Reviews

Most people assume that electronic cigarettes (e-cigarettes) are safer than conventional tobacco products. Nevertheless, we should not encourage addicted smokers to try “vaping” as an alternative to nicotine replacement therapy, and we should discourage never-smokers from taking up vaping as vigorously as we try to discourage them from taking up smoking.

This article examines the prevailing assumptions and the evidence regarding the safety of e-cigarettes and traditional nicotine replacement therapy.

SMOKING IS DECLINING BUT FAR FROM GONE

While smoking rates have been declining over the past 50 years, the burden of disease attributable to tobacco use remains high. In the United States, it is estimated that nearly 6 million of those currently under the age of 18 will die of tobacco-related illnesses.¹ In the 50 years since the US surgeon general first reported on the health concerns related to tobacco, smoking has claimed the lives of nearly 21 million Americans¹ and continues to kill more than 400,000 every year.²

Even though the risks of smoking are well known, smoking remains one of the most difficult habits to quit. Indeed, about half of all smokers attempt to quit each year, but very few succeed.³

NICOTINE REPLACEMENT: GUM, PATCHES…E-CIGARETTES?

Smoking continues to kill more than 400,000 Americans every year

Nicotine replacement therapy was born out of the thought that, though nicotine is responsible for tobacco’s addictive quality, most tobacco-related disease is attributable to the 7,000 other substances found in tobacco smoke.⁴ Nicotine polacrilex gum was approved by the US Food and Drug Administration (FDA) in 1984, and nicotine transdermal film was approved in 1991.⁵

Nicotine replacement therapy, in the form of patches and gum, has been shown to improve the odds of successfully quitting smoking by a factor of nearly 1.5 to 2.⁶ Nicotine patches and gum were initially prescription medications but became available over the counter in 1996.⁷ They quickly became first-line agents for smoking cessation, and their over-the-counter availability softened any potential concerns about the possible deleterious health consequences of nicotine itself.

E-cigarettes—devices that generate a nicotine vapor that can be inhaled in a fashion that mimics the experience of smoking—were introduced in China in 2004.⁸ By 2012, sales of these devices in the United States had reached $500 million and in 2013 were expected to top $1 billion.^9,10

E-cigarette manufacturers make no therapeutic claims about their products, thus allowing them to escape regulation by the FDA as nicotine replacement therapy. A recent FDA proposal, however, is likely to change their “protected” status.¹¹ Despite the lack of regulation up to this point, patients generally assume that e-cigarettes are just another form of nicotine replacement therapy, even though they contain substances other than nicotine.

WHAT’S IN E-CIGARETTES?

Nicotine, which is bad in itself

E-cigarettes contain nicotine in varying amounts (some cartridges contain none at all). Though nicotine replacement therapy is less harmful than tobacco, nicotine by itself is associated with its own health problems, notably cancer, cardiovascular disease, birth defects (possibly), and poisoning.

Carcinogenesis. Nicotine plays a direct role in carcinogenesis through a variety of mechanisms, including increasing the activity of tumor growth-promoting transcription factors, decreasing apoptosis, and increasing angiogenesis in tumors.¹² Additionally, specific types of nicotinic acetylcholine receptors— eg, alpha 7 receptors, which are stimulated by nicotine—are found in many malignant tumors and are thought to play a role in tumor progression.¹² Blockade of alpha 7 nicotinic acetylcholine receptors has been shown to decrease the growth of certain cancers.¹²

However, these findings were from in vitro studies, and the concerns they raised have not been reflected in in vivo studies. Despite having been on the market for 30 years, nicotine replacement therapy has as yet not been associated with any “real world” increase in cancer risk.

Smoking is one of the leading risk factors for cervical cancer, and nicotine itself may play a contributing role. Nicotine has been shown to increase cellular proliferation in cervical cancer.¹³ Some evidence suggests that it may also play a role in the lymphogenic metastasis of cervical cancer.¹³

Cardiovascular disease. Nicotine has been linked to cardiovascular disease. It directly affects the heart’s rate and rhythm via nicotinic acetylcholine receptors in the peripheral autonomic nervous system. It impairs endothelial-dependent dilation of blood vessels in response to nitric oxide, and this inhibition in the coronary arteries may contribute to smoking-related heart disease.^14,15 Nicotine has also been shown to raise serum cholesterol levels, increase clot formation, and contribute to plaque formation by increasing vascular smooth muscle.¹⁴

Possible teratogenic effect. There is some theoretical concern regarding exposure to nicotine in utero, as nicotinic acetylcholine receptors develop before neurons, and nicotine may therefore interfere with the natural influence of acetylcholine on the development of this system.¹⁴

Consuming one vial of e-cigarette fluid could be fatal

Direct toxicity. Nicotine is toxic at high levels. The overdose potential associated with nicotine is particularly worrisome with e-cigarettes, as the nicotine solution they use is typically supplied in 5-mL, 10-mL, or 20-mL vials that range in concentration from 8.5 to 22.2 mg of nicotine per mL.¹⁶ The fatal single dose range of nicotine has been reported at 30 to 60 mg in adults and 10 mg in children and can be achieved by oral, intravenous, or transdermal absorption,¹⁶ so one vial, if consumed orally, could be fatal.

The number of calls to US poison control centers regarding e-cigarettes has increased, closely paralleling their rise in popularity. In 2010, there were only 30 e-cigarette related calls to poison control centers; in 2011 the number increased to 269, and in 2012 it had reached 459 and included one fatality that was deemed a suicide.^17–19 Even though such toxic nicotine overdoses are rare, physicians should exercise caution and avoid recommending e-cigarettes to individuals with mental confusion, psychotic disorders, or suicidality, who might consume an entire vial.

Possible positive effects? Smoking is one of the worst things that people can do to their body, but the picture is complicated by a few possible positive effects. In the brain, although smoking increases the risk of Alzheimer disease, it is associated with a lower risk of Parkinson disease. In the bowel, it increases the risk of Crohn disease but may decrease the risk of ulcerative colitis. Gahring and Rogers²⁰ pointed out that neuronal nicotinic receptors are present in nonneuronal cells throughout the body and proposed that expression of these receptors may play a role in mediating the consequences of nicotine use, both good and bad. The lesson may be that nicotine is very active in the body, its effects are complicated and still incompletely understood, and therefore we should not encourage people to inhale nicotine products ad lib.

Additives

E-cigarettes typically contain propylene glycol, flavorings, and glycerine. One study that analyzed the additive contents of e-cigarettes found that propylene glycol accounted for 66% of the fluid, glycerine 24%, and flavorings less than 0.1%.²¹ Propylene glycol is the substance typically used in theater fog machines and is used to generate the vapor in e-cigarettes. Other substances such as tobacco-specific nitrosamines and diethylene glycol have also been found in e-cigarettes in small amounts.²²

Propylene glycol, ‘generally recognized as safe’

Propylene glycol has been used in theater fog machines for years—think Phantom of the Opera. It is also widely used as a solvent in many consumer products and pharmaceuticals. The FDA classified it as “generally recognized as safe” on the basis of one study conducted in rats and monkeys over 60 years ago.²³ As other authors have noted, however, a major manufacturer of propylene glycol recommends that exposure to propylene glycol mist be avoided.^24,25 Potential concern over propylene glycol mist was heightened when it was discovered that of all industries, the entertainment business ranked first in terms of work-related asthma symptoms and had the fifth-highest rate of wheezing.^26,27

Studies conducted over the last several decades have raised numerous health concerns about the safety of propylene glycol (Table 1).^26,28–36 The studies of propylene glycol fog are particularly important, as they most closely resemble the route of exposure in e-cigarette users.

Wieslander et al²⁸ exposed 27 volunteers to propylene glycol mist for 1 minute in an aircraft simulator under training conditions. Exposures were high, ranging from 176 to 851 mg/m³ (mean = 309 mg/m³). Four volunteers who developed a cough exhibited evidence of airway obstruction as indicated by a 5% decrease in forced expiratory volume in 1 second (FEV₁), while the rest did not exhibit any change in FEV₁.

Moline et al²⁹ conducted a non-peer-reviewed study for the Actors Equity Association and the League of American Theaters and Producers of 439 actors exposed to theater fog. They found statistically significant evidence of throat and vocal cord inflammation with prolonged peak exposure to glycols and recommended that actors not be exposed to glycol concentrations exceeding 40 mg/m³.

Varughese et al²⁶ conducted a study in 101 volunteers at 19 sites. The mean concentration of glycol-based fog was much lower than that in the studies by Wieslander et al²⁸ and Moline et al,²⁹ at 0.49 mg/m³ (the maximum was 3.22 mg/m³). The investigators concluded that glycol-based fog was associated with deleterious respiratory effects and that employees’ exposure should be limited.

The health issues related to propylene glycol are unique to e-cigarettes compared with nicotine replacement therapy. Unfortunately, the most applicable data available are from studies of persons exposed to theater fog, which involved periodic exposure and likely do not emulate the deep inhalation, multiple times daily, of propylene glycol by e-cigarette smokers. A 2014 review of the chemistry of contaminants in e-cigarettes³⁷ concluded that estimated levels of propylene glycol exposure in e-cigarette users come close to the threshold limit value set by the American Conference of Governmental Industrial Hygienists, and should merit concern.

These studies and real-life experience in the theater, while limited in scope, should give physicians pause and should cause increased awareness of the possibility of e-cigarette-induced pulmonary and upper airway complications. If such complications should occur, discontinuation of vaping should be advised.

Contaminants

The issue of adulterants is common to both e-cigarettes and nicotine replacement therapy. Several unlisted substances have been found in analyzed samples of e-cigarette fluid, including tobacco-specific nitrosamines (TSNAs), diethylene glycol (found in only one e-cigarette cartridge), cotinine, anabasine, myosmine, and beta-nicotyrine.²² The tobacco-specific nitrosamines N´-nitrosonornicotine (NNN), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), N-nitrosoanabasine, and nitrosoanatabine have been found in five e-cigarette cartridge samples from two manufacturers in amounts similar to those found in nicotine replacement products.²²

Goniewicz et al³⁸ tested the vapor generated by 12 e-cigarette brands and found trace amounts of TSNAs. NNN was present in the vapor of eight of the samples in concentrations ranging from 0.8 to 4.3 ng per 150 puffs, and NNK in the vapor of nine of the samples in concentrations ranging from 1.1 to 28.3 ng per 150 puffs. Neither NNN nor NNK was found in blank samples nor with the Nicorette inhalator tested in the same study.³⁸

Because TNSAs can be formed from nicotine and its metabolites, there is also concern that cancer-causing nitrosamines may be formed from nicotine after it is absorbed into the body (ie, endogenously). While endogenous formation of NNK from nicotine has never been demonstrated, endogenous formation of NNN has been seen in some nicotine patch users.³⁹ The presence of these nitrosamines has raised concern that e-cigarettes and nicotine replacement therapy may have carcinogenic potential. The amounts of tobacco-specific nitrosamines found in e-cigarettes are also found in some nicotine replacement products.⁴⁰

Investigators have examined a possible connection between e-cigarettes and potentially carcinogenic carbonyl compounds, including formaldehyde, acetaldehyde, and acrolein. Formaldehyde (a known carcinogen) and acetaldehyde (a potential carcinogen) have been detected in e-cigarette cartridges and vapor.^38,41–43 Acrolein, a mucosal irritant, has been found in e-cigarette vapor.^38,43 Goniewicz et al³⁸ suggested that acrolein may be formed by the heating of the glycerin contained in the e-cigarette solution.

An extensive review of the studies of possible contaminant exposures (including polycylic aromatic hydrocarbons, TSNAs, volatile organic compounds, diethylene glycol, and inorganic compounds) with e-cigarette use according to occupational hygiene standards concluded that there was no cause for concern about increased health risk.³⁷ The study by Goniewicz et al also concluded that using e-cigarettes instead of traditional cigarettes may significantly reduce exposure to some tobacco-specific toxins.³⁸

E-CIGARETTES VS NICOTINE REPLACEMENT

Traditional nicotine replacement therapy products are regulated by the FDA and therefore standardized in terms of their contents. E-cigarettes, on the other hand, are unregulated vehicles for supplying nicotine, and may pose other health risks. One such risk is related to exposure to propylene glycol, which has never been studied under conditions (in terms of mode of delivery, frequency of dosing, and total duration of exposure) that approximate the exposure associated with e-cigarettes. Furthermore, the high concentration of nicotine in e-cigarette fluid poses a real risk of toxicity and potentially fatal overdose.

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine if used in a harm-reduction strategy as a maintenance medication. Whether the ongoing nicotine addiction makes it more likely that individuals would switch back and forth between nicotine replacement and tobacco-based products is not clear. Also not known is whether e-cigarettes may serve as the “gateway drug” by which teens enter into nicotine addiction, but we believe that the potential exists, as these products are potentially more appealing in terms of the lack of pungent smell, the perception of safety, and the variety of flavors of e-cigarettes.

The efficacy of nicotine replacement therapy in improving smoking cessation has been reviewed extensively elsewhere³⁷ and is beyond the scope of this article. E-cigarettes may be appealing to many cigarette smokers because they deliver smokeless nicotine, and they more closely emulate the actual experience of smoking compared with traditional nicotine replacement therapy. Though some evidence suggests that e-cigarettes may be modestly effective in helping tobacco smokers quit nicotine, they are not FDA-approved for smoking cessation and are not marketed for that indication.⁴⁴ Medical practitioners should see them for what they are: a new nicotine product with a novel delivery system that is not approved as treatment. Because of the inherent risks involved with e-cigarettes, medical practitioners are best advised to remain neutral on the relative value of e-cigarettes and should continue to motivate patients to discontinue nicotine use altogether.

Most people assume that electronic cigarettes (e-cigarettes) are safer than conventional tobacco products. Nevertheless, we should not encourage addicted smokers to try “vaping” as an alternative to nicotine replacement therapy, and we should discourage never-smokers from taking up vaping as vigorously as we try to discourage them from taking up smoking.

This article examines the prevailing assumptions and the evidence regarding the safety of e-cigarettes and traditional nicotine replacement therapy.

SMOKING IS DECLINING BUT FAR FROM GONE

While smoking rates have been declining over the past 50 years, the burden of disease attributable to tobacco use remains high. In the United States, it is estimated that nearly 6 million of those currently under the age of 18 will die of tobacco-related illnesses.¹ In the 50 years since the US surgeon general first reported on the health concerns related to tobacco, smoking has claimed the lives of nearly 21 million Americans¹ and continues to kill more than 400,000 every year.²

Even though the risks of smoking are well known, smoking remains one of the most difficult habits to quit. Indeed, about half of all smokers attempt to quit each year, but very few succeed.³

NICOTINE REPLACEMENT: GUM, PATCHES…E-CIGARETTES?

Smoking continues to kill more than 400,000 Americans every year

Nicotine replacement therapy was born out of the thought that, though nicotine is responsible for tobacco’s addictive quality, most tobacco-related disease is attributable to the 7,000 other substances found in tobacco smoke.⁴ Nicotine polacrilex gum was approved by the US Food and Drug Administration (FDA) in 1984, and nicotine transdermal film was approved in 1991.⁵

Nicotine replacement therapy, in the form of patches and gum, has been shown to improve the odds of successfully quitting smoking by a factor of nearly 1.5 to 2.⁶ Nicotine patches and gum were initially prescription medications but became available over the counter in 1996.⁷ They quickly became first-line agents for smoking cessation, and their over-the-counter availability softened any potential concerns about the possible deleterious health consequences of nicotine itself.

