Resuming low-dose aspirin after an initial lower gastrointestinal bleed significantly increased the chances of recurrence but protected against serious cardiovascular events, based on a single-center retrospective study published in the August issue of Gastroenterology.

In contrast, “we did not find concomitant use of anticoagulants, antiplatelets, and steroids as a predictor of recurrent lower GI bleeding,” said Dr. Francis Chan of the Prince of Wales Hospital in Hong Kong and his associates. “This may be due to the low percentage of concomitant drug use in both groups. Multicenter studies with a large number of patients will be required to identify additional risk factors for recurrent lower GI bleeding with aspirin use.”

Low-dose aspirin has long been known to help prevent coronary artery and cerebrovascular disease, and more recently has been found to potentially reduce the risk of several types of cancer, the researchers noted. Aspirin is well known to increase the risk of upper GI bleeding, but some studies have also linked it to lower GI bleeding. However, “patients with underlying cardiovascular diseases often require lifelong aspirin,” they added. The risks and benefits of stopping or remaining on aspirin after an initial lower GI bleed are unclear (Gastroenterology 2016 Apr 26. doi: 10.1053/j.gastro.2016.04.013).

Accordingly, the researchers retrospectively studied 295 patients who had an initial aspirin-associated lower GI bleed, defined as 325 mg aspirin a day within a week of bleeding onset. All patients had melena or hematochezia documented by an attending physician and had no endoscopic evidence of upper GI bleeding.

For patients who continued using aspirin at least half the time, the 5-year cumulative incidence of recurrent lower GI bleeding was 19% (95% confidence interval [CI], 13%-25%) – more than double the rate among patients who used aspirin 20% or less of the time (5-year cumulative incidence, 7%; 95% CI, 3%-13%; P = .01). However, the 5-year cumulative incidence of serious cardiovascular events among nonusers was 37% (95% CI, 27%-46%), while the rate among aspirin users was 23% (95% CI, 17%-30%; P = .02). Mortality from noncardiovascular causes was also higher among nonusers (27%) than users (8%; P less than .001), probably because nonusers of aspirin tended to be older than users, but perhaps also because aspirin had a “nonvascular protective effect,” the researchers said.

A multivariate analysis confirmed these findings, linking lower GI bleeding to aspirin but not to use of steroids, anticoagulants, or antiplatelet drugs, or to age, sex, alcohol consumption, smoking, comorbidities, or cardiovascular risks. Indeed, continued aspirin use nearly tripled the chances of a recurrent lower GI bleed (hazard ratio, 2.76; 95% CI, 1.3-6.0; P = .01), but cut the risk of serious cardiovascular events by about 40% (HR, 0.59; 95% CI, 0.4-0.9; P = .02).

Deciding whether to resume aspirin after a severe lower GI bleed “presents a management dilemma for physicians, patients, and their families, particularly in the absence of risk-mitigating therapies and a lack of data on the risks and benefits of resuming aspirin,” the investigators emphasized. Their findings highlight the importance of weighing the cardiovascular benefits of aspirin against GI toxicity, they said. “Since there is substantial risk of recurrent bleeding, physicians should critically evaluate individual patients’ cardiovascular risk before resuming aspirin therapy. Our findings also suggest a need for a composite endpoint to evaluate clinically significant events throughout the GI tract in patients receiving antiplatelet drugs.”

The Chinese University of Hong Kong funded the study. Dr. Chan reported financial ties to Pfizer, Eisai, Takeda, Otsuka, and Astrazeneca.

Resuming low-dose aspirin after an initial lower gastrointestinal bleed significantly increased the chances of recurrence but protected against serious cardiovascular events, based on a single-center retrospective study published in the August issue of Gastroenterology.

In contrast, “we did not find concomitant use of anticoagulants, antiplatelets, and steroids as a predictor of recurrent lower GI bleeding,” said Dr. Francis Chan of the Prince of Wales Hospital in Hong Kong and his associates. “This may be due to the low percentage of concomitant drug use in both groups. Multicenter studies with a large number of patients will be required to identify additional risk factors for recurrent lower GI bleeding with aspirin use.”

Low-dose aspirin has long been known to help prevent coronary artery and cerebrovascular disease, and more recently has been found to potentially reduce the risk of several types of cancer, the researchers noted. Aspirin is well known to increase the risk of upper GI bleeding, but some studies have also linked it to lower GI bleeding. However, “patients with underlying cardiovascular diseases often require lifelong aspirin,” they added. The risks and benefits of stopping or remaining on aspirin after an initial lower GI bleed are unclear (Gastroenterology 2016 Apr 26. doi: 10.1053/j.gastro.2016.04.013).

Accordingly, the researchers retrospectively studied 295 patients who had an initial aspirin-associated lower GI bleed, defined as 325 mg aspirin a day within a week of bleeding onset. All patients had melena or hematochezia documented by an attending physician and had no endoscopic evidence of upper GI bleeding.

For patients who continued using aspirin at least half the time, the 5-year cumulative incidence of recurrent lower GI bleeding was 19% (95% confidence interval [CI], 13%-25%) – more than double the rate among patients who used aspirin 20% or less of the time (5-year cumulative incidence, 7%; 95% CI, 3%-13%; P = .01). However, the 5-year cumulative incidence of serious cardiovascular events among nonusers was 37% (95% CI, 27%-46%), while the rate among aspirin users was 23% (95% CI, 17%-30%; P = .02). Mortality from noncardiovascular causes was also higher among nonusers (27%) than users (8%; P less than .001), probably because nonusers of aspirin tended to be older than users, but perhaps also because aspirin had a “nonvascular protective effect,” the researchers said.

A multivariate analysis confirmed these findings, linking lower GI bleeding to aspirin but not to use of steroids, anticoagulants, or antiplatelet drugs, or to age, sex, alcohol consumption, smoking, comorbidities, or cardiovascular risks. Indeed, continued aspirin use nearly tripled the chances of a recurrent lower GI bleed (hazard ratio, 2.76; 95% CI, 1.3-6.0; P = .01), but cut the risk of serious cardiovascular events by about 40% (HR, 0.59; 95% CI, 0.4-0.9; P = .02).

Deciding whether to resume aspirin after a severe lower GI bleed “presents a management dilemma for physicians, patients, and their families, particularly in the absence of risk-mitigating therapies and a lack of data on the risks and benefits of resuming aspirin,” the investigators emphasized. Their findings highlight the importance of weighing the cardiovascular benefits of aspirin against GI toxicity, they said. “Since there is substantial risk of recurrent bleeding, physicians should critically evaluate individual patients’ cardiovascular risk before resuming aspirin therapy. Our findings also suggest a need for a composite endpoint to evaluate clinically significant events throughout the GI tract in patients receiving antiplatelet drugs.”

The Chinese University of Hong Kong funded the study. Dr. Chan reported financial ties to Pfizer, Eisai, Takeda, Otsuka, and Astrazeneca.

Resuming low-dose aspirin after an initial lower gastrointestinal bleed significantly increased the chances of recurrence but protected against serious cardiovascular events, based on a single-center retrospective study published in the August issue of Gastroenterology.

In contrast, “we did not find concomitant use of anticoagulants, antiplatelets, and steroids as a predictor of recurrent lower GI bleeding,” said Dr. Francis Chan of the Prince of Wales Hospital in Hong Kong and his associates. “This may be due to the low percentage of concomitant drug use in both groups. Multicenter studies with a large number of patients will be required to identify additional risk factors for recurrent lower GI bleeding with aspirin use.”

Low-dose aspirin has long been known to help prevent coronary artery and cerebrovascular disease, and more recently has been found to potentially reduce the risk of several types of cancer, the researchers noted. Aspirin is well known to increase the risk of upper GI bleeding, but some studies have also linked it to lower GI bleeding. However, “patients with underlying cardiovascular diseases often require lifelong aspirin,” they added. The risks and benefits of stopping or remaining on aspirin after an initial lower GI bleed are unclear (Gastroenterology 2016 Apr 26. doi: 10.1053/j.gastro.2016.04.013).

Accordingly, the researchers retrospectively studied 295 patients who had an initial aspirin-associated lower GI bleed, defined as 325 mg aspirin a day within a week of bleeding onset. All patients had melena or hematochezia documented by an attending physician and had no endoscopic evidence of upper GI bleeding.

For patients who continued using aspirin at least half the time, the 5-year cumulative incidence of recurrent lower GI bleeding was 19% (95% confidence interval [CI], 13%-25%) – more than double the rate among patients who used aspirin 20% or less of the time (5-year cumulative incidence, 7%; 95% CI, 3%-13%; P = .01). However, the 5-year cumulative incidence of serious cardiovascular events among nonusers was 37% (95% CI, 27%-46%), while the rate among aspirin users was 23% (95% CI, 17%-30%; P = .02). Mortality from noncardiovascular causes was also higher among nonusers (27%) than users (8%; P less than .001), probably because nonusers of aspirin tended to be older than users, but perhaps also because aspirin had a “nonvascular protective effect,” the researchers said.

A multivariate analysis confirmed these findings, linking lower GI bleeding to aspirin but not to use of steroids, anticoagulants, or antiplatelet drugs, or to age, sex, alcohol consumption, smoking, comorbidities, or cardiovascular risks. Indeed, continued aspirin use nearly tripled the chances of a recurrent lower GI bleed (hazard ratio, 2.76; 95% CI, 1.3-6.0; P = .01), but cut the risk of serious cardiovascular events by about 40% (HR, 0.59; 95% CI, 0.4-0.9; P = .02).

Deciding whether to resume aspirin after a severe lower GI bleed “presents a management dilemma for physicians, patients, and their families, particularly in the absence of risk-mitigating therapies and a lack of data on the risks and benefits of resuming aspirin,” the investigators emphasized. Their findings highlight the importance of weighing the cardiovascular benefits of aspirin against GI toxicity, they said. “Since there is substantial risk of recurrent bleeding, physicians should critically evaluate individual patients’ cardiovascular risk before resuming aspirin therapy. Our findings also suggest a need for a composite endpoint to evaluate clinically significant events throughout the GI tract in patients receiving antiplatelet drugs.”

The Chinese University of Hong Kong funded the study. Dr. Chan reported financial ties to Pfizer, Eisai, Takeda, Otsuka, and Astrazeneca.

A single, 100-mg rectal dose of indomethacin cut the odds of moderate to severe pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP) by 85% in a single-center retrospective study of more than 4,000 patients reported in the August issue of Gastroenterology.

The effect extended to low-risk patients and those with malignant biliary obstruction, who make up the majority of ERCP patients in community practice, said Nikhil R. Thiruvengadam, MD, and his associates at the University of Pennsylvania. “Usage of rectal indomethacin in current clinical practice is low, as most endoscopists outside of referral centers perform ERCP for indications that are considered low-risk for PEP [post-ERCP pancreatitis], and until now, there were no data to support a benefit of rectal NSAIDs in this population,” they wrote in Gastroenterology. Their “real-world analysis” clearly shows the benefits of rectal indomethacin in low-risk patients and supports its increased use after ERCP, they added.

Pancreatitis, the most common complication of ERCP, affected 2%-9% of patients in prior studies and costs about $200 million in the United States annually, the investigators noted. Pancreatic duct stents help prevent post-ERCP pancreatitis, but require experience to place and have their own complications that limit their use in low-risk patients. Past studies of rectal indomethacin after ERCP reported mixed results and mainly focused on high-risk patients, leaving questions about whether to routinely use this NSAID after ERCP, said the researchers (Gastroenterology. 2016 May 20. doi: 10.1053/j.gastro.2016.04.048). Their study included 4,017 patients who underwent ERCP at the University of Pennsylvania between 2009 and 2015. From 2012 onward, nearly all patients received 100 mg rectal indomethacin immediately after the duodenoscope was withdrawn. This indomethacin group included 2,007 patients, while 2,010 patients in the study did not receive rectal indomethacin. In all, 95 (4.73%) untreated patients developed post-ERCP pancreatitis, compared with only 40 (1.99%) patients who received indomethacin, for a 65% reduction in the odds of post-ERCP pancreatitis (odds ratio, 0.35; 95% confidence interval, 0.24-0.51; P less than .001). Rectal indomethacin also led to an 83% drop in the odds of moderate to severe post-ERCP pancreatitis (OR, 0.17; 95% CI, 0.09-0.32; P less than .001) and showed very similar protective effects for patients with malignant obstruction (OR, 0.35; 95% CI, 0.17-0.75; P less than .001] and 0.20; 95% CI, 0.07-0.63; P less than 0.001, respectively).

Rectal indomethacin was particularly beneficial for patients with malignant obstruction and pancreatic adenocarcinoma, the investigators noted. Such patients had post-ERCP rates of 2.31% when they received rectal indomethacin and 7.53% otherwise (P less than .001). They also had a nearly sevenfold lower rate of moderate to severe post-ERCP pancreatitis when they received rectal indomethacin (P = .001).

Treatment did not affect the chances of perforation and did not cause anaphylaxis, but was tied to a slightly higher rate of postprocedural gastrointestinal bleeding among sphincterotomy patients (0.65% with treatment versus 0.45% without; P = .52). However, sphincterotomy patients were much less likely to develop pancreatitis when they received rectal indomethacin than when they did not (0% and 9.58% of patients, respectively; P = .003).

“The majority of ERCPs were performed by experienced endoscopists at a tertiary care center, which may have limited the effects of variable procedural skills on the risk of PEP,” the researchers said. “Therefore, generalizability of our findings to other populations may be limited. However, it should be noted that the overall PEP rate in both the unexposed and indomethacin groups was fairly low and similar to large community-based estimates, suggesting that our overall patient population was of similar overall risk.” The study was not powered to assess the combined effects of rectal indomethacin and pancreatic duct stents, they noted.

The investigators reported no funding sources and had no disclosures.

A single, 100-mg rectal dose of indomethacin cut the odds of moderate to severe pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP) by 85% in a single-center retrospective study of more than 4,000 patients reported in the August issue of Gastroenterology.

The effect extended to low-risk patients and those with malignant biliary obstruction, who make up the majority of ERCP patients in community practice, said Nikhil R. Thiruvengadam, MD, and his associates at the University of Pennsylvania. “Usage of rectal indomethacin in current clinical practice is low, as most endoscopists outside of referral centers perform ERCP for indications that are considered low-risk for PEP [post-ERCP pancreatitis], and until now, there were no data to support a benefit of rectal NSAIDs in this population,” they wrote in Gastroenterology. Their “real-world analysis” clearly shows the benefits of rectal indomethacin in low-risk patients and supports its increased use after ERCP, they added.

Pancreatitis, the most common complication of ERCP, affected 2%-9% of patients in prior studies and costs about $200 million in the United States annually, the investigators noted. Pancreatic duct stents help prevent post-ERCP pancreatitis, but require experience to place and have their own complications that limit their use in low-risk patients. Past studies of rectal indomethacin after ERCP reported mixed results and mainly focused on high-risk patients, leaving questions about whether to routinely use this NSAID after ERCP, said the researchers (Gastroenterology. 2016 May 20. doi: 10.1053/j.gastro.2016.04.048). Their study included 4,017 patients who underwent ERCP at the University of Pennsylvania between 2009 and 2015. From 2012 onward, nearly all patients received 100 mg rectal indomethacin immediately after the duodenoscope was withdrawn. This indomethacin group included 2,007 patients, while 2,010 patients in the study did not receive rectal indomethacin. In all, 95 (4.73%) untreated patients developed post-ERCP pancreatitis, compared with only 40 (1.99%) patients who received indomethacin, for a 65% reduction in the odds of post-ERCP pancreatitis (odds ratio, 0.35; 95% confidence interval, 0.24-0.51; P less than .001). Rectal indomethacin also led to an 83% drop in the odds of moderate to severe post-ERCP pancreatitis (OR, 0.17; 95% CI, 0.09-0.32; P less than .001) and showed very similar protective effects for patients with malignant obstruction (OR, 0.35; 95% CI, 0.17-0.75; P less than .001] and 0.20; 95% CI, 0.07-0.63; P less than 0.001, respectively).

Rectal indomethacin was particularly beneficial for patients with malignant obstruction and pancreatic adenocarcinoma, the investigators noted. Such patients had post-ERCP rates of 2.31% when they received rectal indomethacin and 7.53% otherwise (P less than .001). They also had a nearly sevenfold lower rate of moderate to severe post-ERCP pancreatitis when they received rectal indomethacin (P = .001).

Treatment did not affect the chances of perforation and did not cause anaphylaxis, but was tied to a slightly higher rate of postprocedural gastrointestinal bleeding among sphincterotomy patients (0.65% with treatment versus 0.45% without; P = .52). However, sphincterotomy patients were much less likely to develop pancreatitis when they received rectal indomethacin than when they did not (0% and 9.58% of patients, respectively; P = .003).

“The majority of ERCPs were performed by experienced endoscopists at a tertiary care center, which may have limited the effects of variable procedural skills on the risk of PEP,” the researchers said. “Therefore, generalizability of our findings to other populations may be limited. However, it should be noted that the overall PEP rate in both the unexposed and indomethacin groups was fairly low and similar to large community-based estimates, suggesting that our overall patient population was of similar overall risk.” The study was not powered to assess the combined effects of rectal indomethacin and pancreatic duct stents, they noted.

The investigators reported no funding sources and had no disclosures.

A single, 100-mg rectal dose of indomethacin cut the odds of moderate to severe pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP) by 85% in a single-center retrospective study of more than 4,000 patients reported in the August issue of Gastroenterology.

The effect extended to low-risk patients and those with malignant biliary obstruction, who make up the majority of ERCP patients in community practice, said Nikhil R. Thiruvengadam, MD, and his associates at the University of Pennsylvania. “Usage of rectal indomethacin in current clinical practice is low, as most endoscopists outside of referral centers perform ERCP for indications that are considered low-risk for PEP [post-ERCP pancreatitis], and until now, there were no data to support a benefit of rectal NSAIDs in this population,” they wrote in Gastroenterology. Their “real-world analysis” clearly shows the benefits of rectal indomethacin in low-risk patients and supports its increased use after ERCP, they added.

Pancreatitis, the most common complication of ERCP, affected 2%-9% of patients in prior studies and costs about $200 million in the United States annually, the investigators noted. Pancreatic duct stents help prevent post-ERCP pancreatitis, but require experience to place and have their own complications that limit their use in low-risk patients. Past studies of rectal indomethacin after ERCP reported mixed results and mainly focused on high-risk patients, leaving questions about whether to routinely use this NSAID after ERCP, said the researchers (Gastroenterology. 2016 May 20. doi: 10.1053/j.gastro.2016.04.048). Their study included 4,017 patients who underwent ERCP at the University of Pennsylvania between 2009 and 2015. From 2012 onward, nearly all patients received 100 mg rectal indomethacin immediately after the duodenoscope was withdrawn. This indomethacin group included 2,007 patients, while 2,010 patients in the study did not receive rectal indomethacin. In all, 95 (4.73%) untreated patients developed post-ERCP pancreatitis, compared with only 40 (1.99%) patients who received indomethacin, for a 65% reduction in the odds of post-ERCP pancreatitis (odds ratio, 0.35; 95% confidence interval, 0.24-0.51; P less than .001). Rectal indomethacin also led to an 83% drop in the odds of moderate to severe post-ERCP pancreatitis (OR, 0.17; 95% CI, 0.09-0.32; P less than .001) and showed very similar protective effects for patients with malignant obstruction (OR, 0.35; 95% CI, 0.17-0.75; P less than .001] and 0.20; 95% CI, 0.07-0.63; P less than 0.001, respectively).

Rectal indomethacin was particularly beneficial for patients with malignant obstruction and pancreatic adenocarcinoma, the investigators noted. Such patients had post-ERCP rates of 2.31% when they received rectal indomethacin and 7.53% otherwise (P less than .001). They also had a nearly sevenfold lower rate of moderate to severe post-ERCP pancreatitis when they received rectal indomethacin (P = .001).

Treatment did not affect the chances of perforation and did not cause anaphylaxis, but was tied to a slightly higher rate of postprocedural gastrointestinal bleeding among sphincterotomy patients (0.65% with treatment versus 0.45% without; P = .52). However, sphincterotomy patients were much less likely to develop pancreatitis when they received rectal indomethacin than when they did not (0% and 9.58% of patients, respectively; P = .003).

“The majority of ERCPs were performed by experienced endoscopists at a tertiary care center, which may have limited the effects of variable procedural skills on the risk of PEP,” the researchers said. “Therefore, generalizability of our findings to other populations may be limited. However, it should be noted that the overall PEP rate in both the unexposed and indomethacin groups was fairly low and similar to large community-based estimates, suggesting that our overall patient population was of similar overall risk.” The study was not powered to assess the combined effects of rectal indomethacin and pancreatic duct stents, they noted.

The investigators reported no funding sources and had no disclosures.

To the Editor: I read the excellent overview article on measles by Drs. Kumar and Sabella.¹ However, there are additional important clinical points regarding measles diagnosis that deserve further comment. Prior to definitive diagnosis, measles is a clinical diagnosis. Properly, much attention is given to the rash, but there are important clinical clues besides the rash that are helpful diagnostically.

Some clinical findings are more characteristic of a disease than others, eg, eye findings in measles, and clinicians should specifically look for them. Other findings not characteristic but consistent with the diagnosis are less helpful, eg, measles with diarrhea.² Measles is a systemic infection with several extradermatologic manifestations. Characteristically, measles involves the respiratory tract, manifested by runny nose, dry cough, or shortness of breath, ie, measles pneumonia.²

Gastrointestinal involvement may be manifested as nausea, vomiting, diarrhea, or abdominal pain. Abdominal pain, when located in the right lower quadrant, may mimic acute appendicitis, ie, pseudoappendicitis. In patients undergoing appendectomy, pathologically the appendix is normal but contains multinucleated giant cells (Warthin-Finkeldey cells). Measles pseudoappendicitis can be perplexing because it presents before the rash. Even without appendiceal involvement (pseudoappendicitis), Warthin-Finkeldey cells are also present in the nasal mucosa. If nasal swab cytology shows Warthin-Finkeldey cells, an early diagnosis of measles may be made days before IgM measles antibodies are reported.³

Other often-overlooked important clues are eye findings, eg, watery eyes, lower lid edema, conjunctival suffusion, and keratitis. No other disease in the differential diagnosis of measles presents with watery eyes with bilateral lower lid edema and conjunctival suffusion. Headache, mental confusion or oral ulcers may suggest alternate diagnoses. Typically, nonspecific laboratory abnormalities include leukopenia and thrombocytopenia, and importantly, the erythrocyte sedimentation rate is not elevated (Table 1).^4,5

In measles, much is made of Koplik spots, which are found early on dark red buccal mucosa (vs Fordyce spots).¹ However, if these are missed or not present, clinicians can use other characteristic findings to make a presumptive diagnosis of measles.

To the Editor: I read the excellent overview article on measles by Drs. Kumar and Sabella.¹ However, there are additional important clinical points regarding measles diagnosis that deserve further comment. Prior to definitive diagnosis, measles is a clinical diagnosis. Properly, much attention is given to the rash, but there are important clinical clues besides the rash that are helpful diagnostically.

Some clinical findings are more characteristic of a disease than others, eg, eye findings in measles, and clinicians should specifically look for them. Other findings not characteristic but consistent with the diagnosis are less helpful, eg, measles with diarrhea.² Measles is a systemic infection with several extradermatologic manifestations. Characteristically, measles involves the respiratory tract, manifested by runny nose, dry cough, or shortness of breath, ie, measles pneumonia.²

Gastrointestinal involvement may be manifested as nausea, vomiting, diarrhea, or abdominal pain. Abdominal pain, when located in the right lower quadrant, may mimic acute appendicitis, ie, pseudoappendicitis. In patients undergoing appendectomy, pathologically the appendix is normal but contains multinucleated giant cells (Warthin-Finkeldey cells). Measles pseudoappendicitis can be perplexing because it presents before the rash. Even without appendiceal involvement (pseudoappendicitis), Warthin-Finkeldey cells are also present in the nasal mucosa. If nasal swab cytology shows Warthin-Finkeldey cells, an early diagnosis of measles may be made days before IgM measles antibodies are reported.³

Other often-overlooked important clues are eye findings, eg, watery eyes, lower lid edema, conjunctival suffusion, and keratitis. No other disease in the differential diagnosis of measles presents with watery eyes with bilateral lower lid edema and conjunctival suffusion. Headache, mental confusion or oral ulcers may suggest alternate diagnoses. Typically, nonspecific laboratory abnormalities include leukopenia and thrombocytopenia, and importantly, the erythrocyte sedimentation rate is not elevated (Table 1).^4,5

In measles, much is made of Koplik spots, which are found early on dark red buccal mucosa (vs Fordyce spots).¹ However, if these are missed or not present, clinicians can use other characteristic findings to make a presumptive diagnosis of measles.

To the Editor: I read the excellent overview article on measles by Drs. Kumar and Sabella.¹ However, there are additional important clinical points regarding measles diagnosis that deserve further comment. Prior to definitive diagnosis, measles is a clinical diagnosis. Properly, much attention is given to the rash, but there are important clinical clues besides the rash that are helpful diagnostically.

