B-cell lymphoma. A histomorphologic clue supporting a diagnosis of marginal zone B-cell lymphoma over reactive lymphoid hyperplasia is a B-cell predominate (B- to T-cell ratio of at least 3 to 1) infiltrate that is comprised of marginal zone-type cells. Immunohistochemistry demonstrating fewer differentiated B cells with light chain restriction may provide additional evidence that supports a clonal and potentially malignant process.

Erythematous to violaceous nodules on the head and neck of older individuals are characteristic of both granuloma faciale and CBCL. Histologically, granuloma faciale is characterized by a dense cellular infiltrate, often with a nodular outline, occupying the mid dermis.² Granuloma faciale typically spares the immediate subepidermis and hair follicles, forming a grenz zone. The cellular infiltrate is polymorphic and consists of eosinophils and neutrophils with scattered plasma cells, mast cells, and lymphocytes in a vasculocentric distribution, eventually with chronic concentric fibrosis (Figure 2).

Leukemia cutis demonstrates a dermal infiltrate that contains atypical mononuclear cells (myeloblasts and myelocytes)(Figure 3).³ These markedly atypical mononuclear cells can have kidney bean-shaped nuclei and percolate through the dermal collagen, resembling single-file cells. They have increased nuclear to cytoplasmic ratios and occasionally have prominent nucleoli. Correlation with immunophenotypic and cytochemical studies is required for specific typing of the leukemic infiltrate.

Similar to primary CBCL, lymphomatoid papulosis (LyP) consists of erythematous papules or nodules that can occur anywhere on the body. In contrast to CBCL, the lesions of LyP classically self-resolve. However, approximately 10% to 20% of patients develop a malignant lymphoma, with mycosis fungoides, Hodgkin disease, and anaplastic large cell lymphoma being the most commonly associated.

Histologic examination of lesions of LyP classically demonstrates a wedge-shaped dermal infiltrate with variable epidermal changes (Figure 4). The wedge-shaped infiltrate is composed of large atypical cells. Three main types of lesions have been delineated: types A, B, and C. Type A is characterized by an increased number of cells with large vesicular nuclei with clumped chromatin, prominent nucleoli, and pronounced cytoplasm. Reed-Sternberg-like cells with an admixture of inflammatory cells including small lymphocytes, macrophages, neutrophils, and eosinophils also are present. Type B neoplastic cells vary in size and feature hyperchromatic, convoluted, or cerebriform nuclei. The infiltrate can be dense and bandlike with fewer cells resembling mycosis fungoides; type B LyP has neoplastic cells, not inflammatory cells. Finally, type C demonstrates solid sheets of large atypical cells resembling anaplastic large cell lymphoma. Immunohistochemically, the atypical cells often are CD4⁺ and CD8^- with variable loss of pan-T-cell antigens. The atypical cells of types A and C express CD30 reactivity.⁴

Merkel cell carcinoma (MCC) is a primary neuroendocrine carcinoma of the skin that usually arises on sun-exposed skin in elderly patients with lesions that histologically and clinically resemble cutaneous lymphoma.⁵ It classically is composed of small, round to oval, basophilic cells with a vesicular nucleus and multiple small nucleoli. Apoptotic cells and mitoses often are present.⁶ One key finding that helps to differentiate MCC from lymphoma is the presence of finely dispersed salt-and-pepper chromatin and molded nuclear contour in MCC (Figure 5).

Immunophenotyping is important in the differentiation of these diagnoses. The atypical cells of LyP are positive for CD3, CD4, and CD30 but are negative for CD8. However, in type B LyP, the large CD30⁺ cells seen in the other types are not commonly seen. In contrast, MCC expresses reactivity with cytokeratins, in particular cytokeratin 20 and CAM5.2, classically in a paranuclear dotlike pattern. In keeping with MCC's neuroendocrine differentiation, the tumor cells will demonstrate reactivity with synaptophysin, chromogranin, and CD56. The immunohistochemistry for leukemia cutis varies depending on the type of leukemia. Acute myelomonocytic leukemia is positive for myeloperoxidase, CD13, CD33, and CD68. The immunophenotype of these marginal zone lymphoma cells is as follows: positive for CD20, CD79a, and Bcl-2; negative for Bcl-6, CD5, CD10, CD23, and cyclin D1 (Bcl-1).⁷ 

Study Overview

Objective. To compare the safety and efficacy of basal–bolus insulin therapy with sitagliptin plus insulin glargine in type 2 diabetes patients admitted to general medicine and surgical wards.

Design. Multicenter, prospective, open-label, noninferiority randomized clinical trial.

Setting and participants. Type 2 diabetes patients aged 18 to 80 years admitted to the general medicine and surgery services at one of 5 academic-based US hospitals were recruited. Eligible participants presented with a random blood glucose concentration between 140 and 400 mg/dL and were treated at home with diet, oral agents, or oral agents plus insulin at a maximum daily dose of 0.6 units/kg. Among those excluded were patients recently treated with a dipeptidyl peptidase-4 (DPP-4) inhibitor or glucagon-like peptide-1 (GLP-1) agonist, patients with clinically relevant hepatic disease, patients who were not eating for more than 48 hours, and those with an estimated glomerular filtration rate (eGFR) < 30 mL/min.

Intervention. Participants were randomly assigned to receive basal–bolus insulin therapy (BBI) with glargine once daily plus rapid-acting insulin before meals or sitagliptin plus glargine (SPG) once daily. Those in the SPG group received sitagliptin 100 mg/day if their eGFR was > 50 mL/min and sitagliptin 50 mg/day if their eGFR was 30 to 50 mL/min. If the eGFR fell below 30 mL/min during the hospitalization, sitagliptin was reduced to 25 mg/day. Glargine doses for those in the SPG group were started at 0.2 units/kg if randomization blood glucose was 140–200 mg/dL and 0.25 units/kg if randomization glucose was 201–400 mg/dL. Patients aged 70 or older or with an eGFR < 50 mL/min started with a daily glargine dose of 0.15 units/kg. For the BBI group, a total daily insulin dose of 0.4 units/kg was initiated for those with blood glucose levels between 140 and 200 mg/dL, and 0.5 units/kg for those with randomization glucose between 201 and 400 mg/dL. Half of this daily dose was given as glargine and the other half was distributed evenly across 3 pre-meal doses. Both the BBI and SPG groups received pre-meal and bedtime correction doses of rapid-acting insulin for glucose levels above 140 mg/dL. Blood glucose concentrations were measured fasting, before meals, and at bedtime or every 6 hours for patients who were not eating. Target fasting and pre-meal blood glucose levels were 100 to 140 mg/dL. Investigators and participants were not blinded to group assignment and glucose control was managed by the primary medical or surgical team.

Main outcomes measure. The primary outcome for this trial was noninferiority for differences between the SPG and BBI groups in glycemic control. Secondary endpoints included differences in the number of hypoglycemic and hyperglycemic events, the number of blood glucose values between 70 and 140 mg/dL and between 70 and 180 mg/dL, and the number of treatment failures (defined as 2 consecutive blood glucose values > 240 mg/dL or mean daily glucose > 240 mg/dL), length of hospital stay, total daily dose of insulin, number of insulin injections per day, transfer to the intensive care unit, and hospital complications and mortality.

Main results. A total of 138 patients in the SPG group and 139 patients in the BBI group completed the study and were included in this analysis. Of these 277 patients, 84% were admitted to a medicine ward and 16% were admitted to a surgical ward. The average age of participants was approximately 57 years, the average BMI was approximately 35 kg/m², and the average duration of diabetes was approximately 10 years. These baseline characteristics as well as ethnic origin, sex, and baseline A1c (approximately 40% of patients in both groups had a baseline A1c between 7% and 9%) did not differ between groups. Prior to admission, approximately 40% of patients in both groups were managed with oral drugs alone, approximately 25% were managed with insulin alone, and about 22% were managed with insulin and oral therapy.

With respect to the primary outcome, both groups had similar mean daily blood glucose concentrations (171 mg/dL in SPG and 169 mg/dL in BBI) throughout the hospitalization, meeting the noninferiority threshold for glycemic control between groups. As for secondary outcomes, the mean proportion of blood glucose readings between 70 and 140 mg/dL, 70 and 180 mg/dL, and 100 and 140 mg/dL did not differ between groups. Pre-meal and bedtime blood glucose concentrations were also similar in both groups. There was a significant difference between groups in average daily insulin dose (24 units in SPG versus 34 units in BBI), total units of insulin per kg per day (0.2 units/kg in SPG versus 0.3 units/kg in BBI), and number of insulin injections per day (2.2 in SPG versus 2.9 in BBI). There was no difference in the number of hypoglycemic or hyperglycemic events, length of hospital stay (approximately 4 days in both groups), and rates of complications (including acute respiratory failure, acute kidney injury, and myocardial infarction) between groups.

Conclusion. Inpatient treatment with sitagliptin plus glargine was noninferior to basal–bolus insulin therapy in measurements of glycemic control.

Commentary

Approximately 25% to 30% of adult patients admitted to general medical and surgical wards and critical care units have type 2 diabetes [1]. Maintaining adequate blood sugar control is important, as both hyperglycemia and hypoglycemia have been associated with adverse outcomes. Although group consensus statements differ slightly with respect to recommended target glucose levels, generally the recommended range in a noncritical inpatient setting is 140 to 180 mg/dL [2,3]. Establishing and maintaining these levels can often be very challenging. Barriers to achieving adequate glucose control in the inpatient setting include changes in a patients’ nutrition status, renal function, pain level, the use of glucocorticoids, and the development of infections. In addition, a significant gap in knowledge can exist from provider to provider in terms of how to appropriately initiate and titrate insulin regimens. To circumvent this, many hospitals have created built-in order sets and protocols in the electronic medical record for basal–bolus correction insulin regimens. While these protocols may have improved many parameters of inpatient diabetes management at several institutions, improper initiation and execution of these protocols still occur. Also, at times the priorities of the medical team can shift so that titration of the insulin regimen may not occur frequently enough. Overall, simplification of inpatient glucose management would certainly be a welcomed change.

Unfortunately, there is a dearth of studies that investigate the role of oral therapy in the inpatient setting. In general, oral medications are discontinued upon admission and insulin is the recommended standard of care. In this study, Pasquel and colleagues investigated the use of the DPP-4 inhibitor sitagliptin in the inpatient setting. Unlike some of the other classes of oral agents used in the outpatient setting, DPP-4 inhibitors are generally well tolerated. A major advantage of DPP-4 inhibitors is that, with dose titration, they can also be used in mild to moderate renal failure. However, because DPP-4 inhibitors work in the prandial setting, they are not effective in the NPO patient. In this study, both the SPG group and BBI group had similar average daily blood glucose levels after the first day of therapy and throughout the hospitalization (171 mg/dL in SPG versus 169 mg/dL in BBI). Since the key finding here was noninferiority for blood sugar control between the treatments, the major differences between SPG and BBI therapy should be highlighted.

One benefit of SPG versus BBI therapy is that replacement of bolus insulin injections with a once-daily pill reduces the need for frequent bolus insulin dose titration. Nonetheless, renal function should be monitored frequently, as sitagliptin dose adjustments may be required, and the importance of bedside glucose checks should not be diminished, as some patients may not maintain adequate control on this regimen and will need to betransitioned to BBI therapy. Both treatment groups received correctional insulin doses in the prandial setting if their pre-meal glucose levels met a specific threshold. Overall, the SPG group required significantly fewer total insulin injections per day (2.2 injections in SPG versus 2.9 injections in BBI, P < 0.001). Though this difference is rather small, the need to administer fewer insulin injections would certainly be beneficial to nursing staff, who often care for several type 2 diabetes patients at once. It would have been interesting to know how many patients in each group were free of any correctional insulin doses or how many were adequately controlled with just 1 prandial injection per day. Although it cannot be concluded from this study, it could be expected that the reduced need for bolus insulin dose titration and fewer total insulin injections associated with oral therapy would result in less insulin dosing error and perhaps greater patient satisfaction.

