User login
“Sensitivity analyses suggest that previously reported associations and persistent association with SSRI [selective serotonin reuptake inhibitor] initiation may be due to surveillance bias,” wrote gastroenterologist Hamed Khalili, MD, MPH, of Massachusetts General Hospital, Boston, and colleagues in Annals of Internal Medicine, advising clinicians to carefully balance the benefits of these medication classes against the very low likelihood of a causal relationship with MC.
While two smaller studies had challenged the belief that these medications can cause MC, Khalili told GI & Hepatology News, “the quality of the data that supported or refuted this hypothesis were low. Nevertheless, most in the field consider MC to be largely related to medications so we thought it was important to systematically answer this question.”
While most medications thought to trigger MC were found not to be causally linked, he added, “we did observe a marginal association with SSRIs but could not rule out the possibility that the association is related to residual bias.”
The authors noted that the incidence of MC in older persons is rising rapidly and is thought to account for more than 30% of chronic diarrhea cases in this group.
Despite weak evidence in the literature, the treatment guidelines of several societies, including the American Gastroenterological Association, recommend discontinuing potential pharmacologic triggers as first-line prevention or as an adjunct therapy, particularly in recurrent or refractory MC. But this approach may be ineffective in patients with established disease and could lead to inappropriate discontinuation of medication such as antihypertensives, the authors argued.
As to proposed mechanisms of action, said Khalili, “for PPIs [proton-pump inhibitors,] people thought it was related to changes in the gut microbiome. For NSAIDs [nonsteroidal anti-inflammatory drugs], people thought it could be related to changes in the gut barrier function. But overall, not a single mechanism would have explained all the prior associations that were observed.”
While medications such as PPIs and SSRIs can cause diarrhea in a small subset of users, Khalili added, “most patients generally catch these side effects very quickly and realize that stopping these medications will improve their diarrhea. This is very different than most patients we as gastroenterologists see with a new diagnosis of MC. Many of them may have been on these medications for a long time. We believe that stopping medications in these patients is unnecessary.”
Study Details
The investigators looked at eligible residents in Sweden age 65 years or older in the years 2006 to 2017 (n = 191,482 to 2,634,777). Participants had no history of inflammatory bowel disease and different cohorts were examined for various common medications from calcium channel blockers to statins.
With a primary outcome of biopsy-verified MC, dates of diagnosis were obtained from Sweden’s national histopathology cohort ESPRESSO (Epidemiology Strengthened by Histopathology Reports in Sweden). Among the findings:
- The 12- and 24-month cumulative incidences of MC were less than 0.05% under all treatment strategies.
- Estimated 12-month risk differences were close to null under angiotensin-converting enzyme vs calcium-channel blocker (CCB) initiation, angiotensin-receptor blocker vs CCB initiation, NSAID initiation vs noninitiation, PPI inhibitor initiation vs noninitiation, and statin initiation vs noninitiation.
- The estimated 12-month risk difference was 0.04% (95% CI, 0.03%-0.05%) for SSRIs vs mirtazapine.
- Results were similar for 24-month risk differences. Several medications such as SSRIs were also associated with increased risk for undergoing colonoscopy with a normal colorectal mucosa biopsy result.
“We think it’s unlikely that stopping these medications will improve symptoms of MC,” Khalili said.
Commenting on the paper but not involved in it, Jordan E. Axelrad, MD, MPH, codirector of the Inflammatory Bowel Disease Center at NYU Langone Health in New York City, said, “This study strengthens the argument that MC is an immune-mediated disease, not primarily driven by drug exposures. But future studies in diverse cohorts are required to validate these findings.” He said the study nevertheless provides reassurance that previously reported associations may have been overstated or confounded by factors such as reverse causation and increased healthcare utilization preceding the MC diagnosis.
In the meantime, Axelrad added, the findings “may reduce the inclination to promptly discontinue medications historically associated with MC in newly diagnosed cases. Also, these data help shift the clinical focus away from medication cessation alone and toward a needed and broader MC management strategy. US-based validation would likely highlight these changes in our patients.”
Despite concerns about the study’s unmeasured confounding because of differential healthcare utilization or surveillance, the modest association observed between SSRI and MC is supported by literature linking catecholamine and serotonin to gut innate immunity and microbiota, Khalili’s group wrote. “However, this finding may also be confounded by other factors including persisting surveillance and protopathic bias, especially since an association was also seen for risk for receipt of a colonoscopy with normal mucosa.”
