Reviews

Mr. R, 75, is having difficulty sleeping. When he goes to bed, he lies there for what seems like forever, unable to fall asleep. He feels “so tired” and ends up taking naps during the day, but he cannot break this cycle. He has tried using over-the-counter products with little relief.

Mr. R’s primary care physician prescribes zaleplon, 10 mg/d, and asks him to call the clinic in 2 weeks to discuss his progress. He takes zaleplon as directed for several nights and begins to feel “sluggish” during the day, both mentally and physically, despite reporting an increase in the overall amount of sleep at night.

Sedative-hypnotic drugs are among the most commonly used medica­tions in the United States. Use of these drugs, as well as anxiolytics, has increased from 2.8% between 1988 and 1994 to 4.7% between 2007 and 2010, according to the Department of Health and Human Services.¹ In 2011, drugs categorized as sedative-hypnotics or antipsychotics were involved in 6.1% of all human exposures identified in the American Association of Poison Control Centers’ National Poison Data System.² Therefore, an understanding of clinical and pharmacological variables related to safe and effective use is important for clinicians prescribing and monitoring therapy with these agents.

Neuropsychiatric disorders are prevalent among geriatric patients and are associated with age-related physiologic changes in the CNS.³ Such changes involve:  
   • neuroanatomy (brain atrophy, decreased neuronal density, increased plaque formation)  
   • neurotransmitters (reduced cholinergic transmission, decreased synthesis of dopamine and catecholamines), and   
   • neurophysiology (reduced cerebral blood flow).

These physiologic processes manifest as alterations in mental status, reflexes, sen­sation, gait, balance, and sleep. Examples of sleep changes among geriatric patients include decreased sleep efficiency, more fre­quent awakenings, and more variable sleep duration.^3,4 Sleep disorders also may be related to mental disorders and other medi­cal conditions.⁵ For example, the prevalence of sleep-related respiratory disorders, such as obstructive sleep apnea and central sleep apnea, increases with age.⁶

Sleep disorders are common among geri­atric patients. In a large epidemiologic study of sleep complaints in patients age ≥65, more than one-half of patients had at least 1 sleep complaint (ie, difficulty falling asleep, trou­ble waking up, early awakening, need for naps, and feeling ill-rested).⁷ As many as 34% of patients reported symptoms of insom­nia. In an analysis of National Ambulatory Medical Survey Data over 6 years, 24.8% to 27.9% of sleep-related medical office visits were attributed to patients age ≥65.⁸

Pharmacology in aging
Prescribing sedative-hypnotic drugs is not routinely recommended for older patients with a sleep disorder. Geriatric patients, com­pared with younger patients, are at higher risk of iatrogenic complications because of polypharmacy, comorbidities, relative renal and hepatic insufficiency, and other physi­ologic changes leading to alterations in drug exposure and metabolism (Table 1).^9-12

Aging is associated with changes in body composition, including an increase in total body fat and decrease in lean body mass and total body water. These changes, as well as a prolonged GI transit time, decrease in active gut transporters, decreased blood perfusion, and decrease in plasma proteins such as albumin (because of reduced liver function or malnutrition), may lead to alteration in drug absorption patterns and may increase the volume of distribution for lipophilic drugs. Additionally, the elimination half-life of some drugs may increase with age because of larger volumes of distribution and reduction in hepatic or renal clearance.

The clinical significance of these changes is not well established. Although the process of drug absorption can change with age, the amount of drug absorbed might not be significantly affected. An increase in the vol­ume of distribution and reduction in drug metabolism and clearance might lead to increasing amounts of circulating drug and duration of drug exposure, putting geriat­ric patients at an increased risk for adverse effects and drug toxicity.⁹

Among these mechanisms, Dolder et al¹¹ hypothesized that drug metabolism cata­lyzed by cytochrome P450 (CYP) enzymes and renal excretion may be of greatest con­cern. Although in vitro studies suggest that concentration of CYP enzymes does not decline with age, in vivo studies have dem­onstrated reduced CYP activity in geriatric patients.^11,12 Theoretically, a reduction in CYP activity would increase the bioavail­ability of drugs, especially those that are subject to extensive first-pass (ie, hepatic) metabolism, and may lead to a reduction in systemic clearance.

Independent of metabolic changes, geriatric patients are at risk of reduced renal clearance because of age-related changes in glomerular filtration rate. Pharmacodynamic changes might be observed in older patients and could be a concern even in the setting of unaltered pharmacokinetic factors.9 These changes usually require administering smaller drug dosages.

Sedative-hypnotic medications
Sedative-hypnotic agents include several barbiturates, benzodiazepines (BZDs), non-BZD benzodiazepine-receptor ago­nists (BzRAs), a melatonin-receptor agonist (ie, ramelteon), and an orexin-receptor antagonist (ie, suvorexant).^13,14Table 2^14-29 summarizes selected sedative-hypnotic drugs. Additional drug classes used to treat insomnia include:
   • sedating antidepressants (trazodone, amitriptyline, doxepin, mirtazapine)
   • antiepileptic drugs (gabapentin, tiagabine)
   • atypical antipsychotics (quetiapine, olanzapine).

FDA-approved agents for treating insomnia include amobarbital, butabarbi­tal, pentobarbital, phenobarbital, secobar­bital, chloral hydrate, diphenhydramine, doxylamine, doxepin, estazolam, fluraz­epam, lorazepam, quazepam, temazepam, triazolam, eszopiclone, zaleplon, zolpidem, ramelteon, and suvorexant. Not all of these drugs are recommended for use in geriatric patients. Barbiturates, for example, should be avoided.³⁰

Pharmacokinetic characteristics vary among drugs and drug classes. Choice of pharmacotherapy should account for patient and drug characteristics and the specific sleep complaint. Sleep disorders may be var­iously characterized as difficulty with sleep initiation, duration, consolidation, or qual­ity.¹³ Therefore, onset and duration of effect are important drug-related considerations. Sedative-hypnotic drugs with a short time-to-onset may be ideal for patients with sleep-onset insomnia.

The drugs’ duration of effect (eg, presence of active metabolites, long elimination half-life) also must be reviewed. A long elimi­nation half-life may lead to increased drug exposure and unwanted side effects such as residual daytime drowsiness. Despite this, sedative-hypnotic drugs with a longer duration of effect (eg, intermediate- or long-acting drugs) may be best for patients with insomnia defined by difficulty maintaining sleep.

Benzodiazepines vary in their time to onset of effect, rate of elimination, and metabolism.^15-21 BZDs that are FDA- approved for use as sedative-hypnotics are listed in Table 2.^14-29 These BZDs have different onsets of effect as evidenced by time to achieve maximum plasma con­centration (T_max), ranging from 0.5 hours (flurazepam) to 2 hours (estazolam, quaz­epam, triazolam). The elimination half-life varies widely among these medications, from 1.5 hours (triazolam) to >100 hours (flurazepam). Flurazepam’s long half-life is attributable to its active major metabo­lite. Although most BZDs are metabolized hepatically, temazepam is subject to mini­mal hepatic metabolism.

Benzodiazepine-receptor agonists. There is substantial variation in the phar­macokinetic characteristics of BzRAs.^15,16,22-28 There also are differences among the zolpi­dem dosage forms; sublingual formulations have the shortest onset of effect. Eszopiclone and zaleplon have low protein binding com­pared with zolpidem. Elimination half-lives vary among drugs with the shortest attrib­uted to zaleplon (1 hour) and longest to eszopiclone (6 hours). All BzRAs are subject to extensive hepatic metabolism. 

Ramelteon. Singular in its class, ramelteon is a treatment option for insomnia.²⁹ This drug has a short onset of effect, moder­ate protein binding, and extensive hepatic metabolism. Ramelteon is primarily excreted in the urine as its metabolites, and the drug half-life is relatively short.

Suvorexant is the latest addition to the sedative-hypnotic armamentarium, approved by the FDA in August 2014 for dif­ficulty with sleep onset and/or sleep main­tenance.¹⁴ As an orexin-receptor antagonist, suvorexant represents a novel pharmaco­logic class. Suvorexant exhibits moderately rapid absorption with time to peak concen­tration ranging from 30 minutes to 6 hours in fasting conditions; absorption is delayed when taken with or soon after a meal. The drug is highly protein bound and extensively metabolized, primarily through CYP3A. The manufacturer recommends dose reduc­tion (5 mg at bedtime) in patients taking moderate CYP3A inhibitors and avoiding suvorexant in patients taking strong CYP3A inhibitors. Suvorexant is primarily excreted through feces and the mean half-life is rela­tively long.

Considering these characteristics and age-related physiologic changes, the practi­tioner should be concerned about drugs that undergo extensive hepatic metabolism. Age-related reductions in CYP activity may lead to an increase in drug bioavailability and a decrease in the systemic clearance,¹¹ which might be associated with an increase in elimination half-life and duration of action. Dosage adjustments are recommended for several BZDs (lower initial and maximum dosages for most agents) and BzRAs.^17-28 No dosage adjustments for ramelteon or suvorexant in geriatric patients have been specified^14,29; the manufacturers for both products assert that no differences in safety and efficacy have been observed between older and younger adult patients.

Alternative and complementary medications
Several non-prescription products, includ­ing over-the-counter drugs (eg, diphenhy-dramine, doxylamine) and herbal therapies (eg, melatonin, valerian), are used for their sedative-hypnotic properties. There is a lack of evidence supporting using diphenhydra-mine in patients with chronic insomnia, and tolerance to its hypnotic effect has been reported with repeated use.³¹ Concerns about anticholinergic toxicity and CNS depression limit its use in geriatric patients. Among herbal therapies, melatonin may have the strongest evidence for its ability to allevi­ate sleep disorders in geriatric patients³²; however, meta-analyses have demonstrated small effects of melatonin on sleep latency and minimal differences in wake time after sleep onset and total sleep time.¹³

Clinical practice guidelines
Non-pharmacotherapeutic interventions, such as behavioral (eg, sleep hygiene mea­sures) and psychological therapy, are rec­ommended for initial management of sleep disorders in geriatric patients.^13,33 In con­junction, the American Medical Directors Association (AMDA) recommends address­ ing underlying causes and exacerbating fac­tors (eg, medical condition or medication).³³ The AMDA recommends avoiding long-term pharmacotherapy and advises caution with BZD-hypnotic drugs, tricyclic antide­pressants, and antihistamines. The American Academy of Sleep Medicine (AASM) recom­mends an initial treatment period of 2 to 4 weeks, followed by re-evaluation of con­tinued need for treatment.¹³ The AASM recommends short- or intermediate-acting BzRAs or ramelteon for initial pharmaco­logic management of primary insomnias and insomnias comorbid with other condi­tions. The AASM also recommends specific dosages of BzRAs and BZDs for geriatric patients, which coincide with manufacturer-recommended dosages (Table 2).^14-29

Barbiturates, chloral hydrate, and non-barbiturate, non-BZD drugs such as mep­robamate are not recommended because of potential significant adverse effects and tolerance/dependence, and low therapeu­tic index. The AASM advises caution when using prescription drugs off-label for insom­nia (eg, antidepressants, antiepileptics, antipsychotics) and recommends avoiding them, if possible, because of limited evi­dence supporting their use.¹³

Safety concerns
Two commonly used references contain rec­ommendations for sedative-hypnotic medi­cation use in geriatric patients.^30,34 According to Gallagher et al’s³⁴ Screening Tool of Older Person’s Prescriptions (STOPP), long-term (>1 month) use of long-acting BZDs (eg, flurazepam, diazepam) and prolonged use (>1 week) of first-generation antihistamines (eg, diphenhydramine, doxylamine) should be avoided in patients age ≥65 because of the risk of sedation, confusion, and anti­cholinergic side effects. STOPP recognizes that any use of BZDs, neuroleptics, or first-generation antihistamines may contrib­ute to postural imbalance; therefore these agents are not recommended in older patients at risk for falls.

In the 2012 American Geriatrics Society (AGS) Beers Criteria, the AGS recommends avoiding barbiturates in older adults because of the high rate of physical depen­dence, tolerance to sleep effects, and over­dose risk at low dosages.³⁰ The AGS also recommends avoiding BZDs, stating that older adults have increased sensitivity to these agents and are at an increased risk of cognitive impairment, delirium, falls, frac­tures, and motor vehicle accidents when taking these drugs. Non-BZD BzRAs also should not be prescribed to patients with a history of falls or fractures, unless safer alternatives are not available.

The FDA has issued several advisory reports regarding sedative-hypnotic drugs. In 2007, all manufacturers of sedative-hypnotic drugs were required to modify their product labeling to include stronger language about potential risks.35 Among these changes, warnings for ana­phylaxis and complex sleep-related behav­iors were added. Also, the FDA requested that manufacturers of sedative-hypnotic drugs develop and provide patient medi­cation guides, advising consumers on the potential risks and precautions associated with these drugs. More recently, the FDA announced changes to dosing recommen­dations for zolpidem-containing products because of the risk of impaired mental alertness36; manufacturers were required to lower the recommended dosages for each product.

Manufacturers of FDA-approved sedative-hypnotic drugs urge caution when prescribing these medications for geriatric patients, citing the potential for increased sensitivity, manifesting as marked excite­ment, depression, or confusion (eg, barbi­turates), and greater risk for dosage-related adverse effects (eg, oversedation, dizziness, confusion, impaired psychomotor perfor­mance, ataxia).^17-29

Use in clinical practice
Several variables should be considered when evaluating appropriateness of phar­macotherapy, including characteristics of the drug and the patient. Geriatric patients may be prone to comorbidities resulting from age-related physiologic changes. These diseases may be confounding (ie, contributing to sleep disorders); examples include medical illnesses, such as hyperthyroidism and arthritis, and psychiatric illnesses, such as depression and anxiety.³⁷ Other conditions, such as renal and hepatic dysfunction, may lead to alteration in drug exposure. These condi­tions should be assessed through routine renal function tests (eg, serum creatinine and glomerular filtration rate) and liver function tests (eg, serum albumin and liver transaminases).

Multiple comorbidities suggest a higher likelihood of polypharmacy, leading to other drug-related issues (eg, drug-drug interactions). Although these issues may guide therapy by restricting medication options, their potential contribution to the underlying sleep complaints should be con­sidered.³⁷ Several drugs commonly used by geriatric patients may affect wakefulness (eg, analgesics, antidepressants, and anti­hypertensives [sedating], and thyroid hor­mones, corticosteroids, and CNS stimulants [alerting]).

CASE CONTINUED
In Mr. R’s case, zaleplon was initiated at 10 mg/d. Because of his age and the nature of his sleep disorder, the choice of sedative-hypnotic was suitable; however, the prescribed dosage was inappropriate. The sluggishness Mr. R experienced likely was a manifestation of increased exposure to the drug. According to manufacturer and AASM recommendations, a more appropriate dosage is 5 mg/d.^13,23 Mr. R’s medical history and current medica­tions, and his hepatic and renal function, should be assessed. If Mr. R continues to have issues with sleep initiation, zaleplon, 5 mg at bedtime, should be considered.

Related Resources
• Institute for Safe Medication Practices. www.ismp.org.
• MedWatch: The FDA Safety Information and Adverse Event Reporting Program. www.fda.gov/Safety/MedWatch/default.htm.

Drug Brand Names
Amitriptyline • Elavil                                   Mirtazapine • Remeron
Amobarbital • Amytal                                  Olanzapine • Zyprexa
Butabarbital • Butisol                                  Pentobarbital • Nembutal
Chloral hydrate • Somnote                          Phenobarbital • Luminal
Diazepam • Valium                                     Quazepam • Doral
Diphenhydramine • Benadryl, others            Quetiapine • Seroquel
Doxepin • Silenor                                       Ramelteon • Rozerem
Doxylamine • Unisom, others                      Secobarbital • Seconal
Estazolam • ProSom                                   Suvorexant • Belsomra
Eszopiclone • Lunesta                                 Temazepam • Restoril
Flurazepam • Dalmane                               Tiagabine • Gabitril
Gabapentin • Neurontin,                             Trazodone • Desyrel
Gralise, Horizant                                        Zaleplon • Sonata
Lorazepam • Ativan                                   Zolpidem • Ambien, Edluar,
Meprobamate • Equanil                                 Intermezzo, Zolpimist

Acknowledgement
Vicki L. Ellingrod, PharmD, FCCP, is the series editor of Savvy Psychopharmacology.

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Mr. R, 75, is having difficulty sleeping. When he goes to bed, he lies there for what seems like forever, unable to fall asleep. He feels “so tired” and ends up taking naps during the day, but he cannot break this cycle. He has tried using over-the-counter products with little relief.

Mr. R’s primary care physician prescribes zaleplon, 10 mg/d, and asks him to call the clinic in 2 weeks to discuss his progress. He takes zaleplon as directed for several nights and begins to feel “sluggish” during the day, both mentally and physically, despite reporting an increase in the overall amount of sleep at night.

Sedative-hypnotic drugs are among the most commonly used medica­tions in the United States. Use of these drugs, as well as anxiolytics, has increased from 2.8% between 1988 and 1994 to 4.7% between 2007 and 2010, according to the Department of Health and Human Services.¹ In 2011, drugs categorized as sedative-hypnotics or antipsychotics were involved in 6.1% of all human exposures identified in the American Association of Poison Control Centers’ National Poison Data System.² Therefore, an understanding of clinical and pharmacological variables related to safe and effective use is important for clinicians prescribing and monitoring therapy with these agents.

Neuropsychiatric disorders are prevalent among geriatric patients and are associated with age-related physiologic changes in the CNS.³ Such changes involve:  
   • neuroanatomy (brain atrophy, decreased neuronal density, increased plaque formation)  
   • neurotransmitters (reduced cholinergic transmission, decreased synthesis of dopamine and catecholamines), and   
   • neurophysiology (reduced cerebral blood flow).

These physiologic processes manifest as alterations in mental status, reflexes, sen­sation, gait, balance, and sleep. Examples of sleep changes among geriatric patients include decreased sleep efficiency, more fre­quent awakenings, and more variable sleep duration.^3,4 Sleep disorders also may be related to mental disorders and other medi­cal conditions.⁵ For example, the prevalence of sleep-related respiratory disorders, such as obstructive sleep apnea and central sleep apnea, increases with age.⁶

Sleep disorders are common among geri­atric patients. In a large epidemiologic study of sleep complaints in patients age ≥65, more than one-half of patients had at least 1 sleep complaint (ie, difficulty falling asleep, trou­ble waking up, early awakening, need for naps, and feeling ill-rested).⁷ As many as 34% of patients reported symptoms of insom­nia. In an analysis of National Ambulatory Medical Survey Data over 6 years, 24.8% to 27.9% of sleep-related medical office visits were attributed to patients age ≥65.⁸

Pharmacology in aging
Prescribing sedative-hypnotic drugs is not routinely recommended for older patients with a sleep disorder. Geriatric patients, com­pared with younger patients, are at higher risk of iatrogenic complications because of polypharmacy, comorbidities, relative renal and hepatic insufficiency, and other physi­ologic changes leading to alterations in drug exposure and metabolism (Table 1).^9-12

Aging is associated with changes in body composition, including an increase in total body fat and decrease in lean body mass and total body water. These changes, as well as a prolonged GI transit time, decrease in active gut transporters, decreased blood perfusion, and decrease in plasma proteins such as albumin (because of reduced liver function or malnutrition), may lead to alteration in drug absorption patterns and may increase the volume of distribution for lipophilic drugs. Additionally, the elimination half-life of some drugs may increase with age because of larger volumes of distribution and reduction in hepatic or renal clearance.

The clinical significance of these changes is not well established. Although the process of drug absorption can change with age, the amount of drug absorbed might not be significantly affected. An increase in the vol­ume of distribution and reduction in drug metabolism and clearance might lead to increasing amounts of circulating drug and duration of drug exposure, putting geriat­ric patients at an increased risk for adverse effects and drug toxicity.⁹

Among these mechanisms, Dolder et al¹¹ hypothesized that drug metabolism cata­lyzed by cytochrome P450 (CYP) enzymes and renal excretion may be of greatest con­cern. Although in vitro studies suggest that concentration of CYP enzymes does not decline with age, in vivo studies have dem­onstrated reduced CYP activity in geriatric patients.^11,12 Theoretically, a reduction in CYP activity would increase the bioavail­ability of drugs, especially those that are subject to extensive first-pass (ie, hepatic) metabolism, and may lead to a reduction in systemic clearance.

Independent of metabolic changes, geriatric patients are at risk of reduced renal clearance because of age-related changes in glomerular filtration rate. Pharmacodynamic changes might be observed in older patients and could be a concern even in the setting of unaltered pharmacokinetic factors.9 These changes usually require administering smaller drug dosages.

Sedative-hypnotic medications
Sedative-hypnotic agents include several barbiturates, benzodiazepines (BZDs), non-BZD benzodiazepine-receptor ago­nists (BzRAs), a melatonin-receptor agonist (ie, ramelteon), and an orexin-receptor antagonist (ie, suvorexant).^13,14Table 2^14-29 summarizes selected sedative-hypnotic drugs. Additional drug classes used to treat insomnia include:
   • sedating antidepressants (trazodone, amitriptyline, doxepin, mirtazapine)
   • antiepileptic drugs (gabapentin, tiagabine)
   • atypical antipsychotics (quetiapine, olanzapine).