It is important to keep in mind that initiating a DPP-4 inhibitor with basal insulin may not be an appropriate option for all admitted type 2 diabetes patients. It can be a beneficial alternative to insulin for the select group of patients included in this study: those treated at home with diet alone, oral therapy alone, or oral therapy plus insulin.

While the potential for implementation of SPG therapy in an inpatient setting does exist, there are some limitations to this study that make further investigation necessary. Though the patent on Januvia (sitagliptin’s trade name) expires in 2017, sitagliptin is currently a very expensive drug. Therefore, a cost-benefit analysis of SPG therapy versus insulin therapy alone should be undertaken. Also, this was an unblinded study, which may have resulted in more attentive, prioritized blood sugar management than what would typically occur in an inpatient setting. Also, the providers’ level of expertise on insulin management in this study may not be generalized to all inpatient medical and surgical providers. Despite these limitations, this study may have a profound impact on inpatient diabetes management, since a less labor-intensive alternative to basal–bolus insulin therapy may present a more attractive option for many inpatient providers.

Applications for Clinical Practice

This study could pave the way for a practice-changing method of inpatient glucose management for a select group of patients who do not have severely uncontrolled type 2 diabetes. One should keep in mind that cost could be a barrier to implementation of sitagliptin in hospitals, and that while the bolus dose of insulin can be replaced with sitagliptin, patients may still need correctional doses of insulin to maintain target ranges. Also, a daily assess-ment of glucose control is still necessary in order to determine if a change in management is needed. Therefore, the sitagliptin plus glargine option should not be viewed as a “shortcut” therapy, but rather as a potentially less labor-intensive option that may increase the ability to prioritize blood sugar management in the inpatient setting.

— Lisa Parikh, MD, Yale School of Medicine,
New Haven, CT

Study Overview

Objective. To compare the safety and efficacy of basal–bolus insulin therapy with sitagliptin plus insulin glargine in type 2 diabetes patients admitted to general medicine and surgical wards.

Design. Multicenter, prospective, open-label, noninferiority randomized clinical trial.

Setting and participants. Type 2 diabetes patients aged 18 to 80 years admitted to the general medicine and surgery services at one of 5 academic-based US hospitals were recruited. Eligible participants presented with a random blood glucose concentration between 140 and 400 mg/dL and were treated at home with diet, oral agents, or oral agents plus insulin at a maximum daily dose of 0.6 units/kg. Among those excluded were patients recently treated with a dipeptidyl peptidase-4 (DPP-4) inhibitor or glucagon-like peptide-1 (GLP-1) agonist, patients with clinically relevant hepatic disease, patients who were not eating for more than 48 hours, and those with an estimated glomerular filtration rate (eGFR) < 30 mL/min.

Intervention. Participants were randomly assigned to receive basal–bolus insulin therapy (BBI) with glargine once daily plus rapid-acting insulin before meals or sitagliptin plus glargine (SPG) once daily. Those in the SPG group received sitagliptin 100 mg/day if their eGFR was > 50 mL/min and sitagliptin 50 mg/day if their eGFR was 30 to 50 mL/min. If the eGFR fell below 30 mL/min during the hospitalization, sitagliptin was reduced to 25 mg/day. Glargine doses for those in the SPG group were started at 0.2 units/kg if randomization blood glucose was 140–200 mg/dL and 0.25 units/kg if randomization glucose was 201–400 mg/dL. Patients aged 70 or older or with an eGFR < 50 mL/min started with a daily glargine dose of 0.15 units/kg. For the BBI group, a total daily insulin dose of 0.4 units/kg was initiated for those with blood glucose levels between 140 and 200 mg/dL, and 0.5 units/kg for those with randomization glucose between 201 and 400 mg/dL. Half of this daily dose was given as glargine and the other half was distributed evenly across 3 pre-meal doses. Both the BBI and SPG groups received pre-meal and bedtime correction doses of rapid-acting insulin for glucose levels above 140 mg/dL. Blood glucose concentrations were measured fasting, before meals, and at bedtime or every 6 hours for patients who were not eating. Target fasting and pre-meal blood glucose levels were 100 to 140 mg/dL. Investigators and participants were not blinded to group assignment and glucose control was managed by the primary medical or surgical team.

Main outcomes measure. The primary outcome for this trial was noninferiority for differences between the SPG and BBI groups in glycemic control. Secondary endpoints included differences in the number of hypoglycemic and hyperglycemic events, the number of blood glucose values between 70 and 140 mg/dL and between 70 and 180 mg/dL, and the number of treatment failures (defined as 2 consecutive blood glucose values > 240 mg/dL or mean daily glucose > 240 mg/dL), length of hospital stay, total daily dose of insulin, number of insulin injections per day, transfer to the intensive care unit, and hospital complications and mortality.

Main results. A total of 138 patients in the SPG group and 139 patients in the BBI group completed the study and were included in this analysis. Of these 277 patients, 84% were admitted to a medicine ward and 16% were admitted to a surgical ward. The average age of participants was approximately 57 years, the average BMI was approximately 35 kg/m², and the average duration of diabetes was approximately 10 years. These baseline characteristics as well as ethnic origin, sex, and baseline A1c (approximately 40% of patients in both groups had a baseline A1c between 7% and 9%) did not differ between groups. Prior to admission, approximately 40% of patients in both groups were managed with oral drugs alone, approximately 25% were managed with insulin alone, and about 22% were managed with insulin and oral therapy.

With respect to the primary outcome, both groups had similar mean daily blood glucose concentrations (171 mg/dL in SPG and 169 mg/dL in BBI) throughout the hospitalization, meeting the noninferiority threshold for glycemic control between groups. As for secondary outcomes, the mean proportion of blood glucose readings between 70 and 140 mg/dL, 70 and 180 mg/dL, and 100 and 140 mg/dL did not differ between groups. Pre-meal and bedtime blood glucose concentrations were also similar in both groups. There was a significant difference between groups in average daily insulin dose (24 units in SPG versus 34 units in BBI), total units of insulin per kg per day (0.2 units/kg in SPG versus 0.3 units/kg in BBI), and number of insulin injections per day (2.2 in SPG versus 2.9 in BBI). There was no difference in the number of hypoglycemic or hyperglycemic events, length of hospital stay (approximately 4 days in both groups), and rates of complications (including acute respiratory failure, acute kidney injury, and myocardial infarction) between groups.

Conclusion. Inpatient treatment with sitagliptin plus glargine was noninferior to basal–bolus insulin therapy in measurements of glycemic control.

Commentary

Approximately 25% to 30% of adult patients admitted to general medical and surgical wards and critical care units have type 2 diabetes [1]. Maintaining adequate blood sugar control is important, as both hyperglycemia and hypoglycemia have been associated with adverse outcomes. Although group consensus statements differ slightly with respect to recommended target glucose levels, generally the recommended range in a noncritical inpatient setting is 140 to 180 mg/dL [2,3]. Establishing and maintaining these levels can often be very challenging. Barriers to achieving adequate glucose control in the inpatient setting include changes in a patients’ nutrition status, renal function, pain level, the use of glucocorticoids, and the development of infections. In addition, a significant gap in knowledge can exist from provider to provider in terms of how to appropriately initiate and titrate insulin regimens. To circumvent this, many hospitals have created built-in order sets and protocols in the electronic medical record for basal–bolus correction insulin regimens. While these protocols may have improved many parameters of inpatient diabetes management at several institutions, improper initiation and execution of these protocols still occur. Also, at times the priorities of the medical team can shift so that titration of the insulin regimen may not occur frequently enough. Overall, simplification of inpatient glucose management would certainly be a welcomed change.

Unfortunately, there is a dearth of studies that investigate the role of oral therapy in the inpatient setting. In general, oral medications are discontinued upon admission and insulin is the recommended standard of care. In this study, Pasquel and colleagues investigated the use of the DPP-4 inhibitor sitagliptin in the inpatient setting. Unlike some of the other classes of oral agents used in the outpatient setting, DPP-4 inhibitors are generally well tolerated. A major advantage of DPP-4 inhibitors is that, with dose titration, they can also be used in mild to moderate renal failure. However, because DPP-4 inhibitors work in the prandial setting, they are not effective in the NPO patient. In this study, both the SPG group and BBI group had similar average daily blood glucose levels after the first day of therapy and throughout the hospitalization (171 mg/dL in SPG versus 169 mg/dL in BBI). Since the key finding here was noninferiority for blood sugar control between the treatments, the major differences between SPG and BBI therapy should be highlighted.

One benefit of SPG versus BBI therapy is that replacement of bolus insulin injections with a once-daily pill reduces the need for frequent bolus insulin dose titration. Nonetheless, renal function should be monitored frequently, as sitagliptin dose adjustments may be required, and the importance of bedside glucose checks should not be diminished, as some patients may not maintain adequate control on this regimen and will need to betransitioned to BBI therapy. Both treatment groups received correctional insulin doses in the prandial setting if their pre-meal glucose levels met a specific threshold. Overall, the SPG group required significantly fewer total insulin injections per day (2.2 injections in SPG versus 2.9 injections in BBI, P < 0.001). Though this difference is rather small, the need to administer fewer insulin injections would certainly be beneficial to nursing staff, who often care for several type 2 diabetes patients at once. It would have been interesting to know how many patients in each group were free of any correctional insulin doses or how many were adequately controlled with just 1 prandial injection per day. Although it cannot be concluded from this study, it could be expected that the reduced need for bolus insulin dose titration and fewer total insulin injections associated with oral therapy would result in less insulin dosing error and perhaps greater patient satisfaction.

It is important to keep in mind that initiating a DPP-4 inhibitor with basal insulin may not be an appropriate option for all admitted type 2 diabetes patients. It can be a beneficial alternative to insulin for the select group of patients included in this study: those treated at home with diet alone, oral therapy alone, or oral therapy plus insulin.

While the potential for implementation of SPG therapy in an inpatient setting does exist, there are some limitations to this study that make further investigation necessary. Though the patent on Januvia (sitagliptin’s trade name) expires in 2017, sitagliptin is currently a very expensive drug. Therefore, a cost-benefit analysis of SPG therapy versus insulin therapy alone should be undertaken. Also, this was an unblinded study, which may have resulted in more attentive, prioritized blood sugar management than what would typically occur in an inpatient setting. Also, the providers’ level of expertise on insulin management in this study may not be generalized to all inpatient medical and surgical providers. Despite these limitations, this study may have a profound impact on inpatient diabetes management, since a less labor-intensive alternative to basal–bolus insulin therapy may present a more attractive option for many inpatient providers.

Applications for Clinical Practice

This study could pave the way for a practice-changing method of inpatient glucose management for a select group of patients who do not have severely uncontrolled type 2 diabetes. One should keep in mind that cost could be a barrier to implementation of sitagliptin in hospitals, and that while the bolus dose of insulin can be replaced with sitagliptin, patients may still need correctional doses of insulin to maintain target ranges. Also, a daily assess-ment of glucose control is still necessary in order to determine if a change in management is needed. Therefore, the sitagliptin plus glargine option should not be viewed as a “shortcut” therapy, but rather as a potentially less labor-intensive option that may increase the ability to prioritize blood sugar management in the inpatient setting.

— Lisa Parikh, MD, Yale School of Medicine,
New Haven, CT

Study Overview

Objective. To compare the safety and efficacy of basal–bolus insulin therapy with sitagliptin plus insulin glargine in type 2 diabetes patients admitted to general medicine and surgical wards.