Khalili believes the Swedish results are applicable even to the more diverse US population. He noted that lack of primary care data limited measurement of and adjustment for symptoms and medical diagnoses that increase risk. But according to Axelrad, MC is more prevalent in White, older patients, who are well-represented in Swedish cohorts but to a lesser extent in US populations. “Additionally, environmental factors and medication use patterns differ between Sweden and the US, particularly in regard to over-the-counter medication access.”
The findings have implications for future research in pharmacoepidemiologic studies of gastrointestinal-related outcomes. Since many routinely prescribed medications such as SSRIs were associated with an apparent increased risk for colonoscopies with normal colorectal biopsy results, future studies that examine gastrointestinal-specific adverse events should carefully consider potential surveillance bias.
In the meantime, Khalili stressed, it’s important to highlight that while some of these medications cause diarrhea in a small subset of patients, stopping medications in these patients is unnecessary.
This study was supported by the National Institutes of Health (NIH) and the Swedish Research Council. Khalili disclosed grants from the Crohn’s & Coiltis Foundation, the NIH and the Helmsley CharitableTrust, as well as stock ownership in Cylinder Health. One coauthor is employed by Massachusetts General Hospital. Axelrad had no relevant competing interests.
“Sensitivity analyses suggest that previously reported associations and persistent association with SSRI [selective serotonin reuptake inhibitor] initiation may be due to surveillance bias,” wrote gastroenterologist Hamed Khalili, MD, MPH, of Massachusetts General Hospital, Boston, and colleagues in Annals of Internal Medicine, advising clinicians to carefully balance the benefits of these medication classes against the very low likelihood of a causal relationship with MC.
While two smaller studies had challenged the belief that these medications can cause MC, Khalili told GI & Hepatology News, “the quality of the data that supported or refuted this hypothesis were low. Nevertheless, most in the field consider MC to be largely related to medications so we thought it was important to systematically answer this question.”
While most medications thought to trigger MC were found not to be causally linked, he added, “we did observe a marginal association with SSRIs but could not rule out the possibility that the association is related to residual bias.”
The authors noted that the incidence of MC in older persons is rising rapidly and is thought to account for more than 30% of chronic diarrhea cases in this group.
Despite weak evidence in the literature, the treatment guidelines of several societies, including the American Gastroenterological Association, recommend discontinuing potential pharmacologic triggers as first-line prevention or as an adjunct therapy, particularly in recurrent or refractory MC. But this approach may be ineffective in patients with established disease and could lead to inappropriate discontinuation of medication such as antihypertensives, the authors argued.
As to proposed mechanisms of action, said Khalili, “for PPIs [proton-pump inhibitors,] people thought it was related to changes in the gut microbiome. For NSAIDs [nonsteroidal anti-inflammatory drugs], people thought it could be related to changes in the gut barrier function. But overall, not a single mechanism would have explained all the prior associations that were observed.”
While medications such as PPIs and SSRIs can cause diarrhea in a small subset of users, Khalili added, “most patients generally catch these side effects very quickly and realize that stopping these medications will improve their diarrhea. This is very different than most patients we as gastroenterologists see with a new diagnosis of MC. Many of them may have been on these medications for a long time. We believe that stopping medications in these patients is unnecessary.”
Study Details
The investigators looked at eligible residents in Sweden age 65 years or older in the years 2006 to 2017 (n = 191,482 to 2,634,777). Participants had no history of inflammatory bowel disease and different cohorts were examined for various common medications from calcium channel blockers to statins.
With a primary outcome of biopsy-verified MC, dates of diagnosis were obtained from Sweden’s national histopathology cohort ESPRESSO (Epidemiology Strengthened by Histopathology Reports in Sweden). Among the findings:
- The 12- and 24-month cumulative incidences of MC were less than 0.05% under all treatment strategies.
- Estimated 12-month risk differences were close to null under angiotensin-converting enzyme vs calcium-channel blocker (CCB) initiation, angiotensin-receptor blocker vs CCB initiation, NSAID initiation vs noninitiation, PPI inhibitor initiation vs noninitiation, and statin initiation vs noninitiation.