FDA-approved agents for treating insomnia include amobarbital, butabarbi­tal, pentobarbital, phenobarbital, secobar­bital, chloral hydrate, diphenhydramine, doxylamine, doxepin, estazolam, fluraz­epam, lorazepam, quazepam, temazepam, triazolam, eszopiclone, zaleplon, zolpidem, ramelteon, and suvorexant. Not all of these drugs are recommended for use in geriatric patients. Barbiturates, for example, should be avoided.³⁰

Pharmacokinetic characteristics vary among drugs and drug classes. Choice of pharmacotherapy should account for patient and drug characteristics and the specific sleep complaint. Sleep disorders may be var­iously characterized as difficulty with sleep initiation, duration, consolidation, or qual­ity.¹³ Therefore, onset and duration of effect are important drug-related considerations. Sedative-hypnotic drugs with a short time-to-onset may be ideal for patients with sleep-onset insomnia.

The drugs’ duration of effect (eg, presence of active metabolites, long elimination half-life) also must be reviewed. A long elimi­nation half-life may lead to increased drug exposure and unwanted side effects such as residual daytime drowsiness. Despite this, sedative-hypnotic drugs with a longer duration of effect (eg, intermediate- or long-acting drugs) may be best for patients with insomnia defined by difficulty maintaining sleep.

Benzodiazepines vary in their time to onset of effect, rate of elimination, and metabolism.^15-21 BZDs that are FDA- approved for use as sedative-hypnotics are listed in Table 2.^14-29 These BZDs have different onsets of effect as evidenced by time to achieve maximum plasma con­centration (T_max), ranging from 0.5 hours (flurazepam) to 2 hours (estazolam, quaz­epam, triazolam). The elimination half-life varies widely among these medications, from 1.5 hours (triazolam) to >100 hours (flurazepam). Flurazepam’s long half-life is attributable to its active major metabo­lite. Although most BZDs are metabolized hepatically, temazepam is subject to mini­mal hepatic metabolism.

Benzodiazepine-receptor agonists. There is substantial variation in the phar­macokinetic characteristics of BzRAs.^15,16,22-28 There also are differences among the zolpi­dem dosage forms; sublingual formulations have the shortest onset of effect. Eszopiclone and zaleplon have low protein binding com­pared with zolpidem. Elimination half-lives vary among drugs with the shortest attrib­uted to zaleplon (1 hour) and longest to eszopiclone (6 hours). All BzRAs are subject to extensive hepatic metabolism. 

Ramelteon. Singular in its class, ramelteon is a treatment option for insomnia.²⁹ This drug has a short onset of effect, moder­ate protein binding, and extensive hepatic metabolism. Ramelteon is primarily excreted in the urine as its metabolites, and the drug half-life is relatively short.

Suvorexant is the latest addition to the sedative-hypnotic armamentarium, approved by the FDA in August 2014 for dif­ficulty with sleep onset and/or sleep main­tenance.¹⁴ As an orexin-receptor antagonist, suvorexant represents a novel pharmaco­logic class. Suvorexant exhibits moderately rapid absorption with time to peak concen­tration ranging from 30 minutes to 6 hours in fasting conditions; absorption is delayed when taken with or soon after a meal. The drug is highly protein bound and extensively metabolized, primarily through CYP3A. The manufacturer recommends dose reduc­tion (5 mg at bedtime) in patients taking moderate CYP3A inhibitors and avoiding suvorexant in patients taking strong CYP3A inhibitors. Suvorexant is primarily excreted through feces and the mean half-life is rela­tively long.

Considering these characteristics and age-related physiologic changes, the practi­tioner should be concerned about drugs that undergo extensive hepatic metabolism. Age-related reductions in CYP activity may lead to an increase in drug bioavailability and a decrease in the systemic clearance,¹¹ which might be associated with an increase in elimination half-life and duration of action. Dosage adjustments are recommended for several BZDs (lower initial and maximum dosages for most agents) and BzRAs.^17-28 No dosage adjustments for ramelteon or suvorexant in geriatric patients have been specified^14,29; the manufacturers for both products assert that no differences in safety and efficacy have been observed between older and younger adult patients.

Alternative and complementary medications
Several non-prescription products, includ­ing over-the-counter drugs (eg, diphenhy-dramine, doxylamine) and herbal therapies (eg, melatonin, valerian), are used for their sedative-hypnotic properties. There is a lack of evidence supporting using diphenhydra-mine in patients with chronic insomnia, and tolerance to its hypnotic effect has been reported with repeated use.³¹ Concerns about anticholinergic toxicity and CNS depression limit its use in geriatric patients. Among herbal therapies, melatonin may have the strongest evidence for its ability to allevi­ate sleep disorders in geriatric patients³²; however, meta-analyses have demonstrated small effects of melatonin on sleep latency and minimal differences in wake time after sleep onset and total sleep time.¹³

Clinical practice guidelines
Non-pharmacotherapeutic interventions, such as behavioral (eg, sleep hygiene mea­sures) and psychological therapy, are rec­ommended for initial management of sleep disorders in geriatric patients.^13,33 In con­junction, the American Medical Directors Association (AMDA) recommends address­ ing underlying causes and exacerbating fac­tors (eg, medical condition or medication).³³ The AMDA recommends avoiding long-term pharmacotherapy and advises caution with BZD-hypnotic drugs, tricyclic antide­pressants, and antihistamines. The American Academy of Sleep Medicine (AASM) recom­mends an initial treatment period of 2 to 4 weeks, followed by re-evaluation of con­tinued need for treatment.¹³ The AASM recommends short- or intermediate-acting BzRAs or ramelteon for initial pharmaco­logic management of primary insomnias and insomnias comorbid with other condi­tions. The AASM also recommends specific dosages of BzRAs and BZDs for geriatric patients, which coincide with manufacturer-recommended dosages (Table 2).^14-29

Barbiturates, chloral hydrate, and non-barbiturate, non-BZD drugs such as mep­robamate are not recommended because of potential significant adverse effects and tolerance/dependence, and low therapeu­tic index. The AASM advises caution when using prescription drugs off-label for insom­nia (eg, antidepressants, antiepileptics, antipsychotics) and recommends avoiding them, if possible, because of limited evi­dence supporting their use.¹³

Safety concerns
Two commonly used references contain rec­ommendations for sedative-hypnotic medi­cation use in geriatric patients.^30,34 According to Gallagher et al’s³⁴ Screening Tool of Older Person’s Prescriptions (STOPP), long-term (>1 month) use of long-acting BZDs (eg, flurazepam, diazepam) and prolonged use (>1 week) of first-generation antihistamines (eg, diphenhydramine, doxylamine) should be avoided in patients age ≥65 because of the risk of sedation, confusion, and anti­cholinergic side effects. STOPP recognizes that any use of BZDs, neuroleptics, or first-generation antihistamines may contrib­ute to postural imbalance; therefore these agents are not recommended in older patients at risk for falls.

In the 2012 American Geriatrics Society (AGS) Beers Criteria, the AGS recommends avoiding barbiturates in older adults because of the high rate of physical depen­dence, tolerance to sleep effects, and over­dose risk at low dosages.³⁰ The AGS also recommends avoiding BZDs, stating that older adults have increased sensitivity to these agents and are at an increased risk of cognitive impairment, delirium, falls, frac­tures, and motor vehicle accidents when taking these drugs. Non-BZD BzRAs also should not be prescribed to patients with a history of falls or fractures, unless safer alternatives are not available.

The FDA has issued several advisory reports regarding sedative-hypnotic drugs. In 2007, all manufacturers of sedative-hypnotic drugs were required to modify their product labeling to include stronger language about potential risks.35 Among these changes, warnings for ana­phylaxis and complex sleep-related behav­iors were added. Also, the FDA requested that manufacturers of sedative-hypnotic drugs develop and provide patient medi­cation guides, advising consumers on the potential risks and precautions associated with these drugs. More recently, the FDA announced changes to dosing recommen­dations for zolpidem-containing products because of the risk of impaired mental alertness36; manufacturers were required to lower the recommended dosages for each product.

Manufacturers of FDA-approved sedative-hypnotic drugs urge caution when prescribing these medications for geriatric patients, citing the potential for increased sensitivity, manifesting as marked excite­ment, depression, or confusion (eg, barbi­turates), and greater risk for dosage-related adverse effects (eg, oversedation, dizziness, confusion, impaired psychomotor perfor­mance, ataxia).^17-29

Use in clinical practice
Several variables should be considered when evaluating appropriateness of phar­macotherapy, including characteristics of the drug and the patient. Geriatric patients may be prone to comorbidities resulting from age-related physiologic changes. These diseases may be confounding (ie, contributing to sleep disorders); examples include medical illnesses, such as hyperthyroidism and arthritis, and psychiatric illnesses, such as depression and anxiety.³⁷ Other conditions, such as renal and hepatic dysfunction, may lead to alteration in drug exposure. These condi­tions should be assessed through routine renal function tests (eg, serum creatinine and glomerular filtration rate) and liver function tests (eg, serum albumin and liver transaminases).

Multiple comorbidities suggest a higher likelihood of polypharmacy, leading to other drug-related issues (eg, drug-drug interactions). Although these issues may guide therapy by restricting medication options, their potential contribution to the underlying sleep complaints should be con­sidered.³⁷ Several drugs commonly used by geriatric patients may affect wakefulness (eg, analgesics, antidepressants, and anti­hypertensives [sedating], and thyroid hor­mones, corticosteroids, and CNS stimulants [alerting]).

CASE CONTINUED
In Mr. R’s case, zaleplon was initiated at 10 mg/d. Because of his age and the nature of his sleep disorder, the choice of sedative-hypnotic was suitable; however, the prescribed dosage was inappropriate. The sluggishness Mr. R experienced likely was a manifestation of increased exposure to the drug. According to manufacturer and AASM recommendations, a more appropriate dosage is 5 mg/d.^13,23 Mr. R’s medical history and current medica­tions, and his hepatic and renal function, should be assessed. If Mr. R continues to have issues with sleep initiation, zaleplon, 5 mg at bedtime, should be considered.

Related Resources
• Institute for Safe Medication Practices. www.ismp.org.
• MedWatch: The FDA Safety Information and Adverse Event Reporting Program. www.fda.gov/Safety/MedWatch/default.htm.

Drug Brand Names
Amitriptyline • Elavil                                   Mirtazapine • Remeron
Amobarbital • Amytal                                  Olanzapine • Zyprexa
Butabarbital • Butisol                                  Pentobarbital • Nembutal
Chloral hydrate • Somnote                          Phenobarbital • Luminal
Diazepam • Valium                                     Quazepam • Doral
Diphenhydramine • Benadryl, others            Quetiapine • Seroquel
Doxepin • Silenor                                       Ramelteon • Rozerem
Doxylamine • Unisom, others                      Secobarbital • Seconal
Estazolam • ProSom                                   Suvorexant • Belsomra
Eszopiclone • Lunesta                                 Temazepam • Restoril
Flurazepam • Dalmane                               Tiagabine • Gabitril
Gabapentin • Neurontin,                             Trazodone • Desyrel
Gralise, Horizant                                        Zaleplon • Sonata
Lorazepam • Ativan                                   Zolpidem • Ambien, Edluar,
Meprobamate • Equanil                                 Intermezzo, Zolpimist

Acknowledgement
Vicki L. Ellingrod, PharmD, FCCP, is the series editor of Savvy Psychopharmacology.

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Mr. R, 75, is having difficulty sleeping. When he goes to bed, he lies there for what seems like forever, unable to fall asleep. He feels “so tired” and ends up taking naps during the day, but he cannot break this cycle. He has tried using over-the-counter products with little relief.

Mr. R’s primary care physician prescribes zaleplon, 10 mg/d, and asks him to call the clinic in 2 weeks to discuss his progress. He takes zaleplon as directed for several nights and begins to feel “sluggish” during the day, both mentally and physically, despite reporting an increase in the overall amount of sleep at night.

Sedative-hypnotic drugs are among the most commonly used medica­tions in the United States. Use of these drugs, as well as anxiolytics, has increased from 2.8% between 1988 and 1994 to 4.7% between 2007 and 2010, according to the Department of Health and Human Services.¹ In 2011, drugs categorized as sedative-hypnotics or antipsychotics were involved in 6.1% of all human exposures identified in the American Association of Poison Control Centers’ National Poison Data System.² Therefore, an understanding of clinical and pharmacological variables related to safe and effective use is important for clinicians prescribing and monitoring therapy with these agents.

Neuropsychiatric disorders are prevalent among geriatric patients and are associated with age-related physiologic changes in the CNS.³ Such changes involve:  
   • neuroanatomy (brain atrophy, decreased neuronal density, increased plaque formation)  
   • neurotransmitters (reduced cholinergic transmission, decreased synthesis of dopamine and catecholamines), and   
   • neurophysiology (reduced cerebral blood flow).

These physiologic processes manifest as alterations in mental status, reflexes, sen­sation, gait, balance, and sleep. Examples of sleep changes among geriatric patients include decreased sleep efficiency, more fre­quent awakenings, and more variable sleep duration.^3,4 Sleep disorders also may be related to mental disorders and other medi­cal conditions.⁵ For example, the prevalence of sleep-related respiratory disorders, such as obstructive sleep apnea and central sleep apnea, increases with age.⁶

Sleep disorders are common among geri­atric patients. In a large epidemiologic study of sleep complaints in patients age ≥65, more than one-half of patients had at least 1 sleep complaint (ie, difficulty falling asleep, trou­ble waking up, early awakening, need for naps, and feeling ill-rested).⁷ As many as 34% of patients reported symptoms of insom­nia. In an analysis of National Ambulatory Medical Survey Data over 6 years, 24.8% to 27.9% of sleep-related medical office visits were attributed to patients age ≥65.⁸

Pharmacology in aging
Prescribing sedative-hypnotic drugs is not routinely recommended for older patients with a sleep disorder. Geriatric patients, com­pared with younger patients, are at higher risk of iatrogenic complications because of polypharmacy, comorbidities, relative renal and hepatic insufficiency, and other physi­ologic changes leading to alterations in drug exposure and metabolism (Table 1).^9-12

Aging is associated with changes in body composition, including an increase in total body fat and decrease in lean body mass and total body water. These changes, as well as a prolonged GI transit time, decrease in active gut transporters, decreased blood perfusion, and decrease in plasma proteins such as albumin (because of reduced liver function or malnutrition), may lead to alteration in drug absorption patterns and may increase the volume of distribution for lipophilic drugs. Additionally, the elimination half-life of some drugs may increase with age because of larger volumes of distribution and reduction in hepatic or renal clearance.

The clinical significance of these changes is not well established. Although the process of drug absorption can change with age, the amount of drug absorbed might not be significantly affected. An increase in the vol­ume of distribution and reduction in drug metabolism and clearance might lead to increasing amounts of circulating drug and duration of drug exposure, putting geriat­ric patients at an increased risk for adverse effects and drug toxicity.⁹

Among these mechanisms, Dolder et al¹¹ hypothesized that drug metabolism cata­lyzed by cytochrome P450 (CYP) enzymes and renal excretion may be of greatest con­cern. Although in vitro studies suggest that concentration of CYP enzymes does not decline with age, in vivo studies have dem­onstrated reduced CYP activity in geriatric patients.^11,12 Theoretically, a reduction in CYP activity would increase the bioavail­ability of drugs, especially those that are subject to extensive first-pass (ie, hepatic) metabolism, and may lead to a reduction in systemic clearance.

Independent of metabolic changes, geriatric patients are at risk of reduced renal clearance because of age-related changes in glomerular filtration rate. Pharmacodynamic changes might be observed in older patients and could be a concern even in the setting of unaltered pharmacokinetic factors.9 These changes usually require administering smaller drug dosages.

Sedative-hypnotic medications
Sedative-hypnotic agents include several barbiturates, benzodiazepines (BZDs), non-BZD benzodiazepine-receptor ago­nists (BzRAs), a melatonin-receptor agonist (ie, ramelteon), and an orexin-receptor antagonist (ie, suvorexant).^13,14Table 2^14-29 summarizes selected sedative-hypnotic drugs. Additional drug classes used to treat insomnia include:
   • sedating antidepressants (trazodone, amitriptyline, doxepin, mirtazapine)
   • antiepileptic drugs (gabapentin, tiagabine)
   • atypical antipsychotics (quetiapine, olanzapine).

FDA-approved agents for treating insomnia include amobarbital, butabarbi­tal, pentobarbital, phenobarbital, secobar­bital, chloral hydrate, diphenhydramine, doxylamine, doxepin, estazolam, fluraz­epam, lorazepam, quazepam, temazepam, triazolam, eszopiclone, zaleplon, zolpidem, ramelteon, and suvorexant. Not all of these drugs are recommended for use in geriatric patients. Barbiturates, for example, should be avoided.³⁰

Pharmacokinetic characteristics vary among drugs and drug classes. Choice of pharmacotherapy should account for patient and drug characteristics and the specific sleep complaint. Sleep disorders may be var­iously characterized as difficulty with sleep initiation, duration, consolidation, or qual­ity.¹³ Therefore, onset and duration of effect are important drug-related considerations. Sedative-hypnotic drugs with a short time-to-onset may be ideal for patients with sleep-onset insomnia.

The drugs’ duration of effect (eg, presence of active metabolites, long elimination half-life) also must be reviewed. A long elimi­nation half-life may lead to increased drug exposure and unwanted side effects such as residual daytime drowsiness. Despite this, sedative-hypnotic drugs with a longer duration of effect (eg, intermediate- or long-acting drugs) may be best for patients with insomnia defined by difficulty maintaining sleep.

Benzodiazepines vary in their time to onset of effect, rate of elimination, and metabolism.^15-21 BZDs that are FDA- approved for use as sedative-hypnotics are listed in Table 2.^14-29 These BZDs have different onsets of effect as evidenced by time to achieve maximum plasma con­centration (T_max), ranging from 0.5 hours (flurazepam) to 2 hours (estazolam, quaz­epam, triazolam). The elimination half-life varies widely among these medications, from 1.5 hours (triazolam) to >100 hours (flurazepam). Flurazepam’s long half-life is attributable to its active major metabo­lite. Although most BZDs are metabolized hepatically, temazepam is subject to mini­mal hepatic metabolism.

Benzodiazepine-receptor agonists. There is substantial variation in the phar­macokinetic characteristics of BzRAs.^15,16,22-28 There also are differences among the zolpi­dem dosage forms; sublingual formulations have the shortest onset of effect. Eszopiclone and zaleplon have low protein binding com­pared with zolpidem. Elimination half-lives vary among drugs with the shortest attrib­uted to zaleplon (1 hour) and longest to eszopiclone (6 hours). All BzRAs are subject to extensive hepatic metabolism. 

Ramelteon. Singular in its class, ramelteon is a treatment option for insomnia.²⁹ This drug has a short onset of effect, moder­ate protein binding, and extensive hepatic metabolism. Ramelteon is primarily excreted in the urine as its metabolites, and the drug half-life is relatively short.

Suvorexant is the latest addition to the sedative-hypnotic armamentarium, approved by the FDA in August 2014 for dif­ficulty with sleep onset and/or sleep main­tenance.¹⁴ As an orexin-receptor antagonist, suvorexant represents a novel pharmaco­logic class. Suvorexant exhibits moderately rapid absorption with time to peak concen­tration ranging from 30 minutes to 6 hours in fasting conditions; absorption is delayed when taken with or soon after a meal. The drug is highly protein bound and extensively metabolized, primarily through CYP3A. The manufacturer recommends dose reduc­tion (5 mg at bedtime) in patients taking moderate CYP3A inhibitors and avoiding suvorexant in patients taking strong CYP3A inhibitors. Suvorexant is primarily excreted through feces and the mean half-life is rela­tively long.

Considering these characteristics and age-related physiologic changes, the practi­tioner should be concerned about drugs that undergo extensive hepatic metabolism. Age-related reductions in CYP activity may lead to an increase in drug bioavailability and a decrease in the systemic clearance,¹¹ which might be associated with an increase in elimination half-life and duration of action. Dosage adjustments are recommended for several BZDs (lower initial and maximum dosages for most agents) and BzRAs.^17-28 No dosage adjustments for ramelteon or suvorexant in geriatric patients have been specified^14,29; the manufacturers for both products assert that no differences in safety and efficacy have been observed between older and younger adult patients.