Design. Multicenter, prospective, open-label, noninferiority randomized clinical trial.

Setting and participants. Type 2 diabetes patients aged 18 to 80 years admitted to the general medicine and surgery services at one of 5 academic-based US hospitals were recruited. Eligible participants presented with a random blood glucose concentration between 140 and 400 mg/dL and were treated at home with diet, oral agents, or oral agents plus insulin at a maximum daily dose of 0.6 units/kg. Among those excluded were patients recently treated with a dipeptidyl peptidase-4 (DPP-4) inhibitor or glucagon-like peptide-1 (GLP-1) agonist, patients with clinically relevant hepatic disease, patients who were not eating for more than 48 hours, and those with an estimated glomerular filtration rate (eGFR) < 30 mL/min.

Intervention. Participants were randomly assigned to receive basal–bolus insulin therapy (BBI) with glargine once daily plus rapid-acting insulin before meals or sitagliptin plus glargine (SPG) once daily. Those in the SPG group received sitagliptin 100 mg/day if their eGFR was > 50 mL/min and sitagliptin 50 mg/day if their eGFR was 30 to 50 mL/min. If the eGFR fell below 30 mL/min during the hospitalization, sitagliptin was reduced to 25 mg/day. Glargine doses for those in the SPG group were started at 0.2 units/kg if randomization blood glucose was 140–200 mg/dL and 0.25 units/kg if randomization glucose was 201–400 mg/dL. Patients aged 70 or older or with an eGFR < 50 mL/min started with a daily glargine dose of 0.15 units/kg. For the BBI group, a total daily insulin dose of 0.4 units/kg was initiated for those with blood glucose levels between 140 and 200 mg/dL, and 0.5 units/kg for those with randomization glucose between 201 and 400 mg/dL. Half of this daily dose was given as glargine and the other half was distributed evenly across 3 pre-meal doses. Both the BBI and SPG groups received pre-meal and bedtime correction doses of rapid-acting insulin for glucose levels above 140 mg/dL. Blood glucose concentrations were measured fasting, before meals, and at bedtime or every 6 hours for patients who were not eating. Target fasting and pre-meal blood glucose levels were 100 to 140 mg/dL. Investigators and participants were not blinded to group assignment and glucose control was managed by the primary medical or surgical team.

Main outcomes measure. The primary outcome for this trial was noninferiority for differences between the SPG and BBI groups in glycemic control. Secondary endpoints included differences in the number of hypoglycemic and hyperglycemic events, the number of blood glucose values between 70 and 140 mg/dL and between 70 and 180 mg/dL, and the number of treatment failures (defined as 2 consecutive blood glucose values > 240 mg/dL or mean daily glucose > 240 mg/dL), length of hospital stay, total daily dose of insulin, number of insulin injections per day, transfer to the intensive care unit, and hospital complications and mortality.

Main results. A total of 138 patients in the SPG group and 139 patients in the BBI group completed the study and were included in this analysis. Of these 277 patients, 84% were admitted to a medicine ward and 16% were admitted to a surgical ward. The average age of participants was approximately 57 years, the average BMI was approximately 35 kg/m², and the average duration of diabetes was approximately 10 years. These baseline characteristics as well as ethnic origin, sex, and baseline A1c (approximately 40% of patients in both groups had a baseline A1c between 7% and 9%) did not differ between groups. Prior to admission, approximately 40% of patients in both groups were managed with oral drugs alone, approximately 25% were managed with insulin alone, and about 22% were managed with insulin and oral therapy.

With respect to the primary outcome, both groups had similar mean daily blood glucose concentrations (171 mg/dL in SPG and 169 mg/dL in BBI) throughout the hospitalization, meeting the noninferiority threshold for glycemic control between groups. As for secondary outcomes, the mean proportion of blood glucose readings between 70 and 140 mg/dL, 70 and 180 mg/dL, and 100 and 140 mg/dL did not differ between groups. Pre-meal and bedtime blood glucose concentrations were also similar in both groups. There was a significant difference between groups in average daily insulin dose (24 units in SPG versus 34 units in BBI), total units of insulin per kg per day (0.2 units/kg in SPG versus 0.3 units/kg in BBI), and number of insulin injections per day (2.2 in SPG versus 2.9 in BBI). There was no difference in the number of hypoglycemic or hyperglycemic events, length of hospital stay (approximately 4 days in both groups), and rates of complications (including acute respiratory failure, acute kidney injury, and myocardial infarction) between groups.

Conclusion. Inpatient treatment with sitagliptin plus glargine was noninferior to basal–bolus insulin therapy in measurements of glycemic control.

Commentary

Approximately 25% to 30% of adult patients admitted to general medical and surgical wards and critical care units have type 2 diabetes [1]. Maintaining adequate blood sugar control is important, as both hyperglycemia and hypoglycemia have been associated with adverse outcomes. Although group consensus statements differ slightly with respect to recommended target glucose levels, generally the recommended range in a noncritical inpatient setting is 140 to 180 mg/dL [2,3]. Establishing and maintaining these levels can often be very challenging. Barriers to achieving adequate glucose control in the inpatient setting include changes in a patients’ nutrition status, renal function, pain level, the use of glucocorticoids, and the development of infections. In addition, a significant gap in knowledge can exist from provider to provider in terms of how to appropriately initiate and titrate insulin regimens. To circumvent this, many hospitals have created built-in order sets and protocols in the electronic medical record for basal–bolus correction insulin regimens. While these protocols may have improved many parameters of inpatient diabetes management at several institutions, improper initiation and execution of these protocols still occur. Also, at times the priorities of the medical team can shift so that titration of the insulin regimen may not occur frequently enough. Overall, simplification of inpatient glucose management would certainly be a welcomed change.

Unfortunately, there is a dearth of studies that investigate the role of oral therapy in the inpatient setting. In general, oral medications are discontinued upon admission and insulin is the recommended standard of care. In this study, Pasquel and colleagues investigated the use of the DPP-4 inhibitor sitagliptin in the inpatient setting. Unlike some of the other classes of oral agents used in the outpatient setting, DPP-4 inhibitors are generally well tolerated. A major advantage of DPP-4 inhibitors is that, with dose titration, they can also be used in mild to moderate renal failure. However, because DPP-4 inhibitors work in the prandial setting, they are not effective in the NPO patient. In this study, both the SPG group and BBI group had similar average daily blood glucose levels after the first day of therapy and throughout the hospitalization (171 mg/dL in SPG versus 169 mg/dL in BBI). Since the key finding here was noninferiority for blood sugar control between the treatments, the major differences between SPG and BBI therapy should be highlighted.

One benefit of SPG versus BBI therapy is that replacement of bolus insulin injections with a once-daily pill reduces the need for frequent bolus insulin dose titration. Nonetheless, renal function should be monitored frequently, as sitagliptin dose adjustments may be required, and the importance of bedside glucose checks should not be diminished, as some patients may not maintain adequate control on this regimen and will need to betransitioned to BBI therapy. Both treatment groups received correctional insulin doses in the prandial setting if their pre-meal glucose levels met a specific threshold. Overall, the SPG group required significantly fewer total insulin injections per day (2.2 injections in SPG versus 2.9 injections in BBI, P < 0.001). Though this difference is rather small, the need to administer fewer insulin injections would certainly be beneficial to nursing staff, who often care for several type 2 diabetes patients at once. It would have been interesting to know how many patients in each group were free of any correctional insulin doses or how many were adequately controlled with just 1 prandial injection per day. Although it cannot be concluded from this study, it could be expected that the reduced need for bolus insulin dose titration and fewer total insulin injections associated with oral therapy would result in less insulin dosing error and perhaps greater patient satisfaction.

It is important to keep in mind that initiating a DPP-4 inhibitor with basal insulin may not be an appropriate option for all admitted type 2 diabetes patients. It can be a beneficial alternative to insulin for the select group of patients included in this study: those treated at home with diet alone, oral therapy alone, or oral therapy plus insulin.

While the potential for implementation of SPG therapy in an inpatient setting does exist, there are some limitations to this study that make further investigation necessary. Though the patent on Januvia (sitagliptin’s trade name) expires in 2017, sitagliptin is currently a very expensive drug. Therefore, a cost-benefit analysis of SPG therapy versus insulin therapy alone should be undertaken. Also, this was an unblinded study, which may have resulted in more attentive, prioritized blood sugar management than what would typically occur in an inpatient setting. Also, the providers’ level of expertise on insulin management in this study may not be generalized to all inpatient medical and surgical providers. Despite these limitations, this study may have a profound impact on inpatient diabetes management, since a less labor-intensive alternative to basal–bolus insulin therapy may present a more attractive option for many inpatient providers.

Applications for Clinical Practice

This study could pave the way for a practice-changing method of inpatient glucose management for a select group of patients who do not have severely uncontrolled type 2 diabetes. One should keep in mind that cost could be a barrier to implementation of sitagliptin in hospitals, and that while the bolus dose of insulin can be replaced with sitagliptin, patients may still need correctional doses of insulin to maintain target ranges. Also, a daily assess-ment of glucose control is still necessary in order to determine if a change in management is needed. Therefore, the sitagliptin plus glargine option should not be viewed as a “shortcut” therapy, but rather as a potentially less labor-intensive option that may increase the ability to prioritize blood sugar management in the inpatient setting.

— Lisa Parikh, MD, Yale School of Medicine,
New Haven, CT

With great pleasure we announce a collaboration between Cutis^® and the Skin of Color Society (SOCS) to increase the knowledge available to dermatologists to help improve delivery of care to this underserved population.

Established in 2004 by Susan C. Taylor, MD (who also serves on the Cutis Editorial Board), SOCS (http://www.skinofcolorsociety.org) promotes, supports, and stimulates the development of information related to all aspects of skin of color within the specialty of dermatology, making this information readily available to the general population.

“Although a relatively new organization, SOCS has been essential in supporting and encouraging research and scholarly activity to increase our understanding of the ethnic differences that occur in problems related to hair, skin, and nails of the growing population of darker-skinned individuals in our country,” said Vincent A. DeLeo, MD, Editor-in-Chief of Cutis and a founding member of SOCS. “In addition, SOCS has been essential in mentoring young students and increasing minority participation in dermatology, and Cutis will strive to assist in those endeavors.”

The society also seeks to increase the body of dermatologic literature related to skin of color. To achieve this goal, SOCS will be collaborating with the editors of Cutis to publish quarterly Skin of Color columns to educate dermatologists and residents on basic science and clinical, surgical, and cosmetic research relevant to this patient population.

“SOCS is very excited to collaborate with Cutis in our mutual academic pursuits,” said Seemal R. Desai, MD, current secretary/treasurer of SOCS and president-elect. “It is vitally important to the mission of SOCS that dermatologists and patients be educated with the most up-to-date objective data, studies, and information that is available to most effectively help those suffering from skin disease in the skin of color population.”

Look for Skin of Color columns in upcoming issues of Cutis.

With great pleasure we announce a collaboration between Cutis^® and the Skin of Color Society (SOCS) to increase the knowledge available to dermatologists to help improve delivery of care to this underserved population.

Established in 2004 by Susan C. Taylor, MD (who also serves on the Cutis Editorial Board), SOCS (http://www.skinofcolorsociety.org) promotes, supports, and stimulates the development of information related to all aspects of skin of color within the specialty of dermatology, making this information readily available to the general population.