- The estimated 12-month risk difference was 0.04% (95% CI, 0.03%-0.05%) for SSRIs vs mirtazapine.
- Results were similar for 24-month risk differences. Several medications such as SSRIs were also associated with increased risk for undergoing colonoscopy with a normal colorectal mucosa biopsy result.
“We think it’s unlikely that stopping these medications will improve symptoms of MC,” Khalili said.
Commenting on the paper but not involved in it, Jordan E. Axelrad, MD, MPH, codirector of the Inflammatory Bowel Disease Center at NYU Langone Health in New York City, said, “This study strengthens the argument that MC is an immune-mediated disease, not primarily driven by drug exposures. But future studies in diverse cohorts are required to validate these findings.” He said the study nevertheless provides reassurance that previously reported associations may have been overstated or confounded by factors such as reverse causation and increased healthcare utilization preceding the MC diagnosis.
In the meantime, Axelrad added, the findings “may reduce the inclination to promptly discontinue medications historically associated with MC in newly diagnosed cases. Also, these data help shift the clinical focus away from medication cessation alone and toward a needed and broader MC management strategy. US-based validation would likely highlight these changes in our patients.”
Despite concerns about the study’s unmeasured confounding because of differential healthcare utilization or surveillance, the modest association observed between SSRI and MC is supported by literature linking catecholamine and serotonin to gut innate immunity and microbiota, Khalili’s group wrote. “However, this finding may also be confounded by other factors including persisting surveillance and protopathic bias, especially since an association was also seen for risk for receipt of a colonoscopy with normal mucosa.”
Khalili believes the Swedish results are applicable even to the more diverse US population. He noted that lack of primary care data limited measurement of and adjustment for symptoms and medical diagnoses that increase risk. But according to Axelrad, MC is more prevalent in White, older patients, who are well-represented in Swedish cohorts but to a lesser extent in US populations. “Additionally, environmental factors and medication use patterns differ between Sweden and the US, particularly in regard to over-the-counter medication access.”
The findings have implications for future research in pharmacoepidemiologic studies of gastrointestinal-related outcomes. Since many routinely prescribed medications such as SSRIs were associated with an apparent increased risk for colonoscopies with normal colorectal biopsy results, future studies that examine gastrointestinal-specific adverse events should carefully consider potential surveillance bias.
In the meantime, Khalili stressed, it’s important to highlight that while some of these medications cause diarrhea in a small subset of patients, stopping medications in these patients is unnecessary.
This study was supported by the National Institutes of Health (NIH) and the Swedish Research Council. Khalili disclosed grants from the Crohn’s & Coiltis Foundation, the NIH and the Helmsley CharitableTrust, as well as stock ownership in Cylinder Health. One coauthor is employed by Massachusetts General Hospital. Axelrad had no relevant competing interests.
“Sensitivity analyses suggest that previously reported associations and persistent association with SSRI [selective serotonin reuptake inhibitor] initiation may be due to surveillance bias,” wrote gastroenterologist Hamed Khalili, MD, MPH, of Massachusetts General Hospital, Boston, and colleagues in Annals of Internal Medicine, advising clinicians to carefully balance the benefits of these medication classes against the very low likelihood of a causal relationship with MC.
While two smaller studies had challenged the belief that these medications can cause MC, Khalili told GI & Hepatology News, “the quality of the data that supported or refuted this hypothesis were low. Nevertheless, most in the field consider MC to be largely related to medications so we thought it was important to systematically answer this question.”
While most medications thought to trigger MC were found not to be causally linked, he added, “we did observe a marginal association with SSRIs but could not rule out the possibility that the association is related to residual bias.”
The authors noted that the incidence of MC in older persons is rising rapidly and is thought to account for more than 30% of chronic diarrhea cases in this group.
Despite weak evidence in the literature, the treatment guidelines of several societies, including the American Gastroenterological Association, recommend discontinuing potential pharmacologic triggers as first-line prevention or as an adjunct therapy, particularly in recurrent or refractory MC. But this approach may be ineffective in patients with established disease and could lead to inappropriate discontinuation of medication such as antihypertensives, the authors argued.