Alternative and complementary medications
Several non-prescription products, includ­ing over-the-counter drugs (eg, diphenhy-dramine, doxylamine) and herbal therapies (eg, melatonin, valerian), are used for their sedative-hypnotic properties. There is a lack of evidence supporting using diphenhydra-mine in patients with chronic insomnia, and tolerance to its hypnotic effect has been reported with repeated use.³¹ Concerns about anticholinergic toxicity and CNS depression limit its use in geriatric patients. Among herbal therapies, melatonin may have the strongest evidence for its ability to allevi­ate sleep disorders in geriatric patients³²; however, meta-analyses have demonstrated small effects of melatonin on sleep latency and minimal differences in wake time after sleep onset and total sleep time.¹³

Clinical practice guidelines
Non-pharmacotherapeutic interventions, such as behavioral (eg, sleep hygiene mea­sures) and psychological therapy, are rec­ommended for initial management of sleep disorders in geriatric patients.^13,33 In con­junction, the American Medical Directors Association (AMDA) recommends address­ ing underlying causes and exacerbating fac­tors (eg, medical condition or medication).³³ The AMDA recommends avoiding long-term pharmacotherapy and advises caution with BZD-hypnotic drugs, tricyclic antide­pressants, and antihistamines. The American Academy of Sleep Medicine (AASM) recom­mends an initial treatment period of 2 to 4 weeks, followed by re-evaluation of con­tinued need for treatment.¹³ The AASM recommends short- or intermediate-acting BzRAs or ramelteon for initial pharmaco­logic management of primary insomnias and insomnias comorbid with other condi­tions. The AASM also recommends specific dosages of BzRAs and BZDs for geriatric patients, which coincide with manufacturer-recommended dosages (Table 2).^14-29

Barbiturates, chloral hydrate, and non-barbiturate, non-BZD drugs such as mep­robamate are not recommended because of potential significant adverse effects and tolerance/dependence, and low therapeu­tic index. The AASM advises caution when using prescription drugs off-label for insom­nia (eg, antidepressants, antiepileptics, antipsychotics) and recommends avoiding them, if possible, because of limited evi­dence supporting their use.¹³

Safety concerns
Two commonly used references contain rec­ommendations for sedative-hypnotic medi­cation use in geriatric patients.^30,34 According to Gallagher et al’s³⁴ Screening Tool of Older Person’s Prescriptions (STOPP), long-term (>1 month) use of long-acting BZDs (eg, flurazepam, diazepam) and prolonged use (>1 week) of first-generation antihistamines (eg, diphenhydramine, doxylamine) should be avoided in patients age ≥65 because of the risk of sedation, confusion, and anti­cholinergic side effects. STOPP recognizes that any use of BZDs, neuroleptics, or first-generation antihistamines may contrib­ute to postural imbalance; therefore these agents are not recommended in older patients at risk for falls.

In the 2012 American Geriatrics Society (AGS) Beers Criteria, the AGS recommends avoiding barbiturates in older adults because of the high rate of physical depen­dence, tolerance to sleep effects, and over­dose risk at low dosages.³⁰ The AGS also recommends avoiding BZDs, stating that older adults have increased sensitivity to these agents and are at an increased risk of cognitive impairment, delirium, falls, frac­tures, and motor vehicle accidents when taking these drugs. Non-BZD BzRAs also should not be prescribed to patients with a history of falls or fractures, unless safer alternatives are not available.

The FDA has issued several advisory reports regarding sedative-hypnotic drugs. In 2007, all manufacturers of sedative-hypnotic drugs were required to modify their product labeling to include stronger language about potential risks.35 Among these changes, warnings for ana­phylaxis and complex sleep-related behav­iors were added. Also, the FDA requested that manufacturers of sedative-hypnotic drugs develop and provide patient medi­cation guides, advising consumers on the potential risks and precautions associated with these drugs. More recently, the FDA announced changes to dosing recommen­dations for zolpidem-containing products because of the risk of impaired mental alertness36; manufacturers were required to lower the recommended dosages for each product.

Manufacturers of FDA-approved sedative-hypnotic drugs urge caution when prescribing these medications for geriatric patients, citing the potential for increased sensitivity, manifesting as marked excite­ment, depression, or confusion (eg, barbi­turates), and greater risk for dosage-related adverse effects (eg, oversedation, dizziness, confusion, impaired psychomotor perfor­mance, ataxia).^17-29

Use in clinical practice
Several variables should be considered when evaluating appropriateness of phar­macotherapy, including characteristics of the drug and the patient. Geriatric patients may be prone to comorbidities resulting from age-related physiologic changes. These diseases may be confounding (ie, contributing to sleep disorders); examples include medical illnesses, such as hyperthyroidism and arthritis, and psychiatric illnesses, such as depression and anxiety.³⁷ Other conditions, such as renal and hepatic dysfunction, may lead to alteration in drug exposure. These condi­tions should be assessed through routine renal function tests (eg, serum creatinine and glomerular filtration rate) and liver function tests (eg, serum albumin and liver transaminases).

Multiple comorbidities suggest a higher likelihood of polypharmacy, leading to other drug-related issues (eg, drug-drug interactions). Although these issues may guide therapy by restricting medication options, their potential contribution to the underlying sleep complaints should be con­sidered.³⁷ Several drugs commonly used by geriatric patients may affect wakefulness (eg, analgesics, antidepressants, and anti­hypertensives [sedating], and thyroid hor­mones, corticosteroids, and CNS stimulants [alerting]).

CASE CONTINUED
In Mr. R’s case, zaleplon was initiated at 10 mg/d. Because of his age and the nature of his sleep disorder, the choice of sedative-hypnotic was suitable; however, the prescribed dosage was inappropriate. The sluggishness Mr. R experienced likely was a manifestation of increased exposure to the drug. According to manufacturer and AASM recommendations, a more appropriate dosage is 5 mg/d.^13,23 Mr. R’s medical history and current medica­tions, and his hepatic and renal function, should be assessed. If Mr. R continues to have issues with sleep initiation, zaleplon, 5 mg at bedtime, should be considered.

Related Resources
• Institute for Safe Medication Practices. www.ismp.org.
• MedWatch: The FDA Safety Information and Adverse Event Reporting Program. www.fda.gov/Safety/MedWatch/default.htm.

Drug Brand Names
Amitriptyline • Elavil                                   Mirtazapine • Remeron
Amobarbital • Amytal                                  Olanzapine • Zyprexa
Butabarbital • Butisol                                  Pentobarbital • Nembutal
Chloral hydrate • Somnote                          Phenobarbital • Luminal
Diazepam • Valium                                     Quazepam • Doral
Diphenhydramine • Benadryl, others            Quetiapine • Seroquel
Doxepin • Silenor                                       Ramelteon • Rozerem
Doxylamine • Unisom, others                      Secobarbital • Seconal
Estazolam • ProSom                                   Suvorexant • Belsomra
Eszopiclone • Lunesta                                 Temazepam • Restoril
Flurazepam • Dalmane                               Tiagabine • Gabitril
Gabapentin • Neurontin,                             Trazodone • Desyrel
Gralise, Horizant                                        Zaleplon • Sonata
Lorazepam • Ativan                                   Zolpidem • Ambien, Edluar,
Meprobamate • Equanil                                 Intermezzo, Zolpimist

Acknowledgement
Vicki L. Ellingrod, PharmD, FCCP, is the series editor of Savvy Psychopharmacology.

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Series Editor: Arthur T. Skarin, MD, FACP, FCCP

Neuroendocrine tumors (NETs) are a rare, heterogeneous group of neoplasms that arise from neuroendocrine cells located throughout the body. These tumors are characterized by variable but most often indolent biologic behavior. They are also classically characterized by their ability to secrete peptides, resulting in distinctive hormonal syndromes. Although NETs have been considered rare, recent studies suggest that they are more common than previously suspected. An analysis of the Surveillance, Epidemiology, and End Results (SEER) database demonstrated a significant increase in the incidence of NETs over time with an age-adjusted annual incidence in the United States of 5.25 cases per 100,000 population. The increase in incidence is likely attributable to increasing awareness, improved diagnostic strategies, and possibly other undetermined environmental and genetic factors.

To read the full article in PDF:

Click here

Series Editor: Arthur T. Skarin, MD, FACP, FCCP

Neuroendocrine tumors (NETs) are a rare, heterogeneous group of neoplasms that arise from neuroendocrine cells located throughout the body. These tumors are characterized by variable but most often indolent biologic behavior. They are also classically characterized by their ability to secrete peptides, resulting in distinctive hormonal syndromes. Although NETs have been considered rare, recent studies suggest that they are more common than previously suspected. An analysis of the Surveillance, Epidemiology, and End Results (SEER) database demonstrated a significant increase in the incidence of NETs over time with an age-adjusted annual incidence in the United States of 5.25 cases per 100,000 population. The increase in incidence is likely attributable to increasing awareness, improved diagnostic strategies, and possibly other undetermined environmental and genetic factors.

To read the full article in PDF:

Click here

Series Editor: Arthur T. Skarin, MD, FACP, FCCP

Neuroendocrine tumors (NETs) are a rare, heterogeneous group of neoplasms that arise from neuroendocrine cells located throughout the body. These tumors are characterized by variable but most often indolent biologic behavior. They are also classically characterized by their ability to secrete peptides, resulting in distinctive hormonal syndromes. Although NETs have been considered rare, recent studies suggest that they are more common than previously suspected. An analysis of the Surveillance, Epidemiology, and End Results (SEER) database demonstrated a significant increase in the incidence of NETs over time with an age-adjusted annual incidence in the United States of 5.25 cases per 100,000 population. The increase in incidence is likely attributable to increasing awareness, improved diagnostic strategies, and possibly other undetermined environmental and genetic factors.

To read the full article in PDF:

Click here

In 1984, a worker at a Pennsylvania nuclear power plant triggered the radiation detector as he was getting ready to go home. This would not be unusual for such a facility, but there was no nuclear fuel on site at the time. The alarm went off every time he left work.

One day, he triggered the alarm as he crossed the detector on arriving at the plant, leading him to suspect that he was bringing radiation from home. He eventually convinced the plant’s health physicists to check his home, although at first they were opposed to the idea. The results revealed high concentrations of radon everywhere, especially in his basement.

Radon was already known to be associated with health risks in underground miners at that time. This event revealed that a naturally occurring radioactive gas could be found in households at potentially hazardous concentrations.

The incident captured the public’s attention, and the Environmental Protection Agency (EPA) and the US Centers for Disease Control and Prevention (CDC) recommended that nearly all homes be tested.^1,2 In 1988, the International Agency for Research on Cancer classified radon as a human carcinogen, and Congress passed the Indoor Radon Abatement Act in response to growing concern over health risks.³ This law funded state and federal measures to survey schools and federal buildings for radon levels, to educate citizens, and to develop programs for technical assistance. The long-term goal was to reduce indoor levels nationwide to no more than outdoor levels.

Radon is still considered an important public health hazard. From 15,000 to 21,000 people are estimated to die of lung cancer as a result of radon exposure each year in the United States, making it the second most common cause of lung cancer, behind smoking.⁴

Considering the relevance of this issue, this article will review the unique characteristics of radon as a risk factor for lung cancer.

WHAT IS RADON?

Radon is a noble gas that occurs naturally as a decay product of uranium 238 and thorium 232. It is colorless, tasteless, and imperceptible to our senses. Its most common isotope is radon 222 (²²²Rn), which has a half-life of 3.8 days and decays by emitting an alpha particle to become polonium 218. The decay chain continues through several intermediate steps until the stable isotope lead 206 is formed (Figure 1). Two of the isotopes in this chain, polonium 218 and polonium 214, also emit alpha particles.^5–7

Radon is primarily formed in soil. Its most important precursor, uranium 238, is ubiquitous, found in most soils and rocks in various concentrations. Radon can also be found in surface water, metal mines (uranium, phosphorus, silver, gold), residue of coal combustion, and natural gas.

Outdoor levels are usually much lower than indoor levels, as radon dissipates very quickly. Indoor radon mostly comes from the soil under the house or building, but it can also originate from coal combustion, gas appliances, and water (especially from private wells). In municipal water systems or surface reservoirs, most of the radon dissipates into the air or decays before the water reaches homes.^8,9

Radon’s only commercial application in the United States is in calibrating measuring instruments. In the past, it was used in radiography and to treat cancer but was later replaced by other radiation sources that cost less and pose less hazard of alpha radiation.¹⁰

HOW RADON CAN HARM

Alpha particles, emitted by radon 222 and its progenies polonium 218 and polonium 214, are highly effective in damaging tissues. Although they do not travel far or fast, with their two protons and two neutrons, alpha particles are heavy and therefore can cause considerable damage at short range. Although alpha particles can be stopped by a thin barrier such as a piece of paper or the skin, if the source is inhaled or ingested and lodges against mucosal linings, the alpha particles emitted can destroy cells.¹¹

The main route of radon exposure is by inhalation. Since radon is biologically inert, it is readily exhaled after it reaches the lungs. However, radon’s progenies can also be inhaled, either as free particles or attached to airborne particles such as dust, which they tend to attract as a result of their charged state. This attached fraction is believed to be more carcinogenic because it tends to deposit on the respiratory epithelium, notably in the carinae of bronchi. The smaller the dust particle, the deeper it can travel into the lung. The radiation emissions damage the genetic material of cells lining the airways, with the potential to result in lung cancer if the repair process is incomplete.^5,8,9

Other routes of exposure include ingestion and dermal exposure. Radon and its progenies can be swallowed in drinking water, passing through the stomach walls and bowels and entering the blood.¹² Dermal exposure is not considered a significant route unless the dermis is exposed, since in usual circumstances the skin protects the body from alpha radiation.¹³

Possible biologic mechanisms by which radon exposure might increase the risk of cancer include gene mutations, chromosome aberrations, generation of reactive oxygen species, up- or down-regulation of cytokines, and production of proteins associated with cell-cycle regulation.^14–16

HOW IS RADON MEASURED?

Several devices are commercially available to measure radon levels at home. The most common ones are activated charcoal detectors, electret ion chambers, alpha-track detectors, electronic integrating devices, and continuous monitors. There is no evidence that one device is better than another, but devices that measure radon gas are usually preferred over those that measure decay products because they are simpler to use and more cost-effective. These devices are divided into those used for short-term testing (2–90 days) and long-term testing (Table 1).¹⁷

Radon levels can be expressed as follows:

Working levels. One working level (WL) is any combination of radon progeny in 1 L of air that ultimately releases 1.3 × 10⁵ MeV of alpha energy during decay. In studies of miners, the radon progeny concentrations are generally expressed in WL. The cumulative exposure of an individual to this concentration over a “working month” of 170 hours is defined as a working level month (WLM).

Picocuries per liter. In the United States, the rate of decay is commonly reported in picocuries per liter (pCi/L): 1 pCi/L translates to 0.005 WL under usual conditions. The outdoor radon level is normally around 0.4 pCi/L.

Becquerel per cubic meter (Bq/m³) is an International System unit of measure: 1 WL corresponds to 3.7 × 10³ Bq/m³, and 1 pCi/L is equivalent to 37 Bq/m³.

Different areas have different radon levels

The Indoor Radon Abatement Act of 1988 helped identify areas in the United States that have the potential for elevated indoor radon levels. An estimated 6 million homes have concentrations greater than 4 pCi/L.

To assist in implementing radon-reducing strategies and allocation of resources, the EPA has created a map (Figure 2) that classifies counties according to the predicted indoor level.¹⁸

WHAT IS THE RELATIONSHIP BETWEEN RADON AND LUNG CANCER?

Determining the degree to which radon exposure contributes to lung cancer is a complex task. Radon can be found nearly everywhere, and there are diurnal, seasonal, and random year-to-year variations in the concentration of radon in indoor air.

A minority view

Not everyone agrees that radon is completely bad. For centuries, people have flocked to spas to “take the waters,” and the water at many of these spas has been found to contain radon. In the early 20th century, radiation was touted as having medicinal benefits, and people in many places in the world still go to “radon spas” (some of them in abandoned uranium mines) to help treat conditions such as arthritis and to feel invigorated and energized.

In 2006, a report by Zdrojewicz and Strzelczyk¹⁹ urged the medical community to keep an open mind about the possibility that radon exposure may be beneficial in very low doses, perhaps by stimulating repair mechanisms. This concept, called hormesis, differs from the mainstream view that cancer risk rises linearly with radiation dose, with no minimum threshold level (see below).

As early as in the 16th century, metal miners in central Europe were noted to have a high rate of death from respiratory disease. Radon was discovered in 1900, and in the 20th century lung cancer was linked to high levels of radon detected in uranium mines.

Several small studies of underground miners exposed to high concentrations of radon consistently demonstrated an increased risk of lung cancer.

The Committee on the Biological Effects of Ionizing Radiation (BEIR VI 1999) reviewed 11 major cohort studies of miners. The studies included more than 60,000 miners in Europe, North America, Asia, and Australia, of whom 2,600 died of lung cancer. Lung cancer rates increased linearly with cumulative radon exposure, and the estimated average increase in the lung cancer death rate per WLM in the combined studies was 0.44% (95% confidence interval [CI] 0.20–1.00%). The percentage increase in the lung cancer death rate per WLM varied with time since exposure, with the highest increase in risk during the 5 to 14 years after exposure.^4,17 Furthermore, the increase in risk was higher in younger miners, who were exposed to a relatively low radon concentration.

Risk in the general population

The magnitude of the risk in miners led to concern about radon exposure as a cause of lung cancer in the general population. Statistical models were generated that suggested a causal link between radon exposure and lung cancer. Although extrapolation of the results from miners caused controversy, the BEIR VI estimation of risk was validated by studies in the general population.^7,20–23

Since the 1980s, several small case-control studies with limited power examined the relationship between indoor radon and lung cancer in the general population. In these studies, individuals who had developed lung cancer were compared with controls who had not developed the disease but who otherwise represented the population from which the cases of lung cancer came.

To improve the statistical power, the investigators of the major studies in Europe, North America, and China pooled the results in separate analyses (Table 2).^7,20–23 The average radon concentration to which each individual had been exposed over the previous decades was estimated by measuring the radon concentration at their present and previous homes. On the basis of information from the uranium miners, these studies assumed that the period of exposure was the 30 years ending 5 years before the diagnosis or at death from lung cancer.

The results provided convincing evidence that radon exposure is a cause of lung cancer in the general population and substantiated the extrapolation from the studies of miners. Further, the results of all three pooled analyses were consistent with a linear dose-response relationship with no threshold, suggesting an increased risk of lung cancer even with a radon level below 4 pCi/L (200 Bq/m³), which is the concentration at which mitigation actions are recommended in many countries.¹⁷

The North American pooled analysis included 3,662 cases and 4,966 controls from seven studies in the United States and Canada. When data from all studies were combined, the risk of lung cancer was found to increase by 11% per 100-Bq/m³ (about 2.7-pCi/L) increase in measured radon concentration (95% CI 0%–28%). The estimated increase in lung cancer was independent of age, sex, or smoking history.^7,20

The Chinese pooled data²² demonstrated a 13% (95% CI 1%–36%) increased risk per 100 Bq/m³.

In the European study, the risk of lung cancer increased by 8% per 100 Bq/m³ (95% CI 3%–16%). The European investigators repeated the analysis, taking into account the random year-to-year variability in measured radon concentration, finding the final estimated risk was an increase of 16% per 100 Bq/m³ using long-term average concentration.²¹

The combined estimate^21,24 from the three pooling studies based on measured radon concentration is an increased risk of lung cancer of 10% per 100 Bq/m³.

Synergistic risk with smoking

Radon exposure was independently associated with lung cancer, and the relationship with cigarette smoking is believed to be synergistic. The radon progeny particles attach themselves to smoke and dust and are then deposited in the bronchial epithelium.²⁵

In the pool of European case-control studies, the cancer risk for current smokers of 15 to 24 cigarettes per day relative to that in never-smokers was 25.8 (95% CI 21–31). Assuming that in the same analysis the lung cancer risk increased by 16% per 100 Bq/m³ of usual radon concentration regardless of smoking status, the cumulative absolute risk by age 75 would be 0.67% in those who never smoked and 16% in smokers at usual radon levels of 400 Bq/m³ (11 pCi/L).²¹

Rates of all lung cancer subtypes increased

Radon exposure is not associated with a specific histologic subtype of lung cancer. It has been speculated that the incidence of the small-cell subtype might be slightly increased because radon tends to deposit in the more central bronchial carinae.^20,21 However, all subtypes have been described in association with radon, the most common being adenocarcinoma and squamous cell carcinoma.^26–28

EFFECT OF MITIGATION MEASURES

The US Surgeon General and the EPA recommend that all homes be tested.¹⁸ Short-term tests should be used first, keeping in mind that diurnal and seasonal variations may occur.

The World Health Organization has proposed a reference level of 100 Bq/m³ (2.7 pCi/L) to minimize health hazards from indoor radon exposure.¹⁷ If this level cannot be reached under the country-specific conditions, the chosen reference level should not exceed 300 Bq/m³ (8 pCi/L).

In the United States, if the result of home testing is higher than 4 pCi/L, a follow-up measurement should be done using a different short-term test or a long-term test. If the follow-up result confirms a level of more than 4 pCi/L, mitigating actions are recommended. The goal is to reduce the indoor radon level as much as possible—down to zero or at least comparable to outdoor levels (national average 0.4 pCi/L).¹⁸

A variety of radon mitigation strategies have been used, with different rates of efficacy (Table 3). The optimal strategy depends on the likely source or cause, construction characteristics, soil, and climate.²⁹ Table 4 lists resources for the general public about testing and mitigation measures.

How beneficial is radon mitigation?

Although it is logical to try to reduce the indoor radon concentration, there is no strong evidence yet that this intervention decreases the incidence of lung cancer in the general population.

Using the BEIR VI risk model, Méndez et al³⁰ estimated a 21% reduction in the annual radon-related lung cancer mortality rate by 2100 if all households were compliant with government recommendations (mitigation actions at levels of 4 pCi/L) and assuming that the percentage of cigarette smokers remained constant.

On the other hand, if the number of smokers continues to decline, the benefits from radon mitigation may be less. The expected benefit from mitigation in this scenario is a reduction of 12% in annual radon-related deaths by the year 2100.³⁰ However, it will be challenging to determine whether the expected decline in the incidence of lung cancer and lung cancer deaths is truly attributable to mitigation measures.