Some clinical findings are more characteristic of a disease than others, eg, eye findings in measles, and clinicians should specifically look for them. Other findings not characteristic but consistent with the diagnosis are less helpful, eg, measles with diarrhea.² Measles is a systemic infection with several extradermatologic manifestations. Characteristically, measles involves the respiratory tract, manifested by runny nose, dry cough, or shortness of breath, ie, measles pneumonia.²

Gastrointestinal involvement may be manifested as nausea, vomiting, diarrhea, or abdominal pain. Abdominal pain, when located in the right lower quadrant, may mimic acute appendicitis, ie, pseudoappendicitis. In patients undergoing appendectomy, pathologically the appendix is normal but contains multinucleated giant cells (Warthin-Finkeldey cells). Measles pseudoappendicitis can be perplexing because it presents before the rash. Even without appendiceal involvement (pseudoappendicitis), Warthin-Finkeldey cells are also present in the nasal mucosa. If nasal swab cytology shows Warthin-Finkeldey cells, an early diagnosis of measles may be made days before IgM measles antibodies are reported.³

Other often-overlooked important clues are eye findings, eg, watery eyes, lower lid edema, conjunctival suffusion, and keratitis. No other disease in the differential diagnosis of measles presents with watery eyes with bilateral lower lid edema and conjunctival suffusion. Headache, mental confusion or oral ulcers may suggest alternate diagnoses. Typically, nonspecific laboratory abnormalities include leukopenia and thrombocytopenia, and importantly, the erythrocyte sedimentation rate is not elevated (Table 1).^4,5

In measles, much is made of Koplik spots, which are found early on dark red buccal mucosa (vs Fordyce spots).¹ However, if these are missed or not present, clinicians can use other characteristic findings to make a presumptive diagnosis of measles.

In Reply: We thank Dr. Cunha for his comments and appreciate the opportunity to emphasize important points that he highlights.

We agree that measles is a systemic illness with important extradermatologic manifestations that are critical to the diagnosis, and that the nondermatologic manifestations often precede the rash and serve to distinguish measles from other systemic illnesses. As discussed in our review, the respiratory prodrome of cough, coryza, and conjunctivitis is very distinctive and serves as an important clue to the diagnosis. Likewise, we acknowledge the importance of gastrointestinal findings in measles and note appendicitis as an important complication that is well described. Although Koplik spots are pathognomonic, we do stress that these often are not present at the time of presentation.

Finally, we agree that measles is a clinical diagnosis, and that the clinical manifestations beyond the dermatologic manifestations noted in our review and highlighted by Dr. Cunha are extremely helpful to the clinician in considering the diagnosis.

In Reply: We thank Dr. Cunha for his comments and appreciate the opportunity to emphasize important points that he highlights.

We agree that measles is a systemic illness with important extradermatologic manifestations that are critical to the diagnosis, and that the nondermatologic manifestations often precede the rash and serve to distinguish measles from other systemic illnesses. As discussed in our review, the respiratory prodrome of cough, coryza, and conjunctivitis is very distinctive and serves as an important clue to the diagnosis. Likewise, we acknowledge the importance of gastrointestinal findings in measles and note appendicitis as an important complication that is well described. Although Koplik spots are pathognomonic, we do stress that these often are not present at the time of presentation.

Finally, we agree that measles is a clinical diagnosis, and that the clinical manifestations beyond the dermatologic manifestations noted in our review and highlighted by Dr. Cunha are extremely helpful to the clinician in considering the diagnosis.

In Reply: We thank Dr. Cunha for his comments and appreciate the opportunity to emphasize important points that he highlights.

We agree that measles is a systemic illness with important extradermatologic manifestations that are critical to the diagnosis, and that the nondermatologic manifestations often precede the rash and serve to distinguish measles from other systemic illnesses. As discussed in our review, the respiratory prodrome of cough, coryza, and conjunctivitis is very distinctive and serves as an important clue to the diagnosis. Likewise, we acknowledge the importance of gastrointestinal findings in measles and note appendicitis as an important complication that is well described. Although Koplik spots are pathognomonic, we do stress that these often are not present at the time of presentation.

Finally, we agree that measles is a clinical diagnosis, and that the clinical manifestations beyond the dermatologic manifestations noted in our review and highlighted by Dr. Cunha are extremely helpful to the clinician in considering the diagnosis.

To the Editor: As a primary care physician who has practiced for 31 years, I have a few concerns about the article “Prescribing opioids in primary care: Safely starting, monitoring, and stopping” by Drs. Daniel Tobin, Rebecca Andrews, and William Becker in your March 2016 issue.

Prescribing narcotics and other controlled medications has become a hot topic with legal implications. Many physicians have lost their license, and some have lost their freedom as well.

The article notes that primary care physicians provide most of the care for chronic pain but implies that we don’t know what we are doing. Although we might not have completed a residency in the management of chronic pain, we were required to attend a graduate medical school and to complete residency programs. We are also required to read the medical literature and keep up with our CME requirements. We too strive to keep up with and practice the latest cutting-edge medicine. And we have experience. After seeing thousands of patients and writing millions of prescriptions, I think I do know something about whether or not a medicine is safe, effective, and cost-effective. I have encountered quite a few patients with pain, and not one of them has overdosed by taking the medications as prescribed.

Most of what is being said and published about pain medication focuses on the epidemic of young people who are getting pain medications (prescription narcotics) and street narcotics (heroin) illegally. Directly, it has nothing to do with conscientious physicians prescribing narcotics for pain, but indirectly, it has to do with all adults in society. We are allowing our young people to be destroyed by drugs and by a lack of proper discipline in our homes that spills over into society. Where are our children getting these drugs, and who is bringing them into our neighborhoods? The practice of medicine is not the primary problem. I am for scientific, conscientious regulations concerning controlled substances. Medicine needs to inform all doctors about the changing laws surrounding the prescribing of controlled substances.

Cigarettes and alcohol are sold on every corner, and places selling marijuana are popping up everywhere. The former two drugs are harming and killing more young people than all the illegal drugs combined, and marijuana is fast approaching number three. I feel we need to stop picking on the medical profession and stop trying to blame it for all the woes that our young people are encountering every day. The reality is that legal and illegal drugs are not going to go away. We need to love our children more and better prepare them while in our homes to deal with the evils in our society.

To the Editor: As a primary care physician who has practiced for 31 years, I have a few concerns about the article “Prescribing opioids in primary care: Safely starting, monitoring, and stopping” by Drs. Daniel Tobin, Rebecca Andrews, and William Becker in your March 2016 issue.

Prescribing narcotics and other controlled medications has become a hot topic with legal implications. Many physicians have lost their license, and some have lost their freedom as well.

The article notes that primary care physicians provide most of the care for chronic pain but implies that we don’t know what we are doing. Although we might not have completed a residency in the management of chronic pain, we were required to attend a graduate medical school and to complete residency programs. We are also required to read the medical literature and keep up with our CME requirements. We too strive to keep up with and practice the latest cutting-edge medicine. And we have experience. After seeing thousands of patients and writing millions of prescriptions, I think I do know something about whether or not a medicine is safe, effective, and cost-effective. I have encountered quite a few patients with pain, and not one of them has overdosed by taking the medications as prescribed.

Most of what is being said and published about pain medication focuses on the epidemic of young people who are getting pain medications (prescription narcotics) and street narcotics (heroin) illegally. Directly, it has nothing to do with conscientious physicians prescribing narcotics for pain, but indirectly, it has to do with all adults in society. We are allowing our young people to be destroyed by drugs and by a lack of proper discipline in our homes that spills over into society. Where are our children getting these drugs, and who is bringing them into our neighborhoods? The practice of medicine is not the primary problem. I am for scientific, conscientious regulations concerning controlled substances. Medicine needs to inform all doctors about the changing laws surrounding the prescribing of controlled substances.

Cigarettes and alcohol are sold on every corner, and places selling marijuana are popping up everywhere. The former two drugs are harming and killing more young people than all the illegal drugs combined, and marijuana is fast approaching number three. I feel we need to stop picking on the medical profession and stop trying to blame it for all the woes that our young people are encountering every day. The reality is that legal and illegal drugs are not going to go away. We need to love our children more and better prepare them while in our homes to deal with the evils in our society.

To the Editor: As a primary care physician who has practiced for 31 years, I have a few concerns about the article “Prescribing opioids in primary care: Safely starting, monitoring, and stopping” by Drs. Daniel Tobin, Rebecca Andrews, and William Becker in your March 2016 issue.

Prescribing narcotics and other controlled medications has become a hot topic with legal implications. Many physicians have lost their license, and some have lost their freedom as well.

The article notes that primary care physicians provide most of the care for chronic pain but implies that we don’t know what we are doing. Although we might not have completed a residency in the management of chronic pain, we were required to attend a graduate medical school and to complete residency programs. We are also required to read the medical literature and keep up with our CME requirements. We too strive to keep up with and practice the latest cutting-edge medicine. And we have experience. After seeing thousands of patients and writing millions of prescriptions, I think I do know something about whether or not a medicine is safe, effective, and cost-effective. I have encountered quite a few patients with pain, and not one of them has overdosed by taking the medications as prescribed.

Most of what is being said and published about pain medication focuses on the epidemic of young people who are getting pain medications (prescription narcotics) and street narcotics (heroin) illegally. Directly, it has nothing to do with conscientious physicians prescribing narcotics for pain, but indirectly, it has to do with all adults in society. We are allowing our young people to be destroyed by drugs and by a lack of proper discipline in our homes that spills over into society. Where are our children getting these drugs, and who is bringing them into our neighborhoods? The practice of medicine is not the primary problem. I am for scientific, conscientious regulations concerning controlled substances. Medicine needs to inform all doctors about the changing laws surrounding the prescribing of controlled substances.

Cigarettes and alcohol are sold on every corner, and places selling marijuana are popping up everywhere. The former two drugs are harming and killing more young people than all the illegal drugs combined, and marijuana is fast approaching number three. I feel we need to stop picking on the medical profession and stop trying to blame it for all the woes that our young people are encountering every day. The reality is that legal and illegal drugs are not going to go away. We need to love our children more and better prepare them while in our homes to deal with the evils in our society.

In Reply: We thank Dr. Pettiway for his remarks. The intent of our article was to identify common challenges when prescribing opioids for chronic pain and offer tips to the provider struggling with how to do so safely. We hope these comments will offer additional clarity.

First, as general internists who are essentially “self-trained” in the management of chronic pain, we fully acknowledge the importance of practical experience in learning how to prescribe opioids safely and effectively. Dr. Pettiway is correct that a dedicated physician who keeps up with the medical literature, attends relevant continuing medical education courses, and strives to provide deliberate, rational, and evidence-based care to his or her patients can do so effectively. However, the medical literature suggests that medical school training in the management of chronic pain is sparse; one review found that in 2011 only 5 out of 133 US medical schools required coursework on pain management, and only 13 offered it as an elective.¹ Many primary care providers do feel unprepared to handle this challenge.

Additionally, Dr. Pettiway raises a good question about where misused prescription opioids originate and whether prescribers are responsible. The data show that the majority of misused prescription opioids are obtained from a family member or friend and not directly from a physician.^2,3 However, this supply does generally originate from a prescription. Providers need to educate their patients about the risk for diversion, the need to keep pills safely hidden and locked away, and the importance of safely discarding unused supplies. Responsible prescribers need to anticipate these concerns and educate patients about them.

In summary, we firmly believe that primary care providers are capable of safe, effective, and responsible opioid prescribing and hope that our paper provides additional guidance on how to do so.

In Reply: We thank Dr. Pettiway for his remarks. The intent of our article was to identify common challenges when prescribing opioids for chronic pain and offer tips to the provider struggling with how to do so safely. We hope these comments will offer additional clarity.

First, as general internists who are essentially “self-trained” in the management of chronic pain, we fully acknowledge the importance of practical experience in learning how to prescribe opioids safely and effectively. Dr. Pettiway is correct that a dedicated physician who keeps up with the medical literature, attends relevant continuing medical education courses, and strives to provide deliberate, rational, and evidence-based care to his or her patients can do so effectively. However, the medical literature suggests that medical school training in the management of chronic pain is sparse; one review found that in 2011 only 5 out of 133 US medical schools required coursework on pain management, and only 13 offered it as an elective.¹ Many primary care providers do feel unprepared to handle this challenge.

Additionally, Dr. Pettiway raises a good question about where misused prescription opioids originate and whether prescribers are responsible. The data show that the majority of misused prescription opioids are obtained from a family member or friend and not directly from a physician.^2,3 However, this supply does generally originate from a prescription. Providers need to educate their patients about the risk for diversion, the need to keep pills safely hidden and locked away, and the importance of safely discarding unused supplies. Responsible prescribers need to anticipate these concerns and educate patients about them.

In summary, we firmly believe that primary care providers are capable of safe, effective, and responsible opioid prescribing and hope that our paper provides additional guidance on how to do so.

In Reply: We thank Dr. Pettiway for his remarks. The intent of our article was to identify common challenges when prescribing opioids for chronic pain and offer tips to the provider struggling with how to do so safely. We hope these comments will offer additional clarity.

First, as general internists who are essentially “self-trained” in the management of chronic pain, we fully acknowledge the importance of practical experience in learning how to prescribe opioids safely and effectively. Dr. Pettiway is correct that a dedicated physician who keeps up with the medical literature, attends relevant continuing medical education courses, and strives to provide deliberate, rational, and evidence-based care to his or her patients can do so effectively. However, the medical literature suggests that medical school training in the management of chronic pain is sparse; one review found that in 2011 only 5 out of 133 US medical schools required coursework on pain management, and only 13 offered it as an elective.¹ Many primary care providers do feel unprepared to handle this challenge.

Additionally, Dr. Pettiway raises a good question about where misused prescription opioids originate and whether prescribers are responsible. The data show that the majority of misused prescription opioids are obtained from a family member or friend and not directly from a physician.^2,3 However, this supply does generally originate from a prescription. Providers need to educate their patients about the risk for diversion, the need to keep pills safely hidden and locked away, and the importance of safely discarding unused supplies. Responsible prescribers need to anticipate these concerns and educate patients about them.

In summary, we firmly believe that primary care providers are capable of safe, effective, and responsible opioid prescribing and hope that our paper provides additional guidance on how to do so.

Anemia is a frequent complication of chronic kidney disease, occurring in over 90% of patients receiving renal replacement therapy. It is associated with significant morbidity and mortality. While its pathogenesis is typically multifactorial, the predominant cause is failure of the kidneys to produce enough endogenous erythropoietin. The clinical approval of recombinant human erythropoietin in 1989 dramatically changed the treatment of anemia of chronic kidney disease, but randomized controlled trials yielded disappointing results when erythropoiesis-stimulating agents (ESAs) were used to raise hemoglobin to normal levels.

This article reviews the epidemiology and pathophysiology of anemia of chronic kidney disease and discusses the complicated and conflicting evidence regarding its treatment.

DEFINITION AND PREVALENCE

Anemia is defined as a hemoglobin concentration less than 13.0 g/dL for men and less than 12.0 g/dL for premenopausal women.¹ It is more common in patients with impaired kidney function, especially when the glomerular filtration rate (GFR) falls below 60 mL/min. It is rare at GFRs higher than 80 mL/min,² but as the GFR falls, the severity of the anemia worsens³ and its prevalence increases: almost 90% of patients with a GFR less than 30 mL/min are anemic.⁴

RENAL ANEMIA IS ASSOCIATED WITH BAD OUTCOMES

Anemia in chronic kidney disease is independently associated with risk of death. It is also an all-cause mortality multiplier, ie, it magnifies the risk of death from other disease states.⁵

In observational studies, anemia was associated with faster progression of left ventricular hypertrophy, inflammation, and increased myocardial and peripheral oxygen demand, thereby leading to worse cardiac outcomes with increased risk of myocardial infarction, coronary revascularization, and readmission for heart failure.^6–8 Anemia is also associated with fatigue, depression, reduced exercise tolerance, stroke, and increased risk of rehospitalization.^9–13

RENAL ANEMIA IS MULTIFACTORIAL

Anemia of chronic kidney disease is typically attributed to the decrease of erythropoietin production that accompanies the fall in GFR. However, the process is multifactorial, with several other contributing factors: absolute and functional iron deficiency, folate and vitamin B₁₂ deficiencies, reduced red blood cell life span, and suppression of erythropoiesis by the uremic milieu.¹⁴

While it was once thought that chronic kidney disease leads to loss of erythropoietin-producing cells, it is now known that downregulation of hypoxia-inducible factor (HIF; a transcription factor) is at least partially responsible for the decrease in erythropoietin production^15,16 and that this downregulation is reversible (see below).

ERYTHROPOIETIN, IRON, AND RED BLOOD CELLS

Erythropoietin production is triggered by hypoxia, mediated by HIF

Erythropoietin is produced primarily in the deep cortex and outer medulla of the kidneys by a special population of peritubular interstitial cells.¹⁷ The parenchymal cells of the liver also produce erythropoietin, but much less.¹⁸

The rate of renal erythropoietin synthesis is determined by tissue oxygenation rather than by renal blood flow; production increases as the hemoglobin concentration drops and the arterial oxygen tension decreases (Figure 1).¹⁹

The gene for erythropoietin is located on chromosome 7 and is regulated by HIF. HIF molecules are composed of an alpha subunit, which is unstable at high Po₂, and a beta subunit, constitutively present in the nucleus.²⁰

In hypoxic conditions, the HIF dimer is transcriptionally active and binds to specific DNA recognition sequences called hypoxia-response elements. Gene transcription is upregulated, leading to increased production of erythropoietin.²¹

Under normal oxygen tension, on the other hand, the proline residue of the HIF alpha subunit is hydroxylated. The hydroxylated HIF alpha subunit is then degraded by proteasomal ubiquitylation, which is mediated by the von Hippel-Lindau tumor-suppressor gene pVHL.²² Degradation of HIF alpha prevents formation of the HIF heterodimers. HIF therefore cannot bind to the hypoxia-response elements, and erythropoietin gene transcription does not occur.²³

Thus, in states of hypoxia, erythropoietin production is upregulated, whereas with normal oxygen tension, production is downregulated.

Erythropoietin is essential for terminal maturation of erythrocytes

Erythropoietin is essential for terminal maturation of erythrocytes.²⁴ It is thought to stimulate the growth of erythrogenic progenitors: burst-forming units-erythroid (BFU-E) and colony-forming units-erythroid (CFU-E). In the absence of erythropoietin, BFU-E and CFU-E fail to differentiate into mature erythrocytes.²⁵

Binding of erythropoietin to its receptor sets off a series of downstream signals, the most important being the signal transducer and activator of transcription 5 (STAT5). In animal studies, STAT5 was found to inhibit apoptosis through the early induction of an antiapoptotic gene, Bcl-xL.²⁶

Iron metabolism is controlled by several proteins

Iron is characterized by its capacity to accept or donate electrons. This unique property makes it a crucial element in many biochemical reactions such as enzymatic activity, DNA synthesis, oxygen transport, and cell respiration.

Iron metabolism is under the control of several proteins that play different roles in its absorption, recycling, and loss (Figure 2).²⁷

Dietary iron exists primarily in its poorly soluble trivalent ferric form (Fe³⁺), and it needs to be reduced to its soluble divalent ferrous form (Fe²⁺) by ferric reductase to be absorbed. Ferrous iron is taken up at the apical side of enterocytes by a divalent metal transporter (DMT1) and is transported across the brush border.²⁸

To enter the circulation, iron has to be transported across the basolateral membrane by a transporter called ferroportin.²⁹ Ferroportin is also found in placental syncitiotrophoblasts, where it transfers iron from mother to fetus, and in macrophages, where it allows recycling of iron scavenged from damaged cells back into the circulation.³⁰ Upon its release, the ferrous iron is oxidized to the ferric form and loaded onto transferrin. This oxidation process involves hephaestin, a homologue of the ferroxidase ceruloplasmin.³¹

In the plasma, iron is bound to transferrin, and under normal circumstances one-third of transferrin is saturated with iron.³² Transferrin receptors are present on most cells but are most dense on erythroid precursors. Each transferrin receptor can bind two transferrin molecules. After binding to transferrin, the transferrin receptor is endocytosed, and the iron is released into acidified vacuoles. The transferrin-receptor complex is then recycled to the surface.³³

Ferritin is the cellular storage protein for iron, and it can store up to 4,500 atoms of iron within its spherical cavity.³⁴ The serum level of ferritin reflects overall storage, with 1 ng/mL of ferritin indicating 10 mg of total iron stores.³⁵ Ferritin is also an acute-phase reactant, and plasma levels can increase in inflammatory states such as infection or malignancy. As such, elevated ferritin does not necessarily indicate elevated iron stores.

Iron is lost in sweat, shed skin cells, and sloughed intestinal mucosal cells. However, there is no specific mechanism of iron excretion from the human body. Thus, iron is mainly regulated at the level of intestinal absorption. The iron exporter ferroportin is upregulated by the amount of available iron and is degraded by hepcidin.³⁶

Hepcidin is a small cysteine-rich cationic peptide that is primarily produced in the liver, with some minor production also occurring in the kidneys.³⁷ Transcription of the gene encoding hepcidin is downregulated by anemia and hypoxia and upregulated by inflammation and elevated iron levels.³⁸ Transcription of hepcidin leads to degradation of ferroportin and a decrease in intestinal iron absorption. On the other hand, anemia and hypoxia inhibit hepcidin transcription, which allows ferroportin to facilitate intestinal iron absorption.

TREATMENT OF RENAL ANEMIA

Early enthusiasm for erythropoietin agents

Androgens started to be used to treat anemia of end-stage renal disease in 1970,^39,40 and before the advent of recombinant human erythropoietin, they were a mainstay of nontransfusional therapy for anemic patients on dialysis.

The approval of recombinant human erythropoietin in 1989 drastically shifted the treatment of renal anemia. While the initial goal of treating anemia of chronic kidney disease with erythropoietin was to prevent blood transfusions,⁴¹ subsequent studies showed that the benefits might be far greater. Indeed, an initial observational trial showed that erythropoiesis-stimulating agents (ESAs) were associated with improved quality of life,⁴² improved neurocognitive function,^43,44 and even cost savings.⁴⁵ The benefits also extended to major outcomes such as regression of left ventricular hypertrophy,⁴⁶ improvement in New York Heart Association class and cardiac function,⁴⁷ fewer hospitalizations,⁴⁸ and even reduction of cardiovascular mortality rates.⁴⁹

As a result, ESA use gained popularity, and by 2006 an estimated 90% of dialysis patients were receiving these agents.⁵⁰ The target and achieved hemoglobin levels also increased, with mean hemoglobin levels in hemodialysis patients being raised from 9.7 to 12 g/dL.⁵¹

Disappointing results in clinical trials of ESAs to normalize hemoglobin

To prospectively study the effects of normalized hemoglobin targets, four randomized controlled trials were conducted (Table 1):

• The Normal Hematocrit Study (NHCT)⁵²
• The Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial⁵³
• The Cardiovascular Risk Reduction by Early Anemia Treatment (CREATE) trial⁵⁴
• The Trial to Reduce Cardiovascular Events With Aranesp Therapy (TREAT).⁵⁵

These trials randomized patients to either higher “normal-range” hemoglobin targets or to lower target hemoglobin levels.

Their findings were disappointing and raised several red flags about excessive use of ESAs. The trials found no benefit in higher hemoglobin targets, and in fact, some of them demonstrated harm in patients randomized to higher targets. Notably, higher hemoglobin targets were associated with significant side effects such as access-site  thrombosis,⁵² strokes,⁵⁵ and possibly cardiovascular events.^54,55 Only the CREATE trial was able to demonstrate a quality-of-life benefit for the high-target group.⁵⁴ 

It remains unclear whether these adverse events were from the therapy itself or from an increased morbidity burden in the treated patients. Erythropoietin use is associated with hypertension,⁵⁶ thought to be related to endothelin-mediated vasoconstriction.⁵⁷ In our experience, this is most evident when hemoglobin levels are normalized with ESA therapy. Cycling of erythropoietin levels between extreme levels can lead to vascular remodeling, which may also be related to its cardiovascular effects.⁵⁷

A noticeable finding in several of these trials was that patients failed to achieve the higher hemoglobin target despite the use of very high doses of ESA. Reanalysis of data from the CHOIR and CREATE trials showed that the patients who had worse outcomes were more likely to have required very high doses without achieving their target hemoglobin.^58,59 Indeed, patients who achieved the higher target hemoglobin levels, usually at lower ESA doses, had better outcomes. This suggested that the need for a higher dose was associated with poorer outcomes, either as a marker of comorbidity or due to yet undocumented side effects of such high doses.

General approach to therapy

Before attributing anemia to chronic kidney disease, a thorough evaluation should be conducted to look for any reversible process that could be contributing to the anemia.

The causes of anemia are numerous and beyond the scope of this review. However, among the common causes of anemia in chronic kidney disease are deficiencies of iron, vitamin B₁₂, and folate. Therefore, guidelines recommend checking iron, vitamin B₁₂, and folate levels in the initial evaluation of anemia.⁶⁰

Iron deficiency in particular is very common in chronic kidney disease patients and is present in nearly all dialysis patients.⁶¹ Hemodialysis patients are estimated to lose 1 to 3 g of iron per year as a result of blood loss in the dialysis circuit and increased iron utilization secondary to ESA therapy.⁶²

However, in contrast to the general population, in which the upper limits of normal for iron indices are well defined, high serum ferritin levels appear to be poorly predictive of hemoglobin responsiveness in dialysis patients.^63,64 Thus, the cutoffs that define iron responsiveness are much higher than standard definitions for iron deficiency.^65,66 The Dialysis Patients’ Response to IV Iron With Elevated Ferritin (DRIVE) study showed that dialysis patients benefit from intravenous iron therapy even if their ferritin is as high as 1,200 ng/mL, provided their transferrin saturation is below 30%.⁶⁷

Of note, erythropoietin levels cannot be used to distinguish renal anemia from other causes of anemia. Indeed, patients with renal failure may have “relative erythropoietin deficiency,” ie, “normal” erythropoietin levels that are actually too low in view of the degree of anemia.^68,69 In addition to the decreased production capacity by the kidney, there appears to be a component of resistance to the action of erythropoietin in the bone marrow.