As to proposed mechanisms of action, said Khalili, “for PPIs [proton-pump inhibitors,] people thought it was related to changes in the gut microbiome. For NSAIDs [nonsteroidal anti-inflammatory drugs], people thought it could be related to changes in the gut barrier function. But overall, not a single mechanism would have explained all the prior associations that were observed.”
While medications such as PPIs and SSRIs can cause diarrhea in a small subset of users, Khalili added, “most patients generally catch these side effects very quickly and realize that stopping these medications will improve their diarrhea. This is very different than most patients we as gastroenterologists see with a new diagnosis of MC. Many of them may have been on these medications for a long time. We believe that stopping medications in these patients is unnecessary.”
Study Details
The investigators looked at eligible residents in Sweden age 65 years or older in the years 2006 to 2017 (n = 191,482 to 2,634,777). Participants had no history of inflammatory bowel disease and different cohorts were examined for various common medications from calcium channel blockers to statins.
With a primary outcome of biopsy-verified MC, dates of diagnosis were obtained from Sweden’s national histopathology cohort ESPRESSO (Epidemiology Strengthened by Histopathology Reports in Sweden). Among the findings:
- The 12- and 24-month cumulative incidences of MC were less than 0.05% under all treatment strategies.
- Estimated 12-month risk differences were close to null under angiotensin-converting enzyme vs calcium-channel blocker (CCB) initiation, angiotensin-receptor blocker vs CCB initiation, NSAID initiation vs noninitiation, PPI inhibitor initiation vs noninitiation, and statin initiation vs noninitiation.
- The estimated 12-month risk difference was 0.04% (95% CI, 0.03%-0.05%) for SSRIs vs mirtazapine.
- Results were similar for 24-month risk differences. Several medications such as SSRIs were also associated with increased risk for undergoing colonoscopy with a normal colorectal mucosa biopsy result.
“We think it’s unlikely that stopping these medications will improve symptoms of MC,” Khalili said.
Commenting on the paper but not involved in it, Jordan E. Axelrad, MD, MPH, codirector of the Inflammatory Bowel Disease Center at NYU Langone Health in New York City, said, “This study strengthens the argument that MC is an immune-mediated disease, not primarily driven by drug exposures. But future studies in diverse cohorts are required to validate these findings.” He said the study nevertheless provides reassurance that previously reported associations may have been overstated or confounded by factors such as reverse causation and increased healthcare utilization preceding the MC diagnosis.
In the meantime, Axelrad added, the findings “may reduce the inclination to promptly discontinue medications historically associated with MC in newly diagnosed cases. Also, these data help shift the clinical focus away from medication cessation alone and toward a needed and broader MC management strategy. US-based validation would likely highlight these changes in our patients.”
Despite concerns about the study’s unmeasured confounding because of differential healthcare utilization or surveillance, the modest association observed between SSRI and MC is supported by literature linking catecholamine and serotonin to gut innate immunity and microbiota, Khalili’s group wrote. “However, this finding may also be confounded by other factors including persisting surveillance and protopathic bias, especially since an association was also seen for risk for receipt of a colonoscopy with normal mucosa.”
Khalili believes the Swedish results are applicable even to the more diverse US population. He noted that lack of primary care data limited measurement of and adjustment for symptoms and medical diagnoses that increase risk. But according to Axelrad, MC is more prevalent in White, older patients, who are well-represented in Swedish cohorts but to a lesser extent in US populations. “Additionally, environmental factors and medication use patterns differ between Sweden and the US, particularly in regard to over-the-counter medication access.”
The findings have implications for future research in pharmacoepidemiologic studies of gastrointestinal-related outcomes. Since many routinely prescribed medications such as SSRIs were associated with an apparent increased risk for colonoscopies with normal colorectal biopsy results, future studies that examine gastrointestinal-specific adverse events should carefully consider potential surveillance bias.
In the meantime, Khalili stressed, it’s important to highlight that while some of these medications cause diarrhea in a small subset of patients, stopping medications in these patients is unnecessary.
This study was supported by the National Institutes of Health (NIH) and the Swedish Research Council. Khalili disclosed grants from the Crohn’s & Coiltis Foundation, the NIH and the Helmsley CharitableTrust, as well as stock ownership in Cylinder Health. One coauthor is employed by Massachusetts General Hospital. Axelrad had no relevant competing interests.