E-cigarettes—devices that generate a nicotine vapor that can be inhaled in a fashion that mimics the experience of smoking—were introduced in China in 2004.⁸ By 2012, sales of these devices in the United States had reached $500 million and in 2013 were expected to top $1 billion.^9,10

E-cigarette manufacturers make no therapeutic claims about their products, thus allowing them to escape regulation by the FDA as nicotine replacement therapy. A recent FDA proposal, however, is likely to change their “protected” status.¹¹ Despite the lack of regulation up to this point, patients generally assume that e-cigarettes are just another form of nicotine replacement therapy, even though they contain substances other than nicotine.

WHAT’S IN E-CIGARETTES?

Nicotine, which is bad in itself

E-cigarettes contain nicotine in varying amounts (some cartridges contain none at all). Though nicotine replacement therapy is less harmful than tobacco, nicotine by itself is associated with its own health problems, notably cancer, cardiovascular disease, birth defects (possibly), and poisoning.

Carcinogenesis. Nicotine plays a direct role in carcinogenesis through a variety of mechanisms, including increasing the activity of tumor growth-promoting transcription factors, decreasing apoptosis, and increasing angiogenesis in tumors.¹² Additionally, specific types of nicotinic acetylcholine receptors— eg, alpha 7 receptors, which are stimulated by nicotine—are found in many malignant tumors and are thought to play a role in tumor progression.¹² Blockade of alpha 7 nicotinic acetylcholine receptors has been shown to decrease the growth of certain cancers.¹²

However, these findings were from in vitro studies, and the concerns they raised have not been reflected in in vivo studies. Despite having been on the market for 30 years, nicotine replacement therapy has as yet not been associated with any “real world” increase in cancer risk.

Smoking is one of the leading risk factors for cervical cancer, and nicotine itself may play a contributing role. Nicotine has been shown to increase cellular proliferation in cervical cancer.¹³ Some evidence suggests that it may also play a role in the lymphogenic metastasis of cervical cancer.¹³

Cardiovascular disease. Nicotine has been linked to cardiovascular disease. It directly affects the heart’s rate and rhythm via nicotinic acetylcholine receptors in the peripheral autonomic nervous system. It impairs endothelial-dependent dilation of blood vessels in response to nitric oxide, and this inhibition in the coronary arteries may contribute to smoking-related heart disease.^14,15 Nicotine has also been shown to raise serum cholesterol levels, increase clot formation, and contribute to plaque formation by increasing vascular smooth muscle.¹⁴

Possible teratogenic effect. There is some theoretical concern regarding exposure to nicotine in utero, as nicotinic acetylcholine receptors develop before neurons, and nicotine may therefore interfere with the natural influence of acetylcholine on the development of this system.¹⁴

Consuming one vial of e-cigarette fluid could be fatal

Direct toxicity. Nicotine is toxic at high levels. The overdose potential associated with nicotine is particularly worrisome with e-cigarettes, as the nicotine solution they use is typically supplied in 5-mL, 10-mL, or 20-mL vials that range in concentration from 8.5 to 22.2 mg of nicotine per mL.¹⁶ The fatal single dose range of nicotine has been reported at 30 to 60 mg in adults and 10 mg in children and can be achieved by oral, intravenous, or transdermal absorption,¹⁶ so one vial, if consumed orally, could be fatal.

The number of calls to US poison control centers regarding e-cigarettes has increased, closely paralleling their rise in popularity. In 2010, there were only 30 e-cigarette related calls to poison control centers; in 2011 the number increased to 269, and in 2012 it had reached 459 and included one fatality that was deemed a suicide.^17–19 Even though such toxic nicotine overdoses are rare, physicians should exercise caution and avoid recommending e-cigarettes to individuals with mental confusion, psychotic disorders, or suicidality, who might consume an entire vial.

Possible positive effects? Smoking is one of the worst things that people can do to their body, but the picture is complicated by a few possible positive effects. In the brain, although smoking increases the risk of Alzheimer disease, it is associated with a lower risk of Parkinson disease. In the bowel, it increases the risk of Crohn disease but may decrease the risk of ulcerative colitis. Gahring and Rogers²⁰ pointed out that neuronal nicotinic receptors are present in nonneuronal cells throughout the body and proposed that expression of these receptors may play a role in mediating the consequences of nicotine use, both good and bad. The lesson may be that nicotine is very active in the body, its effects are complicated and still incompletely understood, and therefore we should not encourage people to inhale nicotine products ad lib.

Additives

E-cigarettes typically contain propylene glycol, flavorings, and glycerine. One study that analyzed the additive contents of e-cigarettes found that propylene glycol accounted for 66% of the fluid, glycerine 24%, and flavorings less than 0.1%.²¹ Propylene glycol is the substance typically used in theater fog machines and is used to generate the vapor in e-cigarettes. Other substances such as tobacco-specific nitrosamines and diethylene glycol have also been found in e-cigarettes in small amounts.²²

Propylene glycol, ‘generally recognized as safe’

Propylene glycol has been used in theater fog machines for years—think Phantom of the Opera. It is also widely used as a solvent in many consumer products and pharmaceuticals. The FDA classified it as “generally recognized as safe” on the basis of one study conducted in rats and monkeys over 60 years ago.²³ As other authors have noted, however, a major manufacturer of propylene glycol recommends that exposure to propylene glycol mist be avoided.^24,25 Potential concern over propylene glycol mist was heightened when it was discovered that of all industries, the entertainment business ranked first in terms of work-related asthma symptoms and had the fifth-highest rate of wheezing.^26,27

Studies conducted over the last several decades have raised numerous health concerns about the safety of propylene glycol (Table 1).^26,28–36 The studies of propylene glycol fog are particularly important, as they most closely resemble the route of exposure in e-cigarette users.

Wieslander et al²⁸ exposed 27 volunteers to propylene glycol mist for 1 minute in an aircraft simulator under training conditions. Exposures were high, ranging from 176 to 851 mg/m³ (mean = 309 mg/m³). Four volunteers who developed a cough exhibited evidence of airway obstruction as indicated by a 5% decrease in forced expiratory volume in 1 second (FEV₁), while the rest did not exhibit any change in FEV₁.

Moline et al²⁹ conducted a non-peer-reviewed study for the Actors Equity Association and the League of American Theaters and Producers of 439 actors exposed to theater fog. They found statistically significant evidence of throat and vocal cord inflammation with prolonged peak exposure to glycols and recommended that actors not be exposed to glycol concentrations exceeding 40 mg/m³.

Varughese et al²⁶ conducted a study in 101 volunteers at 19 sites. The mean concentration of glycol-based fog was much lower than that in the studies by Wieslander et al²⁸ and Moline et al,²⁹ at 0.49 mg/m³ (the maximum was 3.22 mg/m³). The investigators concluded that glycol-based fog was associated with deleterious respiratory effects and that employees’ exposure should be limited.

The health issues related to propylene glycol are unique to e-cigarettes compared with nicotine replacement therapy. Unfortunately, the most applicable data available are from studies of persons exposed to theater fog, which involved periodic exposure and likely do not emulate the deep inhalation, multiple times daily, of propylene glycol by e-cigarette smokers. A 2014 review of the chemistry of contaminants in e-cigarettes³⁷ concluded that estimated levels of propylene glycol exposure in e-cigarette users come close to the threshold limit value set by the American Conference of Governmental Industrial Hygienists, and should merit concern.

These studies and real-life experience in the theater, while limited in scope, should give physicians pause and should cause increased awareness of the possibility of e-cigarette-induced pulmonary and upper airway complications. If such complications should occur, discontinuation of vaping should be advised.

Contaminants

The issue of adulterants is common to both e-cigarettes and nicotine replacement therapy. Several unlisted substances have been found in analyzed samples of e-cigarette fluid, including tobacco-specific nitrosamines (TSNAs), diethylene glycol (found in only one e-cigarette cartridge), cotinine, anabasine, myosmine, and beta-nicotyrine.²² The tobacco-specific nitrosamines N´-nitrosonornicotine (NNN), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), N-nitrosoanabasine, and nitrosoanatabine have been found in five e-cigarette cartridge samples from two manufacturers in amounts similar to those found in nicotine replacement products.²²

Goniewicz et al³⁸ tested the vapor generated by 12 e-cigarette brands and found trace amounts of TSNAs. NNN was present in the vapor of eight of the samples in concentrations ranging from 0.8 to 4.3 ng per 150 puffs, and NNK in the vapor of nine of the samples in concentrations ranging from 1.1 to 28.3 ng per 150 puffs. Neither NNN nor NNK was found in blank samples nor with the Nicorette inhalator tested in the same study.³⁸

Because TNSAs can be formed from nicotine and its metabolites, there is also concern that cancer-causing nitrosamines may be formed from nicotine after it is absorbed into the body (ie, endogenously). While endogenous formation of NNK from nicotine has never been demonstrated, endogenous formation of NNN has been seen in some nicotine patch users.³⁹ The presence of these nitrosamines has raised concern that e-cigarettes and nicotine replacement therapy may have carcinogenic potential. The amounts of tobacco-specific nitrosamines found in e-cigarettes are also found in some nicotine replacement products.⁴⁰

Investigators have examined a possible connection between e-cigarettes and potentially carcinogenic carbonyl compounds, including formaldehyde, acetaldehyde, and acrolein. Formaldehyde (a known carcinogen) and acetaldehyde (a potential carcinogen) have been detected in e-cigarette cartridges and vapor.^38,41–43 Acrolein, a mucosal irritant, has been found in e-cigarette vapor.^38,43 Goniewicz et al³⁸ suggested that acrolein may be formed by the heating of the glycerin contained in the e-cigarette solution.

An extensive review of the studies of possible contaminant exposures (including polycylic aromatic hydrocarbons, TSNAs, volatile organic compounds, diethylene glycol, and inorganic compounds) with e-cigarette use according to occupational hygiene standards concluded that there was no cause for concern about increased health risk.³⁷ The study by Goniewicz et al also concluded that using e-cigarettes instead of traditional cigarettes may significantly reduce exposure to some tobacco-specific toxins.³⁸

E-CIGARETTES VS NICOTINE REPLACEMENT

Traditional nicotine replacement therapy products are regulated by the FDA and therefore standardized in terms of their contents. E-cigarettes, on the other hand, are unregulated vehicles for supplying nicotine, and may pose other health risks. One such risk is related to exposure to propylene glycol, which has never been studied under conditions (in terms of mode of delivery, frequency of dosing, and total duration of exposure) that approximate the exposure associated with e-cigarettes. Furthermore, the high concentration of nicotine in e-cigarette fluid poses a real risk of toxicity and potentially fatal overdose.

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine if used in a harm-reduction strategy as a maintenance medication. Whether the ongoing nicotine addiction makes it more likely that individuals would switch back and forth between nicotine replacement and tobacco-based products is not clear. Also not known is whether e-cigarettes may serve as the “gateway drug” by which teens enter into nicotine addiction, but we believe that the potential exists, as these products are potentially more appealing in terms of the lack of pungent smell, the perception of safety, and the variety of flavors of e-cigarettes.

The efficacy of nicotine replacement therapy in improving smoking cessation has been reviewed extensively elsewhere³⁷ and is beyond the scope of this article. E-cigarettes may be appealing to many cigarette smokers because they deliver smokeless nicotine, and they more closely emulate the actual experience of smoking compared with traditional nicotine replacement therapy. Though some evidence suggests that e-cigarettes may be modestly effective in helping tobacco smokers quit nicotine, they are not FDA-approved for smoking cessation and are not marketed for that indication.⁴⁴ Medical practitioners should see them for what they are: a new nicotine product with a novel delivery system that is not approved as treatment. Because of the inherent risks involved with e-cigarettes, medical practitioners are best advised to remain neutral on the relative value of e-cigarettes and should continue to motivate patients to discontinue nicotine use altogether.

Most people assume that electronic cigarettes (e-cigarettes) are safer than conventional tobacco products. Nevertheless, we should not encourage addicted smokers to try “vaping” as an alternative to nicotine replacement therapy, and we should discourage never-smokers from taking up vaping as vigorously as we try to discourage them from taking up smoking.

This article examines the prevailing assumptions and the evidence regarding the safety of e-cigarettes and traditional nicotine replacement therapy.

SMOKING IS DECLINING BUT FAR FROM GONE

While smoking rates have been declining over the past 50 years, the burden of disease attributable to tobacco use remains high. In the United States, it is estimated that nearly 6 million of those currently under the age of 18 will die of tobacco-related illnesses.¹ In the 50 years since the US surgeon general first reported on the health concerns related to tobacco, smoking has claimed the lives of nearly 21 million Americans¹ and continues to kill more than 400,000 every year.²

Even though the risks of smoking are well known, smoking remains one of the most difficult habits to quit. Indeed, about half of all smokers attempt to quit each year, but very few succeed.³

NICOTINE REPLACEMENT: GUM, PATCHES…E-CIGARETTES?

Smoking continues to kill more than 400,000 Americans every year

Nicotine replacement therapy was born out of the thought that, though nicotine is responsible for tobacco’s addictive quality, most tobacco-related disease is attributable to the 7,000 other substances found in tobacco smoke.⁴ Nicotine polacrilex gum was approved by the US Food and Drug Administration (FDA) in 1984, and nicotine transdermal film was approved in 1991.⁵

Nicotine replacement therapy, in the form of patches and gum, has been shown to improve the odds of successfully quitting smoking by a factor of nearly 1.5 to 2.⁶ Nicotine patches and gum were initially prescription medications but became available over the counter in 1996.⁷ They quickly became first-line agents for smoking cessation, and their over-the-counter availability softened any potential concerns about the possible deleterious health consequences of nicotine itself.

E-cigarettes—devices that generate a nicotine vapor that can be inhaled in a fashion that mimics the experience of smoking—were introduced in China in 2004.⁸ By 2012, sales of these devices in the United States had reached $500 million and in 2013 were expected to top $1 billion.^9,10

E-cigarette manufacturers make no therapeutic claims about their products, thus allowing them to escape regulation by the FDA as nicotine replacement therapy. A recent FDA proposal, however, is likely to change their “protected” status.¹¹ Despite the lack of regulation up to this point, patients generally assume that e-cigarettes are just another form of nicotine replacement therapy, even though they contain substances other than nicotine.

WHAT’S IN E-CIGARETTES?

Nicotine, which is bad in itself

E-cigarettes contain nicotine in varying amounts (some cartridges contain none at all). Though nicotine replacement therapy is less harmful than tobacco, nicotine by itself is associated with its own health problems, notably cancer, cardiovascular disease, birth defects (possibly), and poisoning.

Carcinogenesis. Nicotine plays a direct role in carcinogenesis through a variety of mechanisms, including increasing the activity of tumor growth-promoting transcription factors, decreasing apoptosis, and increasing angiogenesis in tumors.¹² Additionally, specific types of nicotinic acetylcholine receptors— eg, alpha 7 receptors, which are stimulated by nicotine—are found in many malignant tumors and are thought to play a role in tumor progression.¹² Blockade of alpha 7 nicotinic acetylcholine receptors has been shown to decrease the growth of certain cancers.¹²

However, these findings were from in vitro studies, and the concerns they raised have not been reflected in in vivo studies. Despite having been on the market for 30 years, nicotine replacement therapy has as yet not been associated with any “real world” increase in cancer risk.