For these reasons, there is no erythropoietin level that can be considered “inadequate” or defining of renal anemia. Thus, measuring erythropoietin levels is not routinely recommended in the evaluation of renal anemia.

Two ESA preparations

The two ESAs that have traditionally been used in the treatment of renal anemia are recombinant human erythropoietin and darbepoietin alfa. They appear to be equivalent in terms of safety and efficacy.⁷⁰ However, darbepoietin alfa has more sialic acid molecules, giving it a higher potency and longer half-life and allowing for less-frequent injections.^71,72

In nondialysis patients, recombinant human erythropoietin is typically given every 1 to 2 weeks, whereas darbepoietin alfa can be given every 2 to 4 weeks. In dialysis patients, recombinant human erythropoietin is typically given 3 times per week with every dialysis treatment, while darbepoietin alfa is given once a week.

Target hemoglobin levels: ≤ 11.5 g/dL

In light of the four trials described in Table 1, the international Kidney Disease: Improving Global Outcomes (KDIGO) guidelines⁶⁰ recommend the following (Table 2):

For patients with chronic kidney disease who are not on dialysis, ESA therapy should not be initiated if the hemoglobin level is higher than 10 g/dL. If the hemoglobin level is lower than 10 g/dL, ESA therapy can be initiated, but the decision needs to be individualized based on the rate of fall of hemoglobin concentration, prior response to iron therapy, the risk of needing a transfusion, the risks related to ESA therapy, and the presence of symptoms attributable to anemia.

For patients on dialysis, ESA therapy should be used when the hemoglobin level is between 9 and 10 g/dL to avoid having the hemoglobin fall below 9 g/dL.

In all adult patients, ESAs should not be used to intentionally increase the hemoglobin level above 13 g/dL but rather to maintain levels no higher than 11.5 g/dL. This target is based on the observation that adverse outcomes were associated with ESA use with hemoglobin targets higher than 13 g/dL (Table 1).

Target iron levels

Regarding iron stores, the guidelines recommend the following:

For adult patients with chronic kidney disease who are not on dialysis, iron should be given to keep transferrin saturation above 20% and ferritin above 100 ng/mL. Transferrin saturation should not exceed 30%, and ferritin levels should not exceed 500 ng/mL.

For adult patients on dialysis, iron should be given to maintain transferrin saturation above 30% and ferritin above 200 ng/mL.

The upper limits of ferritin and transferrin saturation are somewhat controversial, as the safety of intentionally maintaining respective levels greater than 30% and 500 ng/mL has been studied in very few patients. Transferrin saturation should in general not exceed 50%.

High ferritin levels are associated with higher death rates, but whether elevation of ferritin levels is a marker of excessive iron administration rather than a nonspecific acute-phase reactant is not clear. The 2006 guidelines⁶⁰ cited upper ferritin limits of 500 to 800 ng/mL. However, the more recent DRIVE trial⁶⁷ showed that patients with ferritin levels of 500 to 1,200 ng/mL will respond to intravenous administration of iron with an increase in their hemoglobin levels. This has led many clinicians to adopt a higher ferritin limit of 1,200 ng/mL.

Hemosiderosis, or excess iron deposition, was a known consequence of frequent transfusions in patients with end-stage renal disease before ESA therapy was available. However, there have been no documented cases of clinical iron overload from iron therapy using current guidelines.⁷³

These algorithms are nuanced, and the benefit of giving intravenous iron should always be weighed against the risks of short-term acute toxicity and infection. Treatment of renal anemia not only requires in-depth knowledge of the topic, but also familiarity with the patient’s specific situation. As such, it is not recommended that clinicians unfamiliar with the treatment of renal anemia manage its treatment.

PARTICULAR CIRCUMSTANCES

Inflammation and ESA resistance

While ESAs are effective in treating anemia in many cases, in many patients the anemia fails to respond. This is of particular importance, since ESA hyporesponsiveness has been found to be a powerful predictor of cardiovascular events and death.⁷⁴ It is unclear, however, whether high doses of ESA are inherently toxic or whether hyporesponsiveness is a marker of adverse outcomes related to comorbidities.

KDIGO defines initial hyporesponsiveness as having no increase in hemoglobin concentration after the first month of appropriate weight-based dosing, and acquired hyporesponsiveness as requiring two increases in ESA doses up to 50% beyond the dose at which the patient had originally been stable.⁶⁰ Identifying ESA hyporesponsiveness should lead to an intensive search for potentially correctable factors.

The two major factors accounting for the state of hyporesponsiveness are inflammation and iron deficiency.^75,76

Inflammation. High C-reactive protein levels have been shown to predict resistance to erythropoietin in dialysis patients.⁷⁷ The release of cytokines such as tumor necrosis factor alpha, interleukin 1, and interferon gamma has an inhibitory effect on erythropoiesis.⁷⁸ Additionally, inflammation can alter the response to ESAs by disrupting the metabolism of iron⁷⁹ through the release of hepcidin, as previously discussed.³⁸ These reasons likely account for the observed lower response to ESAs in the setting of acute illness and explain why ESAs are not recommended for correcting acute anemia.⁸⁰

Iron deficiency also can blunt the response to ESAs. Large amounts of iron are needed for effective erythropoietic bursts. As such, iron supplementation is now a recognized treatment of renal anemia.⁸¹

Other factors associated with hyporesponsiveness include chronic occult blood loss, aluminum toxicity, cobalamin or folate deficiencies, testosterone deficiency, inadequate dialysis, hyperparathyroidism, and superimposed primary bone marrow disease,^82,83 and these should be addressed in patients whose anemia does not respond as expected to ESA therapy. A summary of the main causes of ESA hyporesponsiveness, their reversibility, and recommended treatments is presented in Table 3.

Antibody-mediated pure red-cell aplasia. Rarely, patients receiving ESA therapy develop antibodies that neutralize both the ESA and endogenous erythropoietin. The resulting syndrome, called antibody-mediated pure red-cell aplasia, is characterized by the sudden development of severe transfusion-dependent anemia. This has historically been connected to epoetin beta, a formulation not in use in the United States. However, cases have been documented with epoetin alfa and darbepoetin. The incidence rate is low with subcutaneous ESA use, estimated at 0.5 cases per 10,000 patient-years⁸⁴ and anecdotal with intravenous ESA.⁸⁵ The definitive diagnosis requires demonstration of neutralizing antibodies against erythropoietin. Parvovirus infection should be excluded as an alternative cause of pure red­cell aplasia.

ANEMIA IN CANCER PATIENTS

ESAs are effective in raising hemoglobin levels and reducing transfusion requirements in patients with chemotherapy-induced anemia.⁸⁶ However, there are data linking the use of ESAs to shortened survival in patients who have a variety of solid tumors.⁸⁷

Several mechanisms have been proposed to explain this rapid disease progression, most notably acceleration in tumor growth^88–90 by stimulation of erythropoietin receptors on the surface of the tumor cells, leading to increased tumor angiogenesis.^91,92

For these reasons, treatment of renal anemia in the setting of active malignancy should be referred to an oncologist.

NOVEL TREATMENTS

Several new agents for treating renal anemia are currently under review.

Continuous erythropoiesis receptor activator

Continuous erythropoiesis receptor activator is a pegylated form of recombinant human erythropoietin that has the ability to repeatedly activate the erythropoietin receptor. It appears to be similar to the other forms of erythropoietin in terms of safety and efficacy in both end-stage renal disease⁹³ and chronic kidney disease.⁹⁴ It has the advantage of an extended serum half-life, which allows for longer dosing intervals, ie, every 2 weeks. Its use is currently gaining popularity in the dialysis community.

HIF stabilizers

Our growing understanding of the physiology of erythropoietin offers new potential treatment targets. As previously described, production of erythropoietin is stimulated by HIFs. In order to be degraded, these HIFs are hydroxylated at their proline residues by a prolyl hydroxylase. A new category of drugs called prolyl-hydroxylase inhibitors (PDIs) offers the advantage of stabilizing the HIFs, leading to an increase in erythropoietin production.

In phase 1 and 2 clinical trials, these agents have been shown to increase hemoglobin in both end-stage renal disease and chronic kidney disease patients^15,16 but not in anephric patients, demonstrating a renal source of the erythropoietin production even in nonfunctioning kidneys. The study of one PDI agent (FG 2216) was halted temporarily after a report of death from fulminant hepatitis, but the other (FG 4592) continues to be studied in a phase 2 clinical trial.^95,96

TAKE-HOME POINTS

• Anemia of renal disease is a common condition that is mainly caused by a decrease in erythropoietin production by the kidneys.
• While anemia of renal disease can be corrected with ESAs, it is necessary to investigate and rule out underlying treatable conditions such as iron or vitamin deficiencies before giving an ESA.
• Anemia of renal disease is associated with significant morbidity such as increased risk of left ventricular hypertrophy, myocardial infarction, and heart failure, and has been described as an all-cause mortality multiplier.
• Unfortunately, the only undisputed benefit of treatment to date remains the avoidance of blood transfusions. Furthermore, the large randomized controlled trials that looked at the benefits of ESA have shown that their use can be associated with increased risk of cardiovascular events. Therefore, use of an ESA in end-stage renal disease should never target a normal hemoglobin levels but rather aim for a hemoglobin level of no more than 11.5 g/dL.
• Use of an ESA in chronic kidney disease should be individualized and is not recommended to be started unless the hemoglobin level is less than 10 g/dL.
• Several newer agents for renal anemia are currently under review. A pegylated form of recombinant human erythropoietin has an extended half-life, and a new and promising category of drugs called HIF stabilizers is currently under study.

Anemia is a frequent complication of chronic kidney disease, occurring in over 90% of patients receiving renal replacement therapy. It is associated with significant morbidity and mortality. While its pathogenesis is typically multifactorial, the predominant cause is failure of the kidneys to produce enough endogenous erythropoietin. The clinical approval of recombinant human erythropoietin in 1989 dramatically changed the treatment of anemia of chronic kidney disease, but randomized controlled trials yielded disappointing results when erythropoiesis-stimulating agents (ESAs) were used to raise hemoglobin to normal levels.

This article reviews the epidemiology and pathophysiology of anemia of chronic kidney disease and discusses the complicated and conflicting evidence regarding its treatment.

DEFINITION AND PREVALENCE

Anemia is defined as a hemoglobin concentration less than 13.0 g/dL for men and less than 12.0 g/dL for premenopausal women.¹ It is more common in patients with impaired kidney function, especially when the glomerular filtration rate (GFR) falls below 60 mL/min. It is rare at GFRs higher than 80 mL/min,² but as the GFR falls, the severity of the anemia worsens³ and its prevalence increases: almost 90% of patients with a GFR less than 30 mL/min are anemic.⁴

RENAL ANEMIA IS ASSOCIATED WITH BAD OUTCOMES

Anemia in chronic kidney disease is independently associated with risk of death. It is also an all-cause mortality multiplier, ie, it magnifies the risk of death from other disease states.⁵

In observational studies, anemia was associated with faster progression of left ventricular hypertrophy, inflammation, and increased myocardial and peripheral oxygen demand, thereby leading to worse cardiac outcomes with increased risk of myocardial infarction, coronary revascularization, and readmission for heart failure.^6–8 Anemia is also associated with fatigue, depression, reduced exercise tolerance, stroke, and increased risk of rehospitalization.^9–13

RENAL ANEMIA IS MULTIFACTORIAL

Anemia of chronic kidney disease is typically attributed to the decrease of erythropoietin production that accompanies the fall in GFR. However, the process is multifactorial, with several other contributing factors: absolute and functional iron deficiency, folate and vitamin B₁₂ deficiencies, reduced red blood cell life span, and suppression of erythropoiesis by the uremic milieu.¹⁴

While it was once thought that chronic kidney disease leads to loss of erythropoietin-producing cells, it is now known that downregulation of hypoxia-inducible factor (HIF; a transcription factor) is at least partially responsible for the decrease in erythropoietin production^15,16 and that this downregulation is reversible (see below).

ERYTHROPOIETIN, IRON, AND RED BLOOD CELLS

Erythropoietin production is triggered by hypoxia, mediated by HIF

Erythropoietin is produced primarily in the deep cortex and outer medulla of the kidneys by a special population of peritubular interstitial cells.¹⁷ The parenchymal cells of the liver also produce erythropoietin, but much less.¹⁸

The rate of renal erythropoietin synthesis is determined by tissue oxygenation rather than by renal blood flow; production increases as the hemoglobin concentration drops and the arterial oxygen tension decreases (Figure 1).¹⁹

The gene for erythropoietin is located on chromosome 7 and is regulated by HIF. HIF molecules are composed of an alpha subunit, which is unstable at high Po₂, and a beta subunit, constitutively present in the nucleus.²⁰

In hypoxic conditions, the HIF dimer is transcriptionally active and binds to specific DNA recognition sequences called hypoxia-response elements. Gene transcription is upregulated, leading to increased production of erythropoietin.²¹

Under normal oxygen tension, on the other hand, the proline residue of the HIF alpha subunit is hydroxylated. The hydroxylated HIF alpha subunit is then degraded by proteasomal ubiquitylation, which is mediated by the von Hippel-Lindau tumor-suppressor gene pVHL.²² Degradation of HIF alpha prevents formation of the HIF heterodimers. HIF therefore cannot bind to the hypoxia-response elements, and erythropoietin gene transcription does not occur.²³

Thus, in states of hypoxia, erythropoietin production is upregulated, whereas with normal oxygen tension, production is downregulated.

Erythropoietin is essential for terminal maturation of erythrocytes

Erythropoietin is essential for terminal maturation of erythrocytes.²⁴ It is thought to stimulate the growth of erythrogenic progenitors: burst-forming units-erythroid (BFU-E) and colony-forming units-erythroid (CFU-E). In the absence of erythropoietin, BFU-E and CFU-E fail to differentiate into mature erythrocytes.²⁵

Binding of erythropoietin to its receptor sets off a series of downstream signals, the most important being the signal transducer and activator of transcription 5 (STAT5). In animal studies, STAT5 was found to inhibit apoptosis through the early induction of an antiapoptotic gene, Bcl-xL.²⁶

Iron metabolism is controlled by several proteins

Iron is characterized by its capacity to accept or donate electrons. This unique property makes it a crucial element in many biochemical reactions such as enzymatic activity, DNA synthesis, oxygen transport, and cell respiration.

Iron metabolism is under the control of several proteins that play different roles in its absorption, recycling, and loss (Figure 2).²⁷

Dietary iron exists primarily in its poorly soluble trivalent ferric form (Fe³⁺), and it needs to be reduced to its soluble divalent ferrous form (Fe²⁺) by ferric reductase to be absorbed. Ferrous iron is taken up at the apical side of enterocytes by a divalent metal transporter (DMT1) and is transported across the brush border.²⁸

To enter the circulation, iron has to be transported across the basolateral membrane by a transporter called ferroportin.²⁹ Ferroportin is also found in placental syncitiotrophoblasts, where it transfers iron from mother to fetus, and in macrophages, where it allows recycling of iron scavenged from damaged cells back into the circulation.³⁰ Upon its release, the ferrous iron is oxidized to the ferric form and loaded onto transferrin. This oxidation process involves hephaestin, a homologue of the ferroxidase ceruloplasmin.³¹

In the plasma, iron is bound to transferrin, and under normal circumstances one-third of transferrin is saturated with iron.³² Transferrin receptors are present on most cells but are most dense on erythroid precursors. Each transferrin receptor can bind two transferrin molecules. After binding to transferrin, the transferrin receptor is endocytosed, and the iron is released into acidified vacuoles. The transferrin-receptor complex is then recycled to the surface.³³

Ferritin is the cellular storage protein for iron, and it can store up to 4,500 atoms of iron within its spherical cavity.³⁴ The serum level of ferritin reflects overall storage, with 1 ng/mL of ferritin indicating 10 mg of total iron stores.³⁵ Ferritin is also an acute-phase reactant, and plasma levels can increase in inflammatory states such as infection or malignancy. As such, elevated ferritin does not necessarily indicate elevated iron stores.

Iron is lost in sweat, shed skin cells, and sloughed intestinal mucosal cells. However, there is no specific mechanism of iron excretion from the human body. Thus, iron is mainly regulated at the level of intestinal absorption. The iron exporter ferroportin is upregulated by the amount of available iron and is degraded by hepcidin.³⁶

Hepcidin is a small cysteine-rich cationic peptide that is primarily produced in the liver, with some minor production also occurring in the kidneys.³⁷ Transcription of the gene encoding hepcidin is downregulated by anemia and hypoxia and upregulated by inflammation and elevated iron levels.³⁸ Transcription of hepcidin leads to degradation of ferroportin and a decrease in intestinal iron absorption. On the other hand, anemia and hypoxia inhibit hepcidin transcription, which allows ferroportin to facilitate intestinal iron absorption.

TREATMENT OF RENAL ANEMIA

Early enthusiasm for erythropoietin agents

Androgens started to be used to treat anemia of end-stage renal disease in 1970,^39,40 and before the advent of recombinant human erythropoietin, they were a mainstay of nontransfusional therapy for anemic patients on dialysis.

The approval of recombinant human erythropoietin in 1989 drastically shifted the treatment of renal anemia. While the initial goal of treating anemia of chronic kidney disease with erythropoietin was to prevent blood transfusions,⁴¹ subsequent studies showed that the benefits might be far greater. Indeed, an initial observational trial showed that erythropoiesis-stimulating agents (ESAs) were associated with improved quality of life,⁴² improved neurocognitive function,^43,44 and even cost savings.⁴⁵ The benefits also extended to major outcomes such as regression of left ventricular hypertrophy,⁴⁶ improvement in New York Heart Association class and cardiac function,⁴⁷ fewer hospitalizations,⁴⁸ and even reduction of cardiovascular mortality rates.⁴⁹

As a result, ESA use gained popularity, and by 2006 an estimated 90% of dialysis patients were receiving these agents.⁵⁰ The target and achieved hemoglobin levels also increased, with mean hemoglobin levels in hemodialysis patients being raised from 9.7 to 12 g/dL.⁵¹

Disappointing results in clinical trials of ESAs to normalize hemoglobin

To prospectively study the effects of normalized hemoglobin targets, four randomized controlled trials were conducted (Table 1):

• The Normal Hematocrit Study (NHCT)⁵²
• The Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial⁵³
• The Cardiovascular Risk Reduction by Early Anemia Treatment (CREATE) trial⁵⁴
• The Trial to Reduce Cardiovascular Events With Aranesp Therapy (TREAT).⁵⁵

These trials randomized patients to either higher “normal-range” hemoglobin targets or to lower target hemoglobin levels.

Their findings were disappointing and raised several red flags about excessive use of ESAs. The trials found no benefit in higher hemoglobin targets, and in fact, some of them demonstrated harm in patients randomized to higher targets. Notably, higher hemoglobin targets were associated with significant side effects such as access-site  thrombosis,⁵² strokes,⁵⁵ and possibly cardiovascular events.^54,55 Only the CREATE trial was able to demonstrate a quality-of-life benefit for the high-target group.⁵⁴ 

It remains unclear whether these adverse events were from the therapy itself or from an increased morbidity burden in the treated patients. Erythropoietin use is associated with hypertension,⁵⁶ thought to be related to endothelin-mediated vasoconstriction.⁵⁷ In our experience, this is most evident when hemoglobin levels are normalized with ESA therapy. Cycling of erythropoietin levels between extreme levels can lead to vascular remodeling, which may also be related to its cardiovascular effects.⁵⁷

A noticeable finding in several of these trials was that patients failed to achieve the higher hemoglobin target despite the use of very high doses of ESA. Reanalysis of data from the CHOIR and CREATE trials showed that the patients who had worse outcomes were more likely to have required very high doses without achieving their target hemoglobin.^58,59 Indeed, patients who achieved the higher target hemoglobin levels, usually at lower ESA doses, had better outcomes. This suggested that the need for a higher dose was associated with poorer outcomes, either as a marker of comorbidity or due to yet undocumented side effects of such high doses.

General approach to therapy

Before attributing anemia to chronic kidney disease, a thorough evaluation should be conducted to look for any reversible process that could be contributing to the anemia.

The causes of anemia are numerous and beyond the scope of this review. However, among the common causes of anemia in chronic kidney disease are deficiencies of iron, vitamin B₁₂, and folate. Therefore, guidelines recommend checking iron, vitamin B₁₂, and folate levels in the initial evaluation of anemia.⁶⁰

Iron deficiency in particular is very common in chronic kidney disease patients and is present in nearly all dialysis patients.⁶¹ Hemodialysis patients are estimated to lose 1 to 3 g of iron per year as a result of blood loss in the dialysis circuit and increased iron utilization secondary to ESA therapy.⁶²

However, in contrast to the general population, in which the upper limits of normal for iron indices are well defined, high serum ferritin levels appear to be poorly predictive of hemoglobin responsiveness in dialysis patients.^63,64 Thus, the cutoffs that define iron responsiveness are much higher than standard definitions for iron deficiency.^65,66 The Dialysis Patients’ Response to IV Iron With Elevated Ferritin (DRIVE) study showed that dialysis patients benefit from intravenous iron therapy even if their ferritin is as high as 1,200 ng/mL, provided their transferrin saturation is below 30%.⁶⁷

Of note, erythropoietin levels cannot be used to distinguish renal anemia from other causes of anemia. Indeed, patients with renal failure may have “relative erythropoietin deficiency,” ie, “normal” erythropoietin levels that are actually too low in view of the degree of anemia.^68,69 In addition to the decreased production capacity by the kidney, there appears to be a component of resistance to the action of erythropoietin in the bone marrow.

For these reasons, there is no erythropoietin level that can be considered “inadequate” or defining of renal anemia. Thus, measuring erythropoietin levels is not routinely recommended in the evaluation of renal anemia.

Two ESA preparations

The two ESAs that have traditionally been used in the treatment of renal anemia are recombinant human erythropoietin and darbepoietin alfa. They appear to be equivalent in terms of safety and efficacy.⁷⁰ However, darbepoietin alfa has more sialic acid molecules, giving it a higher potency and longer half-life and allowing for less-frequent injections.^71,72

In nondialysis patients, recombinant human erythropoietin is typically given every 1 to 2 weeks, whereas darbepoietin alfa can be given every 2 to 4 weeks. In dialysis patients, recombinant human erythropoietin is typically given 3 times per week with every dialysis treatment, while darbepoietin alfa is given once a week.

Target hemoglobin levels: ≤ 11.5 g/dL

In light of the four trials described in Table 1, the international Kidney Disease: Improving Global Outcomes (KDIGO) guidelines⁶⁰ recommend the following (Table 2):

For patients with chronic kidney disease who are not on dialysis, ESA therapy should not be initiated if the hemoglobin level is higher than 10 g/dL. If the hemoglobin level is lower than 10 g/dL, ESA therapy can be initiated, but the decision needs to be individualized based on the rate of fall of hemoglobin concentration, prior response to iron therapy, the risk of needing a transfusion, the risks related to ESA therapy, and the presence of symptoms attributable to anemia.

For patients on dialysis, ESA therapy should be used when the hemoglobin level is between 9 and 10 g/dL to avoid having the hemoglobin fall below 9 g/dL.

In all adult patients, ESAs should not be used to intentionally increase the hemoglobin level above 13 g/dL but rather to maintain levels no higher than 11.5 g/dL. This target is based on the observation that adverse outcomes were associated with ESA use with hemoglobin targets higher than 13 g/dL (Table 1).

Target iron levels

Regarding iron stores, the guidelines recommend the following:

For adult patients with chronic kidney disease who are not on dialysis, iron should be given to keep transferrin saturation above 20% and ferritin above 100 ng/mL. Transferrin saturation should not exceed 30%, and ferritin levels should not exceed 500 ng/mL.

For adult patients on dialysis, iron should be given to maintain transferrin saturation above 30% and ferritin above 200 ng/mL.

The upper limits of ferritin and transferrin saturation are somewhat controversial, as the safety of intentionally maintaining respective levels greater than 30% and 500 ng/mL has been studied in very few patients. Transferrin saturation should in general not exceed 50%.

High ferritin levels are associated with higher death rates, but whether elevation of ferritin levels is a marker of excessive iron administration rather than a nonspecific acute-phase reactant is not clear. The 2006 guidelines⁶⁰ cited upper ferritin limits of 500 to 800 ng/mL. However, the more recent DRIVE trial⁶⁷ showed that patients with ferritin levels of 500 to 1,200 ng/mL will respond to intravenous administration of iron with an increase in their hemoglobin levels. This has led many clinicians to adopt a higher ferritin limit of 1,200 ng/mL.

Hemosiderosis, or excess iron deposition, was a known consequence of frequent transfusions in patients with end-stage renal disease before ESA therapy was available. However, there have been no documented cases of clinical iron overload from iron therapy using current guidelines.⁷³

These algorithms are nuanced, and the benefit of giving intravenous iron should always be weighed against the risks of short-term acute toxicity and infection. Treatment of renal anemia not only requires in-depth knowledge of the topic, but also familiarity with the patient’s specific situation. As such, it is not recommended that clinicians unfamiliar with the treatment of renal anemia manage its treatment.

PARTICULAR CIRCUMSTANCES

Inflammation and ESA resistance

While ESAs are effective in treating anemia in many cases, in many patients the anemia fails to respond. This is of particular importance, since ESA hyporesponsiveness has been found to be a powerful predictor of cardiovascular events and death.⁷⁴ It is unclear, however, whether high doses of ESA are inherently toxic or whether hyporesponsiveness is a marker of adverse outcomes related to comorbidities.

