Proton Pump Inhibitor Use Before and After a Diagnosis of Inflammatory Bowel Disease

Abstract

Background

Proton pump inhibitors (PPIs) have an impact on the gut microbiome. We investigated whether increased use of PPIs was associated with a diagnosis of inflammatory bowel disease (IBD).

Methods

The University of Manitoba IBD Epidemiology Database includes all Manitobans diagnosed with IBD between 1984 and 2018 with age-, sex-, and geography-matched control subjects and comprehensive prescription drug data from April 1995. Subjects were considered to be users if they received 2 PPI prescriptions. We assessed PPI prescriptions prediagnosis and for 3 years postdiagnosis of IBD. The absolute and relative rates were calculated and compared for PPI use pre– and post–IBD diagnosis.

Results

A total of 5920 subjects were diagnosed with IBD after April 1996. Rates of PPI use in control subjects increased gradually from 1.5% to 6.5% over 15 years. Persons with IBD had a higher rate of PPI use, peaking up to 17% within 1 year of IBD diagnosis with a rate ratio (RR) of 3.1 (95% confidence interval [CI], 2.9-3.3). Furthermore, persons with Crohn’s disease (RR, 4.2; 95% CI, 3.7-4.6) were more likely to have been PPI users prediagnosis than persons with ulcerative colitis (RR, 2.4; 95% CI, 2.2-2.7). Important predictors of increased PPI use were older age, year of data collection, and Crohn’s disease diagnosis.

Conclusions

Persons with IBD have higher PPI use preceding their diagnosis. Possibly, the use of a PPI alters the gut microbiome, increasing the risk for IBD diagnosis; or persons with IBD have increased rates of dyspepsia, warranting PPI use; or some IBD symptoms are treated with PPIs whether warranted or not.

Lay Summary

Proton pump inhibitor (PPI) use is increased in persons with IBD for up to years prior to their diagnosis, raising the possibility that their use increases the risk for inflammatory bowel disease (IBD), or alternatively, persons with IBD have more upper gastrointestinal complaints. Another possibility is that early IBD-related symptoms are being managed rightly or wrongly with PPI therapy. Further, there is an increased use of PPI therapy post–IBD diagnosis compared with matched control subjects. While the use of PPI therapy in persons with IBD may be appropriate, physicians need to rigorously review their use at all times, especially in persons with IBD, as in this population their use may be excessive and possibly detrimental.

proton pump inhibitors, administrative health data, population based, inflammatory bowel disease, risk factors

Key Messages

What is already known about this topic: Proton pump inhibitors (PPIs) are among the most commonly used drugs worldwide, both prescribed and used over the counter. As they are used for upper gastrointestinal complaints and can impact the gut microbiome, it is possible that their use is associated with ultimate diagnosis of inflammatory bowel disease (IBD).

What is new here: Rates of PPI use are significantly higher in persons with IBD prior to their diagnosis than in healthy control subjects and higher prior to Crohn’s disease diagnosis compared with ulcerative colitis diagnosis. PPI use postdiagnosis remains higher in persons with IBD relative to persons without IBD.

How can this study help patient care: These data reinforce the importance of not using drugs indiscriminately and/or withdrawing drugs over time so as to determine if they are truly necessary. The increased rates of PPI use prior to IBD diagnosis raise the possibility that PPI use altering the gut microbiome may be an environmental cause of IBD, especially as their use is worldwide.

Introduction

Inflammatory bowel disease (IBD) is a chronic immune-mediated inflammatory disease affecting the gastrointestinal tract.¹ IBD includes Crohn’s disease (CD), which can affect any segment of the gastrointestinal tract from the mouth to the anus, and ulcerative colitis (UC), which is limited to colonic mucosa.² In North America and Europe, there are at least 1.5 million and 2 million people, respectively, affected by the disease¹ with incidence rates increasing worldwide. Considering that IBD affects persons of all races, an environmental factor that is available worldwide may contribute to a pathogenetic cascade. A disease trigger for IBD may be a factor that can directly alter the intestinal immune response, or indirectly through changes in the gut microbiome such as through early-life exposures, lifestyle and hygiene, vaccinations, surgeries, exposure to drugs, and gastrointestinal pathogens.^3,4