Smoking is one of the leading risk factors for cervical cancer, and nicotine itself may play a contributing role. Nicotine has been shown to increase cellular proliferation in cervical cancer.¹³ Some evidence suggests that it may also play a role in the lymphogenic metastasis of cervical cancer.¹³

Cardiovascular disease. Nicotine has been linked to cardiovascular disease. It directly affects the heart’s rate and rhythm via nicotinic acetylcholine receptors in the peripheral autonomic nervous system. It impairs endothelial-dependent dilation of blood vessels in response to nitric oxide, and this inhibition in the coronary arteries may contribute to smoking-related heart disease.^14,15 Nicotine has also been shown to raise serum cholesterol levels, increase clot formation, and contribute to plaque formation by increasing vascular smooth muscle.¹⁴

Possible teratogenic effect. There is some theoretical concern regarding exposure to nicotine in utero, as nicotinic acetylcholine receptors develop before neurons, and nicotine may therefore interfere with the natural influence of acetylcholine on the development of this system.¹⁴

Consuming one vial of e-cigarette fluid could be fatal

Direct toxicity. Nicotine is toxic at high levels. The overdose potential associated with nicotine is particularly worrisome with e-cigarettes, as the nicotine solution they use is typically supplied in 5-mL, 10-mL, or 20-mL vials that range in concentration from 8.5 to 22.2 mg of nicotine per mL.¹⁶ The fatal single dose range of nicotine has been reported at 30 to 60 mg in adults and 10 mg in children and can be achieved by oral, intravenous, or transdermal absorption,¹⁶ so one vial, if consumed orally, could be fatal.

The number of calls to US poison control centers regarding e-cigarettes has increased, closely paralleling their rise in popularity. In 2010, there were only 30 e-cigarette related calls to poison control centers; in 2011 the number increased to 269, and in 2012 it had reached 459 and included one fatality that was deemed a suicide.^17–19 Even though such toxic nicotine overdoses are rare, physicians should exercise caution and avoid recommending e-cigarettes to individuals with mental confusion, psychotic disorders, or suicidality, who might consume an entire vial.

Possible positive effects? Smoking is one of the worst things that people can do to their body, but the picture is complicated by a few possible positive effects. In the brain, although smoking increases the risk of Alzheimer disease, it is associated with a lower risk of Parkinson disease. In the bowel, it increases the risk of Crohn disease but may decrease the risk of ulcerative colitis. Gahring and Rogers²⁰ pointed out that neuronal nicotinic receptors are present in nonneuronal cells throughout the body and proposed that expression of these receptors may play a role in mediating the consequences of nicotine use, both good and bad. The lesson may be that nicotine is very active in the body, its effects are complicated and still incompletely understood, and therefore we should not encourage people to inhale nicotine products ad lib.

Additives

E-cigarettes typically contain propylene glycol, flavorings, and glycerine. One study that analyzed the additive contents of e-cigarettes found that propylene glycol accounted for 66% of the fluid, glycerine 24%, and flavorings less than 0.1%.²¹ Propylene glycol is the substance typically used in theater fog machines and is used to generate the vapor in e-cigarettes. Other substances such as tobacco-specific nitrosamines and diethylene glycol have also been found in e-cigarettes in small amounts.²²

Propylene glycol, ‘generally recognized as safe’

Propylene glycol has been used in theater fog machines for years—think Phantom of the Opera. It is also widely used as a solvent in many consumer products and pharmaceuticals. The FDA classified it as “generally recognized as safe” on the basis of one study conducted in rats and monkeys over 60 years ago.²³ As other authors have noted, however, a major manufacturer of propylene glycol recommends that exposure to propylene glycol mist be avoided.^24,25 Potential concern over propylene glycol mist was heightened when it was discovered that of all industries, the entertainment business ranked first in terms of work-related asthma symptoms and had the fifth-highest rate of wheezing.^26,27

Studies conducted over the last several decades have raised numerous health concerns about the safety of propylene glycol (Table 1).^26,28–36 The studies of propylene glycol fog are particularly important, as they most closely resemble the route of exposure in e-cigarette users.

Wieslander et al²⁸ exposed 27 volunteers to propylene glycol mist for 1 minute in an aircraft simulator under training conditions. Exposures were high, ranging from 176 to 851 mg/m³ (mean = 309 mg/m³). Four volunteers who developed a cough exhibited evidence of airway obstruction as indicated by a 5% decrease in forced expiratory volume in 1 second (FEV₁), while the rest did not exhibit any change in FEV₁.

Moline et al²⁹ conducted a non-peer-reviewed study for the Actors Equity Association and the League of American Theaters and Producers of 439 actors exposed to theater fog. They found statistically significant evidence of throat and vocal cord inflammation with prolonged peak exposure to glycols and recommended that actors not be exposed to glycol concentrations exceeding 40 mg/m³.

Varughese et al²⁶ conducted a study in 101 volunteers at 19 sites. The mean concentration of glycol-based fog was much lower than that in the studies by Wieslander et al²⁸ and Moline et al,²⁹ at 0.49 mg/m³ (the maximum was 3.22 mg/m³). The investigators concluded that glycol-based fog was associated with deleterious respiratory effects and that employees’ exposure should be limited.

The health issues related to propylene glycol are unique to e-cigarettes compared with nicotine replacement therapy. Unfortunately, the most applicable data available are from studies of persons exposed to theater fog, which involved periodic exposure and likely do not emulate the deep inhalation, multiple times daily, of propylene glycol by e-cigarette smokers. A 2014 review of the chemistry of contaminants in e-cigarettes³⁷ concluded that estimated levels of propylene glycol exposure in e-cigarette users come close to the threshold limit value set by the American Conference of Governmental Industrial Hygienists, and should merit concern.

These studies and real-life experience in the theater, while limited in scope, should give physicians pause and should cause increased awareness of the possibility of e-cigarette-induced pulmonary and upper airway complications. If such complications should occur, discontinuation of vaping should be advised.

Contaminants

The issue of adulterants is common to both e-cigarettes and nicotine replacement therapy. Several unlisted substances have been found in analyzed samples of e-cigarette fluid, including tobacco-specific nitrosamines (TSNAs), diethylene glycol (found in only one e-cigarette cartridge), cotinine, anabasine, myosmine, and beta-nicotyrine.²² The tobacco-specific nitrosamines N´-nitrosonornicotine (NNN), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), N-nitrosoanabasine, and nitrosoanatabine have been found in five e-cigarette cartridge samples from two manufacturers in amounts similar to those found in nicotine replacement products.²²

Goniewicz et al³⁸ tested the vapor generated by 12 e-cigarette brands and found trace amounts of TSNAs. NNN was present in the vapor of eight of the samples in concentrations ranging from 0.8 to 4.3 ng per 150 puffs, and NNK in the vapor of nine of the samples in concentrations ranging from 1.1 to 28.3 ng per 150 puffs. Neither NNN nor NNK was found in blank samples nor with the Nicorette inhalator tested in the same study.³⁸

Because TNSAs can be formed from nicotine and its metabolites, there is also concern that cancer-causing nitrosamines may be formed from nicotine after it is absorbed into the body (ie, endogenously). While endogenous formation of NNK from nicotine has never been demonstrated, endogenous formation of NNN has been seen in some nicotine patch users.³⁹ The presence of these nitrosamines has raised concern that e-cigarettes and nicotine replacement therapy may have carcinogenic potential. The amounts of tobacco-specific nitrosamines found in e-cigarettes are also found in some nicotine replacement products.⁴⁰

Investigators have examined a possible connection between e-cigarettes and potentially carcinogenic carbonyl compounds, including formaldehyde, acetaldehyde, and acrolein. Formaldehyde (a known carcinogen) and acetaldehyde (a potential carcinogen) have been detected in e-cigarette cartridges and vapor.^38,41–43 Acrolein, a mucosal irritant, has been found in e-cigarette vapor.^38,43 Goniewicz et al³⁸ suggested that acrolein may be formed by the heating of the glycerin contained in the e-cigarette solution.

An extensive review of the studies of possible contaminant exposures (including polycylic aromatic hydrocarbons, TSNAs, volatile organic compounds, diethylene glycol, and inorganic compounds) with e-cigarette use according to occupational hygiene standards concluded that there was no cause for concern about increased health risk.³⁷ The study by Goniewicz et al also concluded that using e-cigarettes instead of traditional cigarettes may significantly reduce exposure to some tobacco-specific toxins.³⁸

E-CIGARETTES VS NICOTINE REPLACEMENT

Traditional nicotine replacement therapy products are regulated by the FDA and therefore standardized in terms of their contents. E-cigarettes, on the other hand, are unregulated vehicles for supplying nicotine, and may pose other health risks. One such risk is related to exposure to propylene glycol, which has never been studied under conditions (in terms of mode of delivery, frequency of dosing, and total duration of exposure) that approximate the exposure associated with e-cigarettes. Furthermore, the high concentration of nicotine in e-cigarette fluid poses a real risk of toxicity and potentially fatal overdose.

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine

Nicotine replacement therapy and e-cigarettes both maintain addiction to nicotine if used in a harm-reduction strategy as a maintenance medication. Whether the ongoing nicotine addiction makes it more likely that individuals would switch back and forth between nicotine replacement and tobacco-based products is not clear. Also not known is whether e-cigarettes may serve as the “gateway drug” by which teens enter into nicotine addiction, but we believe that the potential exists, as these products are potentially more appealing in terms of the lack of pungent smell, the perception of safety, and the variety of flavors of e-cigarettes.

The efficacy of nicotine replacement therapy in improving smoking cessation has been reviewed extensively elsewhere³⁷ and is beyond the scope of this article. E-cigarettes may be appealing to many cigarette smokers because they deliver smokeless nicotine, and they more closely emulate the actual experience of smoking compared with traditional nicotine replacement therapy. Though some evidence suggests that e-cigarettes may be modestly effective in helping tobacco smokers quit nicotine, they are not FDA-approved for smoking cessation and are not marketed for that indication.⁴⁴ Medical practitioners should see them for what they are: a new nicotine product with a novel delivery system that is not approved as treatment. Because of the inherent risks involved with e-cigarettes, medical practitioners are best advised to remain neutral on the relative value of e-cigarettes and should continue to motivate patients to discontinue nicotine use altogether.

Insulin therapy is one of the most effective tools clinicians can use to help patients reach their individualized hemoglobin A_1c target. However, decisions about when and how to start insulin therapy have to be individualized to the needs and goals of each patient. Many insulin options are available, one of the most common being the addition of basal insulin to oral antidiabetic drugs. Although patients are often reluctant to start insulin, this reluctance can be overcome through patient education and hands-on training.

Here, we review hemoglobin A_1c targets, factors that determine when to start insulin therapy, and the different regimens that can be used.

MOST PATIENTS EVENTUALLY NEED INSULIN

Type 2 diabetes mellitus is a chronic progressive disease associated with insulin resistance, beta-cell dysfunction, and decreased insulin secretion. Consequently, most patients eventually require insulin therapy to reduce the risk of long-term complications.

The efficacy of therapy can be assessed by measuring hemoglobin A_1c, an important marker of the chronic hyperglycemic state. The hemoglobin A_1c value can be reported as a ratio (%) standardized against the results of the Diabetes Control and Complications Trial,¹ or as International Federation of Clinical Chemistry units (mmol/mol).² Table 1 shows the relationship between hemoglobin A_1c and average glucose values.³

WHAT IS AN APPROPRIATE HEMOGLOBIN A_1c TARGET?

The short answer is, “It depends.”

Currently, the American Association of Clinical Endocrinologists (AACE) supports a hemoglobin A_1c goal of less than 6.5% for otherwise healthy patients but states that the goal should be individualized for patients with concurrent illnesses or at risk of hypoglycemia.⁴

On the other hand, the American Diabetes Association (ADA) recommends a higher hemoglobin A_1c target of less than 7% for most adults with type 2 diabetes mellitus.⁵ This value was shown to be associated with a reduction in the microvascular and macrovascular complications of diabetes.

Yet when three large trials^6–8 recently compared intensive and standard glucose control regimens, tighter glucose control failed to improve cardiovascular outcomes. Moreover, in one of the trials,⁷ patients receiving intensive treatment had a higher rate of all-cause mortality. Details:

• Action in Diabetes and Vascular Disease (ADVANCE): 11,140 patients; average hemoglobin A_1c levels 6.5% vs 7.3%⁶
• Action to Control Cardiovascular Risk in Diabetes (ACCORD): 10,251 patients; average hemoglobin A_1c levels 6.4% vs 7.5%⁷
• Veterans Affairs Diabetes Trial (VADT): 1,791 patients; average hemoglobin A_1c levels 6.9% vs 8.4%.⁸

Similarly, a 2013 Cochrane review⁹ that included 28 randomized controlled trials concluded that intensive control (in 18,717 patients) did not decrease all-cause and cardiovascular mortality rates compared with traditional glucose control (in 16,195 patients), and it increased the risk of hypoglycemia and serious adverse events.

The AACE and ADA are moving away from one-size-fits-all and toward individualized recommendations

As a result, the ADA⁵ states that a hemoglobin A_1c target less than 6.5% is optional for patients with a long life expectancy, short duration of diabetes, low risk of hypoglycemia, and no significant cardiovascular disease. The ADA further defines a hemoglobin A_1c goal of less than 8% for patients with a history of severe hypoglycemia, limited life expectancy, advanced microvascular or macrovascular complications, extensive comorbid conditions, and long-standing diabetes.

Therefore, the AACE and ADA are moving away from “one-size-fits-all” goals and toward individualizing their recommendations.

WHEN SHOULD INSULIN BE STARTED?

Physicians should consider the needs and preferences of each patient and individualize the treatment. The most recent recommendations from the ADA⁵ stress the importance of a patient-centered approach, with multiple factors taken into account. These include the patient’s attitude, expected compliance with treatment, risk of hypoglycemia, disease duration, life expectancy, and comorbidities, and the side effects of oral medications and insulin.

Compared with previous guidelines, there are fewer rules on how and when to start insulin therapy. But absolute and relative indications for insulin therapy should be considered in patients with the following:

Absolute indications for insulin

• Ketoacidosis or catabolic symptoms, including ketonuria
• Newly diagnosed type 2 diabetes with pronounced hyperglycemia (glucose ≥ 300 mg/dL or hemoglobin A_1c ≥ 10.0%) with or without severe symptoms, including weight loss, polyuria, or polydipsia¹⁰
• Uncontrolled type 2 diabetes mellitus despite using one, two, or more oral antidiabetic drugs or glucagon-like peptide 1 (GLP-1) receptor agonists
• Gestational diabetes
• Preference for insulin.

Relative indications for insulin

• Hospitalized for surgery or acute illnesses
• Advanced renal or hepatic disease
• Inability to afford the cost or tolerate the side effects of oral antidiabetic drugs and GLP-1 receptor agonists.

Depending on the situation, blood glucose is measured fasting, before meals, or after meals after initiating or adjusting insulin regimens (Table 2).

WHAT ARE THE INSULIN REGIMENS?