MANAGING PATIENTS EXPOSED TO RADON

Screen for lung cancer in smokers only

The National Lung Screening Study (NLST) was a large multicenter trial of annual low-dose computed tomography (CT) to screen for lung cancer in a cohort at high risk: age 55 to 74, at least a 30 pack-year history of smoking in a current smoker, or a former smoker who quit within the past 15 years. The trial demonstrated a 20% reduction in lung cancer deaths in the CT screening group.³¹

Since the publication of the NLST results, many societies have endorsed screening for lung cancer with low-dose CT using the study criteria. The National Comprehensive Cancer Network (NCCN) expanded these criteria and has recommended screening in patients over age 50 who have a history of smoking and one additional risk factor, such as radon exposure.

However, radon exposure has not been incorporated into a lung cancer risk-prediction model, and there is no empirical evidence suggesting that people who have such a history would benefit from screening.^32,33 The joint guidelines of the American College of Chest Physicians and American Society of Clinical Oncology recommend annual low-dose CT screening only for patients who meet the NLST criteria.³⁴

What to do about indeterminate lung nodules

The widely used guidelines from the Fleischner Society³⁵ on how to manage small lung nodules stratify patients into groups at low and high risk of developing lung cancer on the basis of risk factors. The guidelines apply to adults age 35 and older in whom an indeterminate solid nodule was recently detected.

If a patient is at high risk, the recommended approach includes follow-up in shorter intervals depending on the nodule size. History of smoking is recognized as a major risk factor, and the statement also lists family history and exposure to asbestos, uranium, and radon.³⁵

Although the association of radon with lung cancer has been shown in epidemiologic studies, radon exposure has not been included in validated statistical models that assess the probability that an indeterminate lung nodule is malignant. We would expect the risk to be higher in miners, who suffer a more intense exposure to higher levels of radon, than in the general population, which has a low and constantly variable residential exposure. Furthermore, there are no data to support a more aggressive follow-up approach in patients with indeterminate lung nodules and a history of radon exposure.

RADON AND OTHER CANCERS

When a person is exposed to radon, the bronchial epithelium receives the highest dose of ionizing radiation, but other organs such as the kidneys, stomach, and bone marrow may receive doses as well, although lower. Several studies have looked into possible associations, but there is no strong evidence to suggest an increased mortality rate related to radon from cancers other than lung.^24,36 However, there seems to be a positive association between radon and the incidence of lymphoproliferative disorders in uranium miners.^37,38

Radon can be measured in drinking water, and a few studies have looked at a possible association with gastrointestinal malignancies. The results did not reveal a consistent positive correlation.^39,40 The risk of cancer from exposure to radon in the public water supply is likely small and mostly from the transfer of radon particles into the air and not from drinking the water. On the other hand, the risk could be higher with private wells, where radon levels are variable and are possibly higher than from public sources.⁴¹

DATA ARE INSUFFICIENT TO GUIDE MANAGEMENT

Radon is a naturally occurring and ubiquitous radioactive gas that can cause tissue damage. Cohort and case-control studies have demonstrated that radon exposure is associated with increased risk of lung cancer. It is recommended that radon levels be measured in every home in the United States and mitigation measures instituted if levels exceed 4 pCi/L.

There are insufficient data to help guide the management of patients with a history of radon exposure, and prospective studies are needed to better understand the individual risk of developing lung cancer and the appropriate management of such patients.

Smoking cessation is an integral part of lung cancer risk reduction from radon exposure.

In 1984, a worker at a Pennsylvania nuclear power plant triggered the radiation detector as he was getting ready to go home. This would not be unusual for such a facility, but there was no nuclear fuel on site at the time. The alarm went off every time he left work.

One day, he triggered the alarm as he crossed the detector on arriving at the plant, leading him to suspect that he was bringing radiation from home. He eventually convinced the plant’s health physicists to check his home, although at first they were opposed to the idea. The results revealed high concentrations of radon everywhere, especially in his basement.

Radon was already known to be associated with health risks in underground miners at that time. This event revealed that a naturally occurring radioactive gas could be found in households at potentially hazardous concentrations.

The incident captured the public’s attention, and the Environmental Protection Agency (EPA) and the US Centers for Disease Control and Prevention (CDC) recommended that nearly all homes be tested.^1,2 In 1988, the International Agency for Research on Cancer classified radon as a human carcinogen, and Congress passed the Indoor Radon Abatement Act in response to growing concern over health risks.³ This law funded state and federal measures to survey schools and federal buildings for radon levels, to educate citizens, and to develop programs for technical assistance. The long-term goal was to reduce indoor levels nationwide to no more than outdoor levels.

Radon is still considered an important public health hazard. From 15,000 to 21,000 people are estimated to die of lung cancer as a result of radon exposure each year in the United States, making it the second most common cause of lung cancer, behind smoking.⁴

Considering the relevance of this issue, this article will review the unique characteristics of radon as a risk factor for lung cancer.

WHAT IS RADON?

Radon is a noble gas that occurs naturally as a decay product of uranium 238 and thorium 232. It is colorless, tasteless, and imperceptible to our senses. Its most common isotope is radon 222 (²²²Rn), which has a half-life of 3.8 days and decays by emitting an alpha particle to become polonium 218. The decay chain continues through several intermediate steps until the stable isotope lead 206 is formed (Figure 1). Two of the isotopes in this chain, polonium 218 and polonium 214, also emit alpha particles.^5–7

Radon is primarily formed in soil. Its most important precursor, uranium 238, is ubiquitous, found in most soils and rocks in various concentrations. Radon can also be found in surface water, metal mines (uranium, phosphorus, silver, gold), residue of coal combustion, and natural gas.

Outdoor levels are usually much lower than indoor levels, as radon dissipates very quickly. Indoor radon mostly comes from the soil under the house or building, but it can also originate from coal combustion, gas appliances, and water (especially from private wells). In municipal water systems or surface reservoirs, most of the radon dissipates into the air or decays before the water reaches homes.^8,9

Radon’s only commercial application in the United States is in calibrating measuring instruments. In the past, it was used in radiography and to treat cancer but was later replaced by other radiation sources that cost less and pose less hazard of alpha radiation.¹⁰

HOW RADON CAN HARM

Alpha particles, emitted by radon 222 and its progenies polonium 218 and polonium 214, are highly effective in damaging tissues. Although they do not travel far or fast, with their two protons and two neutrons, alpha particles are heavy and therefore can cause considerable damage at short range. Although alpha particles can be stopped by a thin barrier such as a piece of paper or the skin, if the source is inhaled or ingested and lodges against mucosal linings, the alpha particles emitted can destroy cells.¹¹

The main route of radon exposure is by inhalation. Since radon is biologically inert, it is readily exhaled after it reaches the lungs. However, radon’s progenies can also be inhaled, either as free particles or attached to airborne particles such as dust, which they tend to attract as a result of their charged state. This attached fraction is believed to be more carcinogenic because it tends to deposit on the respiratory epithelium, notably in the carinae of bronchi. The smaller the dust particle, the deeper it can travel into the lung. The radiation emissions damage the genetic material of cells lining the airways, with the potential to result in lung cancer if the repair process is incomplete.^5,8,9

Other routes of exposure include ingestion and dermal exposure. Radon and its progenies can be swallowed in drinking water, passing through the stomach walls and bowels and entering the blood.¹² Dermal exposure is not considered a significant route unless the dermis is exposed, since in usual circumstances the skin protects the body from alpha radiation.¹³

Possible biologic mechanisms by which radon exposure might increase the risk of cancer include gene mutations, chromosome aberrations, generation of reactive oxygen species, up- or down-regulation of cytokines, and production of proteins associated with cell-cycle regulation.^14–16

HOW IS RADON MEASURED?

Several devices are commercially available to measure radon levels at home. The most common ones are activated charcoal detectors, electret ion chambers, alpha-track detectors, electronic integrating devices, and continuous monitors. There is no evidence that one device is better than another, but devices that measure radon gas are usually preferred over those that measure decay products because they are simpler to use and more cost-effective. These devices are divided into those used for short-term testing (2–90 days) and long-term testing (Table 1).¹⁷

Radon levels can be expressed as follows:

Working levels. One working level (WL) is any combination of radon progeny in 1 L of air that ultimately releases 1.3 × 10⁵ MeV of alpha energy during decay. In studies of miners, the radon progeny concentrations are generally expressed in WL. The cumulative exposure of an individual to this concentration over a “working month” of 170 hours is defined as a working level month (WLM).

Picocuries per liter. In the United States, the rate of decay is commonly reported in picocuries per liter (pCi/L): 1 pCi/L translates to 0.005 WL under usual conditions. The outdoor radon level is normally around 0.4 pCi/L.

Becquerel per cubic meter (Bq/m³) is an International System unit of measure: 1 WL corresponds to 3.7 × 10³ Bq/m³, and 1 pCi/L is equivalent to 37 Bq/m³.

Different areas have different radon levels

The Indoor Radon Abatement Act of 1988 helped identify areas in the United States that have the potential for elevated indoor radon levels. An estimated 6 million homes have concentrations greater than 4 pCi/L.

To assist in implementing radon-reducing strategies and allocation of resources, the EPA has created a map (Figure 2) that classifies counties according to the predicted indoor level.¹⁸

WHAT IS THE RELATIONSHIP BETWEEN RADON AND LUNG CANCER?

Determining the degree to which radon exposure contributes to lung cancer is a complex task. Radon can be found nearly everywhere, and there are diurnal, seasonal, and random year-to-year variations in the concentration of radon in indoor air.

A minority view

Not everyone agrees that radon is completely bad. For centuries, people have flocked to spas to “take the waters,” and the water at many of these spas has been found to contain radon. In the early 20th century, radiation was touted as having medicinal benefits, and people in many places in the world still go to “radon spas” (some of them in abandoned uranium mines) to help treat conditions such as arthritis and to feel invigorated and energized.

In 2006, a report by Zdrojewicz and Strzelczyk¹⁹ urged the medical community to keep an open mind about the possibility that radon exposure may be beneficial in very low doses, perhaps by stimulating repair mechanisms. This concept, called hormesis, differs from the mainstream view that cancer risk rises linearly with radiation dose, with no minimum threshold level (see below).

As early as in the 16th century, metal miners in central Europe were noted to have a high rate of death from respiratory disease. Radon was discovered in 1900, and in the 20th century lung cancer was linked to high levels of radon detected in uranium mines.

Several small studies of underground miners exposed to high concentrations of radon consistently demonstrated an increased risk of lung cancer.

The Committee on the Biological Effects of Ionizing Radiation (BEIR VI 1999) reviewed 11 major cohort studies of miners. The studies included more than 60,000 miners in Europe, North America, Asia, and Australia, of whom 2,600 died of lung cancer. Lung cancer rates increased linearly with cumulative radon exposure, and the estimated average increase in the lung cancer death rate per WLM in the combined studies was 0.44% (95% confidence interval [CI] 0.20–1.00%). The percentage increase in the lung cancer death rate per WLM varied with time since exposure, with the highest increase in risk during the 5 to 14 years after exposure.^4,17 Furthermore, the increase in risk was higher in younger miners, who were exposed to a relatively low radon concentration.

Risk in the general population

The magnitude of the risk in miners led to concern about radon exposure as a cause of lung cancer in the general population. Statistical models were generated that suggested a causal link between radon exposure and lung cancer. Although extrapolation of the results from miners caused controversy, the BEIR VI estimation of risk was validated by studies in the general population.^7,20–23

Since the 1980s, several small case-control studies with limited power examined the relationship between indoor radon and lung cancer in the general population. In these studies, individuals who had developed lung cancer were compared with controls who had not developed the disease but who otherwise represented the population from which the cases of lung cancer came.

To improve the statistical power, the investigators of the major studies in Europe, North America, and China pooled the results in separate analyses (Table 2).^7,20–23 The average radon concentration to which each individual had been exposed over the previous decades was estimated by measuring the radon concentration at their present and previous homes. On the basis of information from the uranium miners, these studies assumed that the period of exposure was the 30 years ending 5 years before the diagnosis or at death from lung cancer.

The results provided convincing evidence that radon exposure is a cause of lung cancer in the general population and substantiated the extrapolation from the studies of miners. Further, the results of all three pooled analyses were consistent with a linear dose-response relationship with no threshold, suggesting an increased risk of lung cancer even with a radon level below 4 pCi/L (200 Bq/m³), which is the concentration at which mitigation actions are recommended in many countries.¹⁷

The North American pooled analysis included 3,662 cases and 4,966 controls from seven studies in the United States and Canada. When data from all studies were combined, the risk of lung cancer was found to increase by 11% per 100-Bq/m³ (about 2.7-pCi/L) increase in measured radon concentration (95% CI 0%–28%). The estimated increase in lung cancer was independent of age, sex, or smoking history.^7,20

The Chinese pooled data²² demonstrated a 13% (95% CI 1%–36%) increased risk per 100 Bq/m³.

In the European study, the risk of lung cancer increased by 8% per 100 Bq/m³ (95% CI 3%–16%). The European investigators repeated the analysis, taking into account the random year-to-year variability in measured radon concentration, finding the final estimated risk was an increase of 16% per 100 Bq/m³ using long-term average concentration.²¹

The combined estimate^21,24 from the three pooling studies based on measured radon concentration is an increased risk of lung cancer of 10% per 100 Bq/m³.

Synergistic risk with smoking

Radon exposure was independently associated with lung cancer, and the relationship with cigarette smoking is believed to be synergistic. The radon progeny particles attach themselves to smoke and dust and are then deposited in the bronchial epithelium.²⁵

In the pool of European case-control studies, the cancer risk for current smokers of 15 to 24 cigarettes per day relative to that in never-smokers was 25.8 (95% CI 21–31). Assuming that in the same analysis the lung cancer risk increased by 16% per 100 Bq/m³ of usual radon concentration regardless of smoking status, the cumulative absolute risk by age 75 would be 0.67% in those who never smoked and 16% in smokers at usual radon levels of 400 Bq/m³ (11 pCi/L).²¹

Rates of all lung cancer subtypes increased

Radon exposure is not associated with a specific histologic subtype of lung cancer. It has been speculated that the incidence of the small-cell subtype might be slightly increased because radon tends to deposit in the more central bronchial carinae.^20,21 However, all subtypes have been described in association with radon, the most common being adenocarcinoma and squamous cell carcinoma.^26–28

EFFECT OF MITIGATION MEASURES

The US Surgeon General and the EPA recommend that all homes be tested.¹⁸ Short-term tests should be used first, keeping in mind that diurnal and seasonal variations may occur.

The World Health Organization has proposed a reference level of 100 Bq/m³ (2.7 pCi/L) to minimize health hazards from indoor radon exposure.¹⁷ If this level cannot be reached under the country-specific conditions, the chosen reference level should not exceed 300 Bq/m³ (8 pCi/L).

In the United States, if the result of home testing is higher than 4 pCi/L, a follow-up measurement should be done using a different short-term test or a long-term test. If the follow-up result confirms a level of more than 4 pCi/L, mitigating actions are recommended. The goal is to reduce the indoor radon level as much as possible—down to zero or at least comparable to outdoor levels (national average 0.4 pCi/L).¹⁸

A variety of radon mitigation strategies have been used, with different rates of efficacy (Table 3). The optimal strategy depends on the likely source or cause, construction characteristics, soil, and climate.²⁹ Table 4 lists resources for the general public about testing and mitigation measures.

How beneficial is radon mitigation?

Although it is logical to try to reduce the indoor radon concentration, there is no strong evidence yet that this intervention decreases the incidence of lung cancer in the general population.

Using the BEIR VI risk model, Méndez et al³⁰ estimated a 21% reduction in the annual radon-related lung cancer mortality rate by 2100 if all households were compliant with government recommendations (mitigation actions at levels of 4 pCi/L) and assuming that the percentage of cigarette smokers remained constant.

On the other hand, if the number of smokers continues to decline, the benefits from radon mitigation may be less. The expected benefit from mitigation in this scenario is a reduction of 12% in annual radon-related deaths by the year 2100.³⁰ However, it will be challenging to determine whether the expected decline in the incidence of lung cancer and lung cancer deaths is truly attributable to mitigation measures.

MANAGING PATIENTS EXPOSED TO RADON

Screen for lung cancer in smokers only

The National Lung Screening Study (NLST) was a large multicenter trial of annual low-dose computed tomography (CT) to screen for lung cancer in a cohort at high risk: age 55 to 74, at least a 30 pack-year history of smoking in a current smoker, or a former smoker who quit within the past 15 years. The trial demonstrated a 20% reduction in lung cancer deaths in the CT screening group.³¹

Since the publication of the NLST results, many societies have endorsed screening for lung cancer with low-dose CT using the study criteria. The National Comprehensive Cancer Network (NCCN) expanded these criteria and has recommended screening in patients over age 50 who have a history of smoking and one additional risk factor, such as radon exposure.

However, radon exposure has not been incorporated into a lung cancer risk-prediction model, and there is no empirical evidence suggesting that people who have such a history would benefit from screening.^32,33 The joint guidelines of the American College of Chest Physicians and American Society of Clinical Oncology recommend annual low-dose CT screening only for patients who meet the NLST criteria.³⁴

What to do about indeterminate lung nodules

The widely used guidelines from the Fleischner Society³⁵ on how to manage small lung nodules stratify patients into groups at low and high risk of developing lung cancer on the basis of risk factors. The guidelines apply to adults age 35 and older in whom an indeterminate solid nodule was recently detected.

If a patient is at high risk, the recommended approach includes follow-up in shorter intervals depending on the nodule size. History of smoking is recognized as a major risk factor, and the statement also lists family history and exposure to asbestos, uranium, and radon.³⁵

Although the association of radon with lung cancer has been shown in epidemiologic studies, radon exposure has not been included in validated statistical models that assess the probability that an indeterminate lung nodule is malignant. We would expect the risk to be higher in miners, who suffer a more intense exposure to higher levels of radon, than in the general population, which has a low and constantly variable residential exposure. Furthermore, there are no data to support a more aggressive follow-up approach in patients with indeterminate lung nodules and a history of radon exposure.

RADON AND OTHER CANCERS

When a person is exposed to radon, the bronchial epithelium receives the highest dose of ionizing radiation, but other organs such as the kidneys, stomach, and bone marrow may receive doses as well, although lower. Several studies have looked into possible associations, but there is no strong evidence to suggest an increased mortality rate related to radon from cancers other than lung.^24,36 However, there seems to be a positive association between radon and the incidence of lymphoproliferative disorders in uranium miners.^37,38

Radon can be measured in drinking water, and a few studies have looked at a possible association with gastrointestinal malignancies. The results did not reveal a consistent positive correlation.^39,40 The risk of cancer from exposure to radon in the public water supply is likely small and mostly from the transfer of radon particles into the air and not from drinking the water. On the other hand, the risk could be higher with private wells, where radon levels are variable and are possibly higher than from public sources.⁴¹

DATA ARE INSUFFICIENT TO GUIDE MANAGEMENT

Radon is a naturally occurring and ubiquitous radioactive gas that can cause tissue damage. Cohort and case-control studies have demonstrated that radon exposure is associated with increased risk of lung cancer. It is recommended that radon levels be measured in every home in the United States and mitigation measures instituted if levels exceed 4 pCi/L.

There are insufficient data to help guide the management of patients with a history of radon exposure, and prospective studies are needed to better understand the individual risk of developing lung cancer and the appropriate management of such patients.

Smoking cessation is an integral part of lung cancer risk reduction from radon exposure.

In 1984, a worker at a Pennsylvania nuclear power plant triggered the radiation detector as he was getting ready to go home. This would not be unusual for such a facility, but there was no nuclear fuel on site at the time. The alarm went off every time he left work.

One day, he triggered the alarm as he crossed the detector on arriving at the plant, leading him to suspect that he was bringing radiation from home. He eventually convinced the plant’s health physicists to check his home, although at first they were opposed to the idea. The results revealed high concentrations of radon everywhere, especially in his basement.

Radon was already known to be associated with health risks in underground miners at that time. This event revealed that a naturally occurring radioactive gas could be found in households at potentially hazardous concentrations.

The incident captured the public’s attention, and the Environmental Protection Agency (EPA) and the US Centers for Disease Control and Prevention (CDC) recommended that nearly all homes be tested.^1,2 In 1988, the International Agency for Research on Cancer classified radon as a human carcinogen, and Congress passed the Indoor Radon Abatement Act in response to growing concern over health risks.³ This law funded state and federal measures to survey schools and federal buildings for radon levels, to educate citizens, and to develop programs for technical assistance. The long-term goal was to reduce indoor levels nationwide to no more than outdoor levels.

Radon is still considered an important public health hazard. From 15,000 to 21,000 people are estimated to die of lung cancer as a result of radon exposure each year in the United States, making it the second most common cause of lung cancer, behind smoking.⁴

Considering the relevance of this issue, this article will review the unique characteristics of radon as a risk factor for lung cancer.

WHAT IS RADON?

Radon is a noble gas that occurs naturally as a decay product of uranium 238 and thorium 232. It is colorless, tasteless, and imperceptible to our senses. Its most common isotope is radon 222 (²²²Rn), which has a half-life of 3.8 days and decays by emitting an alpha particle to become polonium 218. The decay chain continues through several intermediate steps until the stable isotope lead 206 is formed (Figure 1). Two of the isotopes in this chain, polonium 218 and polonium 214, also emit alpha particles.^5–7

Radon is primarily formed in soil. Its most important precursor, uranium 238, is ubiquitous, found in most soils and rocks in various concentrations. Radon can also be found in surface water, metal mines (uranium, phosphorus, silver, gold), residue of coal combustion, and natural gas.