“Although a relatively new organization, SOCS has been essential in supporting and encouraging research and scholarly activity to increase our understanding of the ethnic differences that occur in problems related to hair, skin, and nails of the growing population of darker-skinned individuals in our country,” said Vincent A. DeLeo, MD, Editor-in-Chief of Cutis and a founding member of SOCS. “In addition, SOCS has been essential in mentoring young students and increasing minority participation in dermatology, and Cutis will strive to assist in those endeavors.”

The society also seeks to increase the body of dermatologic literature related to skin of color. To achieve this goal, SOCS will be collaborating with the editors of Cutis to publish quarterly Skin of Color columns to educate dermatologists and residents on basic science and clinical, surgical, and cosmetic research relevant to this patient population.

“SOCS is very excited to collaborate with Cutis in our mutual academic pursuits,” said Seemal R. Desai, MD, current secretary/treasurer of SOCS and president-elect. “It is vitally important to the mission of SOCS that dermatologists and patients be educated with the most up-to-date objective data, studies, and information that is available to most effectively help those suffering from skin disease in the skin of color population.”

Look for Skin of Color columns in upcoming issues of Cutis.

With great pleasure we announce a collaboration between Cutis^® and the Skin of Color Society (SOCS) to increase the knowledge available to dermatologists to help improve delivery of care to this underserved population.

Established in 2004 by Susan C. Taylor, MD (who also serves on the Cutis Editorial Board), SOCS (http://www.skinofcolorsociety.org) promotes, supports, and stimulates the development of information related to all aspects of skin of color within the specialty of dermatology, making this information readily available to the general population.

“Although a relatively new organization, SOCS has been essential in supporting and encouraging research and scholarly activity to increase our understanding of the ethnic differences that occur in problems related to hair, skin, and nails of the growing population of darker-skinned individuals in our country,” said Vincent A. DeLeo, MD, Editor-in-Chief of Cutis and a founding member of SOCS. “In addition, SOCS has been essential in mentoring young students and increasing minority participation in dermatology, and Cutis will strive to assist in those endeavors.”

The society also seeks to increase the body of dermatologic literature related to skin of color. To achieve this goal, SOCS will be collaborating with the editors of Cutis to publish quarterly Skin of Color columns to educate dermatologists and residents on basic science and clinical, surgical, and cosmetic research relevant to this patient population.

“SOCS is very excited to collaborate with Cutis in our mutual academic pursuits,” said Seemal R. Desai, MD, current secretary/treasurer of SOCS and president-elect. “It is vitally important to the mission of SOCS that dermatologists and patients be educated with the most up-to-date objective data, studies, and information that is available to most effectively help those suffering from skin disease in the skin of color population.”

Look for Skin of Color columns in upcoming issues of Cutis.

MIAMI—Among patients with Parkinson’s disease–associated mild cognitive impairment (MCI), Montreal Cognitive Assessment (MoCA) subscores in visuospatial function, attention, language, and orientation are the most useful in predicting conversion to dementia, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress.

Melissa Mackenzie, of the Division of Neurology at the University of British Columbia in Vancouver, and colleagues conducted a study to evaluate which subscores on the cognitive assessment predict conversion to dementia in patients with Parkinson’s disease–associated MCI.

The investigators searched the Pacific Parkinson’s Research Centre Database to identify patients with a diagnosis of idiopathic Parkinson’s disease who completed an itemized MoCA in the MCI range (ie, they had a corrected total score between 21 and 27) and who completed at least one other MoCA at least one year later. Patients taking potentially cognitive enhancing medications were excluded.

The researchers included in their study 529 assessments from 164 patients. They separated patients into three groups based on their last MoCA score—those who developed dementia (33 patients), those who returned to normal cognition (48 patients), and those who maintained MoCA scores in the MCI range (83 patients).

In a model that predicted future MoCA score categories with 78% accuracy, the most important subscores were visuospatial, attention, language, and orientation, “but, interestingly, not delayed recall,” Dr. Mackenzie and colleagues said.

“A prevailing theory of cognitive decline in Parkinson’s disease postulates that visuospatial ‘posterior-cortical’ impairments are due to Lewy body deposition, whereas frontal executive dysfunction reflects ‘on–off’ state,” the researchers said. “Interestingly, language scores and memory function in delayed recall were the items that improved the most” in patients who converted from MCI to normal cognition.

“Whether the best approach to assess risk of conversion to dementia is to focus exclusively on these MoCA sections, or alternatively, employing multiple tests that target these cognitive domains remains to be seen,” the researchers concluded.

Patients with Parkinson’s disease–associated MCI at any time “should likely be followed more closely for cognitive decline, as they seem to be at increased risk for developing dementia, even if there is interval maintenance of MCI or return to normal cognition.”

—Jake Remaly

Suggested Reading

Pedersen KF, Larsen JP, Tysnes OB, Alves G. Natural course of mild cognitive impairment in Parkinson disease: A 5-year population-based study. Neurology. 2017;88(8):767-774.

MIAMI—Among patients with Parkinson’s disease–associated mild cognitive impairment (MCI), Montreal Cognitive Assessment (MoCA) subscores in visuospatial function, attention, language, and orientation are the most useful in predicting conversion to dementia, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress.

Melissa Mackenzie, of the Division of Neurology at the University of British Columbia in Vancouver, and colleagues conducted a study to evaluate which subscores on the cognitive assessment predict conversion to dementia in patients with Parkinson’s disease–associated MCI.

The investigators searched the Pacific Parkinson’s Research Centre Database to identify patients with a diagnosis of idiopathic Parkinson’s disease who completed an itemized MoCA in the MCI range (ie, they had a corrected total score between 21 and 27) and who completed at least one other MoCA at least one year later. Patients taking potentially cognitive enhancing medications were excluded.

The researchers included in their study 529 assessments from 164 patients. They separated patients into three groups based on their last MoCA score—those who developed dementia (33 patients), those who returned to normal cognition (48 patients), and those who maintained MoCA scores in the MCI range (83 patients).

In a model that predicted future MoCA score categories with 78% accuracy, the most important subscores were visuospatial, attention, language, and orientation, “but, interestingly, not delayed recall,” Dr. Mackenzie and colleagues said.

“A prevailing theory of cognitive decline in Parkinson’s disease postulates that visuospatial ‘posterior-cortical’ impairments are due to Lewy body deposition, whereas frontal executive dysfunction reflects ‘on–off’ state,” the researchers said. “Interestingly, language scores and memory function in delayed recall were the items that improved the most” in patients who converted from MCI to normal cognition.

“Whether the best approach to assess risk of conversion to dementia is to focus exclusively on these MoCA sections, or alternatively, employing multiple tests that target these cognitive domains remains to be seen,” the researchers concluded.

Patients with Parkinson’s disease–associated MCI at any time “should likely be followed more closely for cognitive decline, as they seem to be at increased risk for developing dementia, even if there is interval maintenance of MCI or return to normal cognition.”

—Jake Remaly

Suggested Reading

Pedersen KF, Larsen JP, Tysnes OB, Alves G. Natural course of mild cognitive impairment in Parkinson disease: A 5-year population-based study. Neurology. 2017;88(8):767-774.

MIAMI—Among patients with Parkinson’s disease–associated mild cognitive impairment (MCI), Montreal Cognitive Assessment (MoCA) subscores in visuospatial function, attention, language, and orientation are the most useful in predicting conversion to dementia, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress.

Melissa Mackenzie, of the Division of Neurology at the University of British Columbia in Vancouver, and colleagues conducted a study to evaluate which subscores on the cognitive assessment predict conversion to dementia in patients with Parkinson’s disease–associated MCI.

The investigators searched the Pacific Parkinson’s Research Centre Database to identify patients with a diagnosis of idiopathic Parkinson’s disease who completed an itemized MoCA in the MCI range (ie, they had a corrected total score between 21 and 27) and who completed at least one other MoCA at least one year later. Patients taking potentially cognitive enhancing medications were excluded.

The researchers included in their study 529 assessments from 164 patients. They separated patients into three groups based on their last MoCA score—those who developed dementia (33 patients), those who returned to normal cognition (48 patients), and those who maintained MoCA scores in the MCI range (83 patients).

In a model that predicted future MoCA score categories with 78% accuracy, the most important subscores were visuospatial, attention, language, and orientation, “but, interestingly, not delayed recall,” Dr. Mackenzie and colleagues said.

“A prevailing theory of cognitive decline in Parkinson’s disease postulates that visuospatial ‘posterior-cortical’ impairments are due to Lewy body deposition, whereas frontal executive dysfunction reflects ‘on–off’ state,” the researchers said. “Interestingly, language scores and memory function in delayed recall were the items that improved the most” in patients who converted from MCI to normal cognition.

“Whether the best approach to assess risk of conversion to dementia is to focus exclusively on these MoCA sections, or alternatively, employing multiple tests that target these cognitive domains remains to be seen,” the researchers concluded.

Patients with Parkinson’s disease–associated MCI at any time “should likely be followed more closely for cognitive decline, as they seem to be at increased risk for developing dementia, even if there is interval maintenance of MCI or return to normal cognition.”

—Jake Remaly

Suggested Reading

Pedersen KF, Larsen JP, Tysnes OB, Alves G. Natural course of mild cognitive impairment in Parkinson disease: A 5-year population-based study. Neurology. 2017;88(8):767-774.

MIAMI—Among patients with Parkinson’s disease with well-defined morning off episodes, 83% achieved an on-medication state within 45 minutes of treatment with sublingual apomorphine film, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress. Investigators presented preliminary results from an open-label dose titration phase of a phase III trial.

The sublingual apomorphine film, known as APL-130277, is being developed by Marlborough, Massachusetts-based Sunovion Pharmaceuticals. Apomorphine injected subcutaneously is approved for the acute, intermittent treatment of off episodes, but it is not widely used, possibly because of its parenteral administration, researchers have said. The dissolvable film consists of an apomorphine drug layer and a second layer that is designed to neutralize acid generation and enhance drug permeability. The film appeared to be effective in a phase II open-label study.

Daily Off Episodes

Participants were older than 18, had idiopathic Parkinson’s disease, and had a modified Hoehn and Yahr stage between 1 and 3 on medication. Participants were responsive to levodopa and had more than two hours of off time per day. Patients were receiving stable doses of levodopa–carbidopa.

Investigators excluded patients with psychosis, dementia, or impulse control disorders; mouth cankers or sores; or prior treatment of Parkinson’s disease with a neurosurgical procedure, continuous subcutaneous apomorphine infusion, or levodopa–carbidopa enteral suspension. They also excluded patients who received subcutaneous apomorphine within seven days before screening or were taking 5-HT3 antagonists, dopamine antagonists (other than quetiapine or clozapine), or dopamine-depleting agents.

Three days prior to the dose titration phase, patients initiated treatment with trimethobenzamide or domperidone, which may reduce nausea and vomiting that can occur during the initiation of apomorphine therapy. Patients arrived at a clinic in an off state, having not taken their regular morning levodopa dose or other adjunctive medication later than midnight, and received 10 mg of APL-130277. Investigators assessed patients using the Unified Parkinson’s Disease Rating Scale Part III prior to dosing and 15, 30, 45, 60, and 90 minutes after dosing. Patients who responded with a fully on response (ie, patients and investigators agreed that medication was benefiting mobility, stiffness, and slowness such that patients had adequate motor function to perform their normal daily activities) were considered to have completed the dose titration phase and could proceed to randomization for the maintenance treatment phase. Patients who responded to a dose could try the next highest dose at a subsequent titration visit to assess the potential for an improved response at the higher dose. Doses increased by 5 mg increments up to 35 mg.

Of 76 patients who entered the dose titration phase, 63 (83%) turned fully on with treatment. Among patients who turned fully on, 24 (38%) did so within 15 minutes, and 49 (78%) did so within 30 minutes. The median dose turning patients to fully on was 20 mg. Patients who turned fully on assessed time to onset as between five and 12 minutes.