Basal insulin

In the early stages of type 2 diabetes, metformin alone or in combination with another oral antidiabetic drug or with a GLP-1 receptor agonist is often used along with healthy eating, weight control, and increased physical activity.

When the target hemoglobin A_1c cannot be achieved with one or two noninsulin drugs, the ADA suggests basal insulin be added to metformin or a two-medication regimen that includes metformin (Table 3). However, recent evidence suggests that combining a GLP-1 receptor agonist with basal insulin, in a regimen without metformin, is safe and improves glycemic control without hypoglycemia or weight gain.¹¹

While a total daily dose of insulin of 0.1 to 0.2 units/kg could be initially used in patients with a hemoglobin A_1c level less than 8%, a higher dose of 0.2 to 0.3 units/kg is required if the hemoglobin A_1c level is between 8% and 10%. The dose can be titrated once or twice weekly if the fasting glucose is above the target level (usually < 130 mg/dL). If hypoglycemia develops (glucose < 70 mg/dL), the insulin dose should be reduced by 10% to 20%.¹⁰

Available basal insulins include glargine, detemir, and neutral protamine Hagedorn (NPH) (Table 4).^12–14 Because glargine and detemir offer better pharmacokinetic properties, less variability in response, and less risk of hypoglycemia, they are preferred over NPH. Glargine has a relatively constant plasma concentration over 24 hours, allowing once-daily dosing at any time during the day (Figure 1).¹⁵ The dose should be taken at the same time every day. Detemir and NPH are usually taken once or twice daily.

Adapted from Hirsch IB. Insulin analogues. N Engl J Med 2005; 352:174-183. Copyright 2005, Massachusetts Medical Society.

Patients treated once daily should take the dose with the evening meal or at bedtime. Patients who require a twice-daily regimen can take the first dose with breakfast and the second one with the evening meal, at bedtime, or 12 hours after the morning dose.

The randomized Treat-to-Target trial,¹⁶ in 756 patients, showed that both glargine and NPH, when added to oral therapy in patients with type 2 diabetes, achieve the target hemoglobin A_1c, but NPH is associated with more episodes of nocturnal hypoglycemia. Similar results were found when NPH was compared with detemir insulin.¹⁷

A Cochrane review¹⁸ suggested that glargine and detemir are similar in efficacy and safety. However, detemir often needs to be injected twice daily, in a higher dose, and is associated with less weight gain. Furthermore, a meta-analysis of 46 randomized clinical trials¹⁹ showed that the weight increase at 1 year is less in patients treated with basal than with twice-daily or prandial regimens.

The ADA suggests basal insulin be added to metformin alone or a regimen that includes metformin

A noninterventional longitudinal study²⁰ in 2,179 patients newly started on insulin showed that the mean weight increase at 1 year was 1.78 kg, and 24% of patients gained more than 5 kg. However, the factors independently associated with the weight gain were a higher hemoglobin A_1c at baseline, a higher insulin dose at baseline and at 1 year, and a lower baseline body mass index, but not the type of insulin regimen.

Currently, a new class of ultralong-acting basal insulins is being studied. Insulins in this class are approved in other countries, but the US Food and Drug Administration requires additional data for approval. Ultralong-acting insulins are expected to reduce the risk of hypoglycemia, specifically the risk of nocturnal episodes. Also, given their longer duration of action and stable steady-state pharmacokinetics, they will offer flexibility in the dose timing.²¹

Basal-bolus regimens

Basal insulin often does not control postprandial hyperglycemia. The need for multiple doses of insulin (including one or more preprandial doses) is suggested by postprandial glucose values above target (usually > 180 mg/dL) or by a hemoglobin A_1c above goal despite well-controlled fasting glucose levels. This usually becomes evident when the total daily dose of basal insulin exceeds 0.5 units/kg. Patients newly diagnosed with diabetes who have a hemoglobin A_1c higher than 10% may also respond better to an initial basal-bolus regimen.

Available bolus insulins include lispro, aspart, glulisine, regular insulin, and the newly approved Technosphere inhaled regular insulin (Table 4).^12–14 They can be taken before each meal, and the total bolus dose usually represents 50% of the total daily dose.²² Rapid-acting insulins have faster onset, shorter duration of action, and more predictable pharmacokinetics, which makes them preferable to regular insulin (Figure 1).¹⁵ Inhaled insulin is another option, but it is contraindicated in patients with chronic obstructive pulmonary disease or asthma because of the increased risk of acute bronchospasm.¹²

Alternatively, the transition to a basal-bolus regimen can be accomplished with a single dose of bolus insulin before the main meal, using a dose that represents approximately 10% of the total daily dose. Additional bolus doses can be added later based on the glycemic control. The adjustment of the preprandial insulin dose is done once or twice weekly, based on the postprandial glucose levels.¹⁰

Premixed combinations of long- and short-acting insulins in ratios of 50% to 50%, 70% to 30%, or 75% to 25% can be considered in patients who cannot adhere to a complex insulin regimen. A propensity-matched comparison of different insulin regimens (basal, premixed, mealtime plus basal, and mealtime) in patients with type 2 diabetes revealed that the hemoglobin A_1c reduction was similar between the different groups.²³ However, the number of hypoglycemic episodes was higher in the premixed insulin group, and the weight gain was less in the basal insulin group.

While premixed insulins require fewer injections, they do not provide dosing flexibility. In other words, dose adjustments for premixed insulins lead to increases in both basal and bolus amounts even though a dose adjustment is needed for only one insulin type. Thus, this is a common reason for increased hypoglycemic episodes.

Continuous subcutaneous insulin infusion

Patients who are engaged in their care are more likely to succeed in their treatment

A meta-analysis showed that continuous subcutaneous insulin infusion (ie, use of an insulin pump) was similar to intensive therapy with multiple daily insulin injections in terms of glycemic control and hypoglycemia.²⁴ Since both options can lead to similar glucose control, additional factors to consider when initiating insulin infusion include lifestyle and technical expertise. Some patients may or may not prefer having a pump attached for nearly all daily activities. Additionally, this type of therapy is complex and requires significant training to ensure efficacy and safety.²⁵

WHAT IS THE COST OF INSULIN THERAPY?

A final factor to keep in mind when initiating insulin is cost (Table 4).^12–14 Asking patients to check their prescription insurance formulary is important to ensure that an affordable option is selected. If patients do not have prescription insurance, medication assistance programs could be an option. However, if a patient is considering an insulin pump, insurance coverage is essential. Depending on the manufacturer, insulin pumps cost about $6,000 to $7,000, and the additional monthly supplies for the pump are also expensive.

If patients are engaged when considering and selecting insulin therapy, the likelihood of treatment success is greater.^26–28

Insulin therapy is one of the most effective tools clinicians can use to help patients reach their individualized hemoglobin A_1c target. However, decisions about when and how to start insulin therapy have to be individualized to the needs and goals of each patient. Many insulin options are available, one of the most common being the addition of basal insulin to oral antidiabetic drugs. Although patients are often reluctant to start insulin, this reluctance can be overcome through patient education and hands-on training.

Here, we review hemoglobin A_1c targets, factors that determine when to start insulin therapy, and the different regimens that can be used.

MOST PATIENTS EVENTUALLY NEED INSULIN

Type 2 diabetes mellitus is a chronic progressive disease associated with insulin resistance, beta-cell dysfunction, and decreased insulin secretion. Consequently, most patients eventually require insulin therapy to reduce the risk of long-term complications.

The efficacy of therapy can be assessed by measuring hemoglobin A_1c, an important marker of the chronic hyperglycemic state. The hemoglobin A_1c value can be reported as a ratio (%) standardized against the results of the Diabetes Control and Complications Trial,¹ or as International Federation of Clinical Chemistry units (mmol/mol).² Table 1 shows the relationship between hemoglobin A_1c and average glucose values.³

WHAT IS AN APPROPRIATE HEMOGLOBIN A_1c TARGET?

The short answer is, “It depends.”

Currently, the American Association of Clinical Endocrinologists (AACE) supports a hemoglobin A_1c goal of less than 6.5% for otherwise healthy patients but states that the goal should be individualized for patients with concurrent illnesses or at risk of hypoglycemia.⁴

On the other hand, the American Diabetes Association (ADA) recommends a higher hemoglobin A_1c target of less than 7% for most adults with type 2 diabetes mellitus.⁵ This value was shown to be associated with a reduction in the microvascular and macrovascular complications of diabetes.

Yet when three large trials^6–8 recently compared intensive and standard glucose control regimens, tighter glucose control failed to improve cardiovascular outcomes. Moreover, in one of the trials,⁷ patients receiving intensive treatment had a higher rate of all-cause mortality. Details:

• Action in Diabetes and Vascular Disease (ADVANCE): 11,140 patients; average hemoglobin A_1c levels 6.5% vs 7.3%⁶
• Action to Control Cardiovascular Risk in Diabetes (ACCORD): 10,251 patients; average hemoglobin A_1c levels 6.4% vs 7.5%⁷
• Veterans Affairs Diabetes Trial (VADT): 1,791 patients; average hemoglobin A_1c levels 6.9% vs 8.4%.⁸

Similarly, a 2013 Cochrane review⁹ that included 28 randomized controlled trials concluded that intensive control (in 18,717 patients) did not decrease all-cause and cardiovascular mortality rates compared with traditional glucose control (in 16,195 patients), and it increased the risk of hypoglycemia and serious adverse events.

The AACE and ADA are moving away from one-size-fits-all and toward individualized recommendations

As a result, the ADA⁵ states that a hemoglobin A_1c target less than 6.5% is optional for patients with a long life expectancy, short duration of diabetes, low risk of hypoglycemia, and no significant cardiovascular disease. The ADA further defines a hemoglobin A_1c goal of less than 8% for patients with a history of severe hypoglycemia, limited life expectancy, advanced microvascular or macrovascular complications, extensive comorbid conditions, and long-standing diabetes.

Therefore, the AACE and ADA are moving away from “one-size-fits-all” goals and toward individualizing their recommendations.

WHEN SHOULD INSULIN BE STARTED?

Physicians should consider the needs and preferences of each patient and individualize the treatment. The most recent recommendations from the ADA⁵ stress the importance of a patient-centered approach, with multiple factors taken into account. These include the patient’s attitude, expected compliance with treatment, risk of hypoglycemia, disease duration, life expectancy, and comorbidities, and the side effects of oral medications and insulin.

Compared with previous guidelines, there are fewer rules on how and when to start insulin therapy. But absolute and relative indications for insulin therapy should be considered in patients with the following:

Absolute indications for insulin

• Ketoacidosis or catabolic symptoms, including ketonuria
• Newly diagnosed type 2 diabetes with pronounced hyperglycemia (glucose ≥ 300 mg/dL or hemoglobin A_1c ≥ 10.0%) with or without severe symptoms, including weight loss, polyuria, or polydipsia¹⁰
• Uncontrolled type 2 diabetes mellitus despite using one, two, or more oral antidiabetic drugs or glucagon-like peptide 1 (GLP-1) receptor agonists
• Gestational diabetes
• Preference for insulin.

Relative indications for insulin

• Hospitalized for surgery or acute illnesses
• Advanced renal or hepatic disease
• Inability to afford the cost or tolerate the side effects of oral antidiabetic drugs and GLP-1 receptor agonists.

Depending on the situation, blood glucose is measured fasting, before meals, or after meals after initiating or adjusting insulin regimens (Table 2).

WHAT ARE THE INSULIN REGIMENS?

Basal insulin

In the early stages of type 2 diabetes, metformin alone or in combination with another oral antidiabetic drug or with a GLP-1 receptor agonist is often used along with healthy eating, weight control, and increased physical activity.

When the target hemoglobin A_1c cannot be achieved with one or two noninsulin drugs, the ADA suggests basal insulin be added to metformin or a two-medication regimen that includes metformin (Table 3). However, recent evidence suggests that combining a GLP-1 receptor agonist with basal insulin, in a regimen without metformin, is safe and improves glycemic control without hypoglycemia or weight gain.¹¹

While a total daily dose of insulin of 0.1 to 0.2 units/kg could be initially used in patients with a hemoglobin A_1c level less than 8%, a higher dose of 0.2 to 0.3 units/kg is required if the hemoglobin A_1c level is between 8% and 10%. The dose can be titrated once or twice weekly if the fasting glucose is above the target level (usually < 130 mg/dL). If hypoglycemia develops (glucose < 70 mg/dL), the insulin dose should be reduced by 10% to 20%.¹⁰

Available basal insulins include glargine, detemir, and neutral protamine Hagedorn (NPH) (Table 4).^12–14 Because glargine and detemir offer better pharmacokinetic properties, less variability in response, and less risk of hypoglycemia, they are preferred over NPH. Glargine has a relatively constant plasma concentration over 24 hours, allowing once-daily dosing at any time during the day (Figure 1).¹⁵ The dose should be taken at the same time every day. Detemir and NPH are usually taken once or twice daily.

Adapted from Hirsch IB. Insulin analogues. N Engl J Med 2005; 352:174-183. Copyright 2005, Massachusetts Medical Society.

Patients treated once daily should take the dose with the evening meal or at bedtime. Patients who require a twice-daily regimen can take the first dose with breakfast and the second one with the evening meal, at bedtime, or 12 hours after the morning dose.

The randomized Treat-to-Target trial,¹⁶ in 756 patients, showed that both glargine and NPH, when added to oral therapy in patients with type 2 diabetes, achieve the target hemoglobin A_1c, but NPH is associated with more episodes of nocturnal hypoglycemia. Similar results were found when NPH was compared with detemir insulin.¹⁷

A Cochrane review¹⁸ suggested that glargine and detemir are similar in efficacy and safety. However, detemir often needs to be injected twice daily, in a higher dose, and is associated with less weight gain. Furthermore, a meta-analysis of 46 randomized clinical trials¹⁹ showed that the weight increase at 1 year is less in patients treated with basal than with twice-daily or prandial regimens.

The ADA suggests basal insulin be added to metformin alone or a regimen that includes metformin

A noninterventional longitudinal study²⁰ in 2,179 patients newly started on insulin showed that the mean weight increase at 1 year was 1.78 kg, and 24% of patients gained more than 5 kg. However, the factors independently associated with the weight gain were a higher hemoglobin A_1c at baseline, a higher insulin dose at baseline and at 1 year, and a lower baseline body mass index, but not the type of insulin regimen.

Currently, a new class of ultralong-acting basal insulins is being studied. Insulins in this class are approved in other countries, but the US Food and Drug Administration requires additional data for approval. Ultralong-acting insulins are expected to reduce the risk of hypoglycemia, specifically the risk of nocturnal episodes. Also, given their longer duration of action and stable steady-state pharmacokinetics, they will offer flexibility in the dose timing.²¹

Basal-bolus regimens

Basal insulin often does not control postprandial hyperglycemia. The need for multiple doses of insulin (including one or more preprandial doses) is suggested by postprandial glucose values above target (usually > 180 mg/dL) or by a hemoglobin A_1c above goal despite well-controlled fasting glucose levels. This usually becomes evident when the total daily dose of basal insulin exceeds 0.5 units/kg. Patients newly diagnosed with diabetes who have a hemoglobin A_1c higher than 10% may also respond better to an initial basal-bolus regimen.