Outdoor levels are usually much lower than indoor levels, as radon dissipates very quickly. Indoor radon mostly comes from the soil under the house or building, but it can also originate from coal combustion, gas appliances, and water (especially from private wells). In municipal water systems or surface reservoirs, most of the radon dissipates into the air or decays before the water reaches homes.^8,9

Radon’s only commercial application in the United States is in calibrating measuring instruments. In the past, it was used in radiography and to treat cancer but was later replaced by other radiation sources that cost less and pose less hazard of alpha radiation.¹⁰

HOW RADON CAN HARM

Alpha particles, emitted by radon 222 and its progenies polonium 218 and polonium 214, are highly effective in damaging tissues. Although they do not travel far or fast, with their two protons and two neutrons, alpha particles are heavy and therefore can cause considerable damage at short range. Although alpha particles can be stopped by a thin barrier such as a piece of paper or the skin, if the source is inhaled or ingested and lodges against mucosal linings, the alpha particles emitted can destroy cells.¹¹

The main route of radon exposure is by inhalation. Since radon is biologically inert, it is readily exhaled after it reaches the lungs. However, radon’s progenies can also be inhaled, either as free particles or attached to airborne particles such as dust, which they tend to attract as a result of their charged state. This attached fraction is believed to be more carcinogenic because it tends to deposit on the respiratory epithelium, notably in the carinae of bronchi. The smaller the dust particle, the deeper it can travel into the lung. The radiation emissions damage the genetic material of cells lining the airways, with the potential to result in lung cancer if the repair process is incomplete.^5,8,9

Other routes of exposure include ingestion and dermal exposure. Radon and its progenies can be swallowed in drinking water, passing through the stomach walls and bowels and entering the blood.¹² Dermal exposure is not considered a significant route unless the dermis is exposed, since in usual circumstances the skin protects the body from alpha radiation.¹³

Possible biologic mechanisms by which radon exposure might increase the risk of cancer include gene mutations, chromosome aberrations, generation of reactive oxygen species, up- or down-regulation of cytokines, and production of proteins associated with cell-cycle regulation.^14–16

HOW IS RADON MEASURED?

Several devices are commercially available to measure radon levels at home. The most common ones are activated charcoal detectors, electret ion chambers, alpha-track detectors, electronic integrating devices, and continuous monitors. There is no evidence that one device is better than another, but devices that measure radon gas are usually preferred over those that measure decay products because they are simpler to use and more cost-effective. These devices are divided into those used for short-term testing (2–90 days) and long-term testing (Table 1).¹⁷

Radon levels can be expressed as follows:

Working levels. One working level (WL) is any combination of radon progeny in 1 L of air that ultimately releases 1.3 × 10⁵ MeV of alpha energy during decay. In studies of miners, the radon progeny concentrations are generally expressed in WL. The cumulative exposure of an individual to this concentration over a “working month” of 170 hours is defined as a working level month (WLM).

Picocuries per liter. In the United States, the rate of decay is commonly reported in picocuries per liter (pCi/L): 1 pCi/L translates to 0.005 WL under usual conditions. The outdoor radon level is normally around 0.4 pCi/L.

Becquerel per cubic meter (Bq/m³) is an International System unit of measure: 1 WL corresponds to 3.7 × 10³ Bq/m³, and 1 pCi/L is equivalent to 37 Bq/m³.

Different areas have different radon levels

The Indoor Radon Abatement Act of 1988 helped identify areas in the United States that have the potential for elevated indoor radon levels. An estimated 6 million homes have concentrations greater than 4 pCi/L.

To assist in implementing radon-reducing strategies and allocation of resources, the EPA has created a map (Figure 2) that classifies counties according to the predicted indoor level.¹⁸

WHAT IS THE RELATIONSHIP BETWEEN RADON AND LUNG CANCER?

Determining the degree to which radon exposure contributes to lung cancer is a complex task. Radon can be found nearly everywhere, and there are diurnal, seasonal, and random year-to-year variations in the concentration of radon in indoor air.

A minority view

Not everyone agrees that radon is completely bad. For centuries, people have flocked to spas to “take the waters,” and the water at many of these spas has been found to contain radon. In the early 20th century, radiation was touted as having medicinal benefits, and people in many places in the world still go to “radon spas” (some of them in abandoned uranium mines) to help treat conditions such as arthritis and to feel invigorated and energized.

In 2006, a report by Zdrojewicz and Strzelczyk¹⁹ urged the medical community to keep an open mind about the possibility that radon exposure may be beneficial in very low doses, perhaps by stimulating repair mechanisms. This concept, called hormesis, differs from the mainstream view that cancer risk rises linearly with radiation dose, with no minimum threshold level (see below).

As early as in the 16th century, metal miners in central Europe were noted to have a high rate of death from respiratory disease. Radon was discovered in 1900, and in the 20th century lung cancer was linked to high levels of radon detected in uranium mines.

Several small studies of underground miners exposed to high concentrations of radon consistently demonstrated an increased risk of lung cancer.

The Committee on the Biological Effects of Ionizing Radiation (BEIR VI 1999) reviewed 11 major cohort studies of miners. The studies included more than 60,000 miners in Europe, North America, Asia, and Australia, of whom 2,600 died of lung cancer. Lung cancer rates increased linearly with cumulative radon exposure, and the estimated average increase in the lung cancer death rate per WLM in the combined studies was 0.44% (95% confidence interval [CI] 0.20–1.00%). The percentage increase in the lung cancer death rate per WLM varied with time since exposure, with the highest increase in risk during the 5 to 14 years after exposure.^4,17 Furthermore, the increase in risk was higher in younger miners, who were exposed to a relatively low radon concentration.

Risk in the general population

The magnitude of the risk in miners led to concern about radon exposure as a cause of lung cancer in the general population. Statistical models were generated that suggested a causal link between radon exposure and lung cancer. Although extrapolation of the results from miners caused controversy, the BEIR VI estimation of risk was validated by studies in the general population.^7,20–23

Since the 1980s, several small case-control studies with limited power examined the relationship between indoor radon and lung cancer in the general population. In these studies, individuals who had developed lung cancer were compared with controls who had not developed the disease but who otherwise represented the population from which the cases of lung cancer came.

To improve the statistical power, the investigators of the major studies in Europe, North America, and China pooled the results in separate analyses (Table 2).^7,20–23 The average radon concentration to which each individual had been exposed over the previous decades was estimated by measuring the radon concentration at their present and previous homes. On the basis of information from the uranium miners, these studies assumed that the period of exposure was the 30 years ending 5 years before the diagnosis or at death from lung cancer.

The results provided convincing evidence that radon exposure is a cause of lung cancer in the general population and substantiated the extrapolation from the studies of miners. Further, the results of all three pooled analyses were consistent with a linear dose-response relationship with no threshold, suggesting an increased risk of lung cancer even with a radon level below 4 pCi/L (200 Bq/m³), which is the concentration at which mitigation actions are recommended in many countries.¹⁷

The North American pooled analysis included 3,662 cases and 4,966 controls from seven studies in the United States and Canada. When data from all studies were combined, the risk of lung cancer was found to increase by 11% per 100-Bq/m³ (about 2.7-pCi/L) increase in measured radon concentration (95% CI 0%–28%). The estimated increase in lung cancer was independent of age, sex, or smoking history.^7,20

The Chinese pooled data²² demonstrated a 13% (95% CI 1%–36%) increased risk per 100 Bq/m³.

In the European study, the risk of lung cancer increased by 8% per 100 Bq/m³ (95% CI 3%–16%). The European investigators repeated the analysis, taking into account the random year-to-year variability in measured radon concentration, finding the final estimated risk was an increase of 16% per 100 Bq/m³ using long-term average concentration.²¹

The combined estimate^21,24 from the three pooling studies based on measured radon concentration is an increased risk of lung cancer of 10% per 100 Bq/m³.

Synergistic risk with smoking

Radon exposure was independently associated with lung cancer, and the relationship with cigarette smoking is believed to be synergistic. The radon progeny particles attach themselves to smoke and dust and are then deposited in the bronchial epithelium.²⁵

In the pool of European case-control studies, the cancer risk for current smokers of 15 to 24 cigarettes per day relative to that in never-smokers was 25.8 (95% CI 21–31). Assuming that in the same analysis the lung cancer risk increased by 16% per 100 Bq/m³ of usual radon concentration regardless of smoking status, the cumulative absolute risk by age 75 would be 0.67% in those who never smoked and 16% in smokers at usual radon levels of 400 Bq/m³ (11 pCi/L).²¹

Rates of all lung cancer subtypes increased

Radon exposure is not associated with a specific histologic subtype of lung cancer. It has been speculated that the incidence of the small-cell subtype might be slightly increased because radon tends to deposit in the more central bronchial carinae.^20,21 However, all subtypes have been described in association with radon, the most common being adenocarcinoma and squamous cell carcinoma.^26–28

EFFECT OF MITIGATION MEASURES

The US Surgeon General and the EPA recommend that all homes be tested.¹⁸ Short-term tests should be used first, keeping in mind that diurnal and seasonal variations may occur.

The World Health Organization has proposed a reference level of 100 Bq/m³ (2.7 pCi/L) to minimize health hazards from indoor radon exposure.¹⁷ If this level cannot be reached under the country-specific conditions, the chosen reference level should not exceed 300 Bq/m³ (8 pCi/L).

In the United States, if the result of home testing is higher than 4 pCi/L, a follow-up measurement should be done using a different short-term test or a long-term test. If the follow-up result confirms a level of more than 4 pCi/L, mitigating actions are recommended. The goal is to reduce the indoor radon level as much as possible—down to zero or at least comparable to outdoor levels (national average 0.4 pCi/L).¹⁸

A variety of radon mitigation strategies have been used, with different rates of efficacy (Table 3). The optimal strategy depends on the likely source or cause, construction characteristics, soil, and climate.²⁹ Table 4 lists resources for the general public about testing and mitigation measures.

How beneficial is radon mitigation?

Although it is logical to try to reduce the indoor radon concentration, there is no strong evidence yet that this intervention decreases the incidence of lung cancer in the general population.

Using the BEIR VI risk model, Méndez et al³⁰ estimated a 21% reduction in the annual radon-related lung cancer mortality rate by 2100 if all households were compliant with government recommendations (mitigation actions at levels of 4 pCi/L) and assuming that the percentage of cigarette smokers remained constant.

On the other hand, if the number of smokers continues to decline, the benefits from radon mitigation may be less. The expected benefit from mitigation in this scenario is a reduction of 12% in annual radon-related deaths by the year 2100.³⁰ However, it will be challenging to determine whether the expected decline in the incidence of lung cancer and lung cancer deaths is truly attributable to mitigation measures.

MANAGING PATIENTS EXPOSED TO RADON

Screen for lung cancer in smokers only

The National Lung Screening Study (NLST) was a large multicenter trial of annual low-dose computed tomography (CT) to screen for lung cancer in a cohort at high risk: age 55 to 74, at least a 30 pack-year history of smoking in a current smoker, or a former smoker who quit within the past 15 years. The trial demonstrated a 20% reduction in lung cancer deaths in the CT screening group.³¹

Since the publication of the NLST results, many societies have endorsed screening for lung cancer with low-dose CT using the study criteria. The National Comprehensive Cancer Network (NCCN) expanded these criteria and has recommended screening in patients over age 50 who have a history of smoking and one additional risk factor, such as radon exposure.

However, radon exposure has not been incorporated into a lung cancer risk-prediction model, and there is no empirical evidence suggesting that people who have such a history would benefit from screening.^32,33 The joint guidelines of the American College of Chest Physicians and American Society of Clinical Oncology recommend annual low-dose CT screening only for patients who meet the NLST criteria.³⁴

What to do about indeterminate lung nodules

The widely used guidelines from the Fleischner Society³⁵ on how to manage small lung nodules stratify patients into groups at low and high risk of developing lung cancer on the basis of risk factors. The guidelines apply to adults age 35 and older in whom an indeterminate solid nodule was recently detected.

If a patient is at high risk, the recommended approach includes follow-up in shorter intervals depending on the nodule size. History of smoking is recognized as a major risk factor, and the statement also lists family history and exposure to asbestos, uranium, and radon.³⁵

Although the association of radon with lung cancer has been shown in epidemiologic studies, radon exposure has not been included in validated statistical models that assess the probability that an indeterminate lung nodule is malignant. We would expect the risk to be higher in miners, who suffer a more intense exposure to higher levels of radon, than in the general population, which has a low and constantly variable residential exposure. Furthermore, there are no data to support a more aggressive follow-up approach in patients with indeterminate lung nodules and a history of radon exposure.

RADON AND OTHER CANCERS

When a person is exposed to radon, the bronchial epithelium receives the highest dose of ionizing radiation, but other organs such as the kidneys, stomach, and bone marrow may receive doses as well, although lower. Several studies have looked into possible associations, but there is no strong evidence to suggest an increased mortality rate related to radon from cancers other than lung.^24,36 However, there seems to be a positive association between radon and the incidence of lymphoproliferative disorders in uranium miners.^37,38

Radon can be measured in drinking water, and a few studies have looked at a possible association with gastrointestinal malignancies. The results did not reveal a consistent positive correlation.^39,40 The risk of cancer from exposure to radon in the public water supply is likely small and mostly from the transfer of radon particles into the air and not from drinking the water. On the other hand, the risk could be higher with private wells, where radon levels are variable and are possibly higher than from public sources.⁴¹

DATA ARE INSUFFICIENT TO GUIDE MANAGEMENT

Radon is a naturally occurring and ubiquitous radioactive gas that can cause tissue damage. Cohort and case-control studies have demonstrated that radon exposure is associated with increased risk of lung cancer. It is recommended that radon levels be measured in every home in the United States and mitigation measures instituted if levels exceed 4 pCi/L.

There are insufficient data to help guide the management of patients with a history of radon exposure, and prospective studies are needed to better understand the individual risk of developing lung cancer and the appropriate management of such patients.

Smoking cessation is an integral part of lung cancer risk reduction from radon exposure.

Eighty percent of people with type 2 diabetes mellitus are obese or overweight.¹ Excess adipose tissue can lead to endocrine dysregulation,² contributing to the pathogenesis of type 2 diabetes, and obesity is one of the strongest predictors of this disease.³

For obese people with type 2 diabetes, diet and exercise can lead to weight loss and many other benefits, such as better glycemic control, less insulin resistance, lower risk of diabetes-related comorbidities and complications, fewer diabetic medications needed, and lower health care costs.^4–7 Intensive lifestyle interventions have also been shown to induce partial remission of diabetes and to prevent the onset of type 2 diabetes in people at high risk of it.^5–7

A very-low-calorie diet is one of many dietary options available to patients with type 2 diabetes who are overweight or obese. The protein-sparing modified fast (PSMF) is a type of very-low-calorie diet with a high protein content and simultaneous restriction of carbohydrate and fat.^8,9 It was developed in the 1970s, and since then various permutations have been used in weight loss and health care clinics worldwide.

MOSTLY PROTEIN, VERY LITTLE CARBOHYDRATE AND FAT

The PSMF is a medically supervised diet that provides less than 800 kcal/day during an initial intensive phase of about 6 months, followed by the gradual reintroduction of calories during a refeeding phase of about 6 to 8 weeks.¹⁰

During the intensive phase, patients obtain most of their calories from protein, approximately 1.2 to 1.5 g/kg of ideal body weight per day. At the same time, carbohydrate intake is restricted to less than 20 to 50 g/day; additional fats outside of protein sources are not allowed.⁹ Thus, the PSMF shares features of both very-low-calorie diets and very-low-carbohydrate ketogenic diets (eg, the Atkins diet), though some differences exist among the three (Figure 1).

Patients rapidly lose weight during the intensive phase, typically between 1 and 3 kg per week, with even greater losses during the first 2 weeks.8,9 Weight loss typically plateaus within 6 months, at which point patients begin the refeeding period. During refeeding, complex carbohydrates and low-glycemic, high-fiber cereals, fruits, vegetables, and fats are gradually reintroduced. Meanwhile, protein intake is reduced to individually tailored amounts as part of a weight-maintenance diet.

LIPOLYSIS, KETOSIS, DIURESIS

Modified from Baker S, et al. Effects and clinical potential of very-low-calorie diets (VLCDS) in type 2 diabetes. Diabetes Res Clin Pract 2009; 85:235–242.

The specific macronutrient composition of the PSMF during the intensive phase is designed so that patients enter ketosis and lose as much fat as they can while preserving lean body mass.^9,11 Figure 2 illustrates the mechanisms of ketosis and the metabolic impact of the PSMF.

With dietary carbohydrate restriction, serum glucose and insulin levels decline and glycogen stores are depleted. The drop in serum insulin allows lipolysis to occur, resulting in loss of adipose tissue and production of ketone bodies in the liver. Ketone bodies become the primary source of energy for the brain and other tissues during fasting and have metabolic and neuroprotective benefits.^12,13

Some studies suggest that ketosis also suppresses appetite, helping curb total caloric intake throughout the diet.¹⁴ Protein itself may increase satiety.¹⁵

Glycogen in the liver is bound to water, so the depletion of glycogen also results in loss of attached water. As a result, diuresis contributes significantly to the initial weight loss within the first 2 weeks on the PSMF.⁹

WHO IS A CANDIDATE FOR THE PSMF?

The PSMF is indicated only for adults with a body mass index (BMI) of at least 30 kg/m² or a BMI of at least 27 kg/m² and at least one comorbidity such as type 2 diabetes, hypertension, dyslipidemia, obstructive sleep apnea, osteoarthritis, or fatty liver.¹² Patients must also be sufficiently committed and motivated to make the intensive dietary and behavioral changes the program calls for.

The PSMF should be considered when more conventional low-calorie approaches to weight loss fail or when patients become discouraged by the slower results seen with traditional diets.⁸ Patients undergoing a PSMF are usually encouraged by the initial period of rapid weight loss, and such diets have lower dropout rates.¹⁶

This diet may also be recommended for obese patients who have poorly controlled type 2 diabetes and growing resistance to medications, to bring down the blood glucose level. Another use is before bariatric surgery to reduce the risk of obesity-related complications.⁸ Patients who regain weight after bariatric surgery may also benefit.

MEAL REPLACEMENTS OR A DIET PLAN?

The PSMF program at Cleveland Clinic is based on modified preparation and selection of conventional foods. Details of the program are described in Table 1. Protein sources must be of high biologic value, containing the right mix of essential amino acids (eg, lean meat, fish, poultry, egg whites).⁹

Some commercially available very-low-calorie diets (eg, OPTIFAST, Medifast) that are advertised as PSMFs consist mainly of meal replacements. In the program at Cleveland Clinic, meal replacements in the form of commercial high-protein shakes or bars can be used occasionally for convenience and to maintain adherence to the diet.

However, preparation of PSMF meals from natural, conventional foods is thought to play an important role in long-term behavior modification and so is strongly encouraged. Patients learn low-fat cooking methods, portion control, and how to make appropriate choices in shopping, eating, and dining out. These lessons are valuable for those who struggle with long-term weight loss. Learning these behaviors through the program may help ease the transition to the weight-maintenance phase and beyond. For some patients, cooking is also a source of enjoyment, as is the sight, smell, and taste of nonliquid foods.¹⁰

In addition, patients appreciate being able to eat the same foods as others in their household, except for omitting high-carbohydrate foods. It has also been reported that patients on a food-based PSMF were significantly less hungry and preoccupied with eating than those on a liquid formula diet.¹⁷

CONTRAINDICATIONS AND SAFETY CONCERNS

Contraindications to the PSMF include a BMI less than 27 kg/m², recent myocardial infarction, angina, significant arrhythmia, decompensated congestive heart failure, cerebrovascular insufficiency or recent stroke, end-stage renal disease, liver failure, malignancy, major psychiatric illness, pregnancy or lactation, and wasting disorders. It is also not recommended for patients under age 16 or over age 65.

In view of the risk of diabetic ketoacidosis and the difficulty of titrating required doses ofinsulin, patients with type 1 diabetes mellitus are usually not advised to undergo a low-carbohydrate or very-low-calorie diet.^8,12 However, we and others have found that the PSMF can be used in some obese patients with type 1 diabetes if it is combined with appropriate education and careful monitoring.¹²

Major concerns about the safety of the PSMF stem from experiences with the first very-low-calorie diets in the 1970s, which were associated with fatal cardiac arrhythmias and sudden death.¹⁸ These early diets used liquid formulas with hydrolyzed collagen protein of poor biologic value and were deficient in many vitamins and minerals. Today’s very-low-calorie diets use protein sources of high biologic value (chiefly animal, soy, and egg for the PSMF) and are supplemented with necessary vitamins and minerals, reducing the risk of electrolyte and cardiac abnormalities.^9,19,20 Furthermore, before starting the PSMF all patients must have an electrocardiogram to be sure they have no arrhythmias (eg, heart block, QT interval prolongation) or ischemia.

Relative contraindications

A known history of cholelithiasis is a relative contraindication to a very-low-calorie diet and may be of concern for some patients and providers. While obesity itself is already a risk factor for gallstones, gallstone formation has also been associated with bile stasis, which occurs from rapid weight loss with liquid formula diets of low fat intake (< 10 g/day).²¹ However, in the PSMF, fat intake from protein sources, though low (45–70 g/day), is considered high enough to allow adequate gallbladder contraction, thus decreasing the risk of gallstone formation.²²

Gout is another relative contraindication, as hyperuricemia with risk of gout is also linked to high-protein diets.⁹ Palgi et al²³ found that uric acid levels rose by a mean of 0.4 mg/dL during the diet. The risk of gout, however, seemed to be small, occurring in fewer than 1% of patients in the study. Furthermore, in a recent study by Li et al,²⁴ uric acid levels were found to significantly decrease in patients on a high-protein, very-low-calorie diet. Nonetheless, uric acid levels should be monitored regularly in patients on the PSMF.

