166
Letters to the Editor
administration of the first sedation doses to scope withdrawal, was shown to be associated with higher sedation requirements. In this study, however, the total procedural time was separated into sedative administration time and scoping time to assess the effect of actual procedural time (time from scope insertion to scope withdrawal) on sedation. In the multivariate analysis, we showed that there was no correlation between the length of actual procedural time and sedation requirement. Therefore, procedural time was not chosen to be one of the matched parameters. In the second part of the study, 32 patients who underwent an esophagogastroduodenoscopy before and after RYGB were selected. If patients underwent multiple procedures, only the esophagogastroduodenoscopies that took place right before and after RYGB were chosen. The majority of the first upper endoscopies that took place after RYGB were in fact diagnostic to investigate abdominal pain, nausea, and weight regain. Common etiologies included marginal ulceration and outlet dilatation. When appropriate, patients then were brought back for a therapeutic procedure at the next endoscopy visit. Of 32 procedures performed after RYGB, 91% were diagnostic. This proportion is similar to that of pre-RYGB upper endoscopies (91% were diagnostic). Therefore, procedure type was not a confounding factor in determining the sedation requirement before and after RYGB. The comment regarding a decrease in CYP450 level secondary to anemia after RYGB is interesting. This may be a topic of future study to further understand the possible mechanisms of increased sedation requirements after RYGB. In addition, we agree that the idea of performing a similar study to better understand use of propofol in this patient population would be interesting. PICHAMOL JIRAPINYO, MD Department of Internal Medicine Yale–New Haven Hospital New Haven, Connecticut NITIN KUMAR, MD CHRISTOPHER C. THOMPSON, MD, MSc Division of Gastroenterology, Hepatology and Endoscopy Brigham and Women’s Hospital Department of Internal Medicine Harvard Medical School Boston, Massachusetts
References 1.
Jirapinyo P, et al. Dig Dis Sci 2014;59:2173–2177.
2.
Jirapinyo P, et al. Clin Gastroenterol Hepatol 2015;13: 1432–1436.
Conflicts of interest The authors disclose no conflicts. Most current article http://dx.doi.org/10.1016/j.cgh.2015.09.011
Clinical Gastroenterology and Hepatology Vol. 14, No. 1
Increased Risk of Venous Thromboembolic Events With Corticosteroid Versus Biologic Therapy for Inflammatory Bowel Disease Dear Editor: We read with interest the article by Higgins et al1 in which they studied the association between corticosteroid therapy and venous thromboembolism (VTE) in 15,100 patients with inflammatory bowel disease (IBD) using a large US Commercial Claims Database.1 Compared with corticosteroid monotherapy, biologic monotherapy was associated with a lower odds (adjusted odds ratio, 0.21) of VTE, whereas combination therapy (corticosteroids þ biologics) was associated with a similar odds (adjusted odds ratio, 1.01) of VTE over a 1-year follow-up period, after controlling for several demographic and clinical factors. VTE events were observed exclusively during the first 2 months of treatment in patients receiving biologic monotherapy, whereas they were observed throughout the follow-up period in corticosteroid-treated patients. Higgins et al1 concluded that corticosteroid use is associated independently with the development of VTE in IBD patients. We praise the authors for undertaking a rigorous analysis of this important question in a large IBD cohort. However, there are several issues that limit inference of a causal association between corticosteroid use and VTE risk. First, it is impossible to separate the direct impact of IBD treatment on VTE events from its indirect impact on VTE through its effects on disease activity. Biologic therapy achieves better overall disease control than corticosteroid-based treatment in IBD,2–5 which may have largely dictated the observed associations. The observation that biologic monotherapy was associated with much faster resolution of VTE risk supports this notion because corticosteroid-treated patients are more likely to have had persistently active disease, disputing the argument by Higgins et al1 that the period beyond the first few months should be less reflective of disease activity. Furthermore, because most patients likely would have tapered off corticosteroids within a few months, the observation of VTE events occurring throughout the follow-up period in this group cannot be attributed entirely to medication effects. Adding to this, some patients who received biologic monotherapy may have simply been refractory or intolerant to other maintenance agents or treated primarily for fistulizing or fibrostenotic complications of Crohn’s disease, and may have had relatively milder inflammatory bowel activity at baseline. Conversely, patients treated with corticosteroids are more likely to have been experiencing a moderate to severe inflammatory disease flare. Higgins et al1 proposed that the observed similarity in VTE risk with combination therapy and
January 2016