Available bolus insulins include lispro, aspart, glulisine, regular insulin, and the newly approved Technosphere inhaled regular insulin (Table 4).^12–14 They can be taken before each meal, and the total bolus dose usually represents 50% of the total daily dose.²² Rapid-acting insulins have faster onset, shorter duration of action, and more predictable pharmacokinetics, which makes them preferable to regular insulin (Figure 1).¹⁵ Inhaled insulin is another option, but it is contraindicated in patients with chronic obstructive pulmonary disease or asthma because of the increased risk of acute bronchospasm.¹²

Alternatively, the transition to a basal-bolus regimen can be accomplished with a single dose of bolus insulin before the main meal, using a dose that represents approximately 10% of the total daily dose. Additional bolus doses can be added later based on the glycemic control. The adjustment of the preprandial insulin dose is done once or twice weekly, based on the postprandial glucose levels.¹⁰

Premixed combinations of long- and short-acting insulins in ratios of 50% to 50%, 70% to 30%, or 75% to 25% can be considered in patients who cannot adhere to a complex insulin regimen. A propensity-matched comparison of different insulin regimens (basal, premixed, mealtime plus basal, and mealtime) in patients with type 2 diabetes revealed that the hemoglobin A_1c reduction was similar between the different groups.²³ However, the number of hypoglycemic episodes was higher in the premixed insulin group, and the weight gain was less in the basal insulin group.

While premixed insulins require fewer injections, they do not provide dosing flexibility. In other words, dose adjustments for premixed insulins lead to increases in both basal and bolus amounts even though a dose adjustment is needed for only one insulin type. Thus, this is a common reason for increased hypoglycemic episodes.

Continuous subcutaneous insulin infusion

Patients who are engaged in their care are more likely to succeed in their treatment

A meta-analysis showed that continuous subcutaneous insulin infusion (ie, use of an insulin pump) was similar to intensive therapy with multiple daily insulin injections in terms of glycemic control and hypoglycemia.²⁴ Since both options can lead to similar glucose control, additional factors to consider when initiating insulin infusion include lifestyle and technical expertise. Some patients may or may not prefer having a pump attached for nearly all daily activities. Additionally, this type of therapy is complex and requires significant training to ensure efficacy and safety.²⁵

WHAT IS THE COST OF INSULIN THERAPY?

A final factor to keep in mind when initiating insulin is cost (Table 4).^12–14 Asking patients to check their prescription insurance formulary is important to ensure that an affordable option is selected. If patients do not have prescription insurance, medication assistance programs could be an option. However, if a patient is considering an insulin pump, insurance coverage is essential. Depending on the manufacturer, insulin pumps cost about $6,000 to $7,000, and the additional monthly supplies for the pump are also expensive.

If patients are engaged when considering and selecting insulin therapy, the likelihood of treatment success is greater.^26–28

Insulin therapy is one of the most effective tools clinicians can use to help patients reach their individualized hemoglobin A_1c target. However, decisions about when and how to start insulin therapy have to be individualized to the needs and goals of each patient. Many insulin options are available, one of the most common being the addition of basal insulin to oral antidiabetic drugs. Although patients are often reluctant to start insulin, this reluctance can be overcome through patient education and hands-on training.

Here, we review hemoglobin A_1c targets, factors that determine when to start insulin therapy, and the different regimens that can be used.

MOST PATIENTS EVENTUALLY NEED INSULIN

Type 2 diabetes mellitus is a chronic progressive disease associated with insulin resistance, beta-cell dysfunction, and decreased insulin secretion. Consequently, most patients eventually require insulin therapy to reduce the risk of long-term complications.

The efficacy of therapy can be assessed by measuring hemoglobin A_1c, an important marker of the chronic hyperglycemic state. The hemoglobin A_1c value can be reported as a ratio (%) standardized against the results of the Diabetes Control and Complications Trial,¹ or as International Federation of Clinical Chemistry units (mmol/mol).² Table 1 shows the relationship between hemoglobin A_1c and average glucose values.³

WHAT IS AN APPROPRIATE HEMOGLOBIN A_1c TARGET?

The short answer is, “It depends.”

Currently, the American Association of Clinical Endocrinologists (AACE) supports a hemoglobin A_1c goal of less than 6.5% for otherwise healthy patients but states that the goal should be individualized for patients with concurrent illnesses or at risk of hypoglycemia.⁴

On the other hand, the American Diabetes Association (ADA) recommends a higher hemoglobin A_1c target of less than 7% for most adults with type 2 diabetes mellitus.⁵ This value was shown to be associated with a reduction in the microvascular and macrovascular complications of diabetes.

Yet when three large trials^6–8 recently compared intensive and standard glucose control regimens, tighter glucose control failed to improve cardiovascular outcomes. Moreover, in one of the trials,⁷ patients receiving intensive treatment had a higher rate of all-cause mortality. Details:

• Action in Diabetes and Vascular Disease (ADVANCE): 11,140 patients; average hemoglobin A_1c levels 6.5% vs 7.3%⁶
• Action to Control Cardiovascular Risk in Diabetes (ACCORD): 10,251 patients; average hemoglobin A_1c levels 6.4% vs 7.5%⁷
• Veterans Affairs Diabetes Trial (VADT): 1,791 patients; average hemoglobin A_1c levels 6.9% vs 8.4%.⁸

Similarly, a 2013 Cochrane review⁹ that included 28 randomized controlled trials concluded that intensive control (in 18,717 patients) did not decrease all-cause and cardiovascular mortality rates compared with traditional glucose control (in 16,195 patients), and it increased the risk of hypoglycemia and serious adverse events.

The AACE and ADA are moving away from one-size-fits-all and toward individualized recommendations

As a result, the ADA⁵ states that a hemoglobin A_1c target less than 6.5% is optional for patients with a long life expectancy, short duration of diabetes, low risk of hypoglycemia, and no significant cardiovascular disease. The ADA further defines a hemoglobin A_1c goal of less than 8% for patients with a history of severe hypoglycemia, limited life expectancy, advanced microvascular or macrovascular complications, extensive comorbid conditions, and long-standing diabetes.

Therefore, the AACE and ADA are moving away from “one-size-fits-all” goals and toward individualizing their recommendations.

WHEN SHOULD INSULIN BE STARTED?

Physicians should consider the needs and preferences of each patient and individualize the treatment. The most recent recommendations from the ADA⁵ stress the importance of a patient-centered approach, with multiple factors taken into account. These include the patient’s attitude, expected compliance with treatment, risk of hypoglycemia, disease duration, life expectancy, and comorbidities, and the side effects of oral medications and insulin.

Compared with previous guidelines, there are fewer rules on how and when to start insulin therapy. But absolute and relative indications for insulin therapy should be considered in patients with the following:

Absolute indications for insulin

• Ketoacidosis or catabolic symptoms, including ketonuria
• Newly diagnosed type 2 diabetes with pronounced hyperglycemia (glucose ≥ 300 mg/dL or hemoglobin A_1c ≥ 10.0%) with or without severe symptoms, including weight loss, polyuria, or polydipsia¹⁰
• Uncontrolled type 2 diabetes mellitus despite using one, two, or more oral antidiabetic drugs or glucagon-like peptide 1 (GLP-1) receptor agonists
• Gestational diabetes
• Preference for insulin.

Relative indications for insulin

• Hospitalized for surgery or acute illnesses
• Advanced renal or hepatic disease
• Inability to afford the cost or tolerate the side effects of oral antidiabetic drugs and GLP-1 receptor agonists.

Depending on the situation, blood glucose is measured fasting, before meals, or after meals after initiating or adjusting insulin regimens (Table 2).

WHAT ARE THE INSULIN REGIMENS?

Basal insulin

In the early stages of type 2 diabetes, metformin alone or in combination with another oral antidiabetic drug or with a GLP-1 receptor agonist is often used along with healthy eating, weight control, and increased physical activity.

When the target hemoglobin A_1c cannot be achieved with one or two noninsulin drugs, the ADA suggests basal insulin be added to metformin or a two-medication regimen that includes metformin (Table 3). However, recent evidence suggests that combining a GLP-1 receptor agonist with basal insulin, in a regimen without metformin, is safe and improves glycemic control without hypoglycemia or weight gain.¹¹

While a total daily dose of insulin of 0.1 to 0.2 units/kg could be initially used in patients with a hemoglobin A_1c level less than 8%, a higher dose of 0.2 to 0.3 units/kg is required if the hemoglobin A_1c level is between 8% and 10%. The dose can be titrated once or twice weekly if the fasting glucose is above the target level (usually < 130 mg/dL). If hypoglycemia develops (glucose < 70 mg/dL), the insulin dose should be reduced by 10% to 20%.¹⁰

Available basal insulins include glargine, detemir, and neutral protamine Hagedorn (NPH) (Table 4).^12–14 Because glargine and detemir offer better pharmacokinetic properties, less variability in response, and less risk of hypoglycemia, they are preferred over NPH. Glargine has a relatively constant plasma concentration over 24 hours, allowing once-daily dosing at any time during the day (Figure 1).¹⁵ The dose should be taken at the same time every day. Detemir and NPH are usually taken once or twice daily.

Adapted from Hirsch IB. Insulin analogues. N Engl J Med 2005; 352:174-183. Copyright 2005, Massachusetts Medical Society.

Patients treated once daily should take the dose with the evening meal or at bedtime. Patients who require a twice-daily regimen can take the first dose with breakfast and the second one with the evening meal, at bedtime, or 12 hours after the morning dose.

The randomized Treat-to-Target trial,¹⁶ in 756 patients, showed that both glargine and NPH, when added to oral therapy in patients with type 2 diabetes, achieve the target hemoglobin A_1c, but NPH is associated with more episodes of nocturnal hypoglycemia. Similar results were found when NPH was compared with detemir insulin.¹⁷

A Cochrane review¹⁸ suggested that glargine and detemir are similar in efficacy and safety. However, detemir often needs to be injected twice daily, in a higher dose, and is associated with less weight gain. Furthermore, a meta-analysis of 46 randomized clinical trials¹⁹ showed that the weight increase at 1 year is less in patients treated with basal than with twice-daily or prandial regimens.

The ADA suggests basal insulin be added to metformin alone or a regimen that includes metformin

A noninterventional longitudinal study²⁰ in 2,179 patients newly started on insulin showed that the mean weight increase at 1 year was 1.78 kg, and 24% of patients gained more than 5 kg. However, the factors independently associated with the weight gain were a higher hemoglobin A_1c at baseline, a higher insulin dose at baseline and at 1 year, and a lower baseline body mass index, but not the type of insulin regimen.

Currently, a new class of ultralong-acting basal insulins is being studied. Insulins in this class are approved in other countries, but the US Food and Drug Administration requires additional data for approval. Ultralong-acting insulins are expected to reduce the risk of hypoglycemia, specifically the risk of nocturnal episodes. Also, given their longer duration of action and stable steady-state pharmacokinetics, they will offer flexibility in the dose timing.²¹

Basal-bolus regimens

Basal insulin often does not control postprandial hyperglycemia. The need for multiple doses of insulin (including one or more preprandial doses) is suggested by postprandial glucose values above target (usually > 180 mg/dL) or by a hemoglobin A_1c above goal despite well-controlled fasting glucose levels. This usually becomes evident when the total daily dose of basal insulin exceeds 0.5 units/kg. Patients newly diagnosed with diabetes who have a hemoglobin A_1c higher than 10% may also respond better to an initial basal-bolus regimen.

Available bolus insulins include lispro, aspart, glulisine, regular insulin, and the newly approved Technosphere inhaled regular insulin (Table 4).^12–14 They can be taken before each meal, and the total bolus dose usually represents 50% of the total daily dose.²² Rapid-acting insulins have faster onset, shorter duration of action, and more predictable pharmacokinetics, which makes them preferable to regular insulin (Figure 1).¹⁵ Inhaled insulin is another option, but it is contraindicated in patients with chronic obstructive pulmonary disease or asthma because of the increased risk of acute bronchospasm.¹²

Alternatively, the transition to a basal-bolus regimen can be accomplished with a single dose of bolus insulin before the main meal, using a dose that represents approximately 10% of the total daily dose. Additional bolus doses can be added later based on the glycemic control. The adjustment of the preprandial insulin dose is done once or twice weekly, based on the postprandial glucose levels.¹⁰

Premixed combinations of long- and short-acting insulins in ratios of 50% to 50%, 70% to 30%, or 75% to 25% can be considered in patients who cannot adhere to a complex insulin regimen. A propensity-matched comparison of different insulin regimens (basal, premixed, mealtime plus basal, and mealtime) in patients with type 2 diabetes revealed that the hemoglobin A_1c reduction was similar between the different groups.²³ However, the number of hypoglycemic episodes was higher in the premixed insulin group, and the weight gain was less in the basal insulin group.

While premixed insulins require fewer injections, they do not provide dosing flexibility. In other words, dose adjustments for premixed insulins lead to increases in both basal and bolus amounts even though a dose adjustment is needed for only one insulin type. Thus, this is a common reason for increased hypoglycemic episodes.

Continuous subcutaneous insulin infusion

Patients who are engaged in their care are more likely to succeed in their treatment

A meta-analysis showed that continuous subcutaneous insulin infusion (ie, use of an insulin pump) was similar to intensive therapy with multiple daily insulin injections in terms of glycemic control and hypoglycemia.²⁴ Since both options can lead to similar glucose control, additional factors to consider when initiating insulin infusion include lifestyle and technical expertise. Some patients may or may not prefer having a pump attached for nearly all daily activities. Additionally, this type of therapy is complex and requires significant training to ensure efficacy and safety.²⁵

WHAT IS THE COST OF INSULIN THERAPY?

A final factor to keep in mind when initiating insulin is cost (Table 4).^12–14 Asking patients to check their prescription insurance formulary is important to ensure that an affordable option is selected. If patients do not have prescription insurance, medication assistance programs could be an option. However, if a patient is considering an insulin pump, insurance coverage is essential. Depending on the manufacturer, insulin pumps cost about $6,000 to $7,000, and the additional monthly supplies for the pump are also expensive.

If patients are engaged when considering and selecting insulin therapy, the likelihood of treatment success is greater.^26–28

The thiazide diuretic hydrochlorothiazide and the thiazidelike diuretic chlorthalidone are two old drugs that are still useful. Although similar, they differ in important ways still not fully appreciated more than a half century after they were introduced.

Most hypertension guidelines recommend thiazide diuretics as one of the classes of agents that can be used either as initial antihypertensive drug therapy or as part of combination therapy.^1–3

In the United States, hydrochlorothiazide is used more often than chlorthalidone, but many clinical trials of antihypertensive therapy have used chlorthalidone.^4,5 In recent years, particularly after the publication of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), interest in chlorthalidone has been increasing, and new data are now available comparing these two diuretics.⁶ While current US guidelines do not recommend one over the other, British guidelines prefer chlorthalidone.⁷

This review summarizes the data comparing the two drugs’ pharmacology, antihypertensive effect, and impact on clinical outcomes to help guide clinicians in choosing antihypertensive drug therapy.

PHARMACOLOGY AND MECHANISM OF ACTION

Many of the differences in effectiveness and adverse effects of hydrochlorothiazide and chlorthalidone are thought to be due to their different pharmacodynamic and pharmacokinetic effects.