KDIGO defines initial hyporesponsiveness as having no increase in hemoglobin concentration after the first month of appropriate weight-based dosing, and acquired hyporesponsiveness as requiring two increases in ESA doses up to 50% beyond the dose at which the patient had originally been stable.⁶⁰ Identifying ESA hyporesponsiveness should lead to an intensive search for potentially correctable factors.

The two major factors accounting for the state of hyporesponsiveness are inflammation and iron deficiency.^75,76

Inflammation. High C-reactive protein levels have been shown to predict resistance to erythropoietin in dialysis patients.⁷⁷ The release of cytokines such as tumor necrosis factor alpha, interleukin 1, and interferon gamma has an inhibitory effect on erythropoiesis.⁷⁸ Additionally, inflammation can alter the response to ESAs by disrupting the metabolism of iron⁷⁹ through the release of hepcidin, as previously discussed.³⁸ These reasons likely account for the observed lower response to ESAs in the setting of acute illness and explain why ESAs are not recommended for correcting acute anemia.⁸⁰

Iron deficiency also can blunt the response to ESAs. Large amounts of iron are needed for effective erythropoietic bursts. As such, iron supplementation is now a recognized treatment of renal anemia.⁸¹

Other factors associated with hyporesponsiveness include chronic occult blood loss, aluminum toxicity, cobalamin or folate deficiencies, testosterone deficiency, inadequate dialysis, hyperparathyroidism, and superimposed primary bone marrow disease,^82,83 and these should be addressed in patients whose anemia does not respond as expected to ESA therapy. A summary of the main causes of ESA hyporesponsiveness, their reversibility, and recommended treatments is presented in Table 3.

Antibody-mediated pure red-cell aplasia. Rarely, patients receiving ESA therapy develop antibodies that neutralize both the ESA and endogenous erythropoietin. The resulting syndrome, called antibody-mediated pure red-cell aplasia, is characterized by the sudden development of severe transfusion-dependent anemia. This has historically been connected to epoetin beta, a formulation not in use in the United States. However, cases have been documented with epoetin alfa and darbepoetin. The incidence rate is low with subcutaneous ESA use, estimated at 0.5 cases per 10,000 patient-years⁸⁴ and anecdotal with intravenous ESA.⁸⁵ The definitive diagnosis requires demonstration of neutralizing antibodies against erythropoietin. Parvovirus infection should be excluded as an alternative cause of pure red­cell aplasia.

ANEMIA IN CANCER PATIENTS

ESAs are effective in raising hemoglobin levels and reducing transfusion requirements in patients with chemotherapy-induced anemia.⁸⁶ However, there are data linking the use of ESAs to shortened survival in patients who have a variety of solid tumors.⁸⁷

Several mechanisms have been proposed to explain this rapid disease progression, most notably acceleration in tumor growth^88–90 by stimulation of erythropoietin receptors on the surface of the tumor cells, leading to increased tumor angiogenesis.^91,92

For these reasons, treatment of renal anemia in the setting of active malignancy should be referred to an oncologist.

NOVEL TREATMENTS

Several new agents for treating renal anemia are currently under review.

Continuous erythropoiesis receptor activator

Continuous erythropoiesis receptor activator is a pegylated form of recombinant human erythropoietin that has the ability to repeatedly activate the erythropoietin receptor. It appears to be similar to the other forms of erythropoietin in terms of safety and efficacy in both end-stage renal disease⁹³ and chronic kidney disease.⁹⁴ It has the advantage of an extended serum half-life, which allows for longer dosing intervals, ie, every 2 weeks. Its use is currently gaining popularity in the dialysis community.

HIF stabilizers

Our growing understanding of the physiology of erythropoietin offers new potential treatment targets. As previously described, production of erythropoietin is stimulated by HIFs. In order to be degraded, these HIFs are hydroxylated at their proline residues by a prolyl hydroxylase. A new category of drugs called prolyl-hydroxylase inhibitors (PDIs) offers the advantage of stabilizing the HIFs, leading to an increase in erythropoietin production.

In phase 1 and 2 clinical trials, these agents have been shown to increase hemoglobin in both end-stage renal disease and chronic kidney disease patients^15,16 but not in anephric patients, demonstrating a renal source of the erythropoietin production even in nonfunctioning kidneys. The study of one PDI agent (FG 2216) was halted temporarily after a report of death from fulminant hepatitis, but the other (FG 4592) continues to be studied in a phase 2 clinical trial.^95,96

TAKE-HOME POINTS

• Anemia of renal disease is a common condition that is mainly caused by a decrease in erythropoietin production by the kidneys.
• While anemia of renal disease can be corrected with ESAs, it is necessary to investigate and rule out underlying treatable conditions such as iron or vitamin deficiencies before giving an ESA.
• Anemia of renal disease is associated with significant morbidity such as increased risk of left ventricular hypertrophy, myocardial infarction, and heart failure, and has been described as an all-cause mortality multiplier.
• Unfortunately, the only undisputed benefit of treatment to date remains the avoidance of blood transfusions. Furthermore, the large randomized controlled trials that looked at the benefits of ESA have shown that their use can be associated with increased risk of cardiovascular events. Therefore, use of an ESA in end-stage renal disease should never target a normal hemoglobin levels but rather aim for a hemoglobin level of no more than 11.5 g/dL.
• Use of an ESA in chronic kidney disease should be individualized and is not recommended to be started unless the hemoglobin level is less than 10 g/dL.
• Several newer agents for renal anemia are currently under review. A pegylated form of recombinant human erythropoietin has an extended half-life, and a new and promising category of drugs called HIF stabilizers is currently under study.

Anemia is a frequent complication of chronic kidney disease, occurring in over 90% of patients receiving renal replacement therapy. It is associated with significant morbidity and mortality. While its pathogenesis is typically multifactorial, the predominant cause is failure of the kidneys to produce enough endogenous erythropoietin. The clinical approval of recombinant human erythropoietin in 1989 dramatically changed the treatment of anemia of chronic kidney disease, but randomized controlled trials yielded disappointing results when erythropoiesis-stimulating agents (ESAs) were used to raise hemoglobin to normal levels.

This article reviews the epidemiology and pathophysiology of anemia of chronic kidney disease and discusses the complicated and conflicting evidence regarding its treatment.

DEFINITION AND PREVALENCE

Anemia is defined as a hemoglobin concentration less than 13.0 g/dL for men and less than 12.0 g/dL for premenopausal women.¹ It is more common in patients with impaired kidney function, especially when the glomerular filtration rate (GFR) falls below 60 mL/min. It is rare at GFRs higher than 80 mL/min,² but as the GFR falls, the severity of the anemia worsens³ and its prevalence increases: almost 90% of patients with a GFR less than 30 mL/min are anemic.⁴

RENAL ANEMIA IS ASSOCIATED WITH BAD OUTCOMES

Anemia in chronic kidney disease is independently associated with risk of death. It is also an all-cause mortality multiplier, ie, it magnifies the risk of death from other disease states.⁵

In observational studies, anemia was associated with faster progression of left ventricular hypertrophy, inflammation, and increased myocardial and peripheral oxygen demand, thereby leading to worse cardiac outcomes with increased risk of myocardial infarction, coronary revascularization, and readmission for heart failure.^6–8 Anemia is also associated with fatigue, depression, reduced exercise tolerance, stroke, and increased risk of rehospitalization.^9–13

RENAL ANEMIA IS MULTIFACTORIAL

Anemia of chronic kidney disease is typically attributed to the decrease of erythropoietin production that accompanies the fall in GFR. However, the process is multifactorial, with several other contributing factors: absolute and functional iron deficiency, folate and vitamin B₁₂ deficiencies, reduced red blood cell life span, and suppression of erythropoiesis by the uremic milieu.¹⁴

While it was once thought that chronic kidney disease leads to loss of erythropoietin-producing cells, it is now known that downregulation of hypoxia-inducible factor (HIF; a transcription factor) is at least partially responsible for the decrease in erythropoietin production^15,16 and that this downregulation is reversible (see below).

ERYTHROPOIETIN, IRON, AND RED BLOOD CELLS

Erythropoietin production is triggered by hypoxia, mediated by HIF

Erythropoietin is produced primarily in the deep cortex and outer medulla of the kidneys by a special population of peritubular interstitial cells.¹⁷ The parenchymal cells of the liver also produce erythropoietin, but much less.¹⁸

The rate of renal erythropoietin synthesis is determined by tissue oxygenation rather than by renal blood flow; production increases as the hemoglobin concentration drops and the arterial oxygen tension decreases (Figure 1).¹⁹

The gene for erythropoietin is located on chromosome 7 and is regulated by HIF. HIF molecules are composed of an alpha subunit, which is unstable at high Po₂, and a beta subunit, constitutively present in the nucleus.²⁰

In hypoxic conditions, the HIF dimer is transcriptionally active and binds to specific DNA recognition sequences called hypoxia-response elements. Gene transcription is upregulated, leading to increased production of erythropoietin.²¹

Under normal oxygen tension, on the other hand, the proline residue of the HIF alpha subunit is hydroxylated. The hydroxylated HIF alpha subunit is then degraded by proteasomal ubiquitylation, which is mediated by the von Hippel-Lindau tumor-suppressor gene pVHL.²² Degradation of HIF alpha prevents formation of the HIF heterodimers. HIF therefore cannot bind to the hypoxia-response elements, and erythropoietin gene transcription does not occur.²³

Thus, in states of hypoxia, erythropoietin production is upregulated, whereas with normal oxygen tension, production is downregulated.

Erythropoietin is essential for terminal maturation of erythrocytes

Erythropoietin is essential for terminal maturation of erythrocytes.²⁴ It is thought to stimulate the growth of erythrogenic progenitors: burst-forming units-erythroid (BFU-E) and colony-forming units-erythroid (CFU-E). In the absence of erythropoietin, BFU-E and CFU-E fail to differentiate into mature erythrocytes.²⁵

Binding of erythropoietin to its receptor sets off a series of downstream signals, the most important being the signal transducer and activator of transcription 5 (STAT5). In animal studies, STAT5 was found to inhibit apoptosis through the early induction of an antiapoptotic gene, Bcl-xL.²⁶

Iron metabolism is controlled by several proteins

Iron is characterized by its capacity to accept or donate electrons. This unique property makes it a crucial element in many biochemical reactions such as enzymatic activity, DNA synthesis, oxygen transport, and cell respiration.

Iron metabolism is under the control of several proteins that play different roles in its absorption, recycling, and loss (Figure 2).²⁷

Dietary iron exists primarily in its poorly soluble trivalent ferric form (Fe³⁺), and it needs to be reduced to its soluble divalent ferrous form (Fe²⁺) by ferric reductase to be absorbed. Ferrous iron is taken up at the apical side of enterocytes by a divalent metal transporter (DMT1) and is transported across the brush border.²⁸

To enter the circulation, iron has to be transported across the basolateral membrane by a transporter called ferroportin.²⁹ Ferroportin is also found in placental syncitiotrophoblasts, where it transfers iron from mother to fetus, and in macrophages, where it allows recycling of iron scavenged from damaged cells back into the circulation.³⁰ Upon its release, the ferrous iron is oxidized to the ferric form and loaded onto transferrin. This oxidation process involves hephaestin, a homologue of the ferroxidase ceruloplasmin.³¹

In the plasma, iron is bound to transferrin, and under normal circumstances one-third of transferrin is saturated with iron.³² Transferrin receptors are present on most cells but are most dense on erythroid precursors. Each transferrin receptor can bind two transferrin molecules. After binding to transferrin, the transferrin receptor is endocytosed, and the iron is released into acidified vacuoles. The transferrin-receptor complex is then recycled to the surface.³³

Ferritin is the cellular storage protein for iron, and it can store up to 4,500 atoms of iron within its spherical cavity.³⁴ The serum level of ferritin reflects overall storage, with 1 ng/mL of ferritin indicating 10 mg of total iron stores.³⁵ Ferritin is also an acute-phase reactant, and plasma levels can increase in inflammatory states such as infection or malignancy. As such, elevated ferritin does not necessarily indicate elevated iron stores.

Iron is lost in sweat, shed skin cells, and sloughed intestinal mucosal cells. However, there is no specific mechanism of iron excretion from the human body. Thus, iron is mainly regulated at the level of intestinal absorption. The iron exporter ferroportin is upregulated by the amount of available iron and is degraded by hepcidin.³⁶

Hepcidin is a small cysteine-rich cationic peptide that is primarily produced in the liver, with some minor production also occurring in the kidneys.³⁷ Transcription of the gene encoding hepcidin is downregulated by anemia and hypoxia and upregulated by inflammation and elevated iron levels.³⁸ Transcription of hepcidin leads to degradation of ferroportin and a decrease in intestinal iron absorption. On the other hand, anemia and hypoxia inhibit hepcidin transcription, which allows ferroportin to facilitate intestinal iron absorption.

TREATMENT OF RENAL ANEMIA

Early enthusiasm for erythropoietin agents

Androgens started to be used to treat anemia of end-stage renal disease in 1970,^39,40 and before the advent of recombinant human erythropoietin, they were a mainstay of nontransfusional therapy for anemic patients on dialysis.

The approval of recombinant human erythropoietin in 1989 drastically shifted the treatment of renal anemia. While the initial goal of treating anemia of chronic kidney disease with erythropoietin was to prevent blood transfusions,⁴¹ subsequent studies showed that the benefits might be far greater. Indeed, an initial observational trial showed that erythropoiesis-stimulating agents (ESAs) were associated with improved quality of life,⁴² improved neurocognitive function,^43,44 and even cost savings.⁴⁵ The benefits also extended to major outcomes such as regression of left ventricular hypertrophy,⁴⁶ improvement in New York Heart Association class and cardiac function,⁴⁷ fewer hospitalizations,⁴⁸ and even reduction of cardiovascular mortality rates.⁴⁹

As a result, ESA use gained popularity, and by 2006 an estimated 90% of dialysis patients were receiving these agents.⁵⁰ The target and achieved hemoglobin levels also increased, with mean hemoglobin levels in hemodialysis patients being raised from 9.7 to 12 g/dL.⁵¹

Disappointing results in clinical trials of ESAs to normalize hemoglobin

To prospectively study the effects of normalized hemoglobin targets, four randomized controlled trials were conducted (Table 1):

• The Normal Hematocrit Study (NHCT)⁵²
• The Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial⁵³
• The Cardiovascular Risk Reduction by Early Anemia Treatment (CREATE) trial⁵⁴
• The Trial to Reduce Cardiovascular Events With Aranesp Therapy (TREAT).⁵⁵

These trials randomized patients to either higher “normal-range” hemoglobin targets or to lower target hemoglobin levels.

Their findings were disappointing and raised several red flags about excessive use of ESAs. The trials found no benefit in higher hemoglobin targets, and in fact, some of them demonstrated harm in patients randomized to higher targets. Notably, higher hemoglobin targets were associated with significant side effects such as access-site  thrombosis,⁵² strokes,⁵⁵ and possibly cardiovascular events.^54,55 Only the CREATE trial was able to demonstrate a quality-of-life benefit for the high-target group.⁵⁴ 

It remains unclear whether these adverse events were from the therapy itself or from an increased morbidity burden in the treated patients. Erythropoietin use is associated with hypertension,⁵⁶ thought to be related to endothelin-mediated vasoconstriction.⁵⁷ In our experience, this is most evident when hemoglobin levels are normalized with ESA therapy. Cycling of erythropoietin levels between extreme levels can lead to vascular remodeling, which may also be related to its cardiovascular effects.⁵⁷

A noticeable finding in several of these trials was that patients failed to achieve the higher hemoglobin target despite the use of very high doses of ESA. Reanalysis of data from the CHOIR and CREATE trials showed that the patients who had worse outcomes were more likely to have required very high doses without achieving their target hemoglobin.^58,59 Indeed, patients who achieved the higher target hemoglobin levels, usually at lower ESA doses, had better outcomes. This suggested that the need for a higher dose was associated with poorer outcomes, either as a marker of comorbidity or due to yet undocumented side effects of such high doses.

General approach to therapy

Before attributing anemia to chronic kidney disease, a thorough evaluation should be conducted to look for any reversible process that could be contributing to the anemia.

The causes of anemia are numerous and beyond the scope of this review. However, among the common causes of anemia in chronic kidney disease are deficiencies of iron, vitamin B₁₂, and folate. Therefore, guidelines recommend checking iron, vitamin B₁₂, and folate levels in the initial evaluation of anemia.⁶⁰

Iron deficiency in particular is very common in chronic kidney disease patients and is present in nearly all dialysis patients.⁶¹ Hemodialysis patients are estimated to lose 1 to 3 g of iron per year as a result of blood loss in the dialysis circuit and increased iron utilization secondary to ESA therapy.⁶²

However, in contrast to the general population, in which the upper limits of normal for iron indices are well defined, high serum ferritin levels appear to be poorly predictive of hemoglobin responsiveness in dialysis patients.^63,64 Thus, the cutoffs that define iron responsiveness are much higher than standard definitions for iron deficiency.^65,66 The Dialysis Patients’ Response to IV Iron With Elevated Ferritin (DRIVE) study showed that dialysis patients benefit from intravenous iron therapy even if their ferritin is as high as 1,200 ng/mL, provided their transferrin saturation is below 30%.⁶⁷

Of note, erythropoietin levels cannot be used to distinguish renal anemia from other causes of anemia. Indeed, patients with renal failure may have “relative erythropoietin deficiency,” ie, “normal” erythropoietin levels that are actually too low in view of the degree of anemia.^68,69 In addition to the decreased production capacity by the kidney, there appears to be a component of resistance to the action of erythropoietin in the bone marrow.

For these reasons, there is no erythropoietin level that can be considered “inadequate” or defining of renal anemia. Thus, measuring erythropoietin levels is not routinely recommended in the evaluation of renal anemia.

Two ESA preparations

The two ESAs that have traditionally been used in the treatment of renal anemia are recombinant human erythropoietin and darbepoietin alfa. They appear to be equivalent in terms of safety and efficacy.⁷⁰ However, darbepoietin alfa has more sialic acid molecules, giving it a higher potency and longer half-life and allowing for less-frequent injections.^71,72

In nondialysis patients, recombinant human erythropoietin is typically given every 1 to 2 weeks, whereas darbepoietin alfa can be given every 2 to 4 weeks. In dialysis patients, recombinant human erythropoietin is typically given 3 times per week with every dialysis treatment, while darbepoietin alfa is given once a week.

Target hemoglobin levels: ≤ 11.5 g/dL

In light of the four trials described in Table 1, the international Kidney Disease: Improving Global Outcomes (KDIGO) guidelines⁶⁰ recommend the following (Table 2):

For patients with chronic kidney disease who are not on dialysis, ESA therapy should not be initiated if the hemoglobin level is higher than 10 g/dL. If the hemoglobin level is lower than 10 g/dL, ESA therapy can be initiated, but the decision needs to be individualized based on the rate of fall of hemoglobin concentration, prior response to iron therapy, the risk of needing a transfusion, the risks related to ESA therapy, and the presence of symptoms attributable to anemia.

For patients on dialysis, ESA therapy should be used when the hemoglobin level is between 9 and 10 g/dL to avoid having the hemoglobin fall below 9 g/dL.

In all adult patients, ESAs should not be used to intentionally increase the hemoglobin level above 13 g/dL but rather to maintain levels no higher than 11.5 g/dL. This target is based on the observation that adverse outcomes were associated with ESA use with hemoglobin targets higher than 13 g/dL (Table 1).

Target iron levels

Regarding iron stores, the guidelines recommend the following:

For adult patients with chronic kidney disease who are not on dialysis, iron should be given to keep transferrin saturation above 20% and ferritin above 100 ng/mL. Transferrin saturation should not exceed 30%, and ferritin levels should not exceed 500 ng/mL.

For adult patients on dialysis, iron should be given to maintain transferrin saturation above 30% and ferritin above 200 ng/mL.

The upper limits of ferritin and transferrin saturation are somewhat controversial, as the safety of intentionally maintaining respective levels greater than 30% and 500 ng/mL has been studied in very few patients. Transferrin saturation should in general not exceed 50%.

High ferritin levels are associated with higher death rates, but whether elevation of ferritin levels is a marker of excessive iron administration rather than a nonspecific acute-phase reactant is not clear. The 2006 guidelines⁶⁰ cited upper ferritin limits of 500 to 800 ng/mL. However, the more recent DRIVE trial⁶⁷ showed that patients with ferritin levels of 500 to 1,200 ng/mL will respond to intravenous administration of iron with an increase in their hemoglobin levels. This has led many clinicians to adopt a higher ferritin limit of 1,200 ng/mL.

Hemosiderosis, or excess iron deposition, was a known consequence of frequent transfusions in patients with end-stage renal disease before ESA therapy was available. However, there have been no documented cases of clinical iron overload from iron therapy using current guidelines.⁷³

These algorithms are nuanced, and the benefit of giving intravenous iron should always be weighed against the risks of short-term acute toxicity and infection. Treatment of renal anemia not only requires in-depth knowledge of the topic, but also familiarity with the patient’s specific situation. As such, it is not recommended that clinicians unfamiliar with the treatment of renal anemia manage its treatment.

PARTICULAR CIRCUMSTANCES

Inflammation and ESA resistance

While ESAs are effective in treating anemia in many cases, in many patients the anemia fails to respond. This is of particular importance, since ESA hyporesponsiveness has been found to be a powerful predictor of cardiovascular events and death.⁷⁴ It is unclear, however, whether high doses of ESA are inherently toxic or whether hyporesponsiveness is a marker of adverse outcomes related to comorbidities.

KDIGO defines initial hyporesponsiveness as having no increase in hemoglobin concentration after the first month of appropriate weight-based dosing, and acquired hyporesponsiveness as requiring two increases in ESA doses up to 50% beyond the dose at which the patient had originally been stable.⁶⁰ Identifying ESA hyporesponsiveness should lead to an intensive search for potentially correctable factors.

The two major factors accounting for the state of hyporesponsiveness are inflammation and iron deficiency.^75,76

Inflammation. High C-reactive protein levels have been shown to predict resistance to erythropoietin in dialysis patients.⁷⁷ The release of cytokines such as tumor necrosis factor alpha, interleukin 1, and interferon gamma has an inhibitory effect on erythropoiesis.⁷⁸ Additionally, inflammation can alter the response to ESAs by disrupting the metabolism of iron⁷⁹ through the release of hepcidin, as previously discussed.³⁸ These reasons likely account for the observed lower response to ESAs in the setting of acute illness and explain why ESAs are not recommended for correcting acute anemia.⁸⁰

Iron deficiency also can blunt the response to ESAs. Large amounts of iron are needed for effective erythropoietic bursts. As such, iron supplementation is now a recognized treatment of renal anemia.⁸¹

Other factors associated with hyporesponsiveness include chronic occult blood loss, aluminum toxicity, cobalamin or folate deficiencies, testosterone deficiency, inadequate dialysis, hyperparathyroidism, and superimposed primary bone marrow disease,^82,83 and these should be addressed in patients whose anemia does not respond as expected to ESA therapy. A summary of the main causes of ESA hyporesponsiveness, their reversibility, and recommended treatments is presented in Table 3.

Antibody-mediated pure red-cell aplasia. Rarely, patients receiving ESA therapy develop antibodies that neutralize both the ESA and endogenous erythropoietin. The resulting syndrome, called antibody-mediated pure red-cell aplasia, is characterized by the sudden development of severe transfusion-dependent anemia. This has historically been connected to epoetin beta, a formulation not in use in the United States. However, cases have been documented with epoetin alfa and darbepoetin. The incidence rate is low with subcutaneous ESA use, estimated at 0.5 cases per 10,000 patient-years⁸⁴ and anecdotal with intravenous ESA.⁸⁵ The definitive diagnosis requires demonstration of neutralizing antibodies against erythropoietin. Parvovirus infection should be excluded as an alternative cause of pure red­cell aplasia.

ANEMIA IN CANCER PATIENTS

ESAs are effective in raising hemoglobin levels and reducing transfusion requirements in patients with chemotherapy-induced anemia.⁸⁶ However, there are data linking the use of ESAs to shortened survival in patients who have a variety of solid tumors.⁸⁷

Several mechanisms have been proposed to explain this rapid disease progression, most notably acceleration in tumor growth^88–90 by stimulation of erythropoietin receptors on the surface of the tumor cells, leading to increased tumor angiogenesis.^91,92

For these reasons, treatment of renal anemia in the setting of active malignancy should be referred to an oncologist.

NOVEL TREATMENTS

Several new agents for treating renal anemia are currently under review.

Continuous erythropoiesis receptor activator

Continuous erythropoiesis receptor activator is a pegylated form of recombinant human erythropoietin that has the ability to repeatedly activate the erythropoietin receptor. It appears to be similar to the other forms of erythropoietin in terms of safety and efficacy in both end-stage renal disease⁹³ and chronic kidney disease.⁹⁴ It has the advantage of an extended serum half-life, which allows for longer dosing intervals, ie, every 2 weeks. Its use is currently gaining popularity in the dialysis community.

HIF stabilizers

Our growing understanding of the physiology of erythropoietin offers new potential treatment targets. As previously described, production of erythropoietin is stimulated by HIFs. In order to be degraded, these HIFs are hydroxylated at their proline residues by a prolyl hydroxylase. A new category of drugs called prolyl-hydroxylase inhibitors (PDIs) offers the advantage of stabilizing the HIFs, leading to an increase in erythropoietin production.