Proton pump inhibitors (PPIs) have been used since the 1980s to reduce gastric acid secretions to prevent and treat acid-related gastrointestinal diseases.⁵ Mild and mostly uncommon side effects of PPIs include increased flatus, loose stools, and abdominal pain, and severe side effects include Clostridioides difficile infection, community-acquired pneumonia, and reduced magnesium and vitamin B12 levels.⁵ PPIs have an impact on the gut microbiome and immune function. The low pH of gastric secretions has been recognized as the first line of defense against exogenous pathogens; thus, use of PPIs can alter the gut microbiome.⁶ A pooled analysis of 3 prospective studies included >600 000 individuals followed for a median of 12 years found that PPI users were at a 42% increased risk of IBD compared with nonusers.⁷ Considering that PPIs might be a common trigger of IBD, it is noteworthy that in the United States approximately 8% of adults had a prescription for a PPI within the past 30 days in 2011 to 2012.⁸ Elsewhere, the use of PPIs has grown since the late 1980s. In the United Kingdom, PPI use has doubled since 2007, and in 2017 there were nearly 59 million items dispensed annually.⁹ The goal of our retrospective analysis was to investigate whether increased use of PPIs was associated with a diagnosis of IBD both prior to and postdiagnosis.

Methods

Data Source and Population

Manitoba is a central province in Canada with a population of 1.4 million (2018). The population is ethnically diverse, with approximately 60% of the residents living in urban areas.¹⁰ The University of Manitoba IBD Epidemiology Database includes all persons with IBD in Manitoba from April 1984 to March 31, 2018.¹⁰ The residents of Manitoba have a unique personal health identification number by which all health systems contacts (prescriptions, outpatient visits, procedures, and hospitalizations) can be tracked. The Discharge Abstract Database captures administrative, clinical, and demographic information on hospital discharges (including deaths, discharges, and transfers).¹¹ This database records up to 16 diagnoses using the International Classification of Disease–Ninth Revision–Clinical Modification (ICD-9-CM) until 2004 and subsequently, up to 25 diagnoses have been recorded using the ICD-10-CM. Persons with CD or UC are identified using a validated administrative definition.¹⁰ This definition uses ICD-9-CM code 555 or ICD-10 code K50 for CD and ICD-9 CM code 556 or ICD-10 code K51 for UC. Each IBD index case was then matched to 10 control subjects without IBD on year of birth, sex, and geographic residence by postal area code.

Identification of PPI users from April 1995 to 2018 was from the Drug Program Information Network, the provincial prescription drug data base that links all prescribed medications to individuals by their personal health identification numbers. In July 2019, the Canada Drug Database listed 243 oral PPIs available in Canada. This list was used to identify PPI prescriptions and users. A subset of the University of Manitoba IBD Epidemiology Database was extracted for cases with an IBD diagnosis date of April 1, 1996, or later (ie, 1 year after drug data became available) and at least 1 year of coverage under Manitoba Health prior to IBD diagnosis. A PPI user must have at least 2 prescriptions. The database contained 5919 cases (2611 CD, 3308 UC) and 52 217 control subjects.

Outcomes and Analysis

For the pre-IBD analyses, subjects were followed from April 1, 1995, or the date that they registered with Manitoba Health, if after April 1, 1995. The event being modeled was the first PPI prescription of at least 2. This first prescription, identifying a subject as a PPI user, must be dispensed before the IBD diagnosis date of a case, or the diagnosis date of the associated case if a control subject. All subjects were censored at IBD diagnosis. Logistic regression analysis was performed by stratified case-control group to produce odds ratios (ORs) of usage probability. Rate ratio (RR) analysis was also performed to calculate and compare PPI usage rates before IBD diagnosis. A hazard ratio (HR) regression analysis, also stratified by case-control group, was carried out to examine PPI use after IBD diagnosis but using only subjects with no known use of PPIs prior to IBD diagnosis. Follow-up started at IBD diagnosis and subjects were censored at death, at loss of Manitoba Health coverage, or at March 31, 2018. Analyses were repeated for the 2 disease type subsets (CD and UC). A separate analysis was undertaken for persons with CD and their control subjects and compared with outcomes in persons with UC and their control subjects.