SIDE EFFECTS OF THE DIET

Common side effects of the PSMF include headache, fatigue, orthostatic hypotension, muscle cramps, cold intolerance, constipation, diarrhea, fatigue, halitosis, menstrual changes, and hair thinning. Most of these are transient and may be alleviated by adjusting fluid, salt, and supplement intake. Other side effects may disappear as the patient is weaned off the diet.^8,9

REGULAR FOLLOW-UP WITH HEALTH CARE PROVIDERS

Current PSMF programs are considered safe when used in combination with regular follow-up with health care providers.^8,12

At Cleveland Clinic, patients meet with a dietitian twice in the first month and monthly thereafter (or more frequently if needed) for weight monitoring and education on nutrition and behavior modification (Table 1). Since the PSMF does not provide complete nutrition, daily supplementation with vitamins and minerals is required.

Daily exercise is encouraged throughout the program to increase fitness and to help keep the weight off during the refeeding phase and after.

Patients also meet every 6 to 8 weeks with the referring nurse practitioner or physician for further monitoring and evaluation of vital signs, laboratory results, and side effects. The PSMF protocol at Cleveland Clinic enables both primary care physicians and specialists (including nurse practitioners) within our network to monitor the patient’s status. Use of a common electronic medical record system is particularly valuable for easy communication between providers. If a primary care physician feels unable to appropriately counsel and supervise a patient in the PSMF program, referral to an endocrinologist or weight loss specialist is recommended.

In addition to baseline electrocardiography and monitoring of uric acid levels, a comprehensive metabolic panel is drawn at baseline, twice in the first month, and monthly thereafter to check for electrolyte imbalances and metabolic and tissue dysfunction such as dehydration, excessive protein loss, and liver or kidney injury.

Patients should not attempt the PSMF without medical supervision. Many patients have friends or family members who want to try the PSMF along with them, but this can be dangerous, especially for those with hypertension or type 2 diabetes. The medications prescribed for these conditions can result in hypotension or hypoglycemia during the PSMF.

Although there are no standard guidelines for adjusting medication use before starting a patient on the PSMF, it is logical to taper off or discontinue antihypertensive agents in patients with tightly controlled hypertension to avoid possible dehydration and hypotension during the first few diuresis-inducing weeks of the diet. In particular, diuretic agents should be discontinued to prevent further electrolyte imbalance and fluid shifts.

Similarly, in patients with tightly controlled type 2 diabetes (hemoglobin A_1c < 7.0%), oral hypoglycemic agents and insulin therapy should be reduced before starting the diet to avoid potential hypoglycemia. During the course of the diet, providers should then adjust medication dosages based on follow-up vital signs and laboratory results and daily glucose monitoring.⁸

EFFECTS OF THE PSMF IN PATIENTS WITH TYPE 2 DIABETES

Though few formal studies have been done, the PSMF may have major effects on hyperglycemia, cardiovascular risk factors, and diabetic nephropathy in obese patients with type 2 diabetes, at least in the short term (Table 2).

Weight loss

In one of the first PSMF studies,²³ in 668 patients with or without type 2 diabetes (baseline weight 98 kg), the mean weight loss was 21 kg after the intensive phase and 19 kg by the end of the refeeding phase.

In another observational report,²⁵ 25% to 30% of patients lost even more weight, averaging 38.6 kg of weight loss. Typically, the higher the baseline weight, the greater the weight loss during the PSMF.²³

Patients with type 2 diabetes lost a similar amount of weight (8.5 kg) compared with those without diabetes (9.4 kg, P = .64) in a study of meal-replacement PSMF (using OPTIFAST shakes and bars).²⁶ In a large meal-replacement study of 2,093 patients, Li et al²⁴ found that weight loss was similar between diabetic, prediabetic, and nondiabetic patients. Weight loss was also closely maintained in those patients who stayed on the diet for 12 months.

In a PSMF study in which all the participants had type 2 diabetes, the mean weight loss was 18.6 kg. Although the patients regained some of this weight, at 1 year they still weighed 8.6 kg less than at baseline. However, a conventional, balanced, low-calorie diet resulted in similar amounts of weight loss after 1 year.²⁷ Furthermore, a second round of the PSMF did not result in significant additional weight loss but rather weight maintenance.²⁸

Fat loss and smaller waist circumference

Most of the weight lost during a PSMF is from fat tissue.^11,26 Abdominal (visceral) fat may be lost first, which is desirable for patients with type 2 diabetes, since a higher degree of abdominal fat is linked to insulin resistance.^2,29

After a meal-replacement PSMF, waist circumference decreased significantly in patients both with and without type 2 diabetes.^24,26 However, in one study, less fat was lost per unit of change of BMI in the group with type 2 diabetes than in the nondiabetic group.²⁶ Since insulin inhibits lipolysis, it is possible that exogenous insulin use in diabetic patients may prevent greater reductions in fat mass, though this is likely not the only mechanism.²⁶

Lower fasting serum glucose

Fasting serum glucose levels decreased significantly from baseline in patients with type 2 diabetes after a PSMF in all studies that measured this variable.^{23–28,30,31} Changes in fasting glucose are immediate and are associated with caloric restriction rather than weight loss itself.^30,32 Furthermore, the observed decrease in serum glucose is even more impressive in view of the withdrawal or reduction of doses of insulin and oral hypoglycemic agents before starting the diet.

In a study that compared glycemic control in a PSMF diet vs a balanced low-calorie diet, the fasting serum glucose in the PSMF group declined 46%, from 255.9 mg/dL at baseline to 138.7 mg/dL at 20 weeks (P = .001). After 1 year, it had risen back to 187.4 mg/dL, which was still 27% lower than at baseline (P = .023). These results compared favorably with those in the low-calorie diet group (P < .05), which saw fasting serum glucose decline 27% after 20 weeks (from 230.6 mg/dL at baseline to 167.6 mg/dL) and then rise to 5% over baseline (243.2 mg/dL) after 1 year.²⁷

In a later study, the decrease in fasting serum glucose was not maintained at 1 year, but a significantly higher percentage (55%) of participants in the PSMF group were still able to remain free of diabetic medications compared with those who followed a balanced low-calorie diet (31%, P = .01).²⁸

Decrease in hemoglobin A_1c

Declines in fasting serum glucose corresponded with short-term declines in hemoglobin A_1c in several reports.^27–31 Hemoglobin A_1c declined significantly from an average of 10.4% to 7.3% (P = .001) after PSMF intervention in patients with type 2 diabetes. In contrast, hemoglobin A_1c in the low-calorie diet control group declined from 10.4% to 8.6%.²⁷ One year later, hemoglobin A_1c remained lower than at baseline in the PSMF group (final 9.2%) and continued to compare favorably against the control group (final 11.8%, between-group P = .001). However, these 1-year post-intervention improvements were not seen in a second, more intensive study.²⁸

Less insulin resistance

In several studies, fasting serum insulin levels declined along with serum glucose levels, implying decreased insulin resistance.^{25,27,28,30,31} In addition, insulin output was enhanced during glucose challenge after completion of the PSMF, suggesting possible improved (though still impaired) pancreatic beta-cell capacity.^25,27,30

Improved lipid profile

The most common effect of the PSMF on the lipid profile is a significant decrease in triglycerides in patients both with and without type 2 diabetes.^8,23,24,28 In addition, high-density lipoprotein cholesterol increased in two studies following PSMF intervention or after 1-year of follow-up.^24,27,28 Total cholesterol and low-density lipoprotein cholesterol levels also improved after the PSMF, but these changes were not always maintained at follow-up visits.^8,24,28

Lower blood pressure

Improvements in both systolic and diastolic blood pressure were noted in two studies, with mean decreases of 6 mm Hg to 13 mm Hg systolic and 8 mm Hg diastolic after PSMF intervention.^23,28 In a third study, reductions in blood pressure were less dramatic, and only changes in diastolic but not systolic blood pressure remained significant at 12 months.²⁴ While improvements were not observed in a fourth study, patients in this study also had impaired kidney function caused by diabetic nephropathy, and changes in medication were not taken into account.³¹

Kidney function tests

In a small study, Friedman et al showed that 12 weeks of the PSMF in six patients with advanced diabetic nephropathy (stage 3B or stage 4 chronic kidney disease) led to a loss of 12% of body weight (P = .03) as well as significant reductions in serum creatinine and cystatin C levels (P < .05).³¹ In addition, albuminuria decreased by 30% (P = .08). Side effects were minimal, and the diet was well tolerated despite its high protein content, which is a concern in patients with impaired kidney function.

Thus, weight loss via the PSMF may still be beneficial in type 2 diabetic patients with chronic kidney disease and may even improve the course of progression of diabetic nephropathy.

Long-term weight loss is elusive

Long-term weight loss has been an elusive goal for many diet programs. In a study using a very-low-calorie diet in obese patients with type 2 diabetes, substantial weight loss was maintained in half of the patients at 3 years after the intervention, but nearly all of the patients had regained most of their weight after 5 years.³³

While commitment to behavior modification, maintenance of physical activity, and continued follow-up are all critical factors in sustaining weight loss, new and innovative approaches to battle weight regain are needed.³⁴

Yet despite considerable weight regain in most patients, the Look AHEAD (Action for Health in Diabetes) study showed that participants in intensive lifestyle intervention programs still achieved greater weight loss after 4 years than those receiving standard care.³⁵ Whether this holds true for those in intensive PSMF programs is unknown. In addition, conclusive PSMF studies regarding glycemic control, lipids, and blood pressure beyond 1 year of follow-up are lacking.

A VIABLE OPTION FOR MANY

Adherence to a very-low-calorie, ketogenic PSMF program results in major short-term health benefits for obese patients with type 2 diabetes. These benefits include significant weight loss, often more than 18 kg, within 6 months.^23–28 In addition, significant improvements in fasting glucose^{23–28,30–32} and hemoglobin A_1c levels^27–31 are linked to the caloric and carbohydrate restriction of the PSMF. Insulin resistance was also attenuated, with possible partial restoration of pancreatic beta-cell capacity.^{25,27,28,30,31} In some studies, the PSMF resulted in lower systolic and diastolic blood pressure^23,24,28 and triglyceride levels.^8,23,24,28 One small study also suggested a possible improvement of diabetic nephropathy.³¹ Lastly, improvements in glycemia and hypertension were associated with a reduction in the need for antidiabetic and antihypertensive drugs.³⁶

Still, weight loss and many of the associated improvements partially return to baseline levels 1 year after the intervention. Thus, more long-term studies are needed to explore factors for better weight maintenance after the PSMF.

Also, only a few studies have compared the effect of the PSMF between patients with or without type 2 diabetes. One study suggested that fat loss may be reduced in patients with type 2 diabetes.²⁶

In conclusion, despite some risks and safety concerns, PSMF is a viable option for many obese, type 2 diabetic patients as a method of short-term weight loss, with evidence for improvement of glycemic control and cardiovascular risk factors for up to 1 year. To strengthen support for the PSMF, however, further research is warranted on the diet’s long-term effects in patients with type 2 diabetes and also in nondiabetic patients.
 

Acknowledgments: Many thanks to Cheryl Reitz, RD, LD, CDE, and Dawn Noe, RD, LD, CDE, for providing their expertise on the PSMF protocols carried out at Cleveland Clinic. Additional thanks to Tejas Kashyap for his initial assistance with this review.

Eighty percent of people with type 2 diabetes mellitus are obese or overweight.¹ Excess adipose tissue can lead to endocrine dysregulation,² contributing to the pathogenesis of type 2 diabetes, and obesity is one of the strongest predictors of this disease.³

For obese people with type 2 diabetes, diet and exercise can lead to weight loss and many other benefits, such as better glycemic control, less insulin resistance, lower risk of diabetes-related comorbidities and complications, fewer diabetic medications needed, and lower health care costs.^4–7 Intensive lifestyle interventions have also been shown to induce partial remission of diabetes and to prevent the onset of type 2 diabetes in people at high risk of it.^5–7

A very-low-calorie diet is one of many dietary options available to patients with type 2 diabetes who are overweight or obese. The protein-sparing modified fast (PSMF) is a type of very-low-calorie diet with a high protein content and simultaneous restriction of carbohydrate and fat.^8,9 It was developed in the 1970s, and since then various permutations have been used in weight loss and health care clinics worldwide.

MOSTLY PROTEIN, VERY LITTLE CARBOHYDRATE AND FAT

The PSMF is a medically supervised diet that provides less than 800 kcal/day during an initial intensive phase of about 6 months, followed by the gradual reintroduction of calories during a refeeding phase of about 6 to 8 weeks.¹⁰

During the intensive phase, patients obtain most of their calories from protein, approximately 1.2 to 1.5 g/kg of ideal body weight per day. At the same time, carbohydrate intake is restricted to less than 20 to 50 g/day; additional fats outside of protein sources are not allowed.⁹ Thus, the PSMF shares features of both very-low-calorie diets and very-low-carbohydrate ketogenic diets (eg, the Atkins diet), though some differences exist among the three (Figure 1).

Patients rapidly lose weight during the intensive phase, typically between 1 and 3 kg per week, with even greater losses during the first 2 weeks.8,9 Weight loss typically plateaus within 6 months, at which point patients begin the refeeding period. During refeeding, complex carbohydrates and low-glycemic, high-fiber cereals, fruits, vegetables, and fats are gradually reintroduced. Meanwhile, protein intake is reduced to individually tailored amounts as part of a weight-maintenance diet.

LIPOLYSIS, KETOSIS, DIURESIS

Modified from Baker S, et al. Effects and clinical potential of very-low-calorie diets (VLCDS) in type 2 diabetes. Diabetes Res Clin Pract 2009; 85:235–242.

The specific macronutrient composition of the PSMF during the intensive phase is designed so that patients enter ketosis and lose as much fat as they can while preserving lean body mass.^9,11 Figure 2 illustrates the mechanisms of ketosis and the metabolic impact of the PSMF.

With dietary carbohydrate restriction, serum glucose and insulin levels decline and glycogen stores are depleted. The drop in serum insulin allows lipolysis to occur, resulting in loss of adipose tissue and production of ketone bodies in the liver. Ketone bodies become the primary source of energy for the brain and other tissues during fasting and have metabolic and neuroprotective benefits.^12,13

Some studies suggest that ketosis also suppresses appetite, helping curb total caloric intake throughout the diet.¹⁴ Protein itself may increase satiety.¹⁵

Glycogen in the liver is bound to water, so the depletion of glycogen also results in loss of attached water. As a result, diuresis contributes significantly to the initial weight loss within the first 2 weeks on the PSMF.⁹

WHO IS A CANDIDATE FOR THE PSMF?

The PSMF is indicated only for adults with a body mass index (BMI) of at least 30 kg/m² or a BMI of at least 27 kg/m² and at least one comorbidity such as type 2 diabetes, hypertension, dyslipidemia, obstructive sleep apnea, osteoarthritis, or fatty liver.¹² Patients must also be sufficiently committed and motivated to make the intensive dietary and behavioral changes the program calls for.

The PSMF should be considered when more conventional low-calorie approaches to weight loss fail or when patients become discouraged by the slower results seen with traditional diets.⁸ Patients undergoing a PSMF are usually encouraged by the initial period of rapid weight loss, and such diets have lower dropout rates.¹⁶

This diet may also be recommended for obese patients who have poorly controlled type 2 diabetes and growing resistance to medications, to bring down the blood glucose level. Another use is before bariatric surgery to reduce the risk of obesity-related complications.⁸ Patients who regain weight after bariatric surgery may also benefit.

MEAL REPLACEMENTS OR A DIET PLAN?

The PSMF program at Cleveland Clinic is based on modified preparation and selection of conventional foods. Details of the program are described in Table 1. Protein sources must be of high biologic value, containing the right mix of essential amino acids (eg, lean meat, fish, poultry, egg whites).⁹

Some commercially available very-low-calorie diets (eg, OPTIFAST, Medifast) that are advertised as PSMFs consist mainly of meal replacements. In the program at Cleveland Clinic, meal replacements in the form of commercial high-protein shakes or bars can be used occasionally for convenience and to maintain adherence to the diet.

However, preparation of PSMF meals from natural, conventional foods is thought to play an important role in long-term behavior modification and so is strongly encouraged. Patients learn low-fat cooking methods, portion control, and how to make appropriate choices in shopping, eating, and dining out. These lessons are valuable for those who struggle with long-term weight loss. Learning these behaviors through the program may help ease the transition to the weight-maintenance phase and beyond. For some patients, cooking is also a source of enjoyment, as is the sight, smell, and taste of nonliquid foods.¹⁰

In addition, patients appreciate being able to eat the same foods as others in their household, except for omitting high-carbohydrate foods. It has also been reported that patients on a food-based PSMF were significantly less hungry and preoccupied with eating than those on a liquid formula diet.¹⁷

CONTRAINDICATIONS AND SAFETY CONCERNS

Contraindications to the PSMF include a BMI less than 27 kg/m², recent myocardial infarction, angina, significant arrhythmia, decompensated congestive heart failure, cerebrovascular insufficiency or recent stroke, end-stage renal disease, liver failure, malignancy, major psychiatric illness, pregnancy or lactation, and wasting disorders. It is also not recommended for patients under age 16 or over age 65.

In view of the risk of diabetic ketoacidosis and the difficulty of titrating required doses ofinsulin, patients with type 1 diabetes mellitus are usually not advised to undergo a low-carbohydrate or very-low-calorie diet.^8,12 However, we and others have found that the PSMF can be used in some obese patients with type 1 diabetes if it is combined with appropriate education and careful monitoring.¹²

Major concerns about the safety of the PSMF stem from experiences with the first very-low-calorie diets in the 1970s, which were associated with fatal cardiac arrhythmias and sudden death.¹⁸ These early diets used liquid formulas with hydrolyzed collagen protein of poor biologic value and were deficient in many vitamins and minerals. Today’s very-low-calorie diets use protein sources of high biologic value (chiefly animal, soy, and egg for the PSMF) and are supplemented with necessary vitamins and minerals, reducing the risk of electrolyte and cardiac abnormalities.^9,19,20 Furthermore, before starting the PSMF all patients must have an electrocardiogram to be sure they have no arrhythmias (eg, heart block, QT interval prolongation) or ischemia.

Relative contraindications

A known history of cholelithiasis is a relative contraindication to a very-low-calorie diet and may be of concern for some patients and providers. While obesity itself is already a risk factor for gallstones, gallstone formation has also been associated with bile stasis, which occurs from rapid weight loss with liquid formula diets of low fat intake (< 10 g/day).²¹ However, in the PSMF, fat intake from protein sources, though low (45–70 g/day), is considered high enough to allow adequate gallbladder contraction, thus decreasing the risk of gallstone formation.²²

Gout is another relative contraindication, as hyperuricemia with risk of gout is also linked to high-protein diets.⁹ Palgi et al²³ found that uric acid levels rose by a mean of 0.4 mg/dL during the diet. The risk of gout, however, seemed to be small, occurring in fewer than 1% of patients in the study. Furthermore, in a recent study by Li et al,²⁴ uric acid levels were found to significantly decrease in patients on a high-protein, very-low-calorie diet. Nonetheless, uric acid levels should be monitored regularly in patients on the PSMF.

SIDE EFFECTS OF THE DIET

Common side effects of the PSMF include headache, fatigue, orthostatic hypotension, muscle cramps, cold intolerance, constipation, diarrhea, fatigue, halitosis, menstrual changes, and hair thinning. Most of these are transient and may be alleviated by adjusting fluid, salt, and supplement intake. Other side effects may disappear as the patient is weaned off the diet.^8,9

REGULAR FOLLOW-UP WITH HEALTH CARE PROVIDERS

Current PSMF programs are considered safe when used in combination with regular follow-up with health care providers.^8,12

At Cleveland Clinic, patients meet with a dietitian twice in the first month and monthly thereafter (or more frequently if needed) for weight monitoring and education on nutrition and behavior modification (Table 1). Since the PSMF does not provide complete nutrition, daily supplementation with vitamins and minerals is required.

Daily exercise is encouraged throughout the program to increase fitness and to help keep the weight off during the refeeding phase and after.

Patients also meet every 6 to 8 weeks with the referring nurse practitioner or physician for further monitoring and evaluation of vital signs, laboratory results, and side effects. The PSMF protocol at Cleveland Clinic enables both primary care physicians and specialists (including nurse practitioners) within our network to monitor the patient’s status. Use of a common electronic medical record system is particularly valuable for easy communication between providers. If a primary care physician feels unable to appropriately counsel and supervise a patient in the PSMF program, referral to an endocrinologist or weight loss specialist is recommended.

In addition to baseline electrocardiography and monitoring of uric acid levels, a comprehensive metabolic panel is drawn at baseline, twice in the first month, and monthly thereafter to check for electrolyte imbalances and metabolic and tissue dysfunction such as dehydration, excessive protein loss, and liver or kidney injury.