Safety Data

In a presentation of preliminary safety data from the dose titration phase, Stuart Isaacson, MD, Director of the Parkinson’s Disease and Movement Disorders Center of Boca Raton in Florida, and colleagues reported that five of the 76 patients who entered the dose titration phase discontinued the trial due to adverse events—two due to nausea, one due to somnolence, one due to headache, and one due to presyncope. Another two patients withdrew consent, and nine patients who did not turn on at the 35 mg dose were discontinued from the trial. Other reported adverse events included dizziness, yawning, vomiting, and symptomatic hypotension. Most adverse events were considered mild. “In this preliminary analysis, APL-130277 was well tolerated in patients in the dose titration phase,” Dr. Isaacson and colleagues concluded.

—Jake Remaly

Suggested Reading

Hauser RA, Olanow CW, Dzyngel B, et al. Sublingual apomorphine (APL-130277) for the acute conversion of OFF to ON in Parkinson's disease. Mov Disord. 2016;31(9):1366-1372.

MIAMI—Among patients with Parkinson’s disease with well-defined morning off episodes, 83% achieved an on-medication state within 45 minutes of treatment with sublingual apomorphine film, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress. Investigators presented preliminary results from an open-label dose titration phase of a phase III trial.

The sublingual apomorphine film, known as APL-130277, is being developed by Marlborough, Massachusetts-based Sunovion Pharmaceuticals. Apomorphine injected subcutaneously is approved for the acute, intermittent treatment of off episodes, but it is not widely used, possibly because of its parenteral administration, researchers have said. The dissolvable film consists of an apomorphine drug layer and a second layer that is designed to neutralize acid generation and enhance drug permeability. The film appeared to be effective in a phase II open-label study.

Daily Off Episodes

Participants were older than 18, had idiopathic Parkinson’s disease, and had a modified Hoehn and Yahr stage between 1 and 3 on medication. Participants were responsive to levodopa and had more than two hours of off time per day. Patients were receiving stable doses of levodopa–carbidopa.

Investigators excluded patients with psychosis, dementia, or impulse control disorders; mouth cankers or sores; or prior treatment of Parkinson’s disease with a neurosurgical procedure, continuous subcutaneous apomorphine infusion, or levodopa–carbidopa enteral suspension. They also excluded patients who received subcutaneous apomorphine within seven days before screening or were taking 5-HT3 antagonists, dopamine antagonists (other than quetiapine or clozapine), or dopamine-depleting agents.

Three days prior to the dose titration phase, patients initiated treatment with trimethobenzamide or domperidone, which may reduce nausea and vomiting that can occur during the initiation of apomorphine therapy. Patients arrived at a clinic in an off state, having not taken their regular morning levodopa dose or other adjunctive medication later than midnight, and received 10 mg of APL-130277. Investigators assessed patients using the Unified Parkinson’s Disease Rating Scale Part III prior to dosing and 15, 30, 45, 60, and 90 minutes after dosing. Patients who responded with a fully on response (ie, patients and investigators agreed that medication was benefiting mobility, stiffness, and slowness such that patients had adequate motor function to perform their normal daily activities) were considered to have completed the dose titration phase and could proceed to randomization for the maintenance treatment phase. Patients who responded to a dose could try the next highest dose at a subsequent titration visit to assess the potential for an improved response at the higher dose. Doses increased by 5 mg increments up to 35 mg.

Of 76 patients who entered the dose titration phase, 63 (83%) turned fully on with treatment. Among patients who turned fully on, 24 (38%) did so within 15 minutes, and 49 (78%) did so within 30 minutes. The median dose turning patients to fully on was 20 mg. Patients who turned fully on assessed time to onset as between five and 12 minutes.

Safety Data

In a presentation of preliminary safety data from the dose titration phase, Stuart Isaacson, MD, Director of the Parkinson’s Disease and Movement Disorders Center of Boca Raton in Florida, and colleagues reported that five of the 76 patients who entered the dose titration phase discontinued the trial due to adverse events—two due to nausea, one due to somnolence, one due to headache, and one due to presyncope. Another two patients withdrew consent, and nine patients who did not turn on at the 35 mg dose were discontinued from the trial. Other reported adverse events included dizziness, yawning, vomiting, and symptomatic hypotension. Most adverse events were considered mild. “In this preliminary analysis, APL-130277 was well tolerated in patients in the dose titration phase,” Dr. Isaacson and colleagues concluded.

—Jake Remaly

Suggested Reading

Hauser RA, Olanow CW, Dzyngel B, et al. Sublingual apomorphine (APL-130277) for the acute conversion of OFF to ON in Parkinson's disease. Mov Disord. 2016;31(9):1366-1372.

MIAMI—Among patients with Parkinson’s disease with well-defined morning off episodes, 83% achieved an on-medication state within 45 minutes of treatment with sublingual apomorphine film, according to research presented at the First Pan American Parkinson’s Disease and Movement Disorders Congress. Investigators presented preliminary results from an open-label dose titration phase of a phase III trial.

The sublingual apomorphine film, known as APL-130277, is being developed by Marlborough, Massachusetts-based Sunovion Pharmaceuticals. Apomorphine injected subcutaneously is approved for the acute, intermittent treatment of off episodes, but it is not widely used, possibly because of its parenteral administration, researchers have said. The dissolvable film consists of an apomorphine drug layer and a second layer that is designed to neutralize acid generation and enhance drug permeability. The film appeared to be effective in a phase II open-label study.

Daily Off Episodes

Participants were older than 18, had idiopathic Parkinson’s disease, and had a modified Hoehn and Yahr stage between 1 and 3 on medication. Participants were responsive to levodopa and had more than two hours of off time per day. Patients were receiving stable doses of levodopa–carbidopa.

Investigators excluded patients with psychosis, dementia, or impulse control disorders; mouth cankers or sores; or prior treatment of Parkinson’s disease with a neurosurgical procedure, continuous subcutaneous apomorphine infusion, or levodopa–carbidopa enteral suspension. They also excluded patients who received subcutaneous apomorphine within seven days before screening or were taking 5-HT3 antagonists, dopamine antagonists (other than quetiapine or clozapine), or dopamine-depleting agents.

Three days prior to the dose titration phase, patients initiated treatment with trimethobenzamide or domperidone, which may reduce nausea and vomiting that can occur during the initiation of apomorphine therapy. Patients arrived at a clinic in an off state, having not taken their regular morning levodopa dose or other adjunctive medication later than midnight, and received 10 mg of APL-130277. Investigators assessed patients using the Unified Parkinson’s Disease Rating Scale Part III prior to dosing and 15, 30, 45, 60, and 90 minutes after dosing. Patients who responded with a fully on response (ie, patients and investigators agreed that medication was benefiting mobility, stiffness, and slowness such that patients had adequate motor function to perform their normal daily activities) were considered to have completed the dose titration phase and could proceed to randomization for the maintenance treatment phase. Patients who responded to a dose could try the next highest dose at a subsequent titration visit to assess the potential for an improved response at the higher dose. Doses increased by 5 mg increments up to 35 mg.

Of 76 patients who entered the dose titration phase, 63 (83%) turned fully on with treatment. Among patients who turned fully on, 24 (38%) did so within 15 minutes, and 49 (78%) did so within 30 minutes. The median dose turning patients to fully on was 20 mg. Patients who turned fully on assessed time to onset as between five and 12 minutes.

Safety Data

In a presentation of preliminary safety data from the dose titration phase, Stuart Isaacson, MD, Director of the Parkinson’s Disease and Movement Disorders Center of Boca Raton in Florida, and colleagues reported that five of the 76 patients who entered the dose titration phase discontinued the trial due to adverse events—two due to nausea, one due to somnolence, one due to headache, and one due to presyncope. Another two patients withdrew consent, and nine patients who did not turn on at the 35 mg dose were discontinued from the trial. Other reported adverse events included dizziness, yawning, vomiting, and symptomatic hypotension. Most adverse events were considered mild. “In this preliminary analysis, APL-130277 was well tolerated in patients in the dose titration phase,” Dr. Isaacson and colleagues concluded.

—Jake Remaly

Suggested Reading

Hauser RA, Olanow CW, Dzyngel B, et al. Sublingual apomorphine (APL-130277) for the acute conversion of OFF to ON in Parkinson's disease. Mov Disord. 2016;31(9):1366-1372.

The incidence of acne in adult females is rising,¹ and treatment with combined oral contraceptive pills (OCPs) is becoming an increasingly important therapy for women with acne. Prior reports have indicated that OCPs were as effective as systemic antibiotics in reducing inflammatory, noninflammatory, and total facial acne lesions after 6 months of treatment.^2,3 The acne management guidelines of the American Academy of Dermatology confer OCPs a grade A recommendation based on consistent and good-quality patient-oriented evidence.⁴

The US Food and Drug Administration (FDA) has approved 3 OCPs for the treatment of acne in adult women: norgestimate–ethinyl estradiol in 1997, norethindrone acetate–ethinyl estradiol in 2001, and drospirenone–ethinyl estradiol in 2007.⁵ However, the use of these OCPs is poorly understood by many dermatologists. One study showed that dermatologists prescribed OCPs in only 2% of visits with female patients aged 12 to 55 years who presented for acne treatment, which is less often than obstetrician/gynecologists (36%) and internists (11%),⁶ perhaps due to perceived risks or unfamiliarity with OCP formulations and guidelines among dermatologists.⁷ Adverse effects of OCPs include venous thromboembolism (VTE), myocardial infarction, and hypertension,⁸ but they generally are well tolerated.⁹

Even less is known about dermatologists’ use of drospirenone-containing OCPs (DCOCPs), which contain the only FDA-approved progestin that blocks androgen receptors. In prior studies, treatment with DCOCPs was associated with greater reductions in total lesion count and investigator-graded acne severity compared to early-generation OCPs.^10,11 However, DCOCPs have been associated with a greater risk for VTE (4.0–6.3 times higher than OCP nonuse; 1.0–3.3 times higher than levonorgestrel-containing OCPs),¹² which may explain the decline in DCOCP prescriptions among gynecologists in Germany from 23.8% of OCP prescriptions in 2007 to 11.4% in 2011.¹³

In this study, we surveyed US dermatologists about their knowledge, comfort, and prescribing practices pertaining to the use of OCPs. We compare OCP-prescribing to nonprescribing dermatologists, and those frequently prescribing DCOCPs to those who infrequently prescribe DCOCPs.

Methods

Survey Design

We performed a cross-sectional survey study using convenience sampling. The instrument was designed based on primary literature on OCPs in acne treatment and questionnaires assessing the use of OCPs in other specialties. Topics included prescribing practices, contraindications for OCPs defined by the Centers for Disease Control and Prevention (CDC),¹⁴ VTE risk, patient selection for hormonal acne therapy, comfort with prescribing OCP therapy, and participant demographics.

Skip logic was employed (ie, subsequent questions depended on prior answers). A pilot study surveyed 9 board-certified dermatologists at our home institution (Weill Cornell Medical College, New York, New York).

Data Collection

Eligible participants were board-certified US dermatologists. Data were collected and managed using an electronic data capture tool through the Weill Cornell Medical College Clinical & Translational Science Center. Surveys were distributed electronically to dermatologic society members, university alumni networks, investigators’ professional contacts, and dermatologists whose contact information was purchased from an email marketing company. Chain-referral sampling (ie, participants’ recruitment among their colleagues) was used. Surveys were distributed at a regional dermatology meeting. Responses were collected from November 2014 to April 2015. This study was approved by the institutional review board.

Statistical Analysis

For the descriptive data, all responses including pilot study participants were analyzed regardless of survey completion and were summarized using frequency counts and percentages (N=130).