Pharmacodynamic effects

Hydrochlorothiazide and chlorthalidone differ significantly in chemical structure (Figure 1), but both contain a sulfonamide group that inhibits carbonic anhydrase activity, which may be associated with lower vascular contractility. Both drugs are concentrated in the kidney and secreted into the tubular lumen⁸; therefore, their therapeutic diuretic effects are often achieved with relatively low plasma concentrations.

Both drugs inhibit the sodium-chloride cotransporter in the luminal membrane of the distal convoluted tubule of the ascending loop of Henle, leading to a modest natriuresis and diuresis. The exact mechanism by which they lower blood pressure is not known: while the initial response is from diuresis and volume changes, long-term reduction in blood pressure is through uncertain mechanisms. In addition, chlorthalidone may have beneficial effects on endothelial function and oxidative stress.^9,10

Both drugs also increase secretion of potassium and hydrogen ions and promote increased reabsorption of calcium through increased expression of a sodium-calcium exchange channel.⁸ Chlorthalidone may cause more inhibition of carbonic anhydrase than hydrochlorothiazide, which can lead to lower intracellular pH and cell volume. This effect may in part explain a pleiotropic effect of chlorthalidone, ie, inhibition of platelet function, which in turn may contribute to this drug’s beneficial effect on cardiovascular outcomes.⁹

Pharmacokinetic differences

Hydrochlorothiazide and chlorthalidone have important differences in their pharmacokinetic properties (Table 1).¹¹

Hydrochlorothiazide has its onset of action in about 2 hours, and it reaches its peak in 4 to 6 hours. Though its duration of action is short—up to 12 hours—its pharmacodynamic response can be much longer than predicted by its kinetics, allowing once-daily dosing.⁸

Chlorthalidone has a longer duration of action than hydrochlorothiazide. This may be because it has a very high volume of distribution, since it is taken up into red blood cells and is bound to carbonic anhydrase.¹² This may result in a “drug reservoir” that keeps drug levels higher for a longer time.¹³ Its long duration of action makes it a favorable choice for patients who have difficulty adhering to medication instructions. In addition, a missed dose is unlikely to have a “rebound” effect like that seen with some other antihypertensive agents. However, both chlorthalidone and hydrochlorothiazide are effective if taken once daily.

BLOOD PRESSURE-LOWERING

Both hydrochlorothiazide and chlorthalidone are effective antihypertensive agents. Table 2 summarizes findings from studies that evaluated their blood pressure-lowering effect at various doses.^14–33 However, relatively few studies have directly compared these two agents’ effects on blood pressure.

Ernst et al,³⁴ in a small study (but probably the best one to address this issue), compared chlorthalidone 12.5 mg/day (force-titrated to 25 mg/day) and hydrochlorothiazide 25 mg/day (force-titrated to 50 mg/day) in untreated hypertensive patients. After 8 weeks, ambulatory blood pressure monitoring indicated a greater reduction from baseline in systolic blood pressure with chlorthalidone 25 mg/day than with hydrochlorothiazide 50 mg/day (24-hour mean –12.4 vs –7.4 mm Hg, P = .05). Interestingly, the change in nighttime blood pressure was greater in the chlorthalidone group (–13.5 mm Hg) than in the hydrochlorothiazide group (–6.4 mm Hg; P = .009). These data suggest that at the doses studied, chlorthalidone is more effective than hydrochlorothiazide in lowering systolic blood pressure.

Bakris et al,³⁵ using a different study design, compared the single-pill combination of azilsartan medoxomil and chlorthalidone vs coadministration of azilsartan medoxomil and hydrochlorothiazide in participants with stage 2 primary hypertension (≥ 160/100 mm Hg). Systolic blood pressure, as measured in the clinic, declined more with the chlorthalidone combination (–35.1 mm Hg) than with the hydrochlorothiazide combination (–29.5 mm Hg, mean difference –5.6 mm Hg, P < .001).

Meta-analyses also support the conclusion that chlorthalidone is more potent than hydrochlorothiazide in lowering blood pressure.^35,36 Several studies have shown that chlorthalidone at the same dose is 1.5 to 2 times as potent as hydrochlorothiazide.^33,36,37 Therefore, for clinical purposes, it is reasonable to consider chlorthalidone 12.5 mg daily as similar to 25 mg of hydrochlorothiazide daily.

ADVERSE EFFECTS

Electrolyte disturbances are a common adverse effect of thiazide diuretics.

Hypokalemia. All thiazide diuretics cause potassium wasting. The frequency of hypokalemia depends on the dose, frequency of administration, diet, and other pharmacologic agents used.

Two large clinical trials, the Systolic Hypertension in the Elderly Program and ALLHAT, found that chlorthalidone caused hypokalemia requiring therapy in about 6% to 8% of patients.^38,39 Chlorthalidone therapy was associated with a lowering of serum potassium levels of 0.2 to 0.5 mmol/L.³⁶ In ALLHAT, chlorthalidone was associated with a reduction in potassium levels of approximately 0.2 mmol/L after 4 years.³⁸

All diuretics require monitoring of electrolytes, especially during the first 2 weeks of therapy. Once a steady state is reached, patients are not usually at risk of hypokalemia  unless the dose is increased, extrarenal losses of potassium increase, or dietary potassium is reduced.

Other electrolyte changes. Thiazide and thiazide-like diuretics can cause other metabolic and endocrine abnormalities such as hypochloremic alkalosis, hyponatremia, and hypercalcemia.^40,41 They can also cause photosensitivity and can precipitate gout.⁴²

Observational studies have suggested that metabolic adverse effects such as hypokalemia and hyperuricemia are more common with chlorthalidone than with hydrochlorothiazide.⁴³ It is prudent to monitor laboratory values periodically in patients on diuretic therapy.

DRUG INTERACTIONS

The drug interaction profiles of hydrochlorothiazide and chlorthalidone are also similar. The most common interactions are pharmacodynamic interactions resulting from potassium depletion caused by the diuretics.

Antiarrythymic drugs. Hypokalemia is a risk factor for arrhythmias such as torsades de pointes, and the risk is magnified with concomitant therapy with antiarrhythmic agents that prolong the QT interval independently of serum potassium concentration (eg, sotalol, dronedarone, ibutilide, propafenone). Therefore, combinations of drugs that can cause hypokalemia (eg, diuretics) and antiarrhythmic agents require vigilant monitoring of potassium and appropriate replenishment.⁴⁴

Dofetilide is a class III antiarrhythmic agent and, like other antiarrhythmic drugs, carries a risk of QT prolongation and torsades de pointes, which is magnified by hypokalemia.⁴⁵ In addition, dofetilide undergoes active tubular secretion in the kidney via the cation transport system, which is inhibited by hydrochlorothiazide.⁴⁵ The resulting increase in plasma concentrations of dofetilide may magnify the risk of arrhythmias. Chlorthalidone has not been specifically studied in combination with dofetilide, but thiazide diuretics in general are thought to have a similar effect on tubular secretion and, therefore, should be considered similar to hydrochlorothiazide in this regard.

Digoxin. Similarly, digoxin toxicity may be enhanced in hypokalemia. As with antiarrhythmic agents, serum potassium should be carefully monitored and replenished appropriately when diuretics are used in combination with digoxin.

Lithium is reabsorbed in the proximal tubule along with sodium. Diuretics including hydrochlorothiazide and chlorthalidone that alter sodium reabsorption can also alter lithium absorption.⁴⁶ When sodium reabsorption is decreased, lithium ion reabsorption is increased and may result in lithium toxicity. Although this combination is not contraindicated, monitoring of serum lithium concentrations and clinical signs and symptoms of lithium toxicity is recommended during initiation and dose adjustments of thiazide diuretics.

Nonsteroidal anti-inflammatory drugs can decrease the natriuretic, diuretic, and antihypertensive effects of both hydrochlorothiazide and chlorthalidone.⁴⁷

Renin-angiotensin-aldosterone system antagonists, ie, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and the renin inhibitor aliskiren, have potentially beneficial interactions with hydrochlorothiazide and chlorthalidone, producing additive decreases in blood pressure when coadministered with these diuretics. These effects may be particularly potent early in concomitant therapy and allow use of lower doses of diuretics, typically 12.5 mg of hydrochlorothiazide in combination therapy.

LONG-TERM EFFECTS ON CARDIOVASCULAR EVENTS

The long-term goal in treating hypertension is to lower the risk of cardiovascular disease. Therefore, the clinician needs to consider the effect of antihypertensive drug therapy on long-term clinical outcomes.

Antihypertensive drug therapy based on thiazide diuretics has been shown to lower cardiovascular risk when compared with placebo.⁴⁸ In addition, the effect of chlorthalidone-based antihypertensive therapy was similar to that of other antihypertensive drug classes in preventing most cardiovascular outcomes in ALLHAT.⁴

However, no study has directly compared hydrochlorothiazide and chlorthalidone with the primary outcome of reduction in long-term cardiovascular events. The data to date come from observational studies and meta-analyses. For example, in a retrospective analysis of the Multiple Risk Factor Intervention Trial, cardiovascular events were significantly fewer in those receiving chlorthalidone vs hydrochlorothiazide (P = .0016).⁴³

In a systematic review and meta-analysis, chlorthalidone was associated with a 23% lower risk of heart failure and a 21% lower risk of all cardiovascular events.⁴⁹

However, a Canadian observational study of 29,873 patients found no difference in the composite outcome of death or hospitalization for heart failure, stroke, or myocardial infarction between chlorthalidone recipients (3.2 events per 100 person-years) and hydrochlorothiazide recipients (3.4 events per 100 person-years; adjusted hazard ratio 0.93, 95% confidence interval 0.81–1.06).⁵⁰

In summary, it is unclear whether chlorthalidone or hydrochlorothiazide is superior in preventing cardiovascular events.

SUMMARY

Thiazide and thiazidelike diuretics play an important role in managing hypertension in most patients. The eighth Joint National Committee guidelines do not recommend either hydrochlorothiazide or chlorthalidone over the other. The target dose recommendations are hydrochlorothiazide 25 to 50 mg or chlorthalidone 12.5 to 25 mg daily, with lower doses considered for the elderly.

There are important differences between hydrochlorothiazide and chlorthalidone in pharmacology, potency, and frequency of metabolic side effects. Clinicians should consider these factors to tailor the choice of thiazide diuretic therapy in hypertensive patients.

The thiazide diuretic hydrochlorothiazide and the thiazidelike diuretic chlorthalidone are two old drugs that are still useful. Although similar, they differ in important ways still not fully appreciated more than a half century after they were introduced.

Most hypertension guidelines recommend thiazide diuretics as one of the classes of agents that can be used either as initial antihypertensive drug therapy or as part of combination therapy.^1–3

In the United States, hydrochlorothiazide is used more often than chlorthalidone, but many clinical trials of antihypertensive therapy have used chlorthalidone.^4,5 In recent years, particularly after the publication of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), interest in chlorthalidone has been increasing, and new data are now available comparing these two diuretics.⁶ While current US guidelines do not recommend one over the other, British guidelines prefer chlorthalidone.⁷

This review summarizes the data comparing the two drugs’ pharmacology, antihypertensive effect, and impact on clinical outcomes to help guide clinicians in choosing antihypertensive drug therapy.

PHARMACOLOGY AND MECHANISM OF ACTION

Many of the differences in effectiveness and adverse effects of hydrochlorothiazide and chlorthalidone are thought to be due to their different pharmacodynamic and pharmacokinetic effects.

Pharmacodynamic effects

Hydrochlorothiazide and chlorthalidone differ significantly in chemical structure (Figure 1), but both contain a sulfonamide group that inhibits carbonic anhydrase activity, which may be associated with lower vascular contractility. Both drugs are concentrated in the kidney and secreted into the tubular lumen⁸; therefore, their therapeutic diuretic effects are often achieved with relatively low plasma concentrations.

Both drugs inhibit the sodium-chloride cotransporter in the luminal membrane of the distal convoluted tubule of the ascending loop of Henle, leading to a modest natriuresis and diuresis. The exact mechanism by which they lower blood pressure is not known: while the initial response is from diuresis and volume changes, long-term reduction in blood pressure is through uncertain mechanisms. In addition, chlorthalidone may have beneficial effects on endothelial function and oxidative stress.^9,10

Both drugs also increase secretion of potassium and hydrogen ions and promote increased reabsorption of calcium through increased expression of a sodium-calcium exchange channel.⁸ Chlorthalidone may cause more inhibition of carbonic anhydrase than hydrochlorothiazide, which can lead to lower intracellular pH and cell volume. This effect may in part explain a pleiotropic effect of chlorthalidone, ie, inhibition of platelet function, which in turn may contribute to this drug’s beneficial effect on cardiovascular outcomes.⁹

Pharmacokinetic differences

Hydrochlorothiazide and chlorthalidone have important differences in their pharmacokinetic properties (Table 1).¹¹

Hydrochlorothiazide has its onset of action in about 2 hours, and it reaches its peak in 4 to 6 hours. Though its duration of action is short—up to 12 hours—its pharmacodynamic response can be much longer than predicted by its kinetics, allowing once-daily dosing.⁸

Chlorthalidone has a longer duration of action than hydrochlorothiazide. This may be because it has a very high volume of distribution, since it is taken up into red blood cells and is bound to carbonic anhydrase.¹² This may result in a “drug reservoir” that keeps drug levels higher for a longer time.¹³ Its long duration of action makes it a favorable choice for patients who have difficulty adhering to medication instructions. In addition, a missed dose is unlikely to have a “rebound” effect like that seen with some other antihypertensive agents. However, both chlorthalidone and hydrochlorothiazide are effective if taken once daily.

BLOOD PRESSURE-LOWERING

Both hydrochlorothiazide and chlorthalidone are effective antihypertensive agents. Table 2 summarizes findings from studies that evaluated their blood pressure-lowering effect at various doses.^14–33 However, relatively few studies have directly compared these two agents’ effects on blood pressure.

Ernst et al,³⁴ in a small study (but probably the best one to address this issue), compared chlorthalidone 12.5 mg/day (force-titrated to 25 mg/day) and hydrochlorothiazide 25 mg/day (force-titrated to 50 mg/day) in untreated hypertensive patients. After 8 weeks, ambulatory blood pressure monitoring indicated a greater reduction from baseline in systolic blood pressure with chlorthalidone 25 mg/day than with hydrochlorothiazide 50 mg/day (24-hour mean –12.4 vs –7.4 mm Hg, P = .05). Interestingly, the change in nighttime blood pressure was greater in the chlorthalidone group (–13.5 mm Hg) than in the hydrochlorothiazide group (–6.4 mm Hg; P = .009). These data suggest that at the doses studied, chlorthalidone is more effective than hydrochlorothiazide in lowering systolic blood pressure.

Bakris et al,³⁵ using a different study design, compared the single-pill combination of azilsartan medoxomil and chlorthalidone vs coadministration of azilsartan medoxomil and hydrochlorothiazide in participants with stage 2 primary hypertension (≥ 160/100 mm Hg). Systolic blood pressure, as measured in the clinic, declined more with the chlorthalidone combination (–35.1 mm Hg) than with the hydrochlorothiazide combination (–29.5 mm Hg, mean difference –5.6 mm Hg, P < .001).