In phase 1 and 2 clinical trials, these agents have been shown to increase hemoglobin in both end-stage renal disease and chronic kidney disease patients^15,16 but not in anephric patients, demonstrating a renal source of the erythropoietin production even in nonfunctioning kidneys. The study of one PDI agent (FG 2216) was halted temporarily after a report of death from fulminant hepatitis, but the other (FG 4592) continues to be studied in a phase 2 clinical trial.^95,96

TAKE-HOME POINTS

• Anemia of renal disease is a common condition that is mainly caused by a decrease in erythropoietin production by the kidneys.
• While anemia of renal disease can be corrected with ESAs, it is necessary to investigate and rule out underlying treatable conditions such as iron or vitamin deficiencies before giving an ESA.
• Anemia of renal disease is associated with significant morbidity such as increased risk of left ventricular hypertrophy, myocardial infarction, and heart failure, and has been described as an all-cause mortality multiplier.
• Unfortunately, the only undisputed benefit of treatment to date remains the avoidance of blood transfusions. Furthermore, the large randomized controlled trials that looked at the benefits of ESA have shown that their use can be associated with increased risk of cardiovascular events. Therefore, use of an ESA in end-stage renal disease should never target a normal hemoglobin levels but rather aim for a hemoglobin level of no more than 11.5 g/dL.
• Use of an ESA in chronic kidney disease should be individualized and is not recommended to be started unless the hemoglobin level is less than 10 g/dL.
• Several newer agents for renal anemia are currently under review. A pegylated form of recombinant human erythropoietin has an extended half-life, and a new and promising category of drugs called HIF stabilizers is currently under study.

Accessing, absorbing, organizing, storing, and retrieving potentially useful medical information is a challenge. Physicians used to try meet this challenge with personal libraries of journal articles in their file cabinet. Today, that is inadequate to combat the deluge of digital information. In 2013, the Institute of Medicine acknowledged this problem, stating, “The ever-increasing volume of evidence makes it difficult for clinicians to maintain a working knowledge of new clinical information.”¹

The sheer volume of data has meant that, rather than try to maintain a regular diet of professional reading (proactive scanning²), many of us now seek information only when we need answers to specific clinical questions (reactive searching). This approach promotes lifelong problem-based learning but assumes that we are consciously aware of this need and are aware of the need to search for new information.

We need to constantly scan for new evidence in our area of practice to avoid becoming falsely assured of our knowledge. We also need to be able to find information we have seen before we need to use it. The following are examples of how using this approach could dramatically empower a busy clinician.

On a recent clinic day, a colleague pokes his head into your office and asks, “Have you heard anything about niacin in the news? I have a patient who is asking me if we should discontinue it.” You respond: “Yes, there was something that just came out. I am not sure where I read it or heard about it. Give me a couple of seconds and I can find it.” True to your word, a few seconds later you find and share the latest article on the HPS2-THRIVE trial and a commentary on the results.

Later, your first patient of the day asks you, “When we switched to dabigatran, you mentioned that, unlike warfarin, there was no specific reversal agent. I heard that there is one now?” Instead of being taken aback, you nod your head. “Yes, I saw that recent article showing that the medication rapidly and completely reverses the effect of dabigatran in the majority of patients. While I hope we don’t need it, this is good news, particularly as there did not seem to be any major adverse events.”

Sounds too good to be true? Can this really be you? In this article, we outline an information management strategy (Figure 1) and tools to help busy clinicians stay up to date with new medical evidence in their areas of interest or expertise. In addition, we provide a strategy for leveraging technology to easily retrieve previously viewed information. A future article will specifically show how to best access information at the point of patient care.

THE NEED TO MANAGE INFORMATION

Physicians are expected to practice evidence-based medicine. When faced with a clinical question, we should search for evidence using focused queries of primary and secondary sources such as PubMed or the Cochrane Library. This is an important skill and is appropriate when we take time to look for an evidence-based answer to a specific question. In many cases, it is appropriate to continue with a current practice until newer information has been reviewed and validated.

Unfortunately, indexing and adding new recommendations to these information sources takes time. We may also be unaware that new information is available and may continue to practice as usual until faced with a situation like those outlined above or until we attend a continuing medical education activity, often quite by chance.

Today we can proactively update ourselves in a manner tailored to our own interests and focus and retrieve important information easily when we need it.

A STRATEGY FOR INFORMATION MANAGEMENT

In general, we come across new information in one of three ways:

• Proactive scanning of personalized sources of information—as discussed above, a habit of regular scanning is critical to information management
• Reactive searching for information to answer clinical problems or when doing research
• Incidentally—an e-mail from a colleague, information shared on a social network or encountered while surfing the Internet.

In each case, we may find information that is potentially useful, something we may need to find again in the future. But unless we use this information often, we will not remember the details or may even forget we had seen it. Thus, we need a strategy to store this information so we can retrieve it easily at any time with any device; neuroscientists call this the “externalization” of memory.³ Ideally, even if we forget that we ever saw this information or where we stored it, a search would retrieve the location and details of this formerly viewed information.

In the following sections, we outline steps and tools of a strategy for managing clinical information relevant to your practice.

STEP 1: SETTING UP INFORMATION FEEDS

The first step in this information management strategy is to become aware of relevant new information in your area of practice or research. To do this, you proactively set up feeds of information from reliable and authentic sources. These feeds can be browsed on any computer or smart mobile device.

There are several possibilities for creating these feeds. One option is to subscribe to the table of contents (TOC) of relevant journals via e-mail.

A more versatile and full-featured option is a research site summary (RSS) feed-reader. RSS is a standard for publishing summaries (feeds) of frequently updated content on the World Wide Web, such as journal TOCs and items from medical journal news sites (Table 1 shows what this looks like on screen), as well as aggregators like the American College of Physicians Journal Club. You can subscribe to these using feed-reader software from Feedly (www.feedly.com) or Inoreader (www.inoreader.com), which can be used with any browser on a desktop or laptop. They are also available as apps for mobile devices such as smartphones and tablets. The feed-reader periodically checks for new content and automatically downloads it to the device. Thus, you do not need to check multiple websites for updates or have e-mail inboxes fill with content; the content is delivered to your device for perusal at your convenience. (See online supplement “Information Management for Clinicians” for step-by-step instructions on creating a free Inoreader account and subscribing to feeds.)

RSS feed-readers have several advantages over e-mailed TOCs:

• RSS feeds create a centralized searchable repository of all subscribed information.
• The software keeps track of what you have read and displays only unread items; after a journal TOC e-mail is opened, the entire TOC is marked as having been read.
• You can organize news items into folders by tagging key words.
• Most journals and medical news sites like Medscape and the health section of the New York Times provide RSS feeds at no cost.
• An unlimited number of feed items, or articles, are stored in the cloud and do not affect e-mail storage capacity.
• The feed is automatically updated multiple times a day instead of once a week or once a month.
• In addition, one can create RSS feeds on PubMed for custom searches. Thus, a physician can get automatic regular updates of new articles indexed in MEDLINE in their area of interest.

Users can thus build their own personalized magazine of constantly updated information for access and can search from any web-enabled device. (Note: It is advisable to turn off notifications generated by these apps on mobile devices to reduce distraction.)

STEP 2: BOOKMARKING EVIDENCE

When you find something useful or interesting, bookmarks help you find the information again quickly when you need it. But while the browsers Firefox, Chrome, Internet Explorer, and Safari allow bookmarking, they have significant limitations. Bookmarks may be available only on the device they were created on, and because people use more than one device to go online, they may not remember which device they used to bookmark or view the web page.

Browser bookmarks generally store the address (URL) of the web page and a label that you create, but they do not do much else. Sharing bookmarks with others is also difficult or impossible.

Social bookmarking

Social bookmarking lets you create bookmarks you can share across other devices and with other people. Diigo (www.diigo.com) and Delicious (www.delicious.com) are two social bookmarking services that let you integrate with all popular browsers through a button or toolbar. They allow you to save displayed web pages with labels, descriptions, and tags.

Diigo offers two additional features. It allows web pages to be annotated with highlights and notes. And during a Google search, relevant results from the Diigo library are simultaneously displayed.

If you forget you bookmarked something and saved it in your Diigo library, when you search for the information again on Google, Diigo will automatically display any results from your Diigo library next to the Google search results. This is very helpful as it is much easier to review information you have already read, marked up, and saved than it is to start over.

Bookmarks and annotations are stored in the cloud and can be accessed by any device. (See “Information Management for Clinicians” to learn how to sign up for a free Diigo account, and how to use it.)

STEP 3: STORING YOUR INFORMATION

You may want the option to store full-text information in your personal library. This information was once stored in file cabinets and, more recently, on hard drives and USB flash drives. But information stored with these methods is not available or searchable on multiple devices from any location.

Cloud storage

Cloud storage services meet the need for access to stored information at any time and with any device. Options include Dropbox (www.dropbox.com), Box (www.box.com), Google Drive (drive.google.com), OneDrive (onedrive.live.com), and Evernote (www.evernote.com). Each provides different amounts of free storage and has apps available for most platforms and devices. They provide search tools and the ability to share articles or “folders” with other users. The information on these online drives is “synced” between all devices so that the most up-to-date version is always available to the user regardless of location and device.

Evernote offers multiple folders called notebooks to store and segregate data. The open notebook shown in Figure 2 is named “reference articles.” It has the HPS2-Thrive article from the New England Journal of Medicine (N Engl J Med) tagged with the terms “niacin” and “lipid” to facilitate retrieval. The article was saved from that journal’s website using an Evernote browser extension that allows entire web pages or selections to be saved to Evernote with a single click. Evernote also has a powerful search feature that can find text in images or in PDF documents. In addition, it allows easy sharing of a note or an entire notebook. Once a note or notebook is shared, all parties can add to it. In The Evernote app also allows tablet and smartphone access to the shareable content.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

The other services listed here have similar feature sets. Dropbox is perhaps the easiest to adopt, but it offers the least amount of free storage. If you use Microsoft Office software, OneDrive lets you edit documents online, and an Office 365 subscription includes 1 terabyte of storage. Google Drive is probably the best solution for online collaboration, such as coauthoring a paper. Box is one of the few online storage solutions that complies with the Health Insurance Portability and Accountability privacy rules.

PUTTING IT ALL INTO PRACTICE

Once you have become familiar with Inoreader and Diigo (see Information Management for Clinicians for step-by-step instructions), the following scenario shows how to adapt them into an efficient workflow.

Dr. Smith has a smartphone, a tablet, a laptop at home, and a desktop at work. She signs onto Google Chrome as her preferred browser on all devices. This seamlessly loads her Diigo extension when she is using a laptop or desktop. She has set her RSS feeds for her preferred journal TOCs and medical news sites to be downloaded to Inoreader. (For details on how to add a medical journals feed bundle and a medical news feed bundle, visit Information Management for Clinicians.)

Instead of reading paper journals, Dr. Smith browses her customized up-to-date “magazine” on Inoreader. When she comes across a relevant article, she marks it as “favorite.” If she has more time, she visits the web page, reviews the information, and saves it to her Diigo library with annotations if appropriate.

When searching for information on the web, she uses Google—without having to remember if she bookmarked information related to the search term. The Diigo extension in her browser automatically searches and displays information from her Diigo library next to her Google search results, and she can instantly see her notes from the last time she read the article.

Relating this workflow to the example of the dabigatran story above, Dr. Smith sees an article about dabigatran reversal while viewing her N Engl J Med medical news feed on her feed-reader. She marks it as a favorite and tags it with the key terms “cardiology” and “vascular” (Figure 3).

Dr. Smith later returns to look at her favorite feed items and visits the article on the N Engl J Med website. She annotates the article and saves it to her Diigo library (Figure 4).

Since this information is highly relevant to her practice, she also visits the N Engl J Med website to read the full article and the accompanying editorial (Figure 5). She annotates these and also saves them to her Diigo library.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

Later, if she searches Google for dabigatran (using her default Google Chrome browser with Diigo extension), she will see the usual Google search results and twinned Diigo bookmarks (Figure 6).

If she clicks on one of the links, the browser will load the web page with all the annotations that she made when she first visited.

CONCLUSION

The strategies and tools we describe here let you create a personalized and constantly updated medical news “magazine,” accessible from any of your web-enabled devices. They can transform the Internet into a searchable notebook of personally selected, annotated information, helping you to more easily stay up to date with advances in your field of practice, and to more easily manage the modern information overload.

Accessing, absorbing, organizing, storing, and retrieving potentially useful medical information is a challenge. Physicians used to try meet this challenge with personal libraries of journal articles in their file cabinet. Today, that is inadequate to combat the deluge of digital information. In 2013, the Institute of Medicine acknowledged this problem, stating, “The ever-increasing volume of evidence makes it difficult for clinicians to maintain a working knowledge of new clinical information.”¹

The sheer volume of data has meant that, rather than try to maintain a regular diet of professional reading (proactive scanning²), many of us now seek information only when we need answers to specific clinical questions (reactive searching). This approach promotes lifelong problem-based learning but assumes that we are consciously aware of this need and are aware of the need to search for new information.

We need to constantly scan for new evidence in our area of practice to avoid becoming falsely assured of our knowledge. We also need to be able to find information we have seen before we need to use it. The following are examples of how using this approach could dramatically empower a busy clinician.

On a recent clinic day, a colleague pokes his head into your office and asks, “Have you heard anything about niacin in the news? I have a patient who is asking me if we should discontinue it.” You respond: “Yes, there was something that just came out. I am not sure where I read it or heard about it. Give me a couple of seconds and I can find it.” True to your word, a few seconds later you find and share the latest article on the HPS2-THRIVE trial and a commentary on the results.

Later, your first patient of the day asks you, “When we switched to dabigatran, you mentioned that, unlike warfarin, there was no specific reversal agent. I heard that there is one now?” Instead of being taken aback, you nod your head. “Yes, I saw that recent article showing that the medication rapidly and completely reverses the effect of dabigatran in the majority of patients. While I hope we don’t need it, this is good news, particularly as there did not seem to be any major adverse events.”

Sounds too good to be true? Can this really be you? In this article, we outline an information management strategy (Figure 1) and tools to help busy clinicians stay up to date with new medical evidence in their areas of interest or expertise. In addition, we provide a strategy for leveraging technology to easily retrieve previously viewed information. A future article will specifically show how to best access information at the point of patient care.

THE NEED TO MANAGE INFORMATION

Physicians are expected to practice evidence-based medicine. When faced with a clinical question, we should search for evidence using focused queries of primary and secondary sources such as PubMed or the Cochrane Library. This is an important skill and is appropriate when we take time to look for an evidence-based answer to a specific question. In many cases, it is appropriate to continue with a current practice until newer information has been reviewed and validated.

Unfortunately, indexing and adding new recommendations to these information sources takes time. We may also be unaware that new information is available and may continue to practice as usual until faced with a situation like those outlined above or until we attend a continuing medical education activity, often quite by chance.

Today we can proactively update ourselves in a manner tailored to our own interests and focus and retrieve important information easily when we need it.

A STRATEGY FOR INFORMATION MANAGEMENT

In general, we come across new information in one of three ways:

• Proactive scanning of personalized sources of information—as discussed above, a habit of regular scanning is critical to information management
• Reactive searching for information to answer clinical problems or when doing research
• Incidentally—an e-mail from a colleague, information shared on a social network or encountered while surfing the Internet.

In each case, we may find information that is potentially useful, something we may need to find again in the future. But unless we use this information often, we will not remember the details or may even forget we had seen it. Thus, we need a strategy to store this information so we can retrieve it easily at any time with any device; neuroscientists call this the “externalization” of memory.³ Ideally, even if we forget that we ever saw this information or where we stored it, a search would retrieve the location and details of this formerly viewed information.

In the following sections, we outline steps and tools of a strategy for managing clinical information relevant to your practice.

STEP 1: SETTING UP INFORMATION FEEDS

The first step in this information management strategy is to become aware of relevant new information in your area of practice or research. To do this, you proactively set up feeds of information from reliable and authentic sources. These feeds can be browsed on any computer or smart mobile device.

There are several possibilities for creating these feeds. One option is to subscribe to the table of contents (TOC) of relevant journals via e-mail.

A more versatile and full-featured option is a research site summary (RSS) feed-reader. RSS is a standard for publishing summaries (feeds) of frequently updated content on the World Wide Web, such as journal TOCs and items from medical journal news sites (Table 1 shows what this looks like on screen), as well as aggregators like the American College of Physicians Journal Club. You can subscribe to these using feed-reader software from Feedly (www.feedly.com) or Inoreader (www.inoreader.com), which can be used with any browser on a desktop or laptop. They are also available as apps for mobile devices such as smartphones and tablets. The feed-reader periodically checks for new content and automatically downloads it to the device. Thus, you do not need to check multiple websites for updates or have e-mail inboxes fill with content; the content is delivered to your device for perusal at your convenience. (See online supplement “Information Management for Clinicians” for step-by-step instructions on creating a free Inoreader account and subscribing to feeds.)

RSS feed-readers have several advantages over e-mailed TOCs:

• RSS feeds create a centralized searchable repository of all subscribed information.
• The software keeps track of what you have read and displays only unread items; after a journal TOC e-mail is opened, the entire TOC is marked as having been read.
• You can organize news items into folders by tagging key words.
• Most journals and medical news sites like Medscape and the health section of the New York Times provide RSS feeds at no cost.
• An unlimited number of feed items, or articles, are stored in the cloud and do not affect e-mail storage capacity.
• The feed is automatically updated multiple times a day instead of once a week or once a month.
• In addition, one can create RSS feeds on PubMed for custom searches. Thus, a physician can get automatic regular updates of new articles indexed in MEDLINE in their area of interest.

Users can thus build their own personalized magazine of constantly updated information for access and can search from any web-enabled device. (Note: It is advisable to turn off notifications generated by these apps on mobile devices to reduce distraction.)

STEP 2: BOOKMARKING EVIDENCE

When you find something useful or interesting, bookmarks help you find the information again quickly when you need it. But while the browsers Firefox, Chrome, Internet Explorer, and Safari allow bookmarking, they have significant limitations. Bookmarks may be available only on the device they were created on, and because people use more than one device to go online, they may not remember which device they used to bookmark or view the web page.

Browser bookmarks generally store the address (URL) of the web page and a label that you create, but they do not do much else. Sharing bookmarks with others is also difficult or impossible.

Social bookmarking

Social bookmarking lets you create bookmarks you can share across other devices and with other people. Diigo (www.diigo.com) and Delicious (www.delicious.com) are two social bookmarking services that let you integrate with all popular browsers through a button or toolbar. They allow you to save displayed web pages with labels, descriptions, and tags.

Diigo offers two additional features. It allows web pages to be annotated with highlights and notes. And during a Google search, relevant results from the Diigo library are simultaneously displayed.

If you forget you bookmarked something and saved it in your Diigo library, when you search for the information again on Google, Diigo will automatically display any results from your Diigo library next to the Google search results. This is very helpful as it is much easier to review information you have already read, marked up, and saved than it is to start over.

Bookmarks and annotations are stored in the cloud and can be accessed by any device. (See “Information Management for Clinicians” to learn how to sign up for a free Diigo account, and how to use it.)

STEP 3: STORING YOUR INFORMATION

You may want the option to store full-text information in your personal library. This information was once stored in file cabinets and, more recently, on hard drives and USB flash drives. But information stored with these methods is not available or searchable on multiple devices from any location.

Cloud storage

Cloud storage services meet the need for access to stored information at any time and with any device. Options include Dropbox (www.dropbox.com), Box (www.box.com), Google Drive (drive.google.com), OneDrive (onedrive.live.com), and Evernote (www.evernote.com). Each provides different amounts of free storage and has apps available for most platforms and devices. They provide search tools and the ability to share articles or “folders” with other users. The information on these online drives is “synced” between all devices so that the most up-to-date version is always available to the user regardless of location and device.

Evernote offers multiple folders called notebooks to store and segregate data. The open notebook shown in Figure 2 is named “reference articles.” It has the HPS2-Thrive article from the New England Journal of Medicine (N Engl J Med) tagged with the terms “niacin” and “lipid” to facilitate retrieval. The article was saved from that journal’s website using an Evernote browser extension that allows entire web pages or selections to be saved to Evernote with a single click. Evernote also has a powerful search feature that can find text in images or in PDF documents. In addition, it allows easy sharing of a note or an entire notebook. Once a note or notebook is shared, all parties can add to it. In The Evernote app also allows tablet and smartphone access to the shareable content.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

The other services listed here have similar feature sets. Dropbox is perhaps the easiest to adopt, but it offers the least amount of free storage. If you use Microsoft Office software, OneDrive lets you edit documents online, and an Office 365 subscription includes 1 terabyte of storage. Google Drive is probably the best solution for online collaboration, such as coauthoring a paper. Box is one of the few online storage solutions that complies with the Health Insurance Portability and Accountability privacy rules.

PUTTING IT ALL INTO PRACTICE

Once you have become familiar with Inoreader and Diigo (see Information Management for Clinicians for step-by-step instructions), the following scenario shows how to adapt them into an efficient workflow.

Dr. Smith has a smartphone, a tablet, a laptop at home, and a desktop at work. She signs onto Google Chrome as her preferred browser on all devices. This seamlessly loads her Diigo extension when she is using a laptop or desktop. She has set her RSS feeds for her preferred journal TOCs and medical news sites to be downloaded to Inoreader. (For details on how to add a medical journals feed bundle and a medical news feed bundle, visit Information Management for Clinicians.)

Instead of reading paper journals, Dr. Smith browses her customized up-to-date “magazine” on Inoreader. When she comes across a relevant article, she marks it as “favorite.” If she has more time, she visits the web page, reviews the information, and saves it to her Diigo library with annotations if appropriate.

When searching for information on the web, she uses Google—without having to remember if she bookmarked information related to the search term. The Diigo extension in her browser automatically searches and displays information from her Diigo library next to her Google search results, and she can instantly see her notes from the last time she read the article.

Relating this workflow to the example of the dabigatran story above, Dr. Smith sees an article about dabigatran reversal while viewing her N Engl J Med medical news feed on her feed-reader. She marks it as a favorite and tags it with the key terms “cardiology” and “vascular” (Figure 3).

Dr. Smith later returns to look at her favorite feed items and visits the article on the N Engl J Med website. She annotates the article and saves it to her Diigo library (Figure 4).

Since this information is highly relevant to her practice, she also visits the N Engl J Med website to read the full article and the accompanying editorial (Figure 5). She annotates these and also saves them to her Diigo library.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

Later, if she searches Google for dabigatran (using her default Google Chrome browser with Diigo extension), she will see the usual Google search results and twinned Diigo bookmarks (Figure 6).

If she clicks on one of the links, the browser will load the web page with all the annotations that she made when she first visited.

CONCLUSION

The strategies and tools we describe here let you create a personalized and constantly updated medical news “magazine,” accessible from any of your web-enabled devices. They can transform the Internet into a searchable notebook of personally selected, annotated information, helping you to more easily stay up to date with advances in your field of practice, and to more easily manage the modern information overload.

Accessing, absorbing, organizing, storing, and retrieving potentially useful medical information is a challenge. Physicians used to try meet this challenge with personal libraries of journal articles in their file cabinet. Today, that is inadequate to combat the deluge of digital information. In 2013, the Institute of Medicine acknowledged this problem, stating, “The ever-increasing volume of evidence makes it difficult for clinicians to maintain a working knowledge of new clinical information.”¹

The sheer volume of data has meant that, rather than try to maintain a regular diet of professional reading (proactive scanning²), many of us now seek information only when we need answers to specific clinical questions (reactive searching). This approach promotes lifelong problem-based learning but assumes that we are consciously aware of this need and are aware of the need to search for new information.

We need to constantly scan for new evidence in our area of practice to avoid becoming falsely assured of our knowledge. We also need to be able to find information we have seen before we need to use it. The following are examples of how using this approach could dramatically empower a busy clinician.

On a recent clinic day, a colleague pokes his head into your office and asks, “Have you heard anything about niacin in the news? I have a patient who is asking me if we should discontinue it.” You respond: “Yes, there was something that just came out. I am not sure where I read it or heard about it. Give me a couple of seconds and I can find it.” True to your word, a few seconds later you find and share the latest article on the HPS2-THRIVE trial and a commentary on the results.

Later, your first patient of the day asks you, “When we switched to dabigatran, you mentioned that, unlike warfarin, there was no specific reversal agent. I heard that there is one now?” Instead of being taken aback, you nod your head. “Yes, I saw that recent article showing that the medication rapidly and completely reverses the effect of dabigatran in the majority of patients. While I hope we don’t need it, this is good news, particularly as there did not seem to be any major adverse events.”

Sounds too good to be true? Can this really be you? In this article, we outline an information management strategy (Figure 1) and tools to help busy clinicians stay up to date with new medical evidence in their areas of interest or expertise. In addition, we provide a strategy for leveraging technology to easily retrieve previously viewed information. A future article will specifically show how to best access information at the point of patient care.

THE NEED TO MANAGE INFORMATION

Physicians are expected to practice evidence-based medicine. When faced with a clinical question, we should search for evidence using focused queries of primary and secondary sources such as PubMed or the Cochrane Library. This is an important skill and is appropriate when we take time to look for an evidence-based answer to a specific question. In many cases, it is appropriate to continue with a current practice until newer information has been reviewed and validated.

Unfortunately, indexing and adding new recommendations to these information sources takes time. We may also be unaware that new information is available and may continue to practice as usual until faced with a situation like those outlined above or until we attend a continuing medical education activity, often quite by chance.