For the analysis of OCP prescription predictors, the sample included all respondents answering the demographic questions and indicating if they prescribe OCPs (N=116). One respondent was excluded for answering other for current practice setting. Demographic predictors of OCP prescription were physician characteristics, geographic region, practice location population density, practice attributes, time spent on medical versus pediatric dermatology, number of weekly acne patients, and percentage of total patients who are female. Medical school graduation year was a categorical variable and was categorized as prior to 1997 (when norgestimate–ethinyl estradiol was FDA approved for acne⁵) versus 1997 or later. Respondents’ practice states were analyzed according to US regions—Northeast, Midwest, South, West/Pacific—and population density (persons per square mile) using US Census Bureau data.^15,16

Univariate logistic regressions modeling OCP prescribing probability were performed for each demographic variable; a multivariable logistic model was constructed including all variables significant at α=.20 from univariate modeling.

To compare frequent prescribers versus infrequent prescribers of DCOCPs, we included all respondents answering whether they frequently prescribe DCOCPs and whether they believed the risk for VTE associated with DCOCPs differed from other OCPs (n=68). A univariate logistic regression was performed to model the probability of responding “Yes, they pose a greater risk” versus any of the other 3 responses by whether or not the respondent frequently prescribed DCOCPs for acne, and an unadjusted odds ratio was obtained. All P values were 2-tailed with statistical significance evaluated at α=.05. Ninety-five percent confidence intervals were calculated to assess precision of obtained estimates. Analyses were performed using SAS software version 9.4.

Results

Demographics

Participant demographics as predictors of OCP prescription practices are described in Table 1.

Knowledge

Oral contraceptive pills were endorsed as effective in the treatment of acne in women by 95.4% (124/130) of respondents. Among prescribers of OCPs for acne, 94.2% (65/69) believed OCPs were associated with an increased risk for VTE, no respondents thought OCPs were associated with a decreased VTE risk, 2.9% (2/69) believed OCPs did not affect VTE risk, and 2.9% (2/69) were unsure.

Among prescribers of OCPs for acne, 46.4% (32/69) believed DCOCPs posed a greater VTE risk than other OCPs. Odds of this response did not differ with frequent DCOCP prescribers versus infrequent prescribers (odds ratio, 0.731 [95% confidence interval, 0.272-1.964]; P=.5342). Participant responses on VTE risk and DCOCPs are provided in Table 2.

Dermatologists prescribing OCPs for acne endorsed greater likelihood of doing so in cases of cyclical flares with menstrual cycle (94.2% [65/69]), acne unresponsive to conventional therapy (87.0% [60/69]), acne on the lower half of the face (78.3% [54/69]), diagnosis of polycystic ovary syndrome (PCOS)(76.8% [53/69]), clinical suspicion of PCOS (71.0% [49/69]), concomitant hirsutism (71.0% [49/69]), late- or adult-onset acne (66.7% [46/69]), laboratory evidence of hyperandrogenism (60.9% [42/69]), and concomitant androgenetic alopecia (49.3% [34/69]).

Among dermatologists who prescribed OCPs for acne, CDC-defined absolute contraindications identified correctly were blood pressure of 160/100 mm Hg (59.4% [41/69]) and history of migraine with focal neurologic symptoms (49.3% [34/69]). The CDC-defined relative contraindications identified correctly were history of deep vein thrombosis or pulmonary embolism (1.4% [1/69]), breast cancer history with 5 years of no disease (15.9% [11/69]), hyperlipidemia (42.0% [29/69]), and 36 years or older smoking fewer than 15 cigarettes per day (21.7% [15/69]).

Comfort

Dermatologist self-reported comfort levels in prescribing OCPs for acne are shown in Table 3.

Prescribing Practices

Among all respondents, acne medications prescribed often included oral antibiotics (76.9% [100/130]), isotretinoin (41.5% [54/130]), and spironolactone (40.8% [53/130]).

Overall, 55.4% (72/130) of respondents prescribed OCPs for the following uses: acne (95.8% [69/72]), concomitant treatment with teratogenic medication (48.6% [35/72]), PCOS (34.7% [25/72]), hirsutism (26.4% [19/72]), androgenetic alopecia (19.4% [14/72]), SAHA (seborrhea, acne, hirsutism, alopecia) syndrome (12.5% [9/72]), and HAIR-AN (hyperandrogenism, insulin resistance, acanthosis nigricans) syndrome (11.1% [8/72]). For teratogenic medications, dermatologists prescribing OCPs did so with isotretinoin (77.8% [56/72]), spironolactone (73.6% [53/72]), tetracycline antibiotics (37.5% [27/72]), and other (34.7% [25/72]).

Of dermatologists prescribing OCPs for acne, frequency included often (19% [13/69]), sometimes (45% [31/69]), and rarely (36% [25/69]). The most frequently prescribed OCPs included Ortho Tri-Cyclen (Janssen Pharmaceuticals, Inc)(80% [55/69]), Yaz (Bayer)(64% [44/69]), and Estrostep (Warner Chilcott)(19% [13/69]). Fill-in responses included Desogen (Merck & Co, Inc)(3/69 [4%]), Alesse (Wyeth Pharmaceuticals, Inc)(3/69 [4%]), Lutera (Watson Pharma, Inc)(1/69 [1%]), Loestrin (Warner Chilcott)(1/69 [1%]), and Yasmin (Bayer)(1/69 [1%]).

In univariate regressions, graduation from medical school in 1997 or later (P=.0416), academic practice setting (P=.0130), and low-density practice setting (P=.0034) were significant predictors of prescribing OCPs. In multivariable regression, only academic practice setting (P=.0295) and low-density practice setting (P=.0050) remained significant predictors. Demographic predictors are summarized in Table 1.

Comment

Our results suggest that most dermatologists (95.4%) believe OCPs effectively treat acne; however, only 54% of respondents reported prescribing them. Academic dermatologists were more likely to prescribe OCPs than nonacademic dermatologists, possibly indicating that academic dermatologists are more familiar with the literature on the efficacy and use of OCPs. Nearly half of respondents seeing 25 or more acne patients weekly did not prescribe OCPs, suggesting a notable practice gap. Dermatologists in less dense US regions were more likely to prescribe OCPs, perhaps because dermatologists may be more likely to prescribe OCPs than refer patients in health care access–limited areas, just as primary care providers treat a broader range of conditions in low-density rural areas than urban ones.¹⁷ Exploring all dermatologists’ referral patterns for OCPs is warranted.

A strong knowledge area revealed from this study was hormonal treatment of acne in women, a vital area because appropriate patient selection is key to treatment success.⁸ Weaker knowledge areas included OCP contraindications and differences in VTE risk between formulations containing drospirenone and those not containing drospirenone. Only half the sample identified CDC-defined absolute contraindications, suggesting an education target for dermatologists to ensure patient safety. In contrast, respondents were conservative about relative contraindications, with most identifying deep vein thrombosis or pulmonary embolism, remote breast cancer history, and light smoking at 36 years or older as absolute contraindications. These results could reflect weighing the risk of relative contraindications against the benefit in acne, resulting in appropriately more conservative management than overall guidelines suggest. If so, it may suggest that dermatologists are adapting overall guidelines appropriately for use of OCPs in skin conditions.

Nearly all respondents knew that OCPs are associated with an increased risk for VTE. Approximately half understood that DCOCPs are associated with a greater VTE risk than other OCPs, with no difference between frequent and infrequent prescribers. Comparing these results to the findings on OCP prescribing overall, some dermatologists’ risk-benefit calculation for VTE differs from other specialties because DCOCPs have superior efficacy in acne, whereas DCOCPs have similar contraceptive efficacy to other OCPs.¹⁸ The fact that more dermatologists believed VTE to be an absolute contraindication than hypertension suggests dermatologists have a heightened awareness of VTE risk but prescribe DCOCPs for acne despite it.

Most OCP prescribers felt very comfortable selecting good candidates for OCPs (55.5%) and counseling on treatment initiation (45.8%) and side effects (48.6%). Only 22.2%, by contrast, were very comfortable managing side effects. This finding likely reflects the notion that VTEs are not most appropriately managed by a dermatologist. Exploring if a greater comfort level in managing side effects would make dermatologists more likely to prescribe OCPs is worthwhile. Additionally, exploring why many dermatologists do not prescribe OCPs despite believing they are effective for acne is warranted.

Study limitations included the use of convenience sampling. Additionally, our study did not investigate dermatologists’ reasons for not prescribing OCPs.

Conclusion

This study demonstrates that dermatologists believe OCPs effectively treat acne in women and that most dermatologists prescribing OCPs do so for acne treatment. Academic practice setting was associated with higher odds of prescribing OCPs than a nonacademic setting, but the number of weekly acne patients did not impact the likelihood of prescribing OCPs, which suggests a treatment gap warranting education efforts for dermatologists in nonacademic settings seeing many acne patients. Our study also suggests that awareness of the increased risk for VTE associated with DCOCPs is not associated with lower likelihood of prescribing DCOCPs, suggesting dermatologists may find greater treatment efficacy to be worth the higher risk.

Acknowledgments

We are grateful to the Department of Dermatology at the Weill Cornell College of Medicine (New York, New York) for providing funding to complete this study. We also acknowledge Paul Christos, DrPH, MS (New York, New York), and Xuming Sun, MS (New York, New York), for their assistance with the survey design. We also are indebted to numerous dermatologic professional societies for allowing the survey to be distributed to their membership.

The incidence of acne in adult females is rising,¹ and treatment with combined oral contraceptive pills (OCPs) is becoming an increasingly important therapy for women with acne. Prior reports have indicated that OCPs were as effective as systemic antibiotics in reducing inflammatory, noninflammatory, and total facial acne lesions after 6 months of treatment.^2,3 The acne management guidelines of the American Academy of Dermatology confer OCPs a grade A recommendation based on consistent and good-quality patient-oriented evidence.⁴

The US Food and Drug Administration (FDA) has approved 3 OCPs for the treatment of acne in adult women: norgestimate–ethinyl estradiol in 1997, norethindrone acetate–ethinyl estradiol in 2001, and drospirenone–ethinyl estradiol in 2007.⁵ However, the use of these OCPs is poorly understood by many dermatologists. One study showed that dermatologists prescribed OCPs in only 2% of visits with female patients aged 12 to 55 years who presented for acne treatment, which is less often than obstetrician/gynecologists (36%) and internists (11%),⁶ perhaps due to perceived risks or unfamiliarity with OCP formulations and guidelines among dermatologists.⁷ Adverse effects of OCPs include venous thromboembolism (VTE), myocardial infarction, and hypertension,⁸ but they generally are well tolerated.⁹

Even less is known about dermatologists’ use of drospirenone-containing OCPs (DCOCPs), which contain the only FDA-approved progestin that blocks androgen receptors. In prior studies, treatment with DCOCPs was associated with greater reductions in total lesion count and investigator-graded acne severity compared to early-generation OCPs.^10,11 However, DCOCPs have been associated with a greater risk for VTE (4.0–6.3 times higher than OCP nonuse; 1.0–3.3 times higher than levonorgestrel-containing OCPs),¹² which may explain the decline in DCOCP prescriptions among gynecologists in Germany from 23.8% of OCP prescriptions in 2007 to 11.4% in 2011.¹³

In this study, we surveyed US dermatologists about their knowledge, comfort, and prescribing practices pertaining to the use of OCPs. We compare OCP-prescribing to nonprescribing dermatologists, and those frequently prescribing DCOCPs to those who infrequently prescribe DCOCPs.