corticosteroid monotherapy further supports the notion of an increased prothrombotic effect of corticosteroid therapy. However, it is possible that some patients who received combination therapy had greater baseline disease activity that lead to the use of two agents. Higgins et al1 were only able to control for surrogate measures of disease activity, including recent hospitalizations and IBD-related surgeries, which are unlikely to capture the full breadth of inflammatory activity in these patients. Indeed, many patients may have been hospitalized previously for reasons unrelated to an IBD flare. It also is possible that there was a significant lag in starting biologics in many patients who received combination therapy owing to requirements for pretreatment testing and vaccination and/or obtaining financial coverage, which could have lead to a delay in optimal disease control. Other patients may have received only corticosteroids because they failed to respond to biologics. In any of these scenarios, greater disease activity in the combination therapy group could be the main driver for the failure to show a similar benefit of biologic therapy as observed in the biologic monotherapy group. A number of studies have shown an association between IBD and VTE risk.6 A large population-based study from the United Kingdom also reported a substantially higher risk of VTE during periods of active disease as compared with periods of remission; here too, however, corticosteroid prescription was used as a surrogate marker for disease activity.7 No study directly implicated corticosteroids or other treatment as a causative factor in the development of VTE in IBD patients. Indeed, there is very little literature supporting a role for corticosteroids in the development of VTE, despite its widespread use. Still, the present study is provocative and prospective studies now are required to better define the role of corticosteroid treatment on VTE risk in IBD patients. SANJAY K. MURTHY, MD, MSc, FRCP(C) The Ottawa Hospital IBD Centre University of Ottawa Ottawa, Ontario, Canada GEOFFREY C. NGUYEN, MD, PhD, FRCP(C) Mount Sinai Hospital IBD Centre University of Toronto Toronto, Ontario, Canada
References 1.
Higgins PDR, et al. Clinical Gastroenterol Hepatol 2015; 13:316–321.
2. 3.
Hanauer SB, et al. Lancet 2002;359:1541–1549. Colombel JF, et al. Gastroenterology 2007;132:52–65.
4.
D’haens G, et al. Lancet 2008;371:660–667.
5.
Rutgeerts G, et al. N Engl J Med 2005;353:2462–2476.
6.
Murthy SK, et al. Am J Gastroenterol 2011;106:713–718.
7.
Grainge MJ, et al. Lancet 2010;375:657–663.
Letters to the Editor
167
Conflicts of interest The authors disclose no conflicts. Most current article http://dx.doi.org/10.1016/j.cgh.2015.04.022
Reply. In the letter by Murthy and Nguyen, 5 points are made about our study, as follows1: (1) it is very difficult to separate the effects of inflammatory bowel disease activity from the effects of inflammatory bowel disease therapy on venous thromboembolic event (VTE) risk; (2) they speculated that patients on steroids and biologics have more disease activity, even after controlling for hospitalizations and surgeries; (3) they speculated that patients would have tapered off of steroids (and therefore had less disease activity) after a few months, so that continued VTE events cannot be attributed to steroids; (4) they speculated that a lag in initiating biologic therapy may have worsened disease control; and (5) they claimed that no study has implicated corticosteroids as a causative factor in the development of VTE. I agree that it is difficult to separate the effects of disease activity from the effects of inflammatory bowel disease therapy on VTE risk. It also is difficult to control for all possible confounding variables in an administrative data set. However, by comparing patients on different therapies and controlling for disease activity with proxies of hospitalization and surgery, a significant association between steroid use and VTE was found. I cannot respond to the speculative scenarios presented by Murthy and Nguyen because I have no data in the administrative data set to support any of these. However, the claim that there is no support for this effect in the literature can be refuted by the 2013 Danish nationwide population-based study by Johannesdottir et al2 in JAMA Internal Medicine, which found an adjusted incidence rate ratio of 2.31 for VTE in patients using corticosteroids, with a temporal gradient of increased risk in new use that decayed with time after multiple adjustments for disease severity. New users of corticosteroids had an incidence rate ratio of 3.06 (95% confidence interval, 2.77–3.38), which was roughly comparable with our finding of an odds ratio for VTE of 0.21 in patients receiving biologics without corticosteroids. Clearly more studies are needed, with careful approaches to adjust for disease severity and compare VTE rates across patients on different therapies, to truly determine whether increased VTE risk is yet another adverse effect of corticosteroids. PETER D. R. HIGGINS, MD, PhD, MSc (CRDSA) Division of Gastroenterology Department of Medicine University of Michigan Ann Arbor, Michigan