Patients should not attempt the PSMF without medical supervision. Many patients have friends or family members who want to try the PSMF along with them, but this can be dangerous, especially for those with hypertension or type 2 diabetes. The medications prescribed for these conditions can result in hypotension or hypoglycemia during the PSMF.

Although there are no standard guidelines for adjusting medication use before starting a patient on the PSMF, it is logical to taper off or discontinue antihypertensive agents in patients with tightly controlled hypertension to avoid possible dehydration and hypotension during the first few diuresis-inducing weeks of the diet. In particular, diuretic agents should be discontinued to prevent further electrolyte imbalance and fluid shifts.

Similarly, in patients with tightly controlled type 2 diabetes (hemoglobin A_1c < 7.0%), oral hypoglycemic agents and insulin therapy should be reduced before starting the diet to avoid potential hypoglycemia. During the course of the diet, providers should then adjust medication dosages based on follow-up vital signs and laboratory results and daily glucose monitoring.⁸

EFFECTS OF THE PSMF IN PATIENTS WITH TYPE 2 DIABETES

Though few formal studies have been done, the PSMF may have major effects on hyperglycemia, cardiovascular risk factors, and diabetic nephropathy in obese patients with type 2 diabetes, at least in the short term (Table 2).

Weight loss

In one of the first PSMF studies,²³ in 668 patients with or without type 2 diabetes (baseline weight 98 kg), the mean weight loss was 21 kg after the intensive phase and 19 kg by the end of the refeeding phase.

In another observational report,²⁵ 25% to 30% of patients lost even more weight, averaging 38.6 kg of weight loss. Typically, the higher the baseline weight, the greater the weight loss during the PSMF.²³

Patients with type 2 diabetes lost a similar amount of weight (8.5 kg) compared with those without diabetes (9.4 kg, P = .64) in a study of meal-replacement PSMF (using OPTIFAST shakes and bars).²⁶ In a large meal-replacement study of 2,093 patients, Li et al²⁴ found that weight loss was similar between diabetic, prediabetic, and nondiabetic patients. Weight loss was also closely maintained in those patients who stayed on the diet for 12 months.

In a PSMF study in which all the participants had type 2 diabetes, the mean weight loss was 18.6 kg. Although the patients regained some of this weight, at 1 year they still weighed 8.6 kg less than at baseline. However, a conventional, balanced, low-calorie diet resulted in similar amounts of weight loss after 1 year.²⁷ Furthermore, a second round of the PSMF did not result in significant additional weight loss but rather weight maintenance.²⁸

Fat loss and smaller waist circumference

Most of the weight lost during a PSMF is from fat tissue.^11,26 Abdominal (visceral) fat may be lost first, which is desirable for patients with type 2 diabetes, since a higher degree of abdominal fat is linked to insulin resistance.^2,29

After a meal-replacement PSMF, waist circumference decreased significantly in patients both with and without type 2 diabetes.^24,26 However, in one study, less fat was lost per unit of change of BMI in the group with type 2 diabetes than in the nondiabetic group.²⁶ Since insulin inhibits lipolysis, it is possible that exogenous insulin use in diabetic patients may prevent greater reductions in fat mass, though this is likely not the only mechanism.²⁶

Lower fasting serum glucose

Fasting serum glucose levels decreased significantly from baseline in patients with type 2 diabetes after a PSMF in all studies that measured this variable.^{23–28,30,31} Changes in fasting glucose are immediate and are associated with caloric restriction rather than weight loss itself.^30,32 Furthermore, the observed decrease in serum glucose is even more impressive in view of the withdrawal or reduction of doses of insulin and oral hypoglycemic agents before starting the diet.

In a study that compared glycemic control in a PSMF diet vs a balanced low-calorie diet, the fasting serum glucose in the PSMF group declined 46%, from 255.9 mg/dL at baseline to 138.7 mg/dL at 20 weeks (P = .001). After 1 year, it had risen back to 187.4 mg/dL, which was still 27% lower than at baseline (P = .023). These results compared favorably with those in the low-calorie diet group (P < .05), which saw fasting serum glucose decline 27% after 20 weeks (from 230.6 mg/dL at baseline to 167.6 mg/dL) and then rise to 5% over baseline (243.2 mg/dL) after 1 year.²⁷

In a later study, the decrease in fasting serum glucose was not maintained at 1 year, but a significantly higher percentage (55%) of participants in the PSMF group were still able to remain free of diabetic medications compared with those who followed a balanced low-calorie diet (31%, P = .01).²⁸

Decrease in hemoglobin A_1c

Declines in fasting serum glucose corresponded with short-term declines in hemoglobin A_1c in several reports.^27–31 Hemoglobin A_1c declined significantly from an average of 10.4% to 7.3% (P = .001) after PSMF intervention in patients with type 2 diabetes. In contrast, hemoglobin A_1c in the low-calorie diet control group declined from 10.4% to 8.6%.²⁷ One year later, hemoglobin A_1c remained lower than at baseline in the PSMF group (final 9.2%) and continued to compare favorably against the control group (final 11.8%, between-group P = .001). However, these 1-year post-intervention improvements were not seen in a second, more intensive study.²⁸

Less insulin resistance

In several studies, fasting serum insulin levels declined along with serum glucose levels, implying decreased insulin resistance.^{25,27,28,30,31} In addition, insulin output was enhanced during glucose challenge after completion of the PSMF, suggesting possible improved (though still impaired) pancreatic beta-cell capacity.^25,27,30

Improved lipid profile

The most common effect of the PSMF on the lipid profile is a significant decrease in triglycerides in patients both with and without type 2 diabetes.^8,23,24,28 In addition, high-density lipoprotein cholesterol increased in two studies following PSMF intervention or after 1-year of follow-up.^24,27,28 Total cholesterol and low-density lipoprotein cholesterol levels also improved after the PSMF, but these changes were not always maintained at follow-up visits.^8,24,28

Lower blood pressure

Improvements in both systolic and diastolic blood pressure were noted in two studies, with mean decreases of 6 mm Hg to 13 mm Hg systolic and 8 mm Hg diastolic after PSMF intervention.^23,28 In a third study, reductions in blood pressure were less dramatic, and only changes in diastolic but not systolic blood pressure remained significant at 12 months.²⁴ While improvements were not observed in a fourth study, patients in this study also had impaired kidney function caused by diabetic nephropathy, and changes in medication were not taken into account.³¹

Kidney function tests

In a small study, Friedman et al showed that 12 weeks of the PSMF in six patients with advanced diabetic nephropathy (stage 3B or stage 4 chronic kidney disease) led to a loss of 12% of body weight (P = .03) as well as significant reductions in serum creatinine and cystatin C levels (P < .05).³¹ In addition, albuminuria decreased by 30% (P = .08). Side effects were minimal, and the diet was well tolerated despite its high protein content, which is a concern in patients with impaired kidney function.

Thus, weight loss via the PSMF may still be beneficial in type 2 diabetic patients with chronic kidney disease and may even improve the course of progression of diabetic nephropathy.

Long-term weight loss is elusive

Long-term weight loss has been an elusive goal for many diet programs. In a study using a very-low-calorie diet in obese patients with type 2 diabetes, substantial weight loss was maintained in half of the patients at 3 years after the intervention, but nearly all of the patients had regained most of their weight after 5 years.³³

While commitment to behavior modification, maintenance of physical activity, and continued follow-up are all critical factors in sustaining weight loss, new and innovative approaches to battle weight regain are needed.³⁴

Yet despite considerable weight regain in most patients, the Look AHEAD (Action for Health in Diabetes) study showed that participants in intensive lifestyle intervention programs still achieved greater weight loss after 4 years than those receiving standard care.³⁵ Whether this holds true for those in intensive PSMF programs is unknown. In addition, conclusive PSMF studies regarding glycemic control, lipids, and blood pressure beyond 1 year of follow-up are lacking.

A VIABLE OPTION FOR MANY

Adherence to a very-low-calorie, ketogenic PSMF program results in major short-term health benefits for obese patients with type 2 diabetes. These benefits include significant weight loss, often more than 18 kg, within 6 months.^23–28 In addition, significant improvements in fasting glucose^{23–28,30–32} and hemoglobin A_1c levels^27–31 are linked to the caloric and carbohydrate restriction of the PSMF. Insulin resistance was also attenuated, with possible partial restoration of pancreatic beta-cell capacity.^{25,27,28,30,31} In some studies, the PSMF resulted in lower systolic and diastolic blood pressure^23,24,28 and triglyceride levels.^8,23,24,28 One small study also suggested a possible improvement of diabetic nephropathy.³¹ Lastly, improvements in glycemia and hypertension were associated with a reduction in the need for antidiabetic and antihypertensive drugs.³⁶

Still, weight loss and many of the associated improvements partially return to baseline levels 1 year after the intervention. Thus, more long-term studies are needed to explore factors for better weight maintenance after the PSMF.

Also, only a few studies have compared the effect of the PSMF between patients with or without type 2 diabetes. One study suggested that fat loss may be reduced in patients with type 2 diabetes.²⁶

In conclusion, despite some risks and safety concerns, PSMF is a viable option for many obese, type 2 diabetic patients as a method of short-term weight loss, with evidence for improvement of glycemic control and cardiovascular risk factors for up to 1 year. To strengthen support for the PSMF, however, further research is warranted on the diet’s long-term effects in patients with type 2 diabetes and also in nondiabetic patients.
 

Acknowledgments: Many thanks to Cheryl Reitz, RD, LD, CDE, and Dawn Noe, RD, LD, CDE, for providing their expertise on the PSMF protocols carried out at Cleveland Clinic. Additional thanks to Tejas Kashyap for his initial assistance with this review.

Eighty percent of people with type 2 diabetes mellitus are obese or overweight.¹ Excess adipose tissue can lead to endocrine dysregulation,² contributing to the pathogenesis of type 2 diabetes, and obesity is one of the strongest predictors of this disease.³

For obese people with type 2 diabetes, diet and exercise can lead to weight loss and many other benefits, such as better glycemic control, less insulin resistance, lower risk of diabetes-related comorbidities and complications, fewer diabetic medications needed, and lower health care costs.^4–7 Intensive lifestyle interventions have also been shown to induce partial remission of diabetes and to prevent the onset of type 2 diabetes in people at high risk of it.^5–7

A very-low-calorie diet is one of many dietary options available to patients with type 2 diabetes who are overweight or obese. The protein-sparing modified fast (PSMF) is a type of very-low-calorie diet with a high protein content and simultaneous restriction of carbohydrate and fat.^8,9 It was developed in the 1970s, and since then various permutations have been used in weight loss and health care clinics worldwide.

MOSTLY PROTEIN, VERY LITTLE CARBOHYDRATE AND FAT

The PSMF is a medically supervised diet that provides less than 800 kcal/day during an initial intensive phase of about 6 months, followed by the gradual reintroduction of calories during a refeeding phase of about 6 to 8 weeks.¹⁰

During the intensive phase, patients obtain most of their calories from protein, approximately 1.2 to 1.5 g/kg of ideal body weight per day. At the same time, carbohydrate intake is restricted to less than 20 to 50 g/day; additional fats outside of protein sources are not allowed.⁹ Thus, the PSMF shares features of both very-low-calorie diets and very-low-carbohydrate ketogenic diets (eg, the Atkins diet), though some differences exist among the three (Figure 1).

Patients rapidly lose weight during the intensive phase, typically between 1 and 3 kg per week, with even greater losses during the first 2 weeks.8,9 Weight loss typically plateaus within 6 months, at which point patients begin the refeeding period. During refeeding, complex carbohydrates and low-glycemic, high-fiber cereals, fruits, vegetables, and fats are gradually reintroduced. Meanwhile, protein intake is reduced to individually tailored amounts as part of a weight-maintenance diet.

LIPOLYSIS, KETOSIS, DIURESIS

Modified from Baker S, et al. Effects and clinical potential of very-low-calorie diets (VLCDS) in type 2 diabetes. Diabetes Res Clin Pract 2009; 85:235–242.

The specific macronutrient composition of the PSMF during the intensive phase is designed so that patients enter ketosis and lose as much fat as they can while preserving lean body mass.^9,11 Figure 2 illustrates the mechanisms of ketosis and the metabolic impact of the PSMF.

With dietary carbohydrate restriction, serum glucose and insulin levels decline and glycogen stores are depleted. The drop in serum insulin allows lipolysis to occur, resulting in loss of adipose tissue and production of ketone bodies in the liver. Ketone bodies become the primary source of energy for the brain and other tissues during fasting and have metabolic and neuroprotective benefits.^12,13

Some studies suggest that ketosis also suppresses appetite, helping curb total caloric intake throughout the diet.¹⁴ Protein itself may increase satiety.¹⁵

Glycogen in the liver is bound to water, so the depletion of glycogen also results in loss of attached water. As a result, diuresis contributes significantly to the initial weight loss within the first 2 weeks on the PSMF.⁹

WHO IS A CANDIDATE FOR THE PSMF?

The PSMF is indicated only for adults with a body mass index (BMI) of at least 30 kg/m² or a BMI of at least 27 kg/m² and at least one comorbidity such as type 2 diabetes, hypertension, dyslipidemia, obstructive sleep apnea, osteoarthritis, or fatty liver.¹² Patients must also be sufficiently committed and motivated to make the intensive dietary and behavioral changes the program calls for.

The PSMF should be considered when more conventional low-calorie approaches to weight loss fail or when patients become discouraged by the slower results seen with traditional diets.⁸ Patients undergoing a PSMF are usually encouraged by the initial period of rapid weight loss, and such diets have lower dropout rates.¹⁶

This diet may also be recommended for obese patients who have poorly controlled type 2 diabetes and growing resistance to medications, to bring down the blood glucose level. Another use is before bariatric surgery to reduce the risk of obesity-related complications.⁸ Patients who regain weight after bariatric surgery may also benefit.

MEAL REPLACEMENTS OR A DIET PLAN?

The PSMF program at Cleveland Clinic is based on modified preparation and selection of conventional foods. Details of the program are described in Table 1. Protein sources must be of high biologic value, containing the right mix of essential amino acids (eg, lean meat, fish, poultry, egg whites).⁹

Some commercially available very-low-calorie diets (eg, OPTIFAST, Medifast) that are advertised as PSMFs consist mainly of meal replacements. In the program at Cleveland Clinic, meal replacements in the form of commercial high-protein shakes or bars can be used occasionally for convenience and to maintain adherence to the diet.

However, preparation of PSMF meals from natural, conventional foods is thought to play an important role in long-term behavior modification and so is strongly encouraged. Patients learn low-fat cooking methods, portion control, and how to make appropriate choices in shopping, eating, and dining out. These lessons are valuable for those who struggle with long-term weight loss. Learning these behaviors through the program may help ease the transition to the weight-maintenance phase and beyond. For some patients, cooking is also a source of enjoyment, as is the sight, smell, and taste of nonliquid foods.¹⁰

In addition, patients appreciate being able to eat the same foods as others in their household, except for omitting high-carbohydrate foods. It has also been reported that patients on a food-based PSMF were significantly less hungry and preoccupied with eating than those on a liquid formula diet.¹⁷

CONTRAINDICATIONS AND SAFETY CONCERNS

Contraindications to the PSMF include a BMI less than 27 kg/m², recent myocardial infarction, angina, significant arrhythmia, decompensated congestive heart failure, cerebrovascular insufficiency or recent stroke, end-stage renal disease, liver failure, malignancy, major psychiatric illness, pregnancy or lactation, and wasting disorders. It is also not recommended for patients under age 16 or over age 65.

In view of the risk of diabetic ketoacidosis and the difficulty of titrating required doses ofinsulin, patients with type 1 diabetes mellitus are usually not advised to undergo a low-carbohydrate or very-low-calorie diet.^8,12 However, we and others have found that the PSMF can be used in some obese patients with type 1 diabetes if it is combined with appropriate education and careful monitoring.¹²

Major concerns about the safety of the PSMF stem from experiences with the first very-low-calorie diets in the 1970s, which were associated with fatal cardiac arrhythmias and sudden death.¹⁸ These early diets used liquid formulas with hydrolyzed collagen protein of poor biologic value and were deficient in many vitamins and minerals. Today’s very-low-calorie diets use protein sources of high biologic value (chiefly animal, soy, and egg for the PSMF) and are supplemented with necessary vitamins and minerals, reducing the risk of electrolyte and cardiac abnormalities.^9,19,20 Furthermore, before starting the PSMF all patients must have an electrocardiogram to be sure they have no arrhythmias (eg, heart block, QT interval prolongation) or ischemia.

Relative contraindications

A known history of cholelithiasis is a relative contraindication to a very-low-calorie diet and may be of concern for some patients and providers. While obesity itself is already a risk factor for gallstones, gallstone formation has also been associated with bile stasis, which occurs from rapid weight loss with liquid formula diets of low fat intake (< 10 g/day).²¹ However, in the PSMF, fat intake from protein sources, though low (45–70 g/day), is considered high enough to allow adequate gallbladder contraction, thus decreasing the risk of gallstone formation.²²

Gout is another relative contraindication, as hyperuricemia with risk of gout is also linked to high-protein diets.⁹ Palgi et al²³ found that uric acid levels rose by a mean of 0.4 mg/dL during the diet. The risk of gout, however, seemed to be small, occurring in fewer than 1% of patients in the study. Furthermore, in a recent study by Li et al,²⁴ uric acid levels were found to significantly decrease in patients on a high-protein, very-low-calorie diet. Nonetheless, uric acid levels should be monitored regularly in patients on the PSMF.

SIDE EFFECTS OF THE DIET

Common side effects of the PSMF include headache, fatigue, orthostatic hypotension, muscle cramps, cold intolerance, constipation, diarrhea, fatigue, halitosis, menstrual changes, and hair thinning. Most of these are transient and may be alleviated by adjusting fluid, salt, and supplement intake. Other side effects may disappear as the patient is weaned off the diet.^8,9

REGULAR FOLLOW-UP WITH HEALTH CARE PROVIDERS

Current PSMF programs are considered safe when used in combination with regular follow-up with health care providers.^8,12

At Cleveland Clinic, patients meet with a dietitian twice in the first month and monthly thereafter (or more frequently if needed) for weight monitoring and education on nutrition and behavior modification (Table 1). Since the PSMF does not provide complete nutrition, daily supplementation with vitamins and minerals is required.

Daily exercise is encouraged throughout the program to increase fitness and to help keep the weight off during the refeeding phase and after.

Patients also meet every 6 to 8 weeks with the referring nurse practitioner or physician for further monitoring and evaluation of vital signs, laboratory results, and side effects. The PSMF protocol at Cleveland Clinic enables both primary care physicians and specialists (including nurse practitioners) within our network to monitor the patient’s status. Use of a common electronic medical record system is particularly valuable for easy communication between providers. If a primary care physician feels unable to appropriately counsel and supervise a patient in the PSMF program, referral to an endocrinologist or weight loss specialist is recommended.

In addition to baseline electrocardiography and monitoring of uric acid levels, a comprehensive metabolic panel is drawn at baseline, twice in the first month, and monthly thereafter to check for electrolyte imbalances and metabolic and tissue dysfunction such as dehydration, excessive protein loss, and liver or kidney injury.

Patients should not attempt the PSMF without medical supervision. Many patients have friends or family members who want to try the PSMF along with them, but this can be dangerous, especially for those with hypertension or type 2 diabetes. The medications prescribed for these conditions can result in hypotension or hypoglycemia during the PSMF.

Although there are no standard guidelines for adjusting medication use before starting a patient on the PSMF, it is logical to taper off or discontinue antihypertensive agents in patients with tightly controlled hypertension to avoid possible dehydration and hypotension during the first few diuresis-inducing weeks of the diet. In particular, diuretic agents should be discontinued to prevent further electrolyte imbalance and fluid shifts.

Similarly, in patients with tightly controlled type 2 diabetes (hemoglobin A_1c < 7.0%), oral hypoglycemic agents and insulin therapy should be reduced before starting the diet to avoid potential hypoglycemia. During the course of the diet, providers should then adjust medication dosages based on follow-up vital signs and laboratory results and daily glucose monitoring.⁸

EFFECTS OF THE PSMF IN PATIENTS WITH TYPE 2 DIABETES

Though few formal studies have been done, the PSMF may have major effects on hyperglycemia, cardiovascular risk factors, and diabetic nephropathy in obese patients with type 2 diabetes, at least in the short term (Table 2).

Weight loss

In one of the first PSMF studies,²³ in 668 patients with or without type 2 diabetes (baseline weight 98 kg), the mean weight loss was 21 kg after the intensive phase and 19 kg by the end of the refeeding phase.

In another observational report,²⁵ 25% to 30% of patients lost even more weight, averaging 38.6 kg of weight loss. Typically, the higher the baseline weight, the greater the weight loss during the PSMF.²³

Patients with type 2 diabetes lost a similar amount of weight (8.5 kg) compared with those without diabetes (9.4 kg, P = .64) in a study of meal-replacement PSMF (using OPTIFAST shakes and bars).²⁶ In a large meal-replacement study of 2,093 patients, Li et al²⁴ found that weight loss was similar between diabetic, prediabetic, and nondiabetic patients. Weight loss was also closely maintained in those patients who stayed on the diet for 12 months.