A nested case-control study was undertaken for outcomes that identify more aggressive disease. Persons with at least 1 prescription for PPIs before IBD diagnosis and 2 or more prescriptions in total were successfully matched with up to 4 cases with no PPI use prior to diagnosis. Matching was on sex, disease (CD or UC), year of IBD diagnosis, and age at IBD diagnoses ±5 years. Proportional hazards regression models were used to generate HRs for each of the following dependent variables: anti-tumor necrosis factor use (any use or at least 2 prescriptions), vedolizumab use, hospitalization with IBD as the first diagnosis on the discharge abstract, and IBD-related surgery.

This study was approved by the University of Manitoba Research Ethics Board (#H.S.20058 [H2016:315]).

Results

Rates of PPI use in control subjects increased gradually from 1.5% to 7.0%, and in persons with IBD rates increased from 1.5% to 17% over 15 years (Figure 1). Persons diagnosed with IBD had higher rates of PPI use, peaking up to 17% within 1 year of IBD diagnosis, with a RR of 3.08 (95% CI, 2.9-3.3) prediagnosis. Table 1 shows the OR for PPI use before IBD diagnosis by age distribution. Children with IBD were significantly more likely to be prescribed PPIs pre-IBD diagnosis than any other age group. Furthermore, persons with CD (RR, 4.15; 95% CI, 3.72-4.63) were more likely to have been PPI users (relative to their matched control subjects) prediagnosis than persons with UC (RR, 2.44; 95% CI, 2.12-2.70) (Table 2). PPI use was highest 1 year preceding the IBD diagnosis in those with IBD compared with matched control subjects (RR, 6.99; 95% CI, 6.05-8.08) (Table 2). This RR was higher among those diagnosed with CD compared with their control subjects, at 10.2 (95% CI, 8.33-12.6), as compared with persons with UC, at 4.83 (95% CI, 3.92-5.92) (Table 2). Overall, persons with IBD were 2.5 to 3 times more likely to be on a PPI 1 to 3 years preceding their IBD diagnosis (Figure 2). Persons with IBD had significantly higher use of PPIs post–IBD diagnosis compared with control subjects (Figure 3) over time. The increase in use postdiagnosis was greater in CD (20%) than in UC (14%) patients compared with control subjects (Figure 3). Female subjects were more likely to be PPI users across all age groups (Supplemental Table 1).

Table 1.

Characteristics of patients using a PPI (received any prescription of >30 days) after April 1996 and those using PPIs before a diagnosis of IBD in Manitoba, Canada.

Age	Cohort		PPI Use Before IBD Diagnosis		Odds Ratio	95% CI
	Case	Control	Case	Control
All	5920	52 227	1410	4833	3.41	3.16-3.6
<18 y	610	5641	57	44	11.01	6.35-19.09
18-39 y	2531	21 694	437	1087	4.02	3.51-4.61
40-64 y	2063	18 707	617	2341	3.13	2.79-3.50
>65 y	716	6185	299	1361	2.73	2.29-3.25

Values are n, unless otherwise indicated.

Abbreviations: IBD, inflammatory bowel disease; PPI, proton pump inhibitor.

Table 2.

Rate ratio of PPI users (with a prescription of at least 30 days) and developing IBD (either UC or CD).