Letters to the Editor
administration of the first sedation doses to scope withdrawal, was shown to be associated with higher sedation requirements. In this study, however, the total procedural time was separated into sedative administration time and scoping time to assess the effect of actual procedural time (time from scope insertion to scope withdrawal) on sedation. In the multivariate analysis, we showed that there was no correlation between the length of actual procedural time and sedation requirement. Therefore, procedural time was not chosen to be one of the matched parameters. In the second part of the study, 32 patients who underwent an esophagogastroduodenoscopy before and after RYGB were selected. If patients underwent multiple procedures, only the esophagogastroduodenoscopies that took place right before and after RYGB were chosen. The majority of the first upper endoscopies that took place after RYGB were in fact diagnostic to investigate abdominal pain, nausea, and weight regain. Common etiologies included marginal ulceration and outlet dilatation. When appropriate, patients then were brought back for a therapeutic procedure at the next endoscopy visit. Of 32 procedures performed after RYGB, 91% were diagnostic. This proportion is similar to that of pre-RYGB upper endoscopies (91% were diagnostic). Therefore, procedure type was not a confounding factor in determining the sedation requirement before and after RYGB. The comment regarding a decrease in CYP450 level secondary to anemia after RYGB is interesting. This may be a topic of future study to further understand the possible mechanisms of increased sedation requirements after RYGB. In addition, we agree that the idea of performing a similar study to better understand use of propofol in this patient population would be interesting. PICHAMOL JIRAPINYO, MD Department of Internal Medicine Yale–New Haven Hospital New Haven, Connecticut NITIN KUMAR, MD CHRISTOPHER C. THOMPSON, MD, MSc Division of Gastroenterology, Hepatology and Endoscopy Brigham and Women’s Hospital Department of Internal Medicine Harvard Medical School Boston, Massachusetts
References 1.
Jirapinyo P, et al. Dig Dis Sci 2014;59:2173–2177.
2.
Jirapinyo P, et al. Clin Gastroenterol Hepatol 2015;13: 1432–1436.
Conflicts of interest The authors disclose no conflicts. Most current article http://dx.doi.org/10.1016/j.cgh.2015.09.011
Clinical Gastroenterology and Hepatology Vol. 14, No. 1
Increased Risk of Venous Thromboembolic Events With Corticosteroid Versus Biologic Therapy for Inflammatory Bowel Disease Dear Editor: We read with interest the article by Higgins et al1 in which they studied the association between corticosteroid therapy and venous thromboembolism (VTE) in 15,100 patients with inflammatory bowel disease (IBD) using a large US Commercial Claims Database.1 Compared with corticosteroid monotherapy, biologic monotherapy was associated with a lower odds (adjusted odds ratio, 0.21) of VTE, whereas combination therapy (corticosteroids þ biologics) was associated with a similar odds (adjusted odds ratio, 1.01) of VTE over a 1-year follow-up period, after controlling for several demographic and clinical factors. VTE events were observed exclusively during the first 2 months of treatment in patients receiving biologic monotherapy, whereas they were observed throughout the follow-up period in corticosteroid-treated patients. Higgins et al1 concluded that corticosteroid use is associated independently with the development of VTE in IBD patients. We praise the authors for undertaking a rigorous analysis of this important question in a large IBD cohort. However, there are several issues that limit inference of a causal association between corticosteroid use and VTE risk. First, it is impossible to separate the direct impact of IBD treatment on VTE events from its indirect impact on VTE through its effects on disease activity. Biologic therapy achieves better overall disease control than corticosteroid-based treatment in IBD,2–5 which may have largely dictated the observed associations. The observation that biologic monotherapy was associated with much faster resolution of VTE risk supports this notion because corticosteroid-treated patients are more likely to have had persistently active disease, disputing the argument by Higgins et al1 that the period beyond the first few months should be less reflective of disease activity. Furthermore, because most patients likely would have tapered off corticosteroids within a few months, the observation of VTE events occurring throughout the follow-up period in this group cannot be attributed entirely to medication effects. Adding to this, some patients who received biologic monotherapy may have simply been refractory or intolerant to other maintenance agents or treated primarily for fistulizing or fibrostenotic complications of Crohn’s disease, and may have had relatively milder inflammatory bowel activity at baseline. Conversely, patients treated with corticosteroids are more likely to have been experiencing a moderate to severe inflammatory disease flare. Higgins et al1 proposed that the observed similarity in VTE risk with combination therapy and
January 2016