Today we can proactively update ourselves in a manner tailored to our own interests and focus and retrieve important information easily when we need it.

A STRATEGY FOR INFORMATION MANAGEMENT

In general, we come across new information in one of three ways:

• Proactive scanning of personalized sources of information—as discussed above, a habit of regular scanning is critical to information management
• Reactive searching for information to answer clinical problems or when doing research
• Incidentally—an e-mail from a colleague, information shared on a social network or encountered while surfing the Internet.

In each case, we may find information that is potentially useful, something we may need to find again in the future. But unless we use this information often, we will not remember the details or may even forget we had seen it. Thus, we need a strategy to store this information so we can retrieve it easily at any time with any device; neuroscientists call this the “externalization” of memory.³ Ideally, even if we forget that we ever saw this information or where we stored it, a search would retrieve the location and details of this formerly viewed information.

In the following sections, we outline steps and tools of a strategy for managing clinical information relevant to your practice.

STEP 1: SETTING UP INFORMATION FEEDS

The first step in this information management strategy is to become aware of relevant new information in your area of practice or research. To do this, you proactively set up feeds of information from reliable and authentic sources. These feeds can be browsed on any computer or smart mobile device.

There are several possibilities for creating these feeds. One option is to subscribe to the table of contents (TOC) of relevant journals via e-mail.

A more versatile and full-featured option is a research site summary (RSS) feed-reader. RSS is a standard for publishing summaries (feeds) of frequently updated content on the World Wide Web, such as journal TOCs and items from medical journal news sites (Table 1 shows what this looks like on screen), as well as aggregators like the American College of Physicians Journal Club. You can subscribe to these using feed-reader software from Feedly (www.feedly.com) or Inoreader (www.inoreader.com), which can be used with any browser on a desktop or laptop. They are also available as apps for mobile devices such as smartphones and tablets. The feed-reader periodically checks for new content and automatically downloads it to the device. Thus, you do not need to check multiple websites for updates or have e-mail inboxes fill with content; the content is delivered to your device for perusal at your convenience. (See online supplement “Information Management for Clinicians” for step-by-step instructions on creating a free Inoreader account and subscribing to feeds.)

RSS feed-readers have several advantages over e-mailed TOCs:

• RSS feeds create a centralized searchable repository of all subscribed information.
• The software keeps track of what you have read and displays only unread items; after a journal TOC e-mail is opened, the entire TOC is marked as having been read.
• You can organize news items into folders by tagging key words.
• Most journals and medical news sites like Medscape and the health section of the New York Times provide RSS feeds at no cost.
• An unlimited number of feed items, or articles, are stored in the cloud and do not affect e-mail storage capacity.
• The feed is automatically updated multiple times a day instead of once a week or once a month.
• In addition, one can create RSS feeds on PubMed for custom searches. Thus, a physician can get automatic regular updates of new articles indexed in MEDLINE in their area of interest.

Users can thus build their own personalized magazine of constantly updated information for access and can search from any web-enabled device. (Note: It is advisable to turn off notifications generated by these apps on mobile devices to reduce distraction.)

STEP 2: BOOKMARKING EVIDENCE

When you find something useful or interesting, bookmarks help you find the information again quickly when you need it. But while the browsers Firefox, Chrome, Internet Explorer, and Safari allow bookmarking, they have significant limitations. Bookmarks may be available only on the device they were created on, and because people use more than one device to go online, they may not remember which device they used to bookmark or view the web page.

Browser bookmarks generally store the address (URL) of the web page and a label that you create, but they do not do much else. Sharing bookmarks with others is also difficult or impossible.

Social bookmarking

Social bookmarking lets you create bookmarks you can share across other devices and with other people. Diigo (www.diigo.com) and Delicious (www.delicious.com) are two social bookmarking services that let you integrate with all popular browsers through a button or toolbar. They allow you to save displayed web pages with labels, descriptions, and tags.

Diigo offers two additional features. It allows web pages to be annotated with highlights and notes. And during a Google search, relevant results from the Diigo library are simultaneously displayed.

If you forget you bookmarked something and saved it in your Diigo library, when you search for the information again on Google, Diigo will automatically display any results from your Diigo library next to the Google search results. This is very helpful as it is much easier to review information you have already read, marked up, and saved than it is to start over.

Bookmarks and annotations are stored in the cloud and can be accessed by any device. (See “Information Management for Clinicians” to learn how to sign up for a free Diigo account, and how to use it.)

STEP 3: STORING YOUR INFORMATION

You may want the option to store full-text information in your personal library. This information was once stored in file cabinets and, more recently, on hard drives and USB flash drives. But information stored with these methods is not available or searchable on multiple devices from any location.

Cloud storage

Cloud storage services meet the need for access to stored information at any time and with any device. Options include Dropbox (www.dropbox.com), Box (www.box.com), Google Drive (drive.google.com), OneDrive (onedrive.live.com), and Evernote (www.evernote.com). Each provides different amounts of free storage and has apps available for most platforms and devices. They provide search tools and the ability to share articles or “folders” with other users. The information on these online drives is “synced” between all devices so that the most up-to-date version is always available to the user regardless of location and device.

Evernote offers multiple folders called notebooks to store and segregate data. The open notebook shown in Figure 2 is named “reference articles.” It has the HPS2-Thrive article from the New England Journal of Medicine (N Engl J Med) tagged with the terms “niacin” and “lipid” to facilitate retrieval. The article was saved from that journal’s website using an Evernote browser extension that allows entire web pages or selections to be saved to Evernote with a single click. Evernote also has a powerful search feature that can find text in images or in PDF documents. In addition, it allows easy sharing of a note or an entire notebook. Once a note or notebook is shared, all parties can add to it. In The Evernote app also allows tablet and smartphone access to the shareable content.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

The other services listed here have similar feature sets. Dropbox is perhaps the easiest to adopt, but it offers the least amount of free storage. If you use Microsoft Office software, OneDrive lets you edit documents online, and an Office 365 subscription includes 1 terabyte of storage. Google Drive is probably the best solution for online collaboration, such as coauthoring a paper. Box is one of the few online storage solutions that complies with the Health Insurance Portability and Accountability privacy rules.

PUTTING IT ALL INTO PRACTICE

Once you have become familiar with Inoreader and Diigo (see Information Management for Clinicians for step-by-step instructions), the following scenario shows how to adapt them into an efficient workflow.

Dr. Smith has a smartphone, a tablet, a laptop at home, and a desktop at work. She signs onto Google Chrome as her preferred browser on all devices. This seamlessly loads her Diigo extension when she is using a laptop or desktop. She has set her RSS feeds for her preferred journal TOCs and medical news sites to be downloaded to Inoreader. (For details on how to add a medical journals feed bundle and a medical news feed bundle, visit Information Management for Clinicians.)

Instead of reading paper journals, Dr. Smith browses her customized up-to-date “magazine” on Inoreader. When she comes across a relevant article, she marks it as “favorite.” If she has more time, she visits the web page, reviews the information, and saves it to her Diigo library with annotations if appropriate.

When searching for information on the web, she uses Google—without having to remember if she bookmarked information related to the search term. The Diigo extension in her browser automatically searches and displays information from her Diigo library next to her Google search results, and she can instantly see her notes from the last time she read the article.

Relating this workflow to the example of the dabigatran story above, Dr. Smith sees an article about dabigatran reversal while viewing her N Engl J Med medical news feed on her feed-reader. She marks it as a favorite and tags it with the key terms “cardiology” and “vascular” (Figure 3).

Dr. Smith later returns to look at her favorite feed items and visits the article on the N Engl J Med website. She annotates the article and saves it to her Diigo library (Figure 4).

Since this information is highly relevant to her practice, she also visits the N Engl J Med website to read the full article and the accompanying editorial (Figure 5). She annotates these and also saves them to her Diigo library.

From Landray MJ, et al. Effects of extended-release niacin with laropiprant in high-risk patients. HPS2-THRIVE Collaborative Group. N Engl J Med 2014; 371:203-212. Copyright 2014, Massachusetts Medical Society (MMS). Reprinted with permission from MMS.

Later, if she searches Google for dabigatran (using her default Google Chrome browser with Diigo extension), she will see the usual Google search results and twinned Diigo bookmarks (Figure 6).

If she clicks on one of the links, the browser will load the web page with all the annotations that she made when she first visited.

CONCLUSION

The strategies and tools we describe here let you create a personalized and constantly updated medical news “magazine,” accessible from any of your web-enabled devices. They can transform the Internet into a searchable notebook of personally selected, annotated information, helping you to more easily stay up to date with advances in your field of practice, and to more easily manage the modern information overload.

The answer is less clear for these patients than for patients with acute coronary syndromes. In the latter group, percutaneous or surgical revascularization reduces the rates of morbidity and mortality, whereas in patients with stable ischemic heart disease, benefits may be limited to the improvement of angina. Certain markers and criteria may help us in this decision, and trials are ongoing.

Of importance, all patients with coronary artery disease should receive guideline-directed medical therapy as tolerated, regardless of whether they undergo revascularization.

MEDICAL THERAPY FOR ALL

In all the relevant trials, patients with stable ischemic heart disease in both the revascularization groups and the unrevascularized groups received guideline-directed medical therapy. Current guidelines¹ give class I recommendations (ie, treatment should be given) for:

• Lipid management
• Blood pressure management
• Physical activity
• Weight management
• Smoking cessation
• Antiplatelet therapy
• Beta-blockers for patients with normal left ventricular function after an acute coronary syndrome event, and for those with an ejection fraction of 40% or less
• Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers for patients who have hypertension, diabetes mellitus, a left ventricular ejection fraction of 40% or less, or chronic kidney disease
• Annual influenza vaccination
• Anti-ischemic medications (beta-blockers, calcium channel blockers, nitrates) for relief of symptoms.

REVASCULARIZATION FOR SOME?

Results of the studies outlined below will help in deciding when to use guideline-directed medical therapy alone or medical therapy plus revascularization.

COURAGE trial: No added benefit in patients at low risk

The findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE), published in 2007, suggested that in select patients, percutaneous coronary intervention for stable coronary artery disease was no better than guideline-directed medical therapy alone for reducing the outcomes of death, myocardial infarction, or hospitalization for acute coronary syndrome.²

Of note, however, is that the 2,287 patients included in COURAGE were a low-risk subset of the more than 35,000 patients initially evaluated. The investigators reviewed the patients’ coronary angiograms before enrollment, and thus many patients with complex or high-risk anatomy were likely excluded based on an a priori assessment of angiographic images.

Also, coronary stent technology has substantially improved since COURAGE (which primarily used bare-metal stents and early drug-eluting stents), and this brings into question whether the results are applicable to current patients.

Moreover, in subsequent substudies from COURAGE, revascularization significantly improved symptoms of angina and quality-of-life scores compared with medical therapy alone.^3,4

Also important is that more than one-third of the patients in the medical therapy group crossed over to revascularization during the study, most often for worsening symptoms of angina.

Regardless of its limitations, COURAGE played an important role in delineating the use of guideline-directed medical therapy alone in certain low-risk patients and sparked debate about when and if to revascularize other patients.

BARI 2D trial: CABG may benefit those with diabetes

The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial, published in 2009, aimed to find out if revascularization in patients with stable ischemic heart disease and diabetes was beneficial compared with medical therapy alone.⁵

While it was not designed to directly compare percutaneous coronary intervention vs coronary artery bypass grafting (CABG), it did find that medical therapy plus CABG might reduce the rate of adverse cardiovascular events in this population compared with medical therapy alone or medical therapy plus percutaneous intervention.

As with COURAGE, however, the patients in the medical therapy group in BARI 2D also had a high rate of crossover to revascularization, primarily driven by worsening anginal symptoms.

FREEDOM and the 2014 updated guideline

Based on the findings of BARI 2D and those of FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease),⁶ the American College of Cardiology and American Heart Association updated their recommendations in 2014.⁷ This focused update states that for patients with diabetes and multivessel coronary artery disease, if revascularization is likely to improve survival (for example, in three-vessel disease or complex two-vessel disease involving the proximal left anterior descending artery), then CABG should be performed if a left internal mammary artery graft can be anastomosed to the left anterior descending artery. Otherwise, percutaneous coronary intervention should be reserved for those patients with diabetes and high-risk or complex multivessel coronary artery disease who are not good surgical candidates.

FAME 2 trial: Fractional flow reserve as a guide

The Fractional Flow Reserve Versus Angiography for Multivessel Evaluation 2 (FAME 2) trial,⁸ published in 2012, evaluated whether clinical outcomes differ between patients who undergo percutaneous revascularization plus medical therapy and those who are treated with medical therapy alone, using fractional flow reserve as a means to determine which stenoses should be considered for intervention. Fractional flow reserve performed during invasive angiography determines the ratio of intracoronary pressure to aortic pressure using a wire advanced across a coronary obstruction.

FAME 2 found a markedly lower incidence of the primary composite end point of death, myocardial infarction, and urgent revascularization with randomization to percutaneous revascularization plus medical therapy compared with medical therapy only (4.3% vs 12.7%, P = .001) in patients with a fractional flow reserve less than 0.80 (considered a hemodynamically significant obstruction). The trial was stopped early because of the markedly different outcomes.

Of note, however, the reduction in adverse clinical outcomes was driven primarily by a reduction in urgent revascularizations in those treated with percutaneous coronary intervention in the revascularization arm. Regardless, using fractional flow reserve to guide whether obstructive coronary lesions should be treated with percutaneous coronary intervention has appropriately become a mainstay in interventional cardiology.

Stress testing

Noninvasive stress testing has played a role in helping to guide revascularization decisions in stable ischemic heart disease. In particular, revascularization in the setting of greater than 10% ischemia on perfusion imaging has been associated with a lower risk of cardiac death than in those who were revascularized with an ischemic burden less than 10%.⁹

A substudy of COURAGE found that percutaneous coronary intervention reduced ischemia to a greater degree than medical therapy alone on serial nuclear stress tests in patients with stable ischemic heart disease.¹⁰ In this substudy, when both groups were combined, the investigators also found that there were fewer adverse events in those who had an overall reduction of ischemia regardless of treatment strategy.

ISCHEMIA: Revascularize those with ischemia?

While COURAGE, BARI 2D, and FAME 2 suggested that early revascularization for low-risk patients with coronary artery disease does not confer a benefit over medical treatment alone with regard to hard clinical end points, it remains unclear whether an early revascularization strategy is advantageous in patients with stable ischemic heart disease who have at least a moderate amount of ischemia on noninvasive stress testing.

The ongoing ISCHEMIA (International Study of Comparative Effectiveness With Medical and Invasive Approaches) trial will help to answer that question. In this study, 8,000 patients with stable angina and at least moderate ischemia on noninvasive stress testing are being randomized before coronary angiography either to guideline-directed medical therapy plus revascularization (percutaneous or surgical) or to medical therapy alone.¹¹ The ISCHEMIA study population reflects current practice more closely than the previous studies discussed above in its inclusion of fractional flow reserve and later-generation drug-eluting stents.

The results of ISCHEMIA will be an important piece of the puzzle to answer whether patients with stable ischemic heart disease benefit from revascularization in terms of cardiovascular mortality or myocardial infarction (the primary end point of the study).

Studies in additional subsets

It is important to recognize that there are additional subsets of patients with stable ischemic heart disease (those with multivessel disease, left main coronary disease, or low ejection fractions, for example) who have been studied to help determine when and how to perform revascularization. In addition, there are guidelines¹² for both interventional cardiologists and cardiac surgeons that help delineate which patients should undergo revascularization. While a complete review is beyond the scope of this discussion, three trials are worth mentioning:

The Coronary Artery Surgery Study (CASS)¹³ revealed that revascularization in left main coronary artery disease is associated with lower mortality rates than medical therapy alone. This study, along with others, eventually led to recommendations for revascularization to be performed in all patients with significant left main coronary disease, regardless of symptoms or stress test findings.^14,15

The Surgical Treatment for Ischemic Heart Failure (STICH) trial¹⁶ found that patients with a low ejection fraction (< 35%) and ischemic heart disease had no difference in all-cause mortality rates when treated with CABG plus medical therapy compared with medical therapy alone (although the study’s design has been heavily criticized).

The Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) study¹⁷ found that CABG was associated with fewer adverse events in three-vessel coronary artery disease or complex left main coronary artery disease compared with percutaneous coronary intervention. The study used early-generation paclitaxel drug-eluting stents that are no longer used in contemporary practice. This study established the SYNTAX score, which is often used to help make revascularization decisions. A low SYNTAX score of 0 to 22 (meaning less-severe coronary artery disease) was associated with equivalent outcomes for both percutaneous coronary intervention and CABG. Thus, even if there is multivessel disease or left main disease, if the SYNTAX score is low, then percutaneous coronary intervention is an acceptable method for revascularization with similar results as for CABG.

A TEAM APPROACH

Due to the complexity of stable ischemic heart disease and the subtleties of managing these patients, a multidisciplinary “heart team” approach may be the best way to navigate treating stable ischemic heart disease via revascularization or with medical therapy alone. The heart team approach could take advantage of the particular expertise that the primary care physician, cardiologist, interventional cardiologist, and cardiac surgeon provide.

The upcoming results of studies such as the ISCHEMIA trial will help to provide additional guidance for these teams in long-term management of patients with stable ischemic heart disease.

The answer is less clear for these patients than for patients with acute coronary syndromes. In the latter group, percutaneous or surgical revascularization reduces the rates of morbidity and mortality, whereas in patients with stable ischemic heart disease, benefits may be limited to the improvement of angina. Certain markers and criteria may help us in this decision, and trials are ongoing.

Of importance, all patients with coronary artery disease should receive guideline-directed medical therapy as tolerated, regardless of whether they undergo revascularization.

MEDICAL THERAPY FOR ALL

In all the relevant trials, patients with stable ischemic heart disease in both the revascularization groups and the unrevascularized groups received guideline-directed medical therapy. Current guidelines¹ give class I recommendations (ie, treatment should be given) for:

• Lipid management
• Blood pressure management
• Physical activity
• Weight management
• Smoking cessation
• Antiplatelet therapy
• Beta-blockers for patients with normal left ventricular function after an acute coronary syndrome event, and for those with an ejection fraction of 40% or less
• Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers for patients who have hypertension, diabetes mellitus, a left ventricular ejection fraction of 40% or less, or chronic kidney disease
• Annual influenza vaccination
• Anti-ischemic medications (beta-blockers, calcium channel blockers, nitrates) for relief of symptoms.

REVASCULARIZATION FOR SOME?

Results of the studies outlined below will help in deciding when to use guideline-directed medical therapy alone or medical therapy plus revascularization.

COURAGE trial: No added benefit in patients at low risk

The findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE), published in 2007, suggested that in select patients, percutaneous coronary intervention for stable coronary artery disease was no better than guideline-directed medical therapy alone for reducing the outcomes of death, myocardial infarction, or hospitalization for acute coronary syndrome.²

Of note, however, is that the 2,287 patients included in COURAGE were a low-risk subset of the more than 35,000 patients initially evaluated. The investigators reviewed the patients’ coronary angiograms before enrollment, and thus many patients with complex or high-risk anatomy were likely excluded based on an a priori assessment of angiographic images.

Also, coronary stent technology has substantially improved since COURAGE (which primarily used bare-metal stents and early drug-eluting stents), and this brings into question whether the results are applicable to current patients.

Moreover, in subsequent substudies from COURAGE, revascularization significantly improved symptoms of angina and quality-of-life scores compared with medical therapy alone.^3,4

Also important is that more than one-third of the patients in the medical therapy group crossed over to revascularization during the study, most often for worsening symptoms of angina.

Regardless of its limitations, COURAGE played an important role in delineating the use of guideline-directed medical therapy alone in certain low-risk patients and sparked debate about when and if to revascularize other patients.

BARI 2D trial: CABG may benefit those with diabetes

The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial, published in 2009, aimed to find out if revascularization in patients with stable ischemic heart disease and diabetes was beneficial compared with medical therapy alone.⁵

While it was not designed to directly compare percutaneous coronary intervention vs coronary artery bypass grafting (CABG), it did find that medical therapy plus CABG might reduce the rate of adverse cardiovascular events in this population compared with medical therapy alone or medical therapy plus percutaneous intervention.

As with COURAGE, however, the patients in the medical therapy group in BARI 2D also had a high rate of crossover to revascularization, primarily driven by worsening anginal symptoms.

FREEDOM and the 2014 updated guideline

Based on the findings of BARI 2D and those of FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease),⁶ the American College of Cardiology and American Heart Association updated their recommendations in 2014.⁷ This focused update states that for patients with diabetes and multivessel coronary artery disease, if revascularization is likely to improve survival (for example, in three-vessel disease or complex two-vessel disease involving the proximal left anterior descending artery), then CABG should be performed if a left internal mammary artery graft can be anastomosed to the left anterior descending artery. Otherwise, percutaneous coronary intervention should be reserved for those patients with diabetes and high-risk or complex multivessel coronary artery disease who are not good surgical candidates.

FAME 2 trial: Fractional flow reserve as a guide

The Fractional Flow Reserve Versus Angiography for Multivessel Evaluation 2 (FAME 2) trial,⁸ published in 2012, evaluated whether clinical outcomes differ between patients who undergo percutaneous revascularization plus medical therapy and those who are treated with medical therapy alone, using fractional flow reserve as a means to determine which stenoses should be considered for intervention. Fractional flow reserve performed during invasive angiography determines the ratio of intracoronary pressure to aortic pressure using a wire advanced across a coronary obstruction.

FAME 2 found a markedly lower incidence of the primary composite end point of death, myocardial infarction, and urgent revascularization with randomization to percutaneous revascularization plus medical therapy compared with medical therapy only (4.3% vs 12.7%, P = .001) in patients with a fractional flow reserve less than 0.80 (considered a hemodynamically significant obstruction). The trial was stopped early because of the markedly different outcomes.

Of note, however, the reduction in adverse clinical outcomes was driven primarily by a reduction in urgent revascularizations in those treated with percutaneous coronary intervention in the revascularization arm. Regardless, using fractional flow reserve to guide whether obstructive coronary lesions should be treated with percutaneous coronary intervention has appropriately become a mainstay in interventional cardiology.

Stress testing

Noninvasive stress testing has played a role in helping to guide revascularization decisions in stable ischemic heart disease. In particular, revascularization in the setting of greater than 10% ischemia on perfusion imaging has been associated with a lower risk of cardiac death than in those who were revascularized with an ischemic burden less than 10%.⁹

A substudy of COURAGE found that percutaneous coronary intervention reduced ischemia to a greater degree than medical therapy alone on serial nuclear stress tests in patients with stable ischemic heart disease.¹⁰ In this substudy, when both groups were combined, the investigators also found that there were fewer adverse events in those who had an overall reduction of ischemia regardless of treatment strategy.

ISCHEMIA: Revascularize those with ischemia?

While COURAGE, BARI 2D, and FAME 2 suggested that early revascularization for low-risk patients with coronary artery disease does not confer a benefit over medical treatment alone with regard to hard clinical end points, it remains unclear whether an early revascularization strategy is advantageous in patients with stable ischemic heart disease who have at least a moderate amount of ischemia on noninvasive stress testing.

The ongoing ISCHEMIA (International Study of Comparative Effectiveness With Medical and Invasive Approaches) trial will help to answer that question. In this study, 8,000 patients with stable angina and at least moderate ischemia on noninvasive stress testing are being randomized before coronary angiography either to guideline-directed medical therapy plus revascularization (percutaneous or surgical) or to medical therapy alone.¹¹ The ISCHEMIA study population reflects current practice more closely than the previous studies discussed above in its inclusion of fractional flow reserve and later-generation drug-eluting stents.

The results of ISCHEMIA will be an important piece of the puzzle to answer whether patients with stable ischemic heart disease benefit from revascularization in terms of cardiovascular mortality or myocardial infarction (the primary end point of the study).

Studies in additional subsets

It is important to recognize that there are additional subsets of patients with stable ischemic heart disease (those with multivessel disease, left main coronary disease, or low ejection fractions, for example) who have been studied to help determine when and how to perform revascularization. In addition, there are guidelines¹² for both interventional cardiologists and cardiac surgeons that help delineate which patients should undergo revascularization. While a complete review is beyond the scope of this discussion, three trials are worth mentioning:

The Coronary Artery Surgery Study (CASS)¹³ revealed that revascularization in left main coronary artery disease is associated with lower mortality rates than medical therapy alone. This study, along with others, eventually led to recommendations for revascularization to be performed in all patients with significant left main coronary disease, regardless of symptoms or stress test findings.^14,15

The Surgical Treatment for Ischemic Heart Failure (STICH) trial¹⁶ found that patients with a low ejection fraction (< 35%) and ischemic heart disease had no difference in all-cause mortality rates when treated with CABG plus medical therapy compared with medical therapy alone (although the study’s design has been heavily criticized).

The Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) study¹⁷ found that CABG was associated with fewer adverse events in three-vessel coronary artery disease or complex left main coronary artery disease compared with percutaneous coronary intervention. The study used early-generation paclitaxel drug-eluting stents that are no longer used in contemporary practice. This study established the SYNTAX score, which is often used to help make revascularization decisions. A low SYNTAX score of 0 to 22 (meaning less-severe coronary artery disease) was associated with equivalent outcomes for both percutaneous coronary intervention and CABG. Thus, even if there is multivessel disease or left main disease, if the SYNTAX score is low, then percutaneous coronary intervention is an acceptable method for revascularization with similar results as for CABG.