Disease	Group	Rate per 1000 Persons (95% CI)	Rate Ratio (95% CI)
IBD	Case	30.4 (28.5-32.5)	3.08 (2.86-3.32)
	Control	9.9 (9.6-10.2)
CD	Case	36.9 (33.5-40.6)	4.15 (3.72-4.63)
	Control	8.9 (8.4-9.4)
UC	Case	26.0 (23.7-28.4)	2.44 (2.21-2.70)
	Control	10.6 (10.2-11.1)
PPI start within 1 y pre–IBD diagnosis
IBD	Case	74.3 (66.6-83.0)	6.99 (6.05-8.08)
	Control	10.6 (9.7-11.7)
CD	Rate	102 (88.2-118.1)	10.2 (8.33-12.6)
	Probability	10.0 (8.6-11.5)
UC	Rate	53.9 (45.5-63.8)	4.83 (3.92-5.95)
	Probability	11.2 (9.9-12.6)
PPI start within 1-2 y pre–IBD diagnosis
IBD	Case	29.3 (24.7-34.6)	2.77 (2.28-3.36)
	Control	10.6 (9.6-11.6)
CD	Case	35.6 (28.2-44.9)	4.05 (3.07-5.35)
	Control	8.8 (7.5-10.3)
UC	Case	24.4 (19.1-31.1)	2.03 (1.55-2.66)
	Control	12.0 (10.7-13.5)
PPI start within 4-5 y pre–IBD diagnosis
IBD	Case	16.2 (13.0-20.1)	2.61 (2.04-3.34)
	Control	6.2 (5.5-7.0)
CD	Case	18.0 (13.2-24.5)	3.48 (2.42-5.02)
	Control	5.2 (4.2-6.3)
UC	Case	14.8 (10.9-19.9)	2.11 (1.50-2.95)
	Control	7.0 (6.0-8.1)

Abbreviations: CD, Crohn’s disease; CI, confidence interval; IBD, inflammatory bowel disease; PPI, proton pump inhibitor; UC, ulcerative colitis.

Figure 1.

Percent of persons with diagnosed inflammatory bowel disease (cases) being given a proton pump inhibitor (PPI) at least 6 months prior to their diagnosis compared with control subjects between April 1996 and March 2018 in Manitoba, Canada.

Figure 2.

Odds ratio (OR) of incident cases of inflammatory bowel disease (IBD) using a proton pump inhibitor (PPI) between 30 days to 3 years prior to their IBD (Crohn’s disease [CD] or ulcerative colitis [UC]) diagnosis (Dx).

Figure 3.

Percent of persons with a diagnosed inflammatory bowel disease (IBD) of either Crohn’s disease (CD) or ulcerative colitis (UC) being given a proton pump inhibitor (PPI) at least 2 years prior to their diagnosis compared with control subjects in Manitoba, Canada. d, days.

Persons older than 65 years of age had higher PPI use at 42% in the 3 years prior to their diagnosis of IBD compared to 21% of control subjects (OR, 2.96; 95% CI, 2.46-3.57, P < .0001) (Table 3). Older persons with CD had higher PPI use, at 51%, compared with persons with UC, at 37% (P < .0001).

Table 3.

Percent of persons using a PPI prior to a diagnosis of IBD (CD or UC) who were given a 1095-day prescription prior to their IBD diagnosis, stratified by age.

	Case Group	Control Group	P Value for Case vs Control	CD	UC	P Value for CD vs UC
Cases	4407	33 904	—	1892	2515	—
Persons receiving any PPI prescription, %	28.1	11.7	<.0001	31.1	25.8	.0001
Persons receiving any PPI prescription in the 5 y prior to diagnosis of IBD, %	24.3	9.2	<.0001	27.6	21.8	<.0001
<18 y	10.6	0.9	<.0001	11.4	9.4	.55
18-39 y	17.4	4.8	<.0001	22.2	13.6	<.0001
40-64 y	29.6	11.5	<.0001	34.1	26.5	.0013
>65 y	41.9	21.0	<.0001	51.0	37.2	.0015

Abbreviations: CD, Crohn’s disease; IBD, inflammatory bowel disease; PPI, proton pump inhibitor; UC, ulcerative colitis.

Table 4 shows the use of PPIs post–IBD diagnosis compared with matched control subjects by age of diagnosis of IBD. Persons diagnosed with IBD as a child had the greatest increase in PPI use postdiagnosis compared with control subjects. Persons diagnosed with IBD after 65 years of age were more likely to use PPIs compared with matched control