Meta-analyses also support the conclusion that chlorthalidone is more potent than hydrochlorothiazide in lowering blood pressure.^35,36 Several studies have shown that chlorthalidone at the same dose is 1.5 to 2 times as potent as hydrochlorothiazide.^33,36,37 Therefore, for clinical purposes, it is reasonable to consider chlorthalidone 12.5 mg daily as similar to 25 mg of hydrochlorothiazide daily.

ADVERSE EFFECTS

Electrolyte disturbances are a common adverse effect of thiazide diuretics.

Hypokalemia. All thiazide diuretics cause potassium wasting. The frequency of hypokalemia depends on the dose, frequency of administration, diet, and other pharmacologic agents used.

Two large clinical trials, the Systolic Hypertension in the Elderly Program and ALLHAT, found that chlorthalidone caused hypokalemia requiring therapy in about 6% to 8% of patients.^38,39 Chlorthalidone therapy was associated with a lowering of serum potassium levels of 0.2 to 0.5 mmol/L.³⁶ In ALLHAT, chlorthalidone was associated with a reduction in potassium levels of approximately 0.2 mmol/L after 4 years.³⁸

All diuretics require monitoring of electrolytes, especially during the first 2 weeks of therapy. Once a steady state is reached, patients are not usually at risk of hypokalemia  unless the dose is increased, extrarenal losses of potassium increase, or dietary potassium is reduced.

Other electrolyte changes. Thiazide and thiazide-like diuretics can cause other metabolic and endocrine abnormalities such as hypochloremic alkalosis, hyponatremia, and hypercalcemia.^40,41 They can also cause photosensitivity and can precipitate gout.⁴²

Observational studies have suggested that metabolic adverse effects such as hypokalemia and hyperuricemia are more common with chlorthalidone than with hydrochlorothiazide.⁴³ It is prudent to monitor laboratory values periodically in patients on diuretic therapy.

DRUG INTERACTIONS

The drug interaction profiles of hydrochlorothiazide and chlorthalidone are also similar. The most common interactions are pharmacodynamic interactions resulting from potassium depletion caused by the diuretics.

Antiarrythymic drugs. Hypokalemia is a risk factor for arrhythmias such as torsades de pointes, and the risk is magnified with concomitant therapy with antiarrhythmic agents that prolong the QT interval independently of serum potassium concentration (eg, sotalol, dronedarone, ibutilide, propafenone). Therefore, combinations of drugs that can cause hypokalemia (eg, diuretics) and antiarrhythmic agents require vigilant monitoring of potassium and appropriate replenishment.⁴⁴

Dofetilide is a class III antiarrhythmic agent and, like other antiarrhythmic drugs, carries a risk of QT prolongation and torsades de pointes, which is magnified by hypokalemia.⁴⁵ In addition, dofetilide undergoes active tubular secretion in the kidney via the cation transport system, which is inhibited by hydrochlorothiazide.⁴⁵ The resulting increase in plasma concentrations of dofetilide may magnify the risk of arrhythmias. Chlorthalidone has not been specifically studied in combination with dofetilide, but thiazide diuretics in general are thought to have a similar effect on tubular secretion and, therefore, should be considered similar to hydrochlorothiazide in this regard.

Digoxin. Similarly, digoxin toxicity may be enhanced in hypokalemia. As with antiarrhythmic agents, serum potassium should be carefully monitored and replenished appropriately when diuretics are used in combination with digoxin.

Lithium is reabsorbed in the proximal tubule along with sodium. Diuretics including hydrochlorothiazide and chlorthalidone that alter sodium reabsorption can also alter lithium absorption.⁴⁶ When sodium reabsorption is decreased, lithium ion reabsorption is increased and may result in lithium toxicity. Although this combination is not contraindicated, monitoring of serum lithium concentrations and clinical signs and symptoms of lithium toxicity is recommended during initiation and dose adjustments of thiazide diuretics.

Nonsteroidal anti-inflammatory drugs can decrease the natriuretic, diuretic, and antihypertensive effects of both hydrochlorothiazide and chlorthalidone.⁴⁷

Renin-angiotensin-aldosterone system antagonists, ie, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and the renin inhibitor aliskiren, have potentially beneficial interactions with hydrochlorothiazide and chlorthalidone, producing additive decreases in blood pressure when coadministered with these diuretics. These effects may be particularly potent early in concomitant therapy and allow use of lower doses of diuretics, typically 12.5 mg of hydrochlorothiazide in combination therapy.

LONG-TERM EFFECTS ON CARDIOVASCULAR EVENTS

The long-term goal in treating hypertension is to lower the risk of cardiovascular disease. Therefore, the clinician needs to consider the effect of antihypertensive drug therapy on long-term clinical outcomes.

Antihypertensive drug therapy based on thiazide diuretics has been shown to lower cardiovascular risk when compared with placebo.⁴⁸ In addition, the effect of chlorthalidone-based antihypertensive therapy was similar to that of other antihypertensive drug classes in preventing most cardiovascular outcomes in ALLHAT.⁴

However, no study has directly compared hydrochlorothiazide and chlorthalidone with the primary outcome of reduction in long-term cardiovascular events. The data to date come from observational studies and meta-analyses. For example, in a retrospective analysis of the Multiple Risk Factor Intervention Trial, cardiovascular events were significantly fewer in those receiving chlorthalidone vs hydrochlorothiazide (P = .0016).⁴³

In a systematic review and meta-analysis, chlorthalidone was associated with a 23% lower risk of heart failure and a 21% lower risk of all cardiovascular events.⁴⁹

However, a Canadian observational study of 29,873 patients found no difference in the composite outcome of death or hospitalization for heart failure, stroke, or myocardial infarction between chlorthalidone recipients (3.2 events per 100 person-years) and hydrochlorothiazide recipients (3.4 events per 100 person-years; adjusted hazard ratio 0.93, 95% confidence interval 0.81–1.06).⁵⁰

In summary, it is unclear whether chlorthalidone or hydrochlorothiazide is superior in preventing cardiovascular events.

SUMMARY

Thiazide and thiazidelike diuretics play an important role in managing hypertension in most patients. The eighth Joint National Committee guidelines do not recommend either hydrochlorothiazide or chlorthalidone over the other. The target dose recommendations are hydrochlorothiazide 25 to 50 mg or chlorthalidone 12.5 to 25 mg daily, with lower doses considered for the elderly.

There are important differences between hydrochlorothiazide and chlorthalidone in pharmacology, potency, and frequency of metabolic side effects. Clinicians should consider these factors to tailor the choice of thiazide diuretic therapy in hypertensive patients.

The thiazide diuretic hydrochlorothiazide and the thiazidelike diuretic chlorthalidone are two old drugs that are still useful. Although similar, they differ in important ways still not fully appreciated more than a half century after they were introduced.

Most hypertension guidelines recommend thiazide diuretics as one of the classes of agents that can be used either as initial antihypertensive drug therapy or as part of combination therapy.^1–3

In the United States, hydrochlorothiazide is used more often than chlorthalidone, but many clinical trials of antihypertensive therapy have used chlorthalidone.^4,5 In recent years, particularly after the publication of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), interest in chlorthalidone has been increasing, and new data are now available comparing these two diuretics.⁶ While current US guidelines do not recommend one over the other, British guidelines prefer chlorthalidone.⁷

This review summarizes the data comparing the two drugs’ pharmacology, antihypertensive effect, and impact on clinical outcomes to help guide clinicians in choosing antihypertensive drug therapy.

PHARMACOLOGY AND MECHANISM OF ACTION

Many of the differences in effectiveness and adverse effects of hydrochlorothiazide and chlorthalidone are thought to be due to their different pharmacodynamic and pharmacokinetic effects.

Pharmacodynamic effects

Hydrochlorothiazide and chlorthalidone differ significantly in chemical structure (Figure 1), but both contain a sulfonamide group that inhibits carbonic anhydrase activity, which may be associated with lower vascular contractility. Both drugs are concentrated in the kidney and secreted into the tubular lumen⁸; therefore, their therapeutic diuretic effects are often achieved with relatively low plasma concentrations.

Both drugs inhibit the sodium-chloride cotransporter in the luminal membrane of the distal convoluted tubule of the ascending loop of Henle, leading to a modest natriuresis and diuresis. The exact mechanism by which they lower blood pressure is not known: while the initial response is from diuresis and volume changes, long-term reduction in blood pressure is through uncertain mechanisms. In addition, chlorthalidone may have beneficial effects on endothelial function and oxidative stress.^9,10

Both drugs also increase secretion of potassium and hydrogen ions and promote increased reabsorption of calcium through increased expression of a sodium-calcium exchange channel.⁸ Chlorthalidone may cause more inhibition of carbonic anhydrase than hydrochlorothiazide, which can lead to lower intracellular pH and cell volume. This effect may in part explain a pleiotropic effect of chlorthalidone, ie, inhibition of platelet function, which in turn may contribute to this drug’s beneficial effect on cardiovascular outcomes.⁹

Pharmacokinetic differences

Hydrochlorothiazide and chlorthalidone have important differences in their pharmacokinetic properties (Table 1).¹¹

Hydrochlorothiazide has its onset of action in about 2 hours, and it reaches its peak in 4 to 6 hours. Though its duration of action is short—up to 12 hours—its pharmacodynamic response can be much longer than predicted by its kinetics, allowing once-daily dosing.⁸

Chlorthalidone has a longer duration of action than hydrochlorothiazide. This may be because it has a very high volume of distribution, since it is taken up into red blood cells and is bound to carbonic anhydrase.¹² This may result in a “drug reservoir” that keeps drug levels higher for a longer time.¹³ Its long duration of action makes it a favorable choice for patients who have difficulty adhering to medication instructions. In addition, a missed dose is unlikely to have a “rebound” effect like that seen with some other antihypertensive agents. However, both chlorthalidone and hydrochlorothiazide are effective if taken once daily.

BLOOD PRESSURE-LOWERING

Both hydrochlorothiazide and chlorthalidone are effective antihypertensive agents. Table 2 summarizes findings from studies that evaluated their blood pressure-lowering effect at various doses.^14–33 However, relatively few studies have directly compared these two agents’ effects on blood pressure.

Ernst et al,³⁴ in a small study (but probably the best one to address this issue), compared chlorthalidone 12.5 mg/day (force-titrated to 25 mg/day) and hydrochlorothiazide 25 mg/day (force-titrated to 50 mg/day) in untreated hypertensive patients. After 8 weeks, ambulatory blood pressure monitoring indicated a greater reduction from baseline in systolic blood pressure with chlorthalidone 25 mg/day than with hydrochlorothiazide 50 mg/day (24-hour mean –12.4 vs –7.4 mm Hg, P = .05). Interestingly, the change in nighttime blood pressure was greater in the chlorthalidone group (–13.5 mm Hg) than in the hydrochlorothiazide group (–6.4 mm Hg; P = .009). These data suggest that at the doses studied, chlorthalidone is more effective than hydrochlorothiazide in lowering systolic blood pressure.

Bakris et al,³⁵ using a different study design, compared the single-pill combination of azilsartan medoxomil and chlorthalidone vs coadministration of azilsartan medoxomil and hydrochlorothiazide in participants with stage 2 primary hypertension (≥ 160/100 mm Hg). Systolic blood pressure, as measured in the clinic, declined more with the chlorthalidone combination (–35.1 mm Hg) than with the hydrochlorothiazide combination (–29.5 mm Hg, mean difference –5.6 mm Hg, P < .001).

Meta-analyses also support the conclusion that chlorthalidone is more potent than hydrochlorothiazide in lowering blood pressure.^35,36 Several studies have shown that chlorthalidone at the same dose is 1.5 to 2 times as potent as hydrochlorothiazide.^33,36,37 Therefore, for clinical purposes, it is reasonable to consider chlorthalidone 12.5 mg daily as similar to 25 mg of hydrochlorothiazide daily.

ADVERSE EFFECTS

Electrolyte disturbances are a common adverse effect of thiazide diuretics.

Hypokalemia. All thiazide diuretics cause potassium wasting. The frequency of hypokalemia depends on the dose, frequency of administration, diet, and other pharmacologic agents used.

Two large clinical trials, the Systolic Hypertension in the Elderly Program and ALLHAT, found that chlorthalidone caused hypokalemia requiring therapy in about 6% to 8% of patients.^38,39 Chlorthalidone therapy was associated with a lowering of serum potassium levels of 0.2 to 0.5 mmol/L.³⁶ In ALLHAT, chlorthalidone was associated with a reduction in potassium levels of approximately 0.2 mmol/L after 4 years.³⁸

All diuretics require monitoring of electrolytes, especially during the first 2 weeks of therapy. Once a steady state is reached, patients are not usually at risk of hypokalemia  unless the dose is increased, extrarenal losses of potassium increase, or dietary potassium is reduced.

Other electrolyte changes. Thiazide and thiazide-like diuretics can cause other metabolic and endocrine abnormalities such as hypochloremic alkalosis, hyponatremia, and hypercalcemia.^40,41 They can also cause photosensitivity and can precipitate gout.⁴²

Observational studies have suggested that metabolic adverse effects such as hypokalemia and hyperuricemia are more common with chlorthalidone than with hydrochlorothiazide.⁴³ It is prudent to monitor laboratory values periodically in patients on diuretic therapy.

DRUG INTERACTIONS

The drug interaction profiles of hydrochlorothiazide and chlorthalidone are also similar. The most common interactions are pharmacodynamic interactions resulting from potassium depletion caused by the diuretics.

Antiarrythymic drugs. Hypokalemia is a risk factor for arrhythmias such as torsades de pointes, and the risk is magnified with concomitant therapy with antiarrhythmic agents that prolong the QT interval independently of serum potassium concentration (eg, sotalol, dronedarone, ibutilide, propafenone). Therefore, combinations of drugs that can cause hypokalemia (eg, diuretics) and antiarrhythmic agents require vigilant monitoring of potassium and appropriate replenishment.⁴⁴

Dofetilide is a class III antiarrhythmic agent and, like other antiarrhythmic drugs, carries a risk of QT prolongation and torsades de pointes, which is magnified by hypokalemia.⁴⁵ In addition, dofetilide undergoes active tubular secretion in the kidney via the cation transport system, which is inhibited by hydrochlorothiazide.⁴⁵ The resulting increase in plasma concentrations of dofetilide may magnify the risk of arrhythmias. Chlorthalidone has not been specifically studied in combination with dofetilide, but thiazide diuretics in general are thought to have a similar effect on tubular secretion and, therefore, should be considered similar to hydrochlorothiazide in this regard.

Digoxin. Similarly, digoxin toxicity may be enhanced in hypokalemia. As with antiarrhythmic agents, serum potassium should be carefully monitored and replenished appropriately when diuretics are used in combination with digoxin.

Lithium is reabsorbed in the proximal tubule along with sodium. Diuretics including hydrochlorothiazide and chlorthalidone that alter sodium reabsorption can also alter lithium absorption.⁴⁶ When sodium reabsorption is decreased, lithium ion reabsorption is increased and may result in lithium toxicity. Although this combination is not contraindicated, monitoring of serum lithium concentrations and clinical signs and symptoms of lithium toxicity is recommended during initiation and dose adjustments of thiazide diuretics.