A TEAM APPROACH

Due to the complexity of stable ischemic heart disease and the subtleties of managing these patients, a multidisciplinary “heart team” approach may be the best way to navigate treating stable ischemic heart disease via revascularization or with medical therapy alone. The heart team approach could take advantage of the particular expertise that the primary care physician, cardiologist, interventional cardiologist, and cardiac surgeon provide.

The upcoming results of studies such as the ISCHEMIA trial will help to provide additional guidance for these teams in long-term management of patients with stable ischemic heart disease.

The answer is less clear for these patients than for patients with acute coronary syndromes. In the latter group, percutaneous or surgical revascularization reduces the rates of morbidity and mortality, whereas in patients with stable ischemic heart disease, benefits may be limited to the improvement of angina. Certain markers and criteria may help us in this decision, and trials are ongoing.

Of importance, all patients with coronary artery disease should receive guideline-directed medical therapy as tolerated, regardless of whether they undergo revascularization.

MEDICAL THERAPY FOR ALL

In all the relevant trials, patients with stable ischemic heart disease in both the revascularization groups and the unrevascularized groups received guideline-directed medical therapy. Current guidelines¹ give class I recommendations (ie, treatment should be given) for:

• Lipid management
• Blood pressure management
• Physical activity
• Weight management
• Smoking cessation
• Antiplatelet therapy
• Beta-blockers for patients with normal left ventricular function after an acute coronary syndrome event, and for those with an ejection fraction of 40% or less
• Angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers for patients who have hypertension, diabetes mellitus, a left ventricular ejection fraction of 40% or less, or chronic kidney disease
• Annual influenza vaccination
• Anti-ischemic medications (beta-blockers, calcium channel blockers, nitrates) for relief of symptoms.

REVASCULARIZATION FOR SOME?

Results of the studies outlined below will help in deciding when to use guideline-directed medical therapy alone or medical therapy plus revascularization.

COURAGE trial: No added benefit in patients at low risk

The findings of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE), published in 2007, suggested that in select patients, percutaneous coronary intervention for stable coronary artery disease was no better than guideline-directed medical therapy alone for reducing the outcomes of death, myocardial infarction, or hospitalization for acute coronary syndrome.²

Of note, however, is that the 2,287 patients included in COURAGE were a low-risk subset of the more than 35,000 patients initially evaluated. The investigators reviewed the patients’ coronary angiograms before enrollment, and thus many patients with complex or high-risk anatomy were likely excluded based on an a priori assessment of angiographic images.

Also, coronary stent technology has substantially improved since COURAGE (which primarily used bare-metal stents and early drug-eluting stents), and this brings into question whether the results are applicable to current patients.

Moreover, in subsequent substudies from COURAGE, revascularization significantly improved symptoms of angina and quality-of-life scores compared with medical therapy alone.^3,4

Also important is that more than one-third of the patients in the medical therapy group crossed over to revascularization during the study, most often for worsening symptoms of angina.

Regardless of its limitations, COURAGE played an important role in delineating the use of guideline-directed medical therapy alone in certain low-risk patients and sparked debate about when and if to revascularize other patients.

BARI 2D trial: CABG may benefit those with diabetes

The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial, published in 2009, aimed to find out if revascularization in patients with stable ischemic heart disease and diabetes was beneficial compared with medical therapy alone.⁵

While it was not designed to directly compare percutaneous coronary intervention vs coronary artery bypass grafting (CABG), it did find that medical therapy plus CABG might reduce the rate of adverse cardiovascular events in this population compared with medical therapy alone or medical therapy plus percutaneous intervention.

As with COURAGE, however, the patients in the medical therapy group in BARI 2D also had a high rate of crossover to revascularization, primarily driven by worsening anginal symptoms.

FREEDOM and the 2014 updated guideline

Based on the findings of BARI 2D and those of FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease),⁶ the American College of Cardiology and American Heart Association updated their recommendations in 2014.⁷ This focused update states that for patients with diabetes and multivessel coronary artery disease, if revascularization is likely to improve survival (for example, in three-vessel disease or complex two-vessel disease involving the proximal left anterior descending artery), then CABG should be performed if a left internal mammary artery graft can be anastomosed to the left anterior descending artery. Otherwise, percutaneous coronary intervention should be reserved for those patients with diabetes and high-risk or complex multivessel coronary artery disease who are not good surgical candidates.

FAME 2 trial: Fractional flow reserve as a guide

The Fractional Flow Reserve Versus Angiography for Multivessel Evaluation 2 (FAME 2) trial,⁸ published in 2012, evaluated whether clinical outcomes differ between patients who undergo percutaneous revascularization plus medical therapy and those who are treated with medical therapy alone, using fractional flow reserve as a means to determine which stenoses should be considered for intervention. Fractional flow reserve performed during invasive angiography determines the ratio of intracoronary pressure to aortic pressure using a wire advanced across a coronary obstruction.

FAME 2 found a markedly lower incidence of the primary composite end point of death, myocardial infarction, and urgent revascularization with randomization to percutaneous revascularization plus medical therapy compared with medical therapy only (4.3% vs 12.7%, P = .001) in patients with a fractional flow reserve less than 0.80 (considered a hemodynamically significant obstruction). The trial was stopped early because of the markedly different outcomes.

Of note, however, the reduction in adverse clinical outcomes was driven primarily by a reduction in urgent revascularizations in those treated with percutaneous coronary intervention in the revascularization arm. Regardless, using fractional flow reserve to guide whether obstructive coronary lesions should be treated with percutaneous coronary intervention has appropriately become a mainstay in interventional cardiology.

Stress testing

Noninvasive stress testing has played a role in helping to guide revascularization decisions in stable ischemic heart disease. In particular, revascularization in the setting of greater than 10% ischemia on perfusion imaging has been associated with a lower risk of cardiac death than in those who were revascularized with an ischemic burden less than 10%.⁹

A substudy of COURAGE found that percutaneous coronary intervention reduced ischemia to a greater degree than medical therapy alone on serial nuclear stress tests in patients with stable ischemic heart disease.¹⁰ In this substudy, when both groups were combined, the investigators also found that there were fewer adverse events in those who had an overall reduction of ischemia regardless of treatment strategy.

ISCHEMIA: Revascularize those with ischemia?

While COURAGE, BARI 2D, and FAME 2 suggested that early revascularization for low-risk patients with coronary artery disease does not confer a benefit over medical treatment alone with regard to hard clinical end points, it remains unclear whether an early revascularization strategy is advantageous in patients with stable ischemic heart disease who have at least a moderate amount of ischemia on noninvasive stress testing.

The ongoing ISCHEMIA (International Study of Comparative Effectiveness With Medical and Invasive Approaches) trial will help to answer that question. In this study, 8,000 patients with stable angina and at least moderate ischemia on noninvasive stress testing are being randomized before coronary angiography either to guideline-directed medical therapy plus revascularization (percutaneous or surgical) or to medical therapy alone.¹¹ The ISCHEMIA study population reflects current practice more closely than the previous studies discussed above in its inclusion of fractional flow reserve and later-generation drug-eluting stents.

The results of ISCHEMIA will be an important piece of the puzzle to answer whether patients with stable ischemic heart disease benefit from revascularization in terms of cardiovascular mortality or myocardial infarction (the primary end point of the study).

Studies in additional subsets

It is important to recognize that there are additional subsets of patients with stable ischemic heart disease (those with multivessel disease, left main coronary disease, or low ejection fractions, for example) who have been studied to help determine when and how to perform revascularization. In addition, there are guidelines¹² for both interventional cardiologists and cardiac surgeons that help delineate which patients should undergo revascularization. While a complete review is beyond the scope of this discussion, three trials are worth mentioning:

The Coronary Artery Surgery Study (CASS)¹³ revealed that revascularization in left main coronary artery disease is associated with lower mortality rates than medical therapy alone. This study, along with others, eventually led to recommendations for revascularization to be performed in all patients with significant left main coronary disease, regardless of symptoms or stress test findings.^14,15

The Surgical Treatment for Ischemic Heart Failure (STICH) trial¹⁶ found that patients with a low ejection fraction (< 35%) and ischemic heart disease had no difference in all-cause mortality rates when treated with CABG plus medical therapy compared with medical therapy alone (although the study’s design has been heavily criticized).

The Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) study¹⁷ found that CABG was associated with fewer adverse events in three-vessel coronary artery disease or complex left main coronary artery disease compared with percutaneous coronary intervention. The study used early-generation paclitaxel drug-eluting stents that are no longer used in contemporary practice. This study established the SYNTAX score, which is often used to help make revascularization decisions. A low SYNTAX score of 0 to 22 (meaning less-severe coronary artery disease) was associated with equivalent outcomes for both percutaneous coronary intervention and CABG. Thus, even if there is multivessel disease or left main disease, if the SYNTAX score is low, then percutaneous coronary intervention is an acceptable method for revascularization with similar results as for CABG.

A TEAM APPROACH

Due to the complexity of stable ischemic heart disease and the subtleties of managing these patients, a multidisciplinary “heart team” approach may be the best way to navigate treating stable ischemic heart disease via revascularization or with medical therapy alone. The heart team approach could take advantage of the particular expertise that the primary care physician, cardiologist, interventional cardiologist, and cardiac surgeon provide.

The upcoming results of studies such as the ISCHEMIA trial will help to provide additional guidance for these teams in long-term management of patients with stable ischemic heart disease.

The duration of hormone therapy needs to be an individualized decision, shared between the patient and her physician and assessed annually. Quality of life, vasomotor symptoms, current age, time since menopause, hysterectomy status, personal risks (of osteoporosis, breast cancer, heart disease, stroke,  venous thromboembolism), and patient preferences need to be considered.

The North American Menopause Society (NAMS) and other organizations recommend that the lowest dose of hormone therapy be used for the shortest duration needed to manage menopausal symptoms.^1–4 However, NAMS states that extending the duration of hormone therapy may be appropriate in women who have persistent symptoms or to prevent osteoporosis if the patient cannot tolerate alternative therapies.¹

Forty-two percent of postmenopausal women continue to experience vasomotor symptoms at age 60 to 65.⁵ The median total duration of vasomotor symptoms is 7.4 years, and in black women and women with moderate or severe hot flashes the symptoms typically last 10 years.⁶ Vasomotor symptoms recur in 50% of women who discontinue hormone therapy, regardless of whether it is stopped abruptly or tapered.¹

FACTORS TO CONSIDER WHEN PRESCRIBING HORMONE THERAPY

Bone health

A statement issued in 2013 by seven medical societies said that hormone therapy is effective and appropriate for preventing osteoporosis-related fracture in at-risk women under age 60 or within 10 years of menopause.⁷

The Women’s Health Initiative,⁸ a randomized placebo-controlled trial, showed a statistically significant lower risk of vertebral and nonvertebral fracture after 3 years of use of conjugated equine estrogen with medroxyprogesterone acetate than with placebo:

• Hazard ratio 0.76, 95% confidence interval (CI) 0.69–0.83.

It also showed a mean increase of 3.7% (P < .001) in total hip bone mineral density. By the end of the trial intervention, women receiving either this combined therapy or conjugated equine estrogen alone saw a 33% overall reduction in hip fracture risk. The absolute risk reduction was 5 per 10,000 years of use.⁹

Karim et al,¹⁰ in a large observational study that followed initial hormone therapy users over 6.5 years, found that those who stopped it had a 55% greater risk of hip fracture and experienced significant bone loss as measured by bone mineral density compared with women who continued hormone therapy, and that the protective effects of hormone therapy disappeared as early as 2 years after stopping treatment.¹⁰

NAMS also recommends that women with premature menopause (before age 40) be offered and encouraged to use hormone therapy to preserve bone density and manage vasomotor symptoms until the age of natural menopause (age 51).^1,11

Cardiovascular health

Large observational studies have found that hormone therapy is associated with a 30% to 50% lower cardiovascular risk.¹² Randomized controlled trials of hormone therapy for 7 to 11 years suggest that coronary heart disease risk is modified by age and time since menopause.^13,14

The Women’s Health Initiative and other randomized controlled trials suggest a lower risk of coronary heart disease in women who begin hormone therapy before age 60 and within 10 years of the onset of menopause, but an increased risk for women over age 60 and more than 10 years since menopause. However, several of these trends have not reached statistical significance (Table 1).^13–15

The Women’s Health Initiative⁹ published its long-term follow-up results in 2013, with data on both the intervention phase (median of 7.2 years for estrogen-only therapy and 5.6 years for estrogen-progestin therapy) and the post-stopping phase (median 6.6 years for the estrogen-only group and 8.2 years for the estrogen-progestin group), with a total cumulative follow-up of 13 years. The overall 13-year cumulative absolute risk of coronary heart disease was 4 fewer events per 10,000 years of estrogen-only therapy and 3 additional events per 10,000 years of estrogen-progestin therapy. Neither result was statistically significant:

• Hazard ratio with estrogen-only use 0.94, 95% CI 0.82–1.09
• Hazard ratio with estrogen-progestin use 1.09, 95% CI 0.92–1.24.

The Danish Osteoporosis Study was the first randomized controlled trial of hormone therapy in women ages 45 through 58 who were recently menopausal (average within 7 months of menopause).¹⁵ Women assigned to hormone therapy in the form of oral estradiol with or without norethisterone (known as norethindrone in the United States) had a statistically significant lower risk of the primary composite end point of heart failure and myocardial infarction after 11 years of hormone therapy, and this finding persisted through 16 years of follow-up (Table 1).

Stroke

Overall stroke risk was significantly increased with hormone therapy in the Women’s Health Initiative trial (hazard ratio 1.32, 95% CI 1.12–1.56); however, the absolute increase in risk was small in both estrogen-alone and estrogen-progestin therapy users, 11 and 8 events, respectively, among 10,000 users. Younger women (ages 50–59) saw a nonsignificantly lower risk (2 fewer cases per 10,000 years of use).¹⁴ After 13 years of cumulative follow-up (combined intervention and follow-up phase), the risk of stroke persisted at 5 cases per 10,000 users for both arms, but only the estrogen-progestin results were statistically significant.⁹

The Danish Osteoporosis Study¹⁵ found no increased risk of stroke after 16 years of follow-up in recently menopausal women:

• Hazard ratio 0.89, 95% CI 0.48–1.65.

Venous thromboembolism

Data from both observational and randomized controlled trials demonstrate an increased risk of venous thromboembolism with oral hormone therapy, and the risk appears to be highest during the first few years of use.¹ The pooled cohort from the Women’s Health Initiative had 18 additional cases of venous thromboembolism per 10,000 women in estrogen-progestin users compared with nonusers, and 7 additional cases in those using estrogen-only therapy.

Breast health

Observational studies and randomized controlled trials have provided data on longer use of hormone therapy and breast cancer risk, but the true magnitude of this risk is unclear.

The Danish Osteoporosis Study,¹⁵ in a younger cohort of women, showed no increased risk of breast cancer after 16 years of follow-up:

• Hazard ratio 0.90, 95% CI 0.52–1.57.

The Women’s Health Initiative⁹ showed a statistically nonsignificant lower risk of breast cancer in women of all ages exposed to conjugated equine estrogen alone for 7.1 years (6 fewer cases per 10,000 women-years of use), and after 6 years of follow-up this developed statistical significance:

• Hazard ratio 0.79, 95% CI 0.65–0.97.

In contrast, those using conjugated equine estrogen plus medroxyprogesterone acetate had a statistically nonsignificant increase in the risk of new breast cancer after 3 to 5 years:

• 3-year relative risk 1.26, 95% CI 0.73–2.20
• 5-year relative risk 1.99, 95% CI 1.18–3.35
• Absolute risk 8 cases per 10,000 women-years of use.

The increased risk of breast cancer significantly declined within 3 years after stopping hormone therapy.

However, even after stopping hormone therapy, there remains a statistically small but significant increased risk of breast cancer, as demonstrated in the postintervention 13-year follow-up data on breast cancer risk and estrogen-progestin use from the Women’s Health Initiative⁹:

• Hazard ratio 1.28, 95% CI 1.11–1.48
• Absolute cumulative risk 9 cases per 10,000 women-years of use.

The Nurses’ Health Study, an observational study, prospectively followed 11,508 hysterectomized women on estrogen therapy and found that breast cancer risk increased with longer duration of use. An analysis by Chen et al¹⁶ found a trend toward increased breast cancer risk after 10 years of estrogen therapy, but this did not become statistically significant until 20 years of ongoing estrogen use. The risk of estrogen receptor-positive and progesterone receptor-positive breast cancer became statistically significant earlier, after 15 years. The relative risk associated with using estrogen for more than 15 years was 1.18, and the risk with using it for more than 20 years was 1.42.¹⁶

To put this in perspective, Chen et al¹⁷ found a similar breast cancer risk with alcohol consumption. The relative risk of invasive breast cancer was 1.15 in women who drank 3 to 6 servings of alcohol per week, 1 serving being equivalent to 4 oz of wine, which contains 11 g of alcohol.

Mortality

Studies have suggested that hormone therapy users have a lower mortality rate, even with long-term use.

A meta-analysis¹⁸ of 8 observational trials and 19 randomized controlled trials found that younger women (average age 54) on hormone therapy had a 28% lower total mortality rate compared with women not taking hormone therapy:

• Relative risk 0.72, 95% credible interval 0.62–0.82.

The Women’s Health Initiative¹⁹ suggested that the mortality rate was 30% lower in hormone therapy users younger than age 60 than in similar nonusers, though this difference did not reach statistical significance.

• Relative risk with estrogen-only therapy: 0.71, 95% CI 0.46–1.11
• Relative risk with combined estrogen-progestin therapy 0.69, 95% CI 0.44–1.07.

The Danish Osteoporosis Study,¹⁵ at 16 years of follow-up, similarly demonstrated a 34% lower mortality rate in hormone therapy users, which was not statistically significant:

• Relative risk 0.66, 95% CI 0.41–1.08.

A Cochrane review²⁰ in 2015 found that the subgroup of women who started hormone therapy before age 60 or within 10 years of menopause saw an overall benefit in terms of survival and lower risk of coronary heart disease: RR 0.70, 95% CI 0.52–0.95 (moderate-quality evidence).

TYPE OF FORMULATION

Compared with estrogen-progestin therapy, estrogen-only therapy has a more favorable risk profile in terms of coronary heart disease and breast cancer, although stroke risk remains elevated in users of conjugated equine estrogen with or without medroxyprogesterone acetate.

There is limited evidence directly comparing different formulations of hormone therapy, although they all effectively treat vasomotor symptoms.¹

Oral vs transdermal formulations

Canonico et al,²¹ in a meta-analysis of observational studies, found that oral estrogen was associated with a higher risk of venous thromboembolism than transdermal estrogen:

• Relative risk with oral estrogen 2.5, 95% CI 1.9–3.4
• Relative risk with transdermal estrogen 1.2, 95% CI 0.9–1.7.

The Estrogen and Thromboembolism Risk (ESTHER) study²² was a multicenter case-control study of women ages 45 to 70 that assessed risk of venous thromboembolism in oral vs transdermal estrogen users. Compared with women not taking hormone therapy, current users of oral estrogen had a significantly higher risk of venous thromboembolism, while transdermal estrogen users did not:

• Odds ratio with oral estrogen 4.2, 95% CI 1.5–11.6
• Odds ratio with transdermal estrogen 0.9, 95% CI 0.4–2.1.

The Kronos Early Estrogen Prevention Study (KEEPS)²³ did not support these findings. This 4-year randomized controlled trial, published in 2014, was designed to assess the risk of atherosclerosis progression with early menopause initiation of placebo vs low-dose oral hormone therapy (conjugated equine estrogen 0.45 mg daily with cyclical micronized progesterone) or transdermal hormone therapy (estradiol 50 µg/week with cyclical micronized progesterone).

In the 727 women in the study, there was one transient ischemic attack in the oral hormone therapy group, one unconfirmed stroke in the transdermal hormone therapy group, and one case of venous thromboembolism in each group, findings that were underpowered for statistical significance. Both oral and transdermal hormonal therapy had neutral effects on atherosclerosis progression, as assessed by arterial imaging. Transdermal hormone therapy was associated with improvements in markers of insulin resistance and was not associated with an increase in triglycerides, C-reactive protein, or sex hormone-binding globulin, as would be expected with transdermal circumvention of the first-pass hepatic effect.

BALANCING THE RISKS AND BENEFITS FOR THE PATIENT

The most effective treatment for vasomotor symptoms in women at any age is hormone therapy, and the benefits are more likely to outweigh risks when initiated before age 60 or within 10 years of menopause.⁷ The Women’s Health Initiative randomized study was limited to 5.6 to 7.2 years of hormone therapy (13 years of cumulative follow-up), and the Danish Osteoporosis Study was limited to 11 years of use (16 years cumulative follow-up).

The coronary heart disease outcomes for longer durations of therapy remain uncertain. There is a small but statistically significant increased risk of stroke and venous thromboembolism with oral hormone therapy, and breast cancer risk is associated with long-term estrogen-progestin use.

Patients on hormone therapy should be evaluated annually regarding the need for ongoing therapy. Persistent moderate-severe vasomotor symptoms, quality of life benefits of hormone therapy, contraindications to its use (Table 2), and patient preference need to be assessed as well as baseline risks of cardiovascular disease, breast cancer, and fracture.

Risk calculators may facilitate the shared decision-making process. Examples are:

• The American College of Cardiology/American Heart association risk calculator for cardiovascular disease²⁴ (www.cvriskcalculator.com)
• The World Health Organization Fracture Risk Assessment Tool (FRAX)²⁵
  (www.shef.ac.uk/FRAX/tool.jsp)

• The Gail model for breast cancer risk²⁶ (www.cancer.gov/bcrisktool/).
• MenoPro, a menopause decision-support algorithm and companion mobile app developed by NAMS to help direct treatment decisions based on the 10-year risk of atherosclerotic cardiovascular disease (www.menopause.org/for-professionals/-i-menopro-i-mobile-app).²⁷
  The discussion of the risks of hormone therapy with patients should incorporate the perspective of absolute risk. For example, a woman wishing to continue estrogen-progestin therapy should be told that the Women’s Health Initiative data suggest that, after 5 years of use, breast cancer risk may be increased by 8 additional cases per 10,000 users per year. According to the World Health Organization, this magnitude of risk is defined as rare (less than 1 event per 1,000 women).²⁸

A strategy of prescribing the lowest dose to achieve the desired clinical benefits is prudent and recommended.^1–3 Table 3 outlines the estrogen formulations now available in the United States, with their doses and formulations.

Unless contraindications develop (Table 2), patients may elect to continue hormone therapy if its benefits outweigh its risks. The American College of Obstetricians and Gynecologists (ACOG) 2014 practice recommendations for management of menopausal symptoms³¹ and the 2015 NAMS statement both recommend that hormone therapy not be discontinued based solely on a woman’s age.²⁹

Hormone therapy is on the Beer’s list of potentially inappropriate medications for older adults,³⁰ which remains a hurdle to its long-term use and seems to be at odds with these ACOG and NAMS statements.

Patients who choose to discontinue hormone therapy need to be monitored for persistent bothersome vasomotor symptoms, bone loss, osteoporosis, and the genitourinary syndrome of menopause (previously referred to as vulvovaginal atrophy)³¹ and offered alternative therapies if needed.

The duration of hormone therapy needs to be an individualized decision, shared between the patient and her physician and assessed annually. Quality of life, vasomotor symptoms, current age, time since menopause, hysterectomy status, personal risks (of osteoporosis, breast cancer, heart disease, stroke,  venous thromboembolism), and patient preferences need to be considered.

The North American Menopause Society (NAMS) and other organizations recommend that the lowest dose of hormone therapy be used for the shortest duration needed to manage menopausal symptoms.^1–4 However, NAMS states that extending the duration of hormone therapy may be appropriate in women who have persistent symptoms or to prevent osteoporosis if the patient cannot tolerate alternative therapies.¹

Forty-two percent of postmenopausal women continue to experience vasomotor symptoms at age 60 to 65.⁵ The median total duration of vasomotor symptoms is 7.4 years, and in black women and women with moderate or severe hot flashes the symptoms typically last 10 years.⁶ Vasomotor symptoms recur in 50% of women who discontinue hormone therapy, regardless of whether it is stopped abruptly or tapered.¹

FACTORS TO CONSIDER WHEN PRESCRIBING HORMONE THERAPY

Bone health

A statement issued in 2013 by seven medical societies said that hormone therapy is effective and appropriate for preventing osteoporosis-related fracture in at-risk women under age 60 or within 10 years of menopause.⁷

The Women’s Health Initiative,⁸ a randomized placebo-controlled trial, showed a statistically significant lower risk of vertebral and nonvertebral fracture after 3 years of use of conjugated equine estrogen with medroxyprogesterone acetate than with placebo:

• Hazard ratio 0.76, 95% confidence interval (CI) 0.69–0.83.