In a PSMF study in which all the participants had type 2 diabetes, the mean weight loss was 18.6 kg. Although the patients regained some of this weight, at 1 year they still weighed 8.6 kg less than at baseline. However, a conventional, balanced, low-calorie diet resulted in similar amounts of weight loss after 1 year.²⁷ Furthermore, a second round of the PSMF did not result in significant additional weight loss but rather weight maintenance.²⁸

Fat loss and smaller waist circumference

Most of the weight lost during a PSMF is from fat tissue.^11,26 Abdominal (visceral) fat may be lost first, which is desirable for patients with type 2 diabetes, since a higher degree of abdominal fat is linked to insulin resistance.^2,29

After a meal-replacement PSMF, waist circumference decreased significantly in patients both with and without type 2 diabetes.^24,26 However, in one study, less fat was lost per unit of change of BMI in the group with type 2 diabetes than in the nondiabetic group.²⁶ Since insulin inhibits lipolysis, it is possible that exogenous insulin use in diabetic patients may prevent greater reductions in fat mass, though this is likely not the only mechanism.²⁶

Lower fasting serum glucose

Fasting serum glucose levels decreased significantly from baseline in patients with type 2 diabetes after a PSMF in all studies that measured this variable.^{23–28,30,31} Changes in fasting glucose are immediate and are associated with caloric restriction rather than weight loss itself.^30,32 Furthermore, the observed decrease in serum glucose is even more impressive in view of the withdrawal or reduction of doses of insulin and oral hypoglycemic agents before starting the diet.

In a study that compared glycemic control in a PSMF diet vs a balanced low-calorie diet, the fasting serum glucose in the PSMF group declined 46%, from 255.9 mg/dL at baseline to 138.7 mg/dL at 20 weeks (P = .001). After 1 year, it had risen back to 187.4 mg/dL, which was still 27% lower than at baseline (P = .023). These results compared favorably with those in the low-calorie diet group (P < .05), which saw fasting serum glucose decline 27% after 20 weeks (from 230.6 mg/dL at baseline to 167.6 mg/dL) and then rise to 5% over baseline (243.2 mg/dL) after 1 year.²⁷

In a later study, the decrease in fasting serum glucose was not maintained at 1 year, but a significantly higher percentage (55%) of participants in the PSMF group were still able to remain free of diabetic medications compared with those who followed a balanced low-calorie diet (31%, P = .01).²⁸

Decrease in hemoglobin A_1c

Declines in fasting serum glucose corresponded with short-term declines in hemoglobin A_1c in several reports.^27–31 Hemoglobin A_1c declined significantly from an average of 10.4% to 7.3% (P = .001) after PSMF intervention in patients with type 2 diabetes. In contrast, hemoglobin A_1c in the low-calorie diet control group declined from 10.4% to 8.6%.²⁷ One year later, hemoglobin A_1c remained lower than at baseline in the PSMF group (final 9.2%) and continued to compare favorably against the control group (final 11.8%, between-group P = .001). However, these 1-year post-intervention improvements were not seen in a second, more intensive study.²⁸

Less insulin resistance

In several studies, fasting serum insulin levels declined along with serum glucose levels, implying decreased insulin resistance.^{25,27,28,30,31} In addition, insulin output was enhanced during glucose challenge after completion of the PSMF, suggesting possible improved (though still impaired) pancreatic beta-cell capacity.^25,27,30

Improved lipid profile

The most common effect of the PSMF on the lipid profile is a significant decrease in triglycerides in patients both with and without type 2 diabetes.^8,23,24,28 In addition, high-density lipoprotein cholesterol increased in two studies following PSMF intervention or after 1-year of follow-up.^24,27,28 Total cholesterol and low-density lipoprotein cholesterol levels also improved after the PSMF, but these changes were not always maintained at follow-up visits.^8,24,28

Lower blood pressure

Improvements in both systolic and diastolic blood pressure were noted in two studies, with mean decreases of 6 mm Hg to 13 mm Hg systolic and 8 mm Hg diastolic after PSMF intervention.^23,28 In a third study, reductions in blood pressure were less dramatic, and only changes in diastolic but not systolic blood pressure remained significant at 12 months.²⁴ While improvements were not observed in a fourth study, patients in this study also had impaired kidney function caused by diabetic nephropathy, and changes in medication were not taken into account.³¹

Kidney function tests

In a small study, Friedman et al showed that 12 weeks of the PSMF in six patients with advanced diabetic nephropathy (stage 3B or stage 4 chronic kidney disease) led to a loss of 12% of body weight (P = .03) as well as significant reductions in serum creatinine and cystatin C levels (P < .05).³¹ In addition, albuminuria decreased by 30% (P = .08). Side effects were minimal, and the diet was well tolerated despite its high protein content, which is a concern in patients with impaired kidney function.

Thus, weight loss via the PSMF may still be beneficial in type 2 diabetic patients with chronic kidney disease and may even improve the course of progression of diabetic nephropathy.

Long-term weight loss is elusive

Long-term weight loss has been an elusive goal for many diet programs. In a study using a very-low-calorie diet in obese patients with type 2 diabetes, substantial weight loss was maintained in half of the patients at 3 years after the intervention, but nearly all of the patients had regained most of their weight after 5 years.³³

While commitment to behavior modification, maintenance of physical activity, and continued follow-up are all critical factors in sustaining weight loss, new and innovative approaches to battle weight regain are needed.³⁴

Yet despite considerable weight regain in most patients, the Look AHEAD (Action for Health in Diabetes) study showed that participants in intensive lifestyle intervention programs still achieved greater weight loss after 4 years than those receiving standard care.³⁵ Whether this holds true for those in intensive PSMF programs is unknown. In addition, conclusive PSMF studies regarding glycemic control, lipids, and blood pressure beyond 1 year of follow-up are lacking.

A VIABLE OPTION FOR MANY

Adherence to a very-low-calorie, ketogenic PSMF program results in major short-term health benefits for obese patients with type 2 diabetes. These benefits include significant weight loss, often more than 18 kg, within 6 months.^23–28 In addition, significant improvements in fasting glucose^{23–28,30–32} and hemoglobin A_1c levels^27–31 are linked to the caloric and carbohydrate restriction of the PSMF. Insulin resistance was also attenuated, with possible partial restoration of pancreatic beta-cell capacity.^{25,27,28,30,31} In some studies, the PSMF resulted in lower systolic and diastolic blood pressure^23,24,28 and triglyceride levels.^8,23,24,28 One small study also suggested a possible improvement of diabetic nephropathy.³¹ Lastly, improvements in glycemia and hypertension were associated with a reduction in the need for antidiabetic and antihypertensive drugs.³⁶

Still, weight loss and many of the associated improvements partially return to baseline levels 1 year after the intervention. Thus, more long-term studies are needed to explore factors for better weight maintenance after the PSMF.

Also, only a few studies have compared the effect of the PSMF between patients with or without type 2 diabetes. One study suggested that fat loss may be reduced in patients with type 2 diabetes.²⁶

In conclusion, despite some risks and safety concerns, PSMF is a viable option for many obese, type 2 diabetic patients as a method of short-term weight loss, with evidence for improvement of glycemic control and cardiovascular risk factors for up to 1 year. To strengthen support for the PSMF, however, further research is warranted on the diet’s long-term effects in patients with type 2 diabetes and also in nondiabetic patients.
 

Acknowledgments: Many thanks to Cheryl Reitz, RD, LD, CDE, and Dawn Noe, RD, LD, CDE, for providing their expertise on the PSMF protocols carried out at Cleveland Clinic. Additional thanks to Tejas Kashyap for his initial assistance with this review.

Hand, foot, and mouth disease (HFMD) is typically a benign childhood infection—except when it isn’t so benign or when it occurs in adults.

The usual presentation is in a child with fever, oral ulcerations, and papules on the palms of the hands and the soles of the feet.¹ However, severe complications can occur, including central nervous system involvement and cardiopulmonary failure, and can lead to significant morbidity and even death.² Fortunately, these complications are rare.

Less common in North America than in other regions, HFMD has recurrently broken out in many areas of Southern Asia and the surrounding Pacific region. However, several North American outbreaks have been documented in recent years and have affected unexpected numbers of immunocompetent adults, demonstrating that this disease is of worldwide importance in adults as well as children.³

Because HFMD has the potential to reach epidemic levels in the United States, early recognition is paramount, and primary care physicians need to be familiar with its common signs and symptoms.

USUALLY A SUMMER DISEASE

HFMD occurs all around the world, exhibiting seasonal variation in temperate climates. In these locations, individual cases and regional outbreaks usually occur in the spring, summer, and fall. No sexual predisposition has been documented. Most symptomatic cases are in children under the age of 10.

OUTBREAKS AROUND THE WORLD

The disease was first described more than 40 years ago, with several large outbreaks in the last 16 years.

1998—An outbreak in Taiwan affected more than 1.5 million people, mostly children. Severe cases numbered just over 400, and 78 children died.⁴

2008—China,⁵ Singapore,⁶ Vietnam,⁷ Mongolia,⁸ and Brunei⁹ were stricken with an outbreak that affected 30,000 people and led to more than 50 deaths.

2009—An outbreak in the Henan and Shandong provinces of eastern China killed 35 people.¹⁰

2010—In several southern Chinese regions, more than 70,000 people were infected, with almost 600 deaths.¹¹

2011 to the present. The United States has had several outbreaks in the last 3 years. Although HFMD is not one of the diseases that must be reported to public health authorities in the United States, from November 2011 to February 2012 the US Centers for Disease Control and Prevention (CDC) received reports of 63 possible cases: 38 in Alabama, 17 in Nevada, 7 in California, and 1 in Connecticut.¹ Fifteen of the patients were adults, and more than half had contacts who were sick.

The most recent US outbreak, in Alabama,¹² was atypical because it occurred in the winter.

CAUSED BY ENTEROVIRUSES

HFMD is caused by infection with a variety of viruses in the genus Enterovirus, a large group that in turn is part of the larger Picornaviridae family.¹³ The taxonomy of this genus is complicated and subject to revision; species include coxsackieviruses, polioviruses, enteroviruses, and echoviruses. They are all small, nonenveloped, single-stranded RNA viruses.

The most common strains that cause HFMD are coxsackievirus A16 and enterovirus 71. In addition, coxsackievirus A6 may be emerging, and many other coxsackievirus strains have been directly implicated, including A5, A7, A9, A10, B2, and B5.

Coxsackievirus A16 is the leading cause of HFMD.

Enterovirus 71 is the second most common cause of HFMD and has also caused outbreaks. It usually results in benign disease. However, among the causes of HFMD, it is associated with more prominent central nervous system involvement¹⁴ and is the most common cause of viral meningoencephalitis in children.

Coxsackievirus A6. In December 2011, the California Department of Public Health isolated a strain of coxsackievirus A6 that caused extensive rash and nail shedding.¹⁵ Among the 63 possible cases of HFMD reported to the CDC from November 2011 to February 2012, specimens for clinical testing were obtained in 34, and 25 of those demonstrated coxsackievirus A6 infection.³

FEVER, ORAL ULCERS, RASH ON HANDS AND FEET

The typical clinical manifestations of HFMD are fever, stomatitis with oral ulcers, and an exanthem affecting the palms, soles, and other parts of the body. These last less than 7 to 10 days, usually occur during the spring to fall months, and have a benign course.

The incubation period is 3 to 5 days, with a prodrome that may include fever, malaise, abdominal pain, and myalgia before the onset of oral and cutaneous findings. Painful oral ulcers may precede the exanthem and can result in dehydration.¹⁶

The cutaneous manifestation of HFMD is typically a papulovesicular rash affecting the palms, soles, and buttocks (Figure 1). Other sites may include the knees, elbows, and the dorsal surfaces of the hands and feet. The lesions may be maculopapular and can be either asymptomatic or tender and painful. Desquamation can follow the exanthem, and lesions usually resolve without scarring or secondary infection.^16,17

Table 1 and Table 2 compare HFMD with other common illnesses that can cause similar skin and mucosal findings. In particular, herpangina has the identical clinical presentation as HFMD except that it does not cause skin lesions. It is caused by many of the same enteroviruses linked to HFMD.

Different viruses, different signs?

The numerous viruses that cause HFMD usually cause similar signs and symptoms during bouts of typical, self-limited disease. However, neurologic and cardiopulmonary involvement, which are fortunately rare, are more often associated with enterovirus 71 infection.

Nail manifestations are common in HFMD. Nail separation from the nail matrix (onychomadesis) was associated with coxsackievirus A6 infection during a 2010 outbreak of HFMD in Taiwan and in a 2009 outbreak in Finland.¹⁸ Moreover, this virus was cultured from a nail specimen in one patient, suggesting viral infiltration as the cause of nail-matrix arrest.¹⁹

Perioral skin eruptions, desquamation, and Beau lines have also been associated with coxsackievirus A6.¹⁸ Beau lines are transverse depressions of the nail, most evident in the central nail plate; when seen on multiple nails, they imply a systemic illness causing disruption of nail matrix growth.²⁰

Atypical HFMD and coxsackievirus A6

Atypical HFMD has recently been described in connection with coxsackievirus A6. Lott et al²¹ reported five cases of coxsackievirus A6-associated HFMD in 2013. Atypically, three of the affected patients presented in winter months, two were adults, and two had widespread skin involvement.²¹

Mathes et al²² reported a series of 80 cases of enteroviral infections in which the lesions had a predilection for the antecubital and popliteal fossae and were similar in severity and distribution to those seen in eczema herpeticum or Kaposi varicelliform eruption in patients with and patients without a history of atopic dermatitis. They named this find-clinical finding of pronounced coxsackievirus-associated skin disease at sites previously affected by atopic dermatitis.

Additional cutaneous findings of coxsackievirus A6 infection may include onychomadesis, Beau lines, and vesiculobullous lesions. Patients with atypical, coxsackievirus A6-associated HFMD may not have oral lesions.²³

In the five cases reported by Lott et al,²¹ significant systemic symptoms (fever, chills, diarrhea, and myalgias) led all but one of the patients to seek care in an emergency department. However, atypical HFMD has not been associated with life-threatening illness.

Atypical HFMD associated with coxsackievirus A6 is an emerging entity in the United States, and the acuity of both cutaneous and systemic symptoms poses a diagnostic dilemma. Furthermore, infection has been documented in immunocompetent adults.²³ Familiarity with the clinical findings may expedite appropriate care, prevent spread to contacts, and avoid unnecessary testing.

Neurologic and cardiopulmonary manifestations

Enteroviruses are the most common causes of viral meningoencephalitis in the United States. They mainly affect children and cause serious and potentially chronic disease in those with humoral immunodeficiencies.²⁴ Neurologic and cardiopulmonary manifestations of HFMD are varied and extremely rare in the United States but should always be viewed clinically as signs of concern and severe disease.

Signs of potentially fatal disease that have been observed in young children include tachycardia, tachypnea, hypotension, hypertension, gastrointestinal bleeding, neurologic symptoms, leukocytosis, absence of oral lesions, and vomiting.² Signs of dysautonomia, myoclonus, ataxia, and brainstem involvement may portend fatal disease in which rapid decompensation is the result of cardiogenic shock due to loss of ventricular contractility, causing pulmonary edema and end-organ dysfunction.¹⁶

Neurologic manifestations associated with enterovirus 71 infection include aseptic meningitis, a poliomyelitis-like syndrome, brainstem encephalitis, neurogenic pulmonary edema, opsoclonus-myoclonus syndrome, cerebellar ataxia, Guillain-Barré syndrome, and transverse myelitis.

Because some patients who have neurologic disease respond to treatment with high-dose methylprednisolone and intravenous immune globulin, there is reason to suspect that an autoimmune phenomenon triggered by the culprit enterovirus may be the cause of many of the neurologic symptoms.²⁵

A 2012 meta-analysis²⁶ found that an elevated white blood cell count and hyperglycemia could be clinically useful in distinguishing benign from severe HFMD. In patients with benign HFMD, white blood cell counts and blood glucose values were no different from those in healthy controls.²⁶

DIAGNOSIS IS USUALLY CLINICAL

Most enteroviral infections are asymptomatic, but HFMD is a possibility if a patient has mild illness, fever, and a maculopapular or vesicular rash on the palms of the hands and soles of the feet, sometimes associated with oral ulcers (herpangina). Skin lesions can also be found on the legs, face, buttocks, and trunk.

In the United States, HFMD most commonly occurs in children under age 4 and is usually caused by coxsackievirus A16. Adults can also be affected, especially if they were in contact with children in child care, which was the case in approximately half of nonpediatric patients who tested positive for HFMD during an outbreak in several states between November 2011 and February 2012.³

The clinical characteristics of HFMD caused by enterovirus 71 may be somewhat different, with smaller vesicles, diffuse erythema of the trunk and limbs, and higher fever (temperature ≥ 39°C [102.2°F] for more than 3 days).²⁷ However, the rash of coxsackievirus A16 HFMD may be more extensive and severe.

Other clinical manifestations of HFMD include nail dystrophies such as Beau lines and nail shedding, hyperglycemia, dehydration, and more serious and potentially life-threatening complications such as pulmonary edema²⁸ and viral meningoencephalitis.²⁹

Laboratory testing

In mild cases of HFMD, particularly in patients with a high probability of having the disease based on their clinical characteristics and sick contacts, laboratory testing is not necessary. Testing is usually reserved for severe cases and public health investigation of outbreaks.

Viral culture is the gold standard for diagnosing HFMD, but the final results can take nearly a week.

Polymerase chain reaction testing is faster, with a turnaround time of less than 1 day. It identifies viral RNA and is highly sensitive for detecting central nervous system infection.³⁰

Where should samples be collected? Serum viremia precedes invasion of the skin and mucous membranes, so plasma can be tested. Inside the body, enteroviruses initially replicate in the gastrointestinal tract, although collecting a rectal swab or a stool sample is somewhat invasive. Further, in an enterovirus 71 epidemic in Taiwan, 93% of the patients had positive throat swabs, but only 30% tested positive by rectal swabs or analysis of the feces.²⁷ At present, throat and vesicle specimens are considered to be the most useful sources for diagnostic purposes.¹⁶

ELISAs. Newly developed IgM-capture enzyme-linked immunosorbent assays (ELISAs) for coxsackievirus A16 and enterovirus 71 appear quite promising for diagnosing HFMD. These tests are inexpensive and detect IgM antibodies early and in a high percentage of patients. In the first week of the disease, the IgM detection rate was found to be 90.2% for enterovirus 71 and 68% for coxsackievirus A16.³¹

Cross-reactivity between these two viruses was a problem with ELISA testing in the past, causing false-positive results for enterovirus 71 in patients who in fact had coxsackievirus A16. The problem appears to be resolved in new versions that use specific enterovirus 71 proteins, eg, VP1.³²

RECOGNITION AND PREVENTION ARE THE BEST MEDICINE

Recognizing HFMD early is crucial, because making the clinical diagnosis can identify patients who have signs of severe disease and can help protect future contacts and decrease the risk of an epidemic.

Infected patients continue to shed the virus for a long time, making hand hygiene and environmental control measures in health care settings and daycare centers of vital importance, to prevent spread of the infection.

Enteroviruses are stable in the environment and therefore capable of fecal-oral and oral-oral transmission. Humans are the only known natural hosts. No chemoprophylaxis or vaccination has been established to prevent HFMD. The recurrence of large-scale epidemics in the developing world is perhaps explained by ineffective sewage treatment and limited access to clean drinking water, especially in light of the fecal-oral spread of the virus. Intrafamilial spread of HFMD has been shown to be an important means of disease transmission, and asymptomatic adult carriers of these viruses may spread it to young children.³³

The different viruses that cause HFMD result in a similar clinical presentation in most patients. Therefore, identifying HFMD caused by enterovirus 71, which carries a risk of severe and even fatal disease in young children vs a virus such as coxsackievirus A16, can be very difficult in practice without virologic testing. Thus, when diagnosed with HFMD, patients should be counseled to control all variables that could lead to further spread of the disease.

An analysis of epidemics in Asia suggested that public health awareness may have averted deaths in successive epidemics, highlighting the need to identify HFMD epidemics in communities and to educate patients and families about measures to prevent further spread of the virus in addition to standard supportive care.³⁴

The CDC recommends³⁵:

• Frequent hand-washing after toileting and changing diapers
• Disinfecting frequently used surfaces and objects, including toys
• Avoiding close contact with infected individuals and sharing of personal items such as utensils and cups.

These measures should be recommended to all affected patients.³⁵

NO PROVEN ANTIVIRAL TREATMENT

No proven antiviral treatment exists for HFMD. Thus, the goals of treatment are typically supportive, as for any self-limited viral syndrome.¹⁶

Does acyclovir help? Shelley et al³⁶ treated 13 patients (12 children and 1 adult) with acyclovir within 1 to 2 days of the onset of the HFMD rash and reported that it was beneficial, with significant relief of fever and skin lesions within 24 hours of starting therapy. These anecdotal results have not been replicated, and acyclovir is not an established treatment for HFMD.

If acyclovir does help, how does it work? Acyclovir, like other common antiviral medications, inactivates thymidine kinase, an enzyme produced by herpesviruses but not by HFMD-causing viruses like coxsackievirus A16. Shelley et al proposed that acyclovir may enhance the antiviral effect of the patient’s own interferon.³⁶

Intravenous immunoglobulin has been used in severe cases during outbreaks in Asia, with retrospective data showing a potential ability to halt disease progression if used before the development of cardiopulmonary failure. However, this has not been studied prospectively and is not currently recommended.¹⁶

Acknowledgment: We would like to thank Dr. Salvador Alvarez of the Mayo Clinic Department of Infectious Disease and Dr. Donald Lookingbill of the Mayo Clinic Department of Dermatology for their collaboration.