Nonsteroidal anti-inflammatory drugs can decrease the natriuretic, diuretic, and antihypertensive effects of both hydrochlorothiazide and chlorthalidone.⁴⁷

Renin-angiotensin-aldosterone system antagonists, ie, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and the renin inhibitor aliskiren, have potentially beneficial interactions with hydrochlorothiazide and chlorthalidone, producing additive decreases in blood pressure when coadministered with these diuretics. These effects may be particularly potent early in concomitant therapy and allow use of lower doses of diuretics, typically 12.5 mg of hydrochlorothiazide in combination therapy.

LONG-TERM EFFECTS ON CARDIOVASCULAR EVENTS

The long-term goal in treating hypertension is to lower the risk of cardiovascular disease. Therefore, the clinician needs to consider the effect of antihypertensive drug therapy on long-term clinical outcomes.

Antihypertensive drug therapy based on thiazide diuretics has been shown to lower cardiovascular risk when compared with placebo.⁴⁸ In addition, the effect of chlorthalidone-based antihypertensive therapy was similar to that of other antihypertensive drug classes in preventing most cardiovascular outcomes in ALLHAT.⁴

However, no study has directly compared hydrochlorothiazide and chlorthalidone with the primary outcome of reduction in long-term cardiovascular events. The data to date come from observational studies and meta-analyses. For example, in a retrospective analysis of the Multiple Risk Factor Intervention Trial, cardiovascular events were significantly fewer in those receiving chlorthalidone vs hydrochlorothiazide (P = .0016).⁴³

In a systematic review and meta-analysis, chlorthalidone was associated with a 23% lower risk of heart failure and a 21% lower risk of all cardiovascular events.⁴⁹

However, a Canadian observational study of 29,873 patients found no difference in the composite outcome of death or hospitalization for heart failure, stroke, or myocardial infarction between chlorthalidone recipients (3.2 events per 100 person-years) and hydrochlorothiazide recipients (3.4 events per 100 person-years; adjusted hazard ratio 0.93, 95% confidence interval 0.81–1.06).⁵⁰

In summary, it is unclear whether chlorthalidone or hydrochlorothiazide is superior in preventing cardiovascular events.

SUMMARY

Thiazide and thiazidelike diuretics play an important role in managing hypertension in most patients. The eighth Joint National Committee guidelines do not recommend either hydrochlorothiazide or chlorthalidone over the other. The target dose recommendations are hydrochlorothiazide 25 to 50 mg or chlorthalidone 12.5 to 25 mg daily, with lower doses considered for the elderly.

There are important differences between hydrochlorothiazide and chlorthalidone in pharmacology, potency, and frequency of metabolic side effects. Clinicians should consider these factors to tailor the choice of thiazide diuretic therapy in hypertensive patients.

In a widely anticipated judgment on the Affordable Care Act (ACA), the US Supreme Court ruled 6-3 in favor of the law on June 26, 2015. The case at hand, King v. Burwell, challenged whether individuals purchasing health insurance through federal exchanges were eligible for federal premium subsidies. This ruling cemented the ACA into law and avoided a potential calamity in the private health insurance market. Let’s take a closer look.

What the case was about
The ACA allows states to set up their own health insurance exchanges or participate in a federally run exchange. Although the drafters of the ACA had expected each state to set up its own exchange, two-thirds of the states declined to do so, many in opposition to the ACA. As a result, 7 million citizens in 34 states now purchase their health ­insurance through federally created exchanges.

The plaintiffs in King v. Burwell argued that, because the legislation refers to those enrolled “through an Exchange established by the State,” individuals in states with federally run exchanges are not eligible for subsidies.

The law states:

The Supreme Court was asked to decide whether to adhere to those exact words or to honor Congress’ intent to allow individuals to purchase subsidized insurance on any type of exchange.

What might have happened
We’ve explored in previous articles the interconnectedness of many sections of the ACA. Nowhere is that interconnectedness more clearly demonstrated than here. In order to ensure that private health insurers provide better coverage, the law requires them to abide by important consumer protections, including the elimination of “preexisting condition” exclusions. In order to prevent adverse selection and keep insurers solvent under these new rules, all individuals are required to have health care coverage—the individual mandate. If everyone is required to purchase health insurance, it has to be affordable, so lower-income individuals were promised subsidies, paid for 100% by the federal government, to help them cover their premiums when insurance is purchased through an exchange. Take away the subsidies and the whole thing starts to unravel.

The Urban Institute estimated that a Supreme Court ruling in favor of King, which would have eliminated the subsidies in states using a federal exchange, would have reduced federal tax subsidies by $29 billion in 2016, making coverage unaffordable for many and increasing the ranks of the uninsured by 8.2 million people.¹

Louise Sheiner and Brendan Mochoruk of the Brookings Institute speculated that healthy individuals would disproportionately leave the marketplace, triggering 35% increases in insurance premiums for those remaining, as well as significant increases in premiums for those who just lost their subsidies.² Many observers, including these experts, forecast that insurance companies would exit the federal exchanges altogether, triggering a health insurance “death spiral”: As premiums rise, the healthiest customers leave the marketplace, causing premiums to rise more, causing more healthy people to leave, and so on.

Clearly, this Supreme Court decision has had dramatic, long-term, real-world effects on millions of Americans. On the national level, 6,387,789 individuals were at risk of losing their tax credits if the Supreme Court had ruled in favor of King. That number represents more than $1.7 billion in total monthly tax credits. For a look at how a judgment in favor of King would have affected subsidies on a state-by-state basis, see TABLE 1.

How premium subsidies work
Premium subsidies are actually tax credits. Individuals and families can qualify for them to purchase any type of health insurance offered on an exchange, except catastrophic coverage. To receive the premium tax credit for coverage starting in 2015, a marketplace enrollee must:

The premium tax credit caps the amount that an individual or family must spend on their monthly payments for health insurance. The cap depends on the family’s income; lower-income families have a lower cap. The amount of the tax credit remains the same, so a person who purchases a more expensive plan pays the cost difference (TABLE 2).

The ruling’s effect on women’s health
On June 26, American College of Obstetricians and Gynecologists President Mark S. DeFrancesco, MD, MBA, hailed the Supreme Court decision, saying, “Importantly, recent data have shown that newly insured adults under the ACA were more likely to be ­women. Those who did gain coverage through the ACA reported better access to health care and better financial security from medical costs.”

“Without question, many women enrollees were able to purchase health insurance coverage due, in part, to the ACA subsidies that helped make this purchase affordable. In fact, government data have suggested that roughly 85% of health exchange enrollees received subsidies,” Dr. DeFrancesco said.

“If the Supreme Court had overturned this important assistance, approximately 4.8 million women would have been unable to afford the coverage that they need. The impact also would have been widespread; as these women were forced to leave the insurance marketplace, it is likely that premiums throughout the marketplace would have risen dramatically,” he continued.

“Instead, patients—especially the low- and moderate-income American women who have particularly benefited from ACA subsidies—will continue to have the peace of mind that comes with insurance coverage.”

Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.

In a widely anticipated judgment on the Affordable Care Act (ACA), the US Supreme Court ruled 6-3 in favor of the law on June 26, 2015. The case at hand, King v. Burwell, challenged whether individuals purchasing health insurance through federal exchanges were eligible for federal premium subsidies. This ruling cemented the ACA into law and avoided a potential calamity in the private health insurance market. Let’s take a closer look.

What the case was about
The ACA allows states to set up their own health insurance exchanges or participate in a federally run exchange. Although the drafters of the ACA had expected each state to set up its own exchange, two-thirds of the states declined to do so, many in opposition to the ACA. As a result, 7 million citizens in 34 states now purchase their health ­insurance through federally created exchanges.

The plaintiffs in King v. Burwell argued that, because the legislation refers to those enrolled “through an Exchange established by the State,” individuals in states with federally run exchanges are not eligible for subsidies.

The law states:

The Supreme Court was asked to decide whether to adhere to those exact words or to honor Congress’ intent to allow individuals to purchase subsidized insurance on any type of exchange.

What might have happened
We’ve explored in previous articles the interconnectedness of many sections of the ACA. Nowhere is that interconnectedness more clearly demonstrated than here. In order to ensure that private health insurers provide better coverage, the law requires them to abide by important consumer protections, including the elimination of “preexisting condition” exclusions. In order to prevent adverse selection and keep insurers solvent under these new rules, all individuals are required to have health care coverage—the individual mandate. If everyone is required to purchase health insurance, it has to be affordable, so lower-income individuals were promised subsidies, paid for 100% by the federal government, to help them cover their premiums when insurance is purchased through an exchange. Take away the subsidies and the whole thing starts to unravel.

The Urban Institute estimated that a Supreme Court ruling in favor of King, which would have eliminated the subsidies in states using a federal exchange, would have reduced federal tax subsidies by $29 billion in 2016, making coverage unaffordable for many and increasing the ranks of the uninsured by 8.2 million people.¹

Louise Sheiner and Brendan Mochoruk of the Brookings Institute speculated that healthy individuals would disproportionately leave the marketplace, triggering 35% increases in insurance premiums for those remaining, as well as significant increases in premiums for those who just lost their subsidies.² Many observers, including these experts, forecast that insurance companies would exit the federal exchanges altogether, triggering a health insurance “death spiral”: As premiums rise, the healthiest customers leave the marketplace, causing premiums to rise more, causing more healthy people to leave, and so on.

Clearly, this Supreme Court decision has had dramatic, long-term, real-world effects on millions of Americans. On the national level, 6,387,789 individuals were at risk of losing their tax credits if the Supreme Court had ruled in favor of King. That number represents more than $1.7 billion in total monthly tax credits. For a look at how a judgment in favor of King would have affected subsidies on a state-by-state basis, see TABLE 1.

How premium subsidies work
Premium subsidies are actually tax credits. Individuals and families can qualify for them to purchase any type of health insurance offered on an exchange, except catastrophic coverage. To receive the premium tax credit for coverage starting in 2015, a marketplace enrollee must:

The premium tax credit caps the amount that an individual or family must spend on their monthly payments for health insurance. The cap depends on the family’s income; lower-income families have a lower cap. The amount of the tax credit remains the same, so a person who purchases a more expensive plan pays the cost difference (TABLE 2).

The ruling’s effect on women’s health
On June 26, American College of Obstetricians and Gynecologists President Mark S. DeFrancesco, MD, MBA, hailed the Supreme Court decision, saying, “Importantly, recent data have shown that newly insured adults under the ACA were more likely to be ­women. Those who did gain coverage through the ACA reported better access to health care and better financial security from medical costs.”

“Without question, many women enrollees were able to purchase health insurance coverage due, in part, to the ACA subsidies that helped make this purchase affordable. In fact, government data have suggested that roughly 85% of health exchange enrollees received subsidies,” Dr. DeFrancesco said.

“If the Supreme Court had overturned this important assistance, approximately 4.8 million women would have been unable to afford the coverage that they need. The impact also would have been widespread; as these women were forced to leave the insurance marketplace, it is likely that premiums throughout the marketplace would have risen dramatically,” he continued.

“Instead, patients—especially the low- and moderate-income American women who have particularly benefited from ACA subsidies—will continue to have the peace of mind that comes with insurance coverage.”

Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.

In a widely anticipated judgment on the Affordable Care Act (ACA), the US Supreme Court ruled 6-3 in favor of the law on June 26, 2015. The case at hand, King v. Burwell, challenged whether individuals purchasing health insurance through federal exchanges were eligible for federal premium subsidies. This ruling cemented the ACA into law and avoided a potential calamity in the private health insurance market. Let’s take a closer look.

What the case was about
The ACA allows states to set up their own health insurance exchanges or participate in a federally run exchange. Although the drafters of the ACA had expected each state to set up its own exchange, two-thirds of the states declined to do so, many in opposition to the ACA. As a result, 7 million citizens in 34 states now purchase their health ­insurance through federally created exchanges.

The plaintiffs in King v. Burwell argued that, because the legislation refers to those enrolled “through an Exchange established by the State,” individuals in states with federally run exchanges are not eligible for subsidies.

The law states:

The Supreme Court was asked to decide whether to adhere to those exact words or to honor Congress’ intent to allow individuals to purchase subsidized insurance on any type of exchange.

What might have happened
We’ve explored in previous articles the interconnectedness of many sections of the ACA. Nowhere is that interconnectedness more clearly demonstrated than here. In order to ensure that private health insurers provide better coverage, the law requires them to abide by important consumer protections, including the elimination of “preexisting condition” exclusions. In order to prevent adverse selection and keep insurers solvent under these new rules, all individuals are required to have health care coverage—the individual mandate. If everyone is required to purchase health insurance, it has to be affordable, so lower-income individuals were promised subsidies, paid for 100% by the federal government, to help them cover their premiums when insurance is purchased through an exchange. Take away the subsidies and the whole thing starts to unravel.

The Urban Institute estimated that a Supreme Court ruling in favor of King, which would have eliminated the subsidies in states using a federal exchange, would have reduced federal tax subsidies by $29 billion in 2016, making coverage unaffordable for many and increasing the ranks of the uninsured by 8.2 million people.¹

Louise Sheiner and Brendan Mochoruk of the Brookings Institute speculated that healthy individuals would disproportionately leave the marketplace, triggering 35% increases in insurance premiums for those remaining, as well as significant increases in premiums for those who just lost their subsidies.² Many observers, including these experts, forecast that insurance companies would exit the federal exchanges altogether, triggering a health insurance “death spiral”: As premiums rise, the healthiest customers leave the marketplace, causing premiums to rise more, causing more healthy people to leave, and so on.

Clearly, this Supreme Court decision has had dramatic, long-term, real-world effects on millions of Americans. On the national level, 6,387,789 individuals were at risk of losing their tax credits if the Supreme Court had ruled in favor of King. That number represents more than $1.7 billion in total monthly tax credits. For a look at how a judgment in favor of King would have affected subsidies on a state-by-state basis, see TABLE 1.

How premium subsidies work
Premium subsidies are actually tax credits. Individuals and families can qualify for them to purchase any type of health insurance offered on an exchange, except catastrophic coverage. To receive the premium tax credit for coverage starting in 2015, a marketplace enrollee must:

The premium tax credit caps the amount that an individual or family must spend on their monthly payments for health insurance. The cap depends on the family’s income; lower-income families have a lower cap. The amount of the tax credit remains the same, so a person who purchases a more expensive plan pays the cost difference (TABLE 2).

The ruling’s effect on women’s health
On June 26, American College of Obstetricians and Gynecologists President Mark S. DeFrancesco, MD, MBA, hailed the Supreme Court decision, saying, “Importantly, recent data have shown that newly insured adults under the ACA were more likely to be ­women. Those who did gain coverage through the ACA reported better access to health care and better financial security from medical costs.”

“Without question, many women enrollees were able to purchase health insurance coverage due, in part, to the ACA subsidies that helped make this purchase affordable. In fact, government data have suggested that roughly 85% of health exchange enrollees received subsidies,” Dr. DeFrancesco said.

“If the Supreme Court had overturned this important assistance, approximately 4.8 million women would have been unable to afford the coverage that they need. The impact also would have been widespread; as these women were forced to leave the insurance marketplace, it is likely that premiums throughout the marketplace would have risen dramatically,” he continued.

“Instead, patients—especially the low- and moderate-income American women who have particularly benefited from ACA subsidies—will continue to have the peace of mind that comes with insurance coverage.”

Share your thoughts! Send your Letter to the Editor to rbarbieri@frontlinemedcom.com. Please include your name and the city and state in which you practice.