It also showed a mean increase of 3.7% (P < .001) in total hip bone mineral density. By the end of the trial intervention, women receiving either this combined therapy or conjugated equine estrogen alone saw a 33% overall reduction in hip fracture risk. The absolute risk reduction was 5 per 10,000 years of use.⁹

Karim et al,¹⁰ in a large observational study that followed initial hormone therapy users over 6.5 years, found that those who stopped it had a 55% greater risk of hip fracture and experienced significant bone loss as measured by bone mineral density compared with women who continued hormone therapy, and that the protective effects of hormone therapy disappeared as early as 2 years after stopping treatment.¹⁰

NAMS also recommends that women with premature menopause (before age 40) be offered and encouraged to use hormone therapy to preserve bone density and manage vasomotor symptoms until the age of natural menopause (age 51).^1,11

Cardiovascular health

Large observational studies have found that hormone therapy is associated with a 30% to 50% lower cardiovascular risk.¹² Randomized controlled trials of hormone therapy for 7 to 11 years suggest that coronary heart disease risk is modified by age and time since menopause.^13,14

The Women’s Health Initiative and other randomized controlled trials suggest a lower risk of coronary heart disease in women who begin hormone therapy before age 60 and within 10 years of the onset of menopause, but an increased risk for women over age 60 and more than 10 years since menopause. However, several of these trends have not reached statistical significance (Table 1).^13–15

The Women’s Health Initiative⁹ published its long-term follow-up results in 2013, with data on both the intervention phase (median of 7.2 years for estrogen-only therapy and 5.6 years for estrogen-progestin therapy) and the post-stopping phase (median 6.6 years for the estrogen-only group and 8.2 years for the estrogen-progestin group), with a total cumulative follow-up of 13 years. The overall 13-year cumulative absolute risk of coronary heart disease was 4 fewer events per 10,000 years of estrogen-only therapy and 3 additional events per 10,000 years of estrogen-progestin therapy. Neither result was statistically significant:

• Hazard ratio with estrogen-only use 0.94, 95% CI 0.82–1.09
• Hazard ratio with estrogen-progestin use 1.09, 95% CI 0.92–1.24.

The Danish Osteoporosis Study was the first randomized controlled trial of hormone therapy in women ages 45 through 58 who were recently menopausal (average within 7 months of menopause).¹⁵ Women assigned to hormone therapy in the form of oral estradiol with or without norethisterone (known as norethindrone in the United States) had a statistically significant lower risk of the primary composite end point of heart failure and myocardial infarction after 11 years of hormone therapy, and this finding persisted through 16 years of follow-up (Table 1).

Stroke

Overall stroke risk was significantly increased with hormone therapy in the Women’s Health Initiative trial (hazard ratio 1.32, 95% CI 1.12–1.56); however, the absolute increase in risk was small in both estrogen-alone and estrogen-progestin therapy users, 11 and 8 events, respectively, among 10,000 users. Younger women (ages 50–59) saw a nonsignificantly lower risk (2 fewer cases per 10,000 years of use).¹⁴ After 13 years of cumulative follow-up (combined intervention and follow-up phase), the risk of stroke persisted at 5 cases per 10,000 users for both arms, but only the estrogen-progestin results were statistically significant.⁹

The Danish Osteoporosis Study¹⁵ found no increased risk of stroke after 16 years of follow-up in recently menopausal women:

• Hazard ratio 0.89, 95% CI 0.48–1.65.

Venous thromboembolism

Data from both observational and randomized controlled trials demonstrate an increased risk of venous thromboembolism with oral hormone therapy, and the risk appears to be highest during the first few years of use.¹ The pooled cohort from the Women’s Health Initiative had 18 additional cases of venous thromboembolism per 10,000 women in estrogen-progestin users compared with nonusers, and 7 additional cases in those using estrogen-only therapy.

Breast health

Observational studies and randomized controlled trials have provided data on longer use of hormone therapy and breast cancer risk, but the true magnitude of this risk is unclear.

The Danish Osteoporosis Study,¹⁵ in a younger cohort of women, showed no increased risk of breast cancer after 16 years of follow-up:

• Hazard ratio 0.90, 95% CI 0.52–1.57.

The Women’s Health Initiative⁹ showed a statistically nonsignificant lower risk of breast cancer in women of all ages exposed to conjugated equine estrogen alone for 7.1 years (6 fewer cases per 10,000 women-years of use), and after 6 years of follow-up this developed statistical significance:

• Hazard ratio 0.79, 95% CI 0.65–0.97.

In contrast, those using conjugated equine estrogen plus medroxyprogesterone acetate had a statistically nonsignificant increase in the risk of new breast cancer after 3 to 5 years:

• 3-year relative risk 1.26, 95% CI 0.73–2.20
• 5-year relative risk 1.99, 95% CI 1.18–3.35
• Absolute risk 8 cases per 10,000 women-years of use.

The increased risk of breast cancer significantly declined within 3 years after stopping hormone therapy.

However, even after stopping hormone therapy, there remains a statistically small but significant increased risk of breast cancer, as demonstrated in the postintervention 13-year follow-up data on breast cancer risk and estrogen-progestin use from the Women’s Health Initiative⁹:

• Hazard ratio 1.28, 95% CI 1.11–1.48
• Absolute cumulative risk 9 cases per 10,000 women-years of use.

The Nurses’ Health Study, an observational study, prospectively followed 11,508 hysterectomized women on estrogen therapy and found that breast cancer risk increased with longer duration of use. An analysis by Chen et al¹⁶ found a trend toward increased breast cancer risk after 10 years of estrogen therapy, but this did not become statistically significant until 20 years of ongoing estrogen use. The risk of estrogen receptor-positive and progesterone receptor-positive breast cancer became statistically significant earlier, after 15 years. The relative risk associated with using estrogen for more than 15 years was 1.18, and the risk with using it for more than 20 years was 1.42.¹⁶

To put this in perspective, Chen et al¹⁷ found a similar breast cancer risk with alcohol consumption. The relative risk of invasive breast cancer was 1.15 in women who drank 3 to 6 servings of alcohol per week, 1 serving being equivalent to 4 oz of wine, which contains 11 g of alcohol.

Mortality

Studies have suggested that hormone therapy users have a lower mortality rate, even with long-term use.

A meta-analysis¹⁸ of 8 observational trials and 19 randomized controlled trials found that younger women (average age 54) on hormone therapy had a 28% lower total mortality rate compared with women not taking hormone therapy:

• Relative risk 0.72, 95% credible interval 0.62–0.82.

The Women’s Health Initiative¹⁹ suggested that the mortality rate was 30% lower in hormone therapy users younger than age 60 than in similar nonusers, though this difference did not reach statistical significance.

• Relative risk with estrogen-only therapy: 0.71, 95% CI 0.46–1.11
• Relative risk with combined estrogen-progestin therapy 0.69, 95% CI 0.44–1.07.

The Danish Osteoporosis Study,¹⁵ at 16 years of follow-up, similarly demonstrated a 34% lower mortality rate in hormone therapy users, which was not statistically significant:

• Relative risk 0.66, 95% CI 0.41–1.08.

A Cochrane review²⁰ in 2015 found that the subgroup of women who started hormone therapy before age 60 or within 10 years of menopause saw an overall benefit in terms of survival and lower risk of coronary heart disease: RR 0.70, 95% CI 0.52–0.95 (moderate-quality evidence).

TYPE OF FORMULATION

Compared with estrogen-progestin therapy, estrogen-only therapy has a more favorable risk profile in terms of coronary heart disease and breast cancer, although stroke risk remains elevated in users of conjugated equine estrogen with or without medroxyprogesterone acetate.

There is limited evidence directly comparing different formulations of hormone therapy, although they all effectively treat vasomotor symptoms.¹

Oral vs transdermal formulations

Canonico et al,²¹ in a meta-analysis of observational studies, found that oral estrogen was associated with a higher risk of venous thromboembolism than transdermal estrogen:

• Relative risk with oral estrogen 2.5, 95% CI 1.9–3.4
• Relative risk with transdermal estrogen 1.2, 95% CI 0.9–1.7.

The Estrogen and Thromboembolism Risk (ESTHER) study²² was a multicenter case-control study of women ages 45 to 70 that assessed risk of venous thromboembolism in oral vs transdermal estrogen users. Compared with women not taking hormone therapy, current users of oral estrogen had a significantly higher risk of venous thromboembolism, while transdermal estrogen users did not:

• Odds ratio with oral estrogen 4.2, 95% CI 1.5–11.6
• Odds ratio with transdermal estrogen 0.9, 95% CI 0.4–2.1.

The Kronos Early Estrogen Prevention Study (KEEPS)²³ did not support these findings. This 4-year randomized controlled trial, published in 2014, was designed to assess the risk of atherosclerosis progression with early menopause initiation of placebo vs low-dose oral hormone therapy (conjugated equine estrogen 0.45 mg daily with cyclical micronized progesterone) or transdermal hormone therapy (estradiol 50 µg/week with cyclical micronized progesterone).

In the 727 women in the study, there was one transient ischemic attack in the oral hormone therapy group, one unconfirmed stroke in the transdermal hormone therapy group, and one case of venous thromboembolism in each group, findings that were underpowered for statistical significance. Both oral and transdermal hormonal therapy had neutral effects on atherosclerosis progression, as assessed by arterial imaging. Transdermal hormone therapy was associated with improvements in markers of insulin resistance and was not associated with an increase in triglycerides, C-reactive protein, or sex hormone-binding globulin, as would be expected with transdermal circumvention of the first-pass hepatic effect.

BALANCING THE RISKS AND BENEFITS FOR THE PATIENT

The most effective treatment for vasomotor symptoms in women at any age is hormone therapy, and the benefits are more likely to outweigh risks when initiated before age 60 or within 10 years of menopause.⁷ The Women’s Health Initiative randomized study was limited to 5.6 to 7.2 years of hormone therapy (13 years of cumulative follow-up), and the Danish Osteoporosis Study was limited to 11 years of use (16 years cumulative follow-up).

The coronary heart disease outcomes for longer durations of therapy remain uncertain. There is a small but statistically significant increased risk of stroke and venous thromboembolism with oral hormone therapy, and breast cancer risk is associated with long-term estrogen-progestin use.

Patients on hormone therapy should be evaluated annually regarding the need for ongoing therapy. Persistent moderate-severe vasomotor symptoms, quality of life benefits of hormone therapy, contraindications to its use (Table 2), and patient preference need to be assessed as well as baseline risks of cardiovascular disease, breast cancer, and fracture.

Risk calculators may facilitate the shared decision-making process. Examples are:

• The American College of Cardiology/American Heart association risk calculator for cardiovascular disease²⁴ (www.cvriskcalculator.com)
• The World Health Organization Fracture Risk Assessment Tool (FRAX)²⁵
  (www.shef.ac.uk/FRAX/tool.jsp)

• The Gail model for breast cancer risk²⁶ (www.cancer.gov/bcrisktool/).
• MenoPro, a menopause decision-support algorithm and companion mobile app developed by NAMS to help direct treatment decisions based on the 10-year risk of atherosclerotic cardiovascular disease (www.menopause.org/for-professionals/-i-menopro-i-mobile-app).²⁷
  The discussion of the risks of hormone therapy with patients should incorporate the perspective of absolute risk. For example, a woman wishing to continue estrogen-progestin therapy should be told that the Women’s Health Initiative data suggest that, after 5 years of use, breast cancer risk may be increased by 8 additional cases per 10,000 users per year. According to the World Health Organization, this magnitude of risk is defined as rare (less than 1 event per 1,000 women).²⁸

A strategy of prescribing the lowest dose to achieve the desired clinical benefits is prudent and recommended.^1–3 Table 3 outlines the estrogen formulations now available in the United States, with their doses and formulations.

Unless contraindications develop (Table 2), patients may elect to continue hormone therapy if its benefits outweigh its risks. The American College of Obstetricians and Gynecologists (ACOG) 2014 practice recommendations for management of menopausal symptoms³¹ and the 2015 NAMS statement both recommend that hormone therapy not be discontinued based solely on a woman’s age.²⁹

Hormone therapy is on the Beer’s list of potentially inappropriate medications for older adults,³⁰ which remains a hurdle to its long-term use and seems to be at odds with these ACOG and NAMS statements.

Patients who choose to discontinue hormone therapy need to be monitored for persistent bothersome vasomotor symptoms, bone loss, osteoporosis, and the genitourinary syndrome of menopause (previously referred to as vulvovaginal atrophy)³¹ and offered alternative therapies if needed.

The duration of hormone therapy needs to be an individualized decision, shared between the patient and her physician and assessed annually. Quality of life, vasomotor symptoms, current age, time since menopause, hysterectomy status, personal risks (of osteoporosis, breast cancer, heart disease, stroke,  venous thromboembolism), and patient preferences need to be considered.

The North American Menopause Society (NAMS) and other organizations recommend that the lowest dose of hormone therapy be used for the shortest duration needed to manage menopausal symptoms.^1–4 However, NAMS states that extending the duration of hormone therapy may be appropriate in women who have persistent symptoms or to prevent osteoporosis if the patient cannot tolerate alternative therapies.¹

Forty-two percent of postmenopausal women continue to experience vasomotor symptoms at age 60 to 65.⁵ The median total duration of vasomotor symptoms is 7.4 years, and in black women and women with moderate or severe hot flashes the symptoms typically last 10 years.⁶ Vasomotor symptoms recur in 50% of women who discontinue hormone therapy, regardless of whether it is stopped abruptly or tapered.¹

FACTORS TO CONSIDER WHEN PRESCRIBING HORMONE THERAPY

Bone health

A statement issued in 2013 by seven medical societies said that hormone therapy is effective and appropriate for preventing osteoporosis-related fracture in at-risk women under age 60 or within 10 years of menopause.⁷

The Women’s Health Initiative,⁸ a randomized placebo-controlled trial, showed a statistically significant lower risk of vertebral and nonvertebral fracture after 3 years of use of conjugated equine estrogen with medroxyprogesterone acetate than with placebo:

• Hazard ratio 0.76, 95% confidence interval (CI) 0.69–0.83.

It also showed a mean increase of 3.7% (P < .001) in total hip bone mineral density. By the end of the trial intervention, women receiving either this combined therapy or conjugated equine estrogen alone saw a 33% overall reduction in hip fracture risk. The absolute risk reduction was 5 per 10,000 years of use.⁹

Karim et al,¹⁰ in a large observational study that followed initial hormone therapy users over 6.5 years, found that those who stopped it had a 55% greater risk of hip fracture and experienced significant bone loss as measured by bone mineral density compared with women who continued hormone therapy, and that the protective effects of hormone therapy disappeared as early as 2 years after stopping treatment.¹⁰

NAMS also recommends that women with premature menopause (before age 40) be offered and encouraged to use hormone therapy to preserve bone density and manage vasomotor symptoms until the age of natural menopause (age 51).^1,11

Cardiovascular health

Large observational studies have found that hormone therapy is associated with a 30% to 50% lower cardiovascular risk.¹² Randomized controlled trials of hormone therapy for 7 to 11 years suggest that coronary heart disease risk is modified by age and time since menopause.^13,14

The Women’s Health Initiative and other randomized controlled trials suggest a lower risk of coronary heart disease in women who begin hormone therapy before age 60 and within 10 years of the onset of menopause, but an increased risk for women over age 60 and more than 10 years since menopause. However, several of these trends have not reached statistical significance (Table 1).^13–15

The Women’s Health Initiative⁹ published its long-term follow-up results in 2013, with data on both the intervention phase (median of 7.2 years for estrogen-only therapy and 5.6 years for estrogen-progestin therapy) and the post-stopping phase (median 6.6 years for the estrogen-only group and 8.2 years for the estrogen-progestin group), with a total cumulative follow-up of 13 years. The overall 13-year cumulative absolute risk of coronary heart disease was 4 fewer events per 10,000 years of estrogen-only therapy and 3 additional events per 10,000 years of estrogen-progestin therapy. Neither result was statistically significant:

• Hazard ratio with estrogen-only use 0.94, 95% CI 0.82–1.09
• Hazard ratio with estrogen-progestin use 1.09, 95% CI 0.92–1.24.

The Danish Osteoporosis Study was the first randomized controlled trial of hormone therapy in women ages 45 through 58 who were recently menopausal (average within 7 months of menopause).¹⁵ Women assigned to hormone therapy in the form of oral estradiol with or without norethisterone (known as norethindrone in the United States) had a statistically significant lower risk of the primary composite end point of heart failure and myocardial infarction after 11 years of hormone therapy, and this finding persisted through 16 years of follow-up (Table 1).

Stroke

Overall stroke risk was significantly increased with hormone therapy in the Women’s Health Initiative trial (hazard ratio 1.32, 95% CI 1.12–1.56); however, the absolute increase in risk was small in both estrogen-alone and estrogen-progestin therapy users, 11 and 8 events, respectively, among 10,000 users. Younger women (ages 50–59) saw a nonsignificantly lower risk (2 fewer cases per 10,000 years of use).¹⁴ After 13 years of cumulative follow-up (combined intervention and follow-up phase), the risk of stroke persisted at 5 cases per 10,000 users for both arms, but only the estrogen-progestin results were statistically significant.⁹

The Danish Osteoporosis Study¹⁵ found no increased risk of stroke after 16 years of follow-up in recently menopausal women:

• Hazard ratio 0.89, 95% CI 0.48–1.65.

Venous thromboembolism

Data from both observational and randomized controlled trials demonstrate an increased risk of venous thromboembolism with oral hormone therapy, and the risk appears to be highest during the first few years of use.¹ The pooled cohort from the Women’s Health Initiative had 18 additional cases of venous thromboembolism per 10,000 women in estrogen-progestin users compared with nonusers, and 7 additional cases in those using estrogen-only therapy.

Breast health

Observational studies and randomized controlled trials have provided data on longer use of hormone therapy and breast cancer risk, but the true magnitude of this risk is unclear.

The Danish Osteoporosis Study,¹⁵ in a younger cohort of women, showed no increased risk of breast cancer after 16 years of follow-up:

• Hazard ratio 0.90, 95% CI 0.52–1.57.

The Women’s Health Initiative⁹ showed a statistically nonsignificant lower risk of breast cancer in women of all ages exposed to conjugated equine estrogen alone for 7.1 years (6 fewer cases per 10,000 women-years of use), and after 6 years of follow-up this developed statistical significance:

• Hazard ratio 0.79, 95% CI 0.65–0.97.

In contrast, those using conjugated equine estrogen plus medroxyprogesterone acetate had a statistically nonsignificant increase in the risk of new breast cancer after 3 to 5 years:

• 3-year relative risk 1.26, 95% CI 0.73–2.20
• 5-year relative risk 1.99, 95% CI 1.18–3.35
• Absolute risk 8 cases per 10,000 women-years of use.

The increased risk of breast cancer significantly declined within 3 years after stopping hormone therapy.

However, even after stopping hormone therapy, there remains a statistically small but significant increased risk of breast cancer, as demonstrated in the postintervention 13-year follow-up data on breast cancer risk and estrogen-progestin use from the Women’s Health Initiative⁹:

• Hazard ratio 1.28, 95% CI 1.11–1.48
• Absolute cumulative risk 9 cases per 10,000 women-years of use.

The Nurses’ Health Study, an observational study, prospectively followed 11,508 hysterectomized women on estrogen therapy and found that breast cancer risk increased with longer duration of use. An analysis by Chen et al¹⁶ found a trend toward increased breast cancer risk after 10 years of estrogen therapy, but this did not become statistically significant until 20 years of ongoing estrogen use. The risk of estrogen receptor-positive and progesterone receptor-positive breast cancer became statistically significant earlier, after 15 years. The relative risk associated with using estrogen for more than 15 years was 1.18, and the risk with using it for more than 20 years was 1.42.¹⁶

To put this in perspective, Chen et al¹⁷ found a similar breast cancer risk with alcohol consumption. The relative risk of invasive breast cancer was 1.15 in women who drank 3 to 6 servings of alcohol per week, 1 serving being equivalent to 4 oz of wine, which contains 11 g of alcohol.

Mortality

Studies have suggested that hormone therapy users have a lower mortality rate, even with long-term use.

A meta-analysis¹⁸ of 8 observational trials and 19 randomized controlled trials found that younger women (average age 54) on hormone therapy had a 28% lower total mortality rate compared with women not taking hormone therapy:

• Relative risk 0.72, 95% credible interval 0.62–0.82.

The Women’s Health Initiative¹⁹ suggested that the mortality rate was 30% lower in hormone therapy users younger than age 60 than in similar nonusers, though this difference did not reach statistical significance.

• Relative risk with estrogen-only therapy: 0.71, 95% CI 0.46–1.11
• Relative risk with combined estrogen-progestin therapy 0.69, 95% CI 0.44–1.07.

The Danish Osteoporosis Study,¹⁵ at 16 years of follow-up, similarly demonstrated a 34% lower mortality rate in hormone therapy users, which was not statistically significant:

• Relative risk 0.66, 95% CI 0.41–1.08.

A Cochrane review²⁰ in 2015 found that the subgroup of women who started hormone therapy before age 60 or within 10 years of menopause saw an overall benefit in terms of survival and lower risk of coronary heart disease: RR 0.70, 95% CI 0.52–0.95 (moderate-quality evidence).

TYPE OF FORMULATION

Compared with estrogen-progestin therapy, estrogen-only therapy has a more favorable risk profile in terms of coronary heart disease and breast cancer, although stroke risk remains elevated in users of conjugated equine estrogen with or without medroxyprogesterone acetate.

There is limited evidence directly comparing different formulations of hormone therapy, although they all effectively treat vasomotor symptoms.¹

Oral vs transdermal formulations

Canonico et al,²¹ in a meta-analysis of observational studies, found that oral estrogen was associated with a higher risk of venous thromboembolism than transdermal estrogen:

• Relative risk with oral estrogen 2.5, 95% CI 1.9–3.4
• Relative risk with transdermal estrogen 1.2, 95% CI 0.9–1.7.

The Estrogen and Thromboembolism Risk (ESTHER) study²² was a multicenter case-control study of women ages 45 to 70 that assessed risk of venous thromboembolism in oral vs transdermal estrogen users. Compared with women not taking hormone therapy, current users of oral estrogen had a significantly higher risk of venous thromboembolism, while transdermal estrogen users did not:

• Odds ratio with oral estrogen 4.2, 95% CI 1.5–11.6
• Odds ratio with transdermal estrogen 0.9, 95% CI 0.4–2.1.

The Kronos Early Estrogen Prevention Study (KEEPS)²³ did not support these findings. This 4-year randomized controlled trial, published in 2014, was designed to assess the risk of atherosclerosis progression with early menopause initiation of placebo vs low-dose oral hormone therapy (conjugated equine estrogen 0.45 mg daily with cyclical micronized progesterone) or transdermal hormone therapy (estradiol 50 µg/week with cyclical micronized progesterone).

In the 727 women in the study, there was one transient ischemic attack in the oral hormone therapy group, one unconfirmed stroke in the transdermal hormone therapy group, and one case of venous thromboembolism in each group, findings that were underpowered for statistical significance. Both oral and transdermal hormonal therapy had neutral effects on atherosclerosis progression, as assessed by arterial imaging. Transdermal hormone therapy was associated with improvements in markers of insulin resistance and was not associated with an increase in triglycerides, C-reactive protein, or sex hormone-binding globulin, as would be expected with transdermal circumvention of the first-pass hepatic effect.

BALANCING THE RISKS AND BENEFITS FOR THE PATIENT

The most effective treatment for vasomotor symptoms in women at any age is hormone therapy, and the benefits are more likely to outweigh risks when initiated before age 60 or within 10 years of menopause.⁷ The Women’s Health Initiative randomized study was limited to 5.6 to 7.2 years of hormone therapy (13 years of cumulative follow-up), and the Danish Osteoporosis Study was limited to 11 years of use (16 years cumulative follow-up).

The coronary heart disease outcomes for longer durations of therapy remain uncertain. There is a small but statistically significant increased risk of stroke and venous thromboembolism with oral hormone therapy, and breast cancer risk is associated with long-term estrogen-progestin use.

Patients on hormone therapy should be evaluated annually regarding the need for ongoing therapy. Persistent moderate-severe vasomotor symptoms, quality of life benefits of hormone therapy, contraindications to its use (Table 2), and patient preference need to be assessed as well as baseline risks of cardiovascular disease, breast cancer, and fracture.

Risk calculators may facilitate the shared decision-making process. Examples are:

• The American College of Cardiology/American Heart association risk calculator for cardiovascular disease²⁴ (www.cvriskcalculator.com)
• The World Health Organization Fracture Risk Assessment Tool (FRAX)²⁵
  (www.shef.ac.uk/FRAX/tool.jsp)

• The Gail model for breast cancer risk²⁶ (www.cancer.gov/bcrisktool/).
• MenoPro, a menopause decision-support algorithm and companion mobile app developed by NAMS to help direct treatment decisions based on the 10-year risk of atherosclerotic cardiovascular disease (www.menopause.org/for-professionals/-i-menopro-i-mobile-app).²⁷
  The discussion of the risks of hormone therapy with patients should incorporate the perspective of absolute risk. For example, a woman wishing to continue estrogen-progestin therapy should be told that the Women’s Health Initiative data suggest that, after 5 years of use, breast cancer risk may be increased by 8 additional cases per 10,000 users per year. According to the World Health Organization, this magnitude of risk is defined as rare (less than 1 event per 1,000 women).²⁸

A strategy of prescribing the lowest dose to achieve the desired clinical benefits is prudent and recommended.^1–3 Table 3 outlines the estrogen formulations now available in the United States, with their doses and formulations.

Unless contraindications develop (Table 2), patients may elect to continue hormone therapy if its benefits outweigh its risks. The American College of Obstetricians and Gynecologists (ACOG) 2014 practice recommendations for management of menopausal symptoms³¹ and the 2015 NAMS statement both recommend that hormone therapy not be discontinued based solely on a woman’s age.²⁹

Hormone therapy is on the Beer’s list of potentially inappropriate medications for older adults,³⁰ which remains a hurdle to its long-term use and seems to be at odds with these ACOG and NAMS statements.

Patients who choose to discontinue hormone therapy need to be monitored for persistent bothersome vasomotor symptoms, bone loss, osteoporosis, and the genitourinary syndrome of menopause (previously referred to as vulvovaginal atrophy)³¹ and offered alternative therapies if needed.