Methods

Survey Design

We performed a cross-sectional survey study using convenience sampling. The instrument was designed based on primary literature on OCPs in acne treatment and questionnaires assessing the use of OCPs in other specialties. Topics included prescribing practices, contraindications for OCPs defined by the Centers for Disease Control and Prevention (CDC),¹⁴ VTE risk, patient selection for hormonal acne therapy, comfort with prescribing OCP therapy, and participant demographics.

Skip logic was employed (ie, subsequent questions depended on prior answers). A pilot study surveyed 9 board-certified dermatologists at our home institution (Weill Cornell Medical College, New York, New York).

Data Collection

Eligible participants were board-certified US dermatologists. Data were collected and managed using an electronic data capture tool through the Weill Cornell Medical College Clinical & Translational Science Center. Surveys were distributed electronically to dermatologic society members, university alumni networks, investigators’ professional contacts, and dermatologists whose contact information was purchased from an email marketing company. Chain-referral sampling (ie, participants’ recruitment among their colleagues) was used. Surveys were distributed at a regional dermatology meeting. Responses were collected from November 2014 to April 2015. This study was approved by the institutional review board.

Statistical Analysis

For the descriptive data, all responses including pilot study participants were analyzed regardless of survey completion and were summarized using frequency counts and percentages (N=130).

For the analysis of OCP prescription predictors, the sample included all respondents answering the demographic questions and indicating if they prescribe OCPs (N=116). One respondent was excluded for answering other for current practice setting. Demographic predictors of OCP prescription were physician characteristics, geographic region, practice location population density, practice attributes, time spent on medical versus pediatric dermatology, number of weekly acne patients, and percentage of total patients who are female. Medical school graduation year was a categorical variable and was categorized as prior to 1997 (when norgestimate–ethinyl estradiol was FDA approved for acne⁵) versus 1997 or later. Respondents’ practice states were analyzed according to US regions—Northeast, Midwest, South, West/Pacific—and population density (persons per square mile) using US Census Bureau data.^15,16

Univariate logistic regressions modeling OCP prescribing probability were performed for each demographic variable; a multivariable logistic model was constructed including all variables significant at α=.20 from univariate modeling.

To compare frequent prescribers versus infrequent prescribers of DCOCPs, we included all respondents answering whether they frequently prescribe DCOCPs and whether they believed the risk for VTE associated with DCOCPs differed from other OCPs (n=68). A univariate logistic regression was performed to model the probability of responding “Yes, they pose a greater risk” versus any of the other 3 responses by whether or not the respondent frequently prescribed DCOCPs for acne, and an unadjusted odds ratio was obtained. All P values were 2-tailed with statistical significance evaluated at α=.05. Ninety-five percent confidence intervals were calculated to assess precision of obtained estimates. Analyses were performed using SAS software version 9.4.

Results

Demographics

Participant demographics as predictors of OCP prescription practices are described in Table 1.

Knowledge

Oral contraceptive pills were endorsed as effective in the treatment of acne in women by 95.4% (124/130) of respondents. Among prescribers of OCPs for acne, 94.2% (65/69) believed OCPs were associated with an increased risk for VTE, no respondents thought OCPs were associated with a decreased VTE risk, 2.9% (2/69) believed OCPs did not affect VTE risk, and 2.9% (2/69) were unsure.

Among prescribers of OCPs for acne, 46.4% (32/69) believed DCOCPs posed a greater VTE risk than other OCPs. Odds of this response did not differ with frequent DCOCP prescribers versus infrequent prescribers (odds ratio, 0.731 [95% confidence interval, 0.272-1.964]; P=.5342). Participant responses on VTE risk and DCOCPs are provided in Table 2.

Dermatologists prescribing OCPs for acne endorsed greater likelihood of doing so in cases of cyclical flares with menstrual cycle (94.2% [65/69]), acne unresponsive to conventional therapy (87.0% [60/69]), acne on the lower half of the face (78.3% [54/69]), diagnosis of polycystic ovary syndrome (PCOS)(76.8% [53/69]), clinical suspicion of PCOS (71.0% [49/69]), concomitant hirsutism (71.0% [49/69]), late- or adult-onset acne (66.7% [46/69]), laboratory evidence of hyperandrogenism (60.9% [42/69]), and concomitant androgenetic alopecia (49.3% [34/69]).

Among dermatologists who prescribed OCPs for acne, CDC-defined absolute contraindications identified correctly were blood pressure of 160/100 mm Hg (59.4% [41/69]) and history of migraine with focal neurologic symptoms (49.3% [34/69]). The CDC-defined relative contraindications identified correctly were history of deep vein thrombosis or pulmonary embolism (1.4% [1/69]), breast cancer history with 5 years of no disease (15.9% [11/69]), hyperlipidemia (42.0% [29/69]), and 36 years or older smoking fewer than 15 cigarettes per day (21.7% [15/69]).

Comfort

Dermatologist self-reported comfort levels in prescribing OCPs for acne are shown in Table 3.

Prescribing Practices

Among all respondents, acne medications prescribed often included oral antibiotics (76.9% [100/130]), isotretinoin (41.5% [54/130]), and spironolactone (40.8% [53/130]).

Overall, 55.4% (72/130) of respondents prescribed OCPs for the following uses: acne (95.8% [69/72]), concomitant treatment with teratogenic medication (48.6% [35/72]), PCOS (34.7% [25/72]), hirsutism (26.4% [19/72]), androgenetic alopecia (19.4% [14/72]), SAHA (seborrhea, acne, hirsutism, alopecia) syndrome (12.5% [9/72]), and HAIR-AN (hyperandrogenism, insulin resistance, acanthosis nigricans) syndrome (11.1% [8/72]). For teratogenic medications, dermatologists prescribing OCPs did so with isotretinoin (77.8% [56/72]), spironolactone (73.6% [53/72]), tetracycline antibiotics (37.5% [27/72]), and other (34.7% [25/72]).

Of dermatologists prescribing OCPs for acne, frequency included often (19% [13/69]), sometimes (45% [31/69]), and rarely (36% [25/69]). The most frequently prescribed OCPs included Ortho Tri-Cyclen (Janssen Pharmaceuticals, Inc)(80% [55/69]), Yaz (Bayer)(64% [44/69]), and Estrostep (Warner Chilcott)(19% [13/69]). Fill-in responses included Desogen (Merck & Co, Inc)(3/69 [4%]), Alesse (Wyeth Pharmaceuticals, Inc)(3/69 [4%]), Lutera (Watson Pharma, Inc)(1/69 [1%]), Loestrin (Warner Chilcott)(1/69 [1%]), and Yasmin (Bayer)(1/69 [1%]).

In univariate regressions, graduation from medical school in 1997 or later (P=.0416), academic practice setting (P=.0130), and low-density practice setting (P=.0034) were significant predictors of prescribing OCPs. In multivariable regression, only academic practice setting (P=.0295) and low-density practice setting (P=.0050) remained significant predictors. Demographic predictors are summarized in Table 1.

Comment

Our results suggest that most dermatologists (95.4%) believe OCPs effectively treat acne; however, only 54% of respondents reported prescribing them. Academic dermatologists were more likely to prescribe OCPs than nonacademic dermatologists, possibly indicating that academic dermatologists are more familiar with the literature on the efficacy and use of OCPs. Nearly half of respondents seeing 25 or more acne patients weekly did not prescribe OCPs, suggesting a notable practice gap. Dermatologists in less dense US regions were more likely to prescribe OCPs, perhaps because dermatologists may be more likely to prescribe OCPs than refer patients in health care access–limited areas, just as primary care providers treat a broader range of conditions in low-density rural areas than urban ones.¹⁷ Exploring all dermatologists’ referral patterns for OCPs is warranted.

A strong knowledge area revealed from this study was hormonal treatment of acne in women, a vital area because appropriate patient selection is key to treatment success.⁸ Weaker knowledge areas included OCP contraindications and differences in VTE risk between formulations containing drospirenone and those not containing drospirenone. Only half the sample identified CDC-defined absolute contraindications, suggesting an education target for dermatologists to ensure patient safety. In contrast, respondents were conservative about relative contraindications, with most identifying deep vein thrombosis or pulmonary embolism, remote breast cancer history, and light smoking at 36 years or older as absolute contraindications. These results could reflect weighing the risk of relative contraindications against the benefit in acne, resulting in appropriately more conservative management than overall guidelines suggest. If so, it may suggest that dermatologists are adapting overall guidelines appropriately for use of OCPs in skin conditions.

Nearly all respondents knew that OCPs are associated with an increased risk for VTE. Approximately half understood that DCOCPs are associated with a greater VTE risk than other OCPs, with no difference between frequent and infrequent prescribers. Comparing these results to the findings on OCP prescribing overall, some dermatologists’ risk-benefit calculation for VTE differs from other specialties because DCOCPs have superior efficacy in acne, whereas DCOCPs have similar contraceptive efficacy to other OCPs.¹⁸ The fact that more dermatologists believed VTE to be an absolute contraindication than hypertension suggests dermatologists have a heightened awareness of VTE risk but prescribe DCOCPs for acne despite it.

Most OCP prescribers felt very comfortable selecting good candidates for OCPs (55.5%) and counseling on treatment initiation (45.8%) and side effects (48.6%). Only 22.2%, by contrast, were very comfortable managing side effects. This finding likely reflects the notion that VTEs are not most appropriately managed by a dermatologist. Exploring if a greater comfort level in managing side effects would make dermatologists more likely to prescribe OCPs is worthwhile. Additionally, exploring why many dermatologists do not prescribe OCPs despite believing they are effective for acne is warranted.

Study limitations included the use of convenience sampling. Additionally, our study did not investigate dermatologists’ reasons for not prescribing OCPs.

Conclusion

This study demonstrates that dermatologists believe OCPs effectively treat acne in women and that most dermatologists prescribing OCPs do so for acne treatment. Academic practice setting was associated with higher odds of prescribing OCPs than a nonacademic setting, but the number of weekly acne patients did not impact the likelihood of prescribing OCPs, which suggests a treatment gap warranting education efforts for dermatologists in nonacademic settings seeing many acne patients. Our study also suggests that awareness of the increased risk for VTE associated with DCOCPs is not associated with lower likelihood of prescribing DCOCPs, suggesting dermatologists may find greater treatment efficacy to be worth the higher risk.

Acknowledgments

We are grateful to the Department of Dermatology at the Weill Cornell College of Medicine (New York, New York) for providing funding to complete this study. We also acknowledge Paul Christos, DrPH, MS (New York, New York), and Xuming Sun, MS (New York, New York), for their assistance with the survey design. We also are indebted to numerous dermatologic professional societies for allowing the survey to be distributed to their membership.

The incidence of acne in adult females is rising,¹ and treatment with combined oral contraceptive pills (OCPs) is becoming an increasingly important therapy for women with acne. Prior reports have indicated that OCPs were as effective as systemic antibiotics in reducing inflammatory, noninflammatory, and total facial acne lesions after 6 months of treatment.^2,3 The acne management guidelines of the American Academy of Dermatology confer OCPs a grade A recommendation based on consistent and good-quality patient-oriented evidence.⁴

The US Food and Drug Administration (FDA) has approved 3 OCPs for the treatment of acne in adult women: norgestimate–ethinyl estradiol in 1997, norethindrone acetate–ethinyl estradiol in 2001, and drospirenone–ethinyl estradiol in 2007.⁵ However, the use of these OCPs is poorly understood by many dermatologists. One study showed that dermatologists prescribed OCPs in only 2% of visits with female patients aged 12 to 55 years who presented for acne treatment, which is less often than obstetrician/gynecologists (36%) and internists (11%),⁶ perhaps due to perceived risks or unfamiliarity with OCP formulations and guidelines among dermatologists.⁷ Adverse effects of OCPs include venous thromboembolism (VTE), myocardial infarction, and hypertension,⁸ but they generally are well tolerated.⁹

Even less is known about dermatologists’ use of drospirenone-containing OCPs (DCOCPs), which contain the only FDA-approved progestin that blocks androgen receptors. In prior studies, treatment with DCOCPs was associated with greater reductions in total lesion count and investigator-graded acne severity compared to early-generation OCPs.^10,11 However, DCOCPs have been associated with a greater risk for VTE (4.0–6.3 times higher than OCP nonuse; 1.0–3.3 times higher than levonorgestrel-containing OCPs),¹² which may explain the decline in DCOCP prescriptions among gynecologists in Germany from 23.8% of OCP prescriptions in 2007 to 11.4% in 2011.¹³

In this study, we surveyed US dermatologists about their knowledge, comfort, and prescribing practices pertaining to the use of OCPs. We compare OCP-prescribing to nonprescribing dermatologists, and those frequently prescribing DCOCPs to those who infrequently prescribe DCOCPs.