corticosteroid monotherapy further supports the notion of an increased prothrombotic effect of corticosteroid therapy. However, it is possible that some patients who received combination therapy had greater baseline disease activity that lead to the use of two agents. Higgins et al1 were only able to control for surrogate measures of disease activity, including recent hospitalizations and IBD-related surgeries, which are unlikely to capture the full breadth of inflammatory activity in these patients. Indeed, many patients may have been hospitalized previously for reasons unrelated to an IBD flare. It also is possible that there was a significant lag in starting biologics in many patients who received combination therapy owing to requirements for pretreatment testing and vaccination and/or obtaining financial coverage, which could have lead to a delay in optimal disease control. Other patients may have received only corticosteroids because they failed to respond to biologics. In any of these scenarios, greater disease activity in the combination therapy group could be the main driver for the failure to show a similar benefit of biologic therapy as observed in the biologic monotherapy group. A number of studies have shown an association between IBD and VTE risk.6 A large population-based study from the United Kingdom also reported a substantially higher risk of VTE during periods of active disease as compared with periods of remission; here too, however, corticosteroid prescription was used as a surrogate marker for disease activity.7 No study directly implicated corticosteroids or other treatment as a causative factor in the development of VTE in IBD patients. Indeed, there is very little literature supporting a role for corticosteroids in the development of VTE, despite its widespread use. Still, the present study is provocative and prospective studies now are required to better define the role of corticosteroid treatment on VTE risk in IBD patients. SANJAY K. MURTHY, MD, MSc, FRCP(C) The Ottawa Hospital IBD Centre University of Ottawa Ottawa, Ontario, Canada GEOFFREY C. NGUYEN, MD, PhD, FRCP(C) Mount Sinai Hospital IBD Centre University of Toronto Toronto, Ontario, Canada
References 1.
Higgins PDR, et al. Clinical Gastroenterol Hepatol 2015; 13:316–321.
2. 3.
Hanauer SB, et al. Lancet 2002;359:1541–1549. Colombel JF, et al. Gastroenterology 2007;132:52–65.
4.
D’haens G, et al. Lancet 2008;371:660–667.
5.
Rutgeerts G, et al. N Engl J Med 2005;353:2462–2476.
6.
Murthy SK, et al. Am J Gastroenterol 2011;106:713–718.
7.
Grainge MJ, et al. Lancet 2010;375:657–663.
Letters to the Editor
167
Conflicts of interest The authors disclose no conflicts. Most current article http://dx.doi.org/10.1016/j.cgh.2015.04.022
Reply. In the letter by Murthy and Nguyen, 5 points are made about our study, as follows1: (1) it is very difficult to separate the effects of inflammatory bowel disease activity from the effects of inflammatory bowel disease therapy on venous thromboembolic event (VTE) risk; (2) they speculated that patients on steroids and biologics have more disease activity, even after controlling for hospitalizations and surgeries; (3) they speculated that patients would have tapered off of steroids (and therefore had less disease activity) after a few months, so that continued VTE events cannot be attributed to steroids; (4) they speculated that a lag in initiating biologic therapy may have worsened disease control; and (5) they claimed that no study has implicated corticosteroids as a causative factor in the development of VTE. I agree that it is difficult to separate the effects of disease activity from the effects of inflammatory bowel disease therapy on VTE risk. It also is difficult to control for all possible confounding variables in an administrative data set. However, by comparing patients on different therapies and controlling for disease activity with proxies of hospitalization and surgery, a significant association between steroid use and VTE was found. I cannot respond to the speculative scenarios presented by Murthy and Nguyen because I have no data in the administrative data set to support any of these. However, the claim that there is no support for this effect in the literature can be refuted by the 2013 Danish nationwide population-based study by Johannesdottir et al2 in JAMA Internal Medicine, which found an adjusted incidence rate ratio of 2.31 for VTE in patients using corticosteroids, with a temporal gradient of increased risk in new use that decayed with time after multiple adjustments for disease severity. New users of corticosteroids had an incidence rate ratio of 3.06 (95% confidence interval, 2.77–3.38), which was roughly comparable with our finding of an odds ratio for VTE of 0.21 in patients receiving biologics without corticosteroids. Clearly more studies are needed, with careful approaches to adjust for disease severity and compare VTE rates across patients on different therapies, to truly determine whether increased VTE risk is yet another adverse effect of corticosteroids. PETER D. R. HIGGINS, MD, PhD, MSc (CRDSA) Division of Gastroenterology Department of Medicine University of Michigan Ann Arbor, Michigan