Methods

Survey Design

We performed a cross-sectional survey study using convenience sampling. The instrument was designed based on primary literature on OCPs in acne treatment and questionnaires assessing the use of OCPs in other specialties. Topics included prescribing practices, contraindications for OCPs defined by the Centers for Disease Control and Prevention (CDC),¹⁴ VTE risk, patient selection for hormonal acne therapy, comfort with prescribing OCP therapy, and participant demographics.

Skip logic was employed (ie, subsequent questions depended on prior answers). A pilot study surveyed 9 board-certified dermatologists at our home institution (Weill Cornell Medical College, New York, New York).

Data Collection

Eligible participants were board-certified US dermatologists. Data were collected and managed using an electronic data capture tool through the Weill Cornell Medical College Clinical & Translational Science Center. Surveys were distributed electronically to dermatologic society members, university alumni networks, investigators’ professional contacts, and dermatologists whose contact information was purchased from an email marketing company. Chain-referral sampling (ie, participants’ recruitment among their colleagues) was used. Surveys were distributed at a regional dermatology meeting. Responses were collected from November 2014 to April 2015. This study was approved by the institutional review board.

Statistical Analysis

For the descriptive data, all responses including pilot study participants were analyzed regardless of survey completion and were summarized using frequency counts and percentages (N=130).

For the analysis of OCP prescription predictors, the sample included all respondents answering the demographic questions and indicating if they prescribe OCPs (N=116). One respondent was excluded for answering other for current practice setting. Demographic predictors of OCP prescription were physician characteristics, geographic region, practice location population density, practice attributes, time spent on medical versus pediatric dermatology, number of weekly acne patients, and percentage of total patients who are female. Medical school graduation year was a categorical variable and was categorized as prior to 1997 (when norgestimate–ethinyl estradiol was FDA approved for acne⁵) versus 1997 or later. Respondents’ practice states were analyzed according to US regions—Northeast, Midwest, South, West/Pacific—and population density (persons per square mile) using US Census Bureau data.^15,16

Univariate logistic regressions modeling OCP prescribing probability were performed for each demographic variable; a multivariable logistic model was constructed including all variables significant at α=.20 from univariate modeling.

To compare frequent prescribers versus infrequent prescribers of DCOCPs, we included all respondents answering whether they frequently prescribe DCOCPs and whether they believed the risk for VTE associated with DCOCPs differed from other OCPs (n=68). A univariate logistic regression was performed to model the probability of responding “Yes, they pose a greater risk” versus any of the other 3 responses by whether or not the respondent frequently prescribed DCOCPs for acne, and an unadjusted odds ratio was obtained. All P values were 2-tailed with statistical significance evaluated at α=.05. Ninety-five percent confidence intervals were calculated to assess precision of obtained estimates. Analyses were performed using SAS software version 9.4.

Results

Demographics

Participant demographics as predictors of OCP prescription practices are described in Table 1.

Knowledge

Oral contraceptive pills were endorsed as effective in the treatment of acne in women by 95.4% (124/130) of respondents. Among prescribers of OCPs for acne, 94.2% (65/69) believed OCPs were associated with an increased risk for VTE, no respondents thought OCPs were associated with a decreased VTE risk, 2.9% (2/69) believed OCPs did not affect VTE risk, and 2.9% (2/69) were unsure.

Among prescribers of OCPs for acne, 46.4% (32/69) believed DCOCPs posed a greater VTE risk than other OCPs. Odds of this response did not differ with frequent DCOCP prescribers versus infrequent prescribers (odds ratio, 0.731 [95% confidence interval, 0.272-1.964]; P=.5342). Participant responses on VTE risk and DCOCPs are provided in Table 2.

Dermatologists prescribing OCPs for acne endorsed greater likelihood of doing so in cases of cyclical flares with menstrual cycle (94.2% [65/69]), acne unresponsive to conventional therapy (87.0% [60/69]), acne on the lower half of the face (78.3% [54/69]), diagnosis of polycystic ovary syndrome (PCOS)(76.8% [53/69]), clinical suspicion of PCOS (71.0% [49/69]), concomitant hirsutism (71.0% [49/69]), late- or adult-onset acne (66.7% [46/69]), laboratory evidence of hyperandrogenism (60.9% [42/69]), and concomitant androgenetic alopecia (49.3% [34/69]).

Among dermatologists who prescribed OCPs for acne, CDC-defined absolute contraindications identified correctly were blood pressure of 160/100 mm Hg (59.4% [41/69]) and history of migraine with focal neurologic symptoms (49.3% [34/69]). The CDC-defined relative contraindications identified correctly were history of deep vein thrombosis or pulmonary embolism (1.4% [1/69]), breast cancer history with 5 years of no disease (15.9% [11/69]), hyperlipidemia (42.0% [29/69]), and 36 years or older smoking fewer than 15 cigarettes per day (21.7% [15/69]).

Comfort

Dermatologist self-reported comfort levels in prescribing OCPs for acne are shown in Table 3.

Prescribing Practices

Among all respondents, acne medications prescribed often included oral antibiotics (76.9% [100/130]), isotretinoin (41.5% [54/130]), and spironolactone (40.8% [53/130]).

Overall, 55.4% (72/130) of respondents prescribed OCPs for the following uses: acne (95.8% [69/72]), concomitant treatment with teratogenic medication (48.6% [35/72]), PCOS (34.7% [25/72]), hirsutism (26.4% [19/72]), androgenetic alopecia (19.4% [14/72]), SAHA (seborrhea, acne, hirsutism, alopecia) syndrome (12.5% [9/72]), and HAIR-AN (hyperandrogenism, insulin resistance, acanthosis nigricans) syndrome (11.1% [8/72]). For teratogenic medications, dermatologists prescribing OCPs did so with isotretinoin (77.8% [56/72]), spironolactone (73.6% [53/72]), tetracycline antibiotics (37.5% [27/72]), and other (34.7% [25/72]).

Of dermatologists prescribing OCPs for acne, frequency included often (19% [13/69]), sometimes (45% [31/69]), and rarely (36% [25/69]). The most frequently prescribed OCPs included Ortho Tri-Cyclen (Janssen Pharmaceuticals, Inc)(80% [55/69]), Yaz (Bayer)(64% [44/69]), and Estrostep (Warner Chilcott)(19% [13/69]). Fill-in responses included Desogen (Merck & Co, Inc)(3/69 [4%]), Alesse (Wyeth Pharmaceuticals, Inc)(3/69 [4%]), Lutera (Watson Pharma, Inc)(1/69 [1%]), Loestrin (Warner Chilcott)(1/69 [1%]), and Yasmin (Bayer)(1/69 [1%]).

In univariate regressions, graduation from medical school in 1997 or later (P=.0416), academic practice setting (P=.0130), and low-density practice setting (P=.0034) were significant predictors of prescribing OCPs. In multivariable regression, only academic practice setting (P=.0295) and low-density practice setting (P=.0050) remained significant predictors. Demographic predictors are summarized in Table 1.

Comment

Our results suggest that most dermatologists (95.4%) believe OCPs effectively treat acne; however, only 54% of respondents reported prescribing them. Academic dermatologists were more likely to prescribe OCPs than nonacademic dermatologists, possibly indicating that academic dermatologists are more familiar with the literature on the efficacy and use of OCPs. Nearly half of respondents seeing 25 or more acne patients weekly did not prescribe OCPs, suggesting a notable practice gap. Dermatologists in less dense US regions were more likely to prescribe OCPs, perhaps because dermatologists may be more likely to prescribe OCPs than refer patients in health care access–limited areas, just as primary care providers treat a broader range of conditions in low-density rural areas than urban ones.¹⁷ Exploring all dermatologists’ referral patterns for OCPs is warranted.

A strong knowledge area revealed from this study was hormonal treatment of acne in women, a vital area because appropriate patient selection is key to treatment success.⁸ Weaker knowledge areas included OCP contraindications and differences in VTE risk between formulations containing drospirenone and those not containing drospirenone. Only half the sample identified CDC-defined absolute contraindications, suggesting an education target for dermatologists to ensure patient safety. In contrast, respondents were conservative about relative contraindications, with most identifying deep vein thrombosis or pulmonary embolism, remote breast cancer history, and light smoking at 36 years or older as absolute contraindications. These results could reflect weighing the risk of relative contraindications against the benefit in acne, resulting in appropriately more conservative management than overall guidelines suggest. If so, it may suggest that dermatologists are adapting overall guidelines appropriately for use of OCPs in skin conditions.

Nearly all respondents knew that OCPs are associated with an increased risk for VTE. Approximately half understood that DCOCPs are associated with a greater VTE risk than other OCPs, with no difference between frequent and infrequent prescribers. Comparing these results to the findings on OCP prescribing overall, some dermatologists’ risk-benefit calculation for VTE differs from other specialties because DCOCPs have superior efficacy in acne, whereas DCOCPs have similar contraceptive efficacy to other OCPs.¹⁸ The fact that more dermatologists believed VTE to be an absolute contraindication than hypertension suggests dermatologists have a heightened awareness of VTE risk but prescribe DCOCPs for acne despite it.

Most OCP prescribers felt very comfortable selecting good candidates for OCPs (55.5%) and counseling on treatment initiation (45.8%) and side effects (48.6%). Only 22.2%, by contrast, were very comfortable managing side effects. This finding likely reflects the notion that VTEs are not most appropriately managed by a dermatologist. Exploring if a greater comfort level in managing side effects would make dermatologists more likely to prescribe OCPs is worthwhile. Additionally, exploring why many dermatologists do not prescribe OCPs despite believing they are effective for acne is warranted.

Study limitations included the use of convenience sampling. Additionally, our study did not investigate dermatologists’ reasons for not prescribing OCPs.

Conclusion

This study demonstrates that dermatologists believe OCPs effectively treat acne in women and that most dermatologists prescribing OCPs do so for acne treatment. Academic practice setting was associated with higher odds of prescribing OCPs than a nonacademic setting, but the number of weekly acne patients did not impact the likelihood of prescribing OCPs, which suggests a treatment gap warranting education efforts for dermatologists in nonacademic settings seeing many acne patients. Our study also suggests that awareness of the increased risk for VTE associated with DCOCPs is not associated with lower likelihood of prescribing DCOCPs, suggesting dermatologists may find greater treatment efficacy to be worth the higher risk.

Acknowledgments

We are grateful to the Department of Dermatology at the Weill Cornell College of Medicine (New York, New York) for providing funding to complete this study. We also acknowledge Paul Christos, DrPH, MS (New York, New York), and Xuming Sun, MS (New York, New York), for their assistance with the survey design. We also are indebted to numerous dermatologic professional societies for allowing the survey to be distributed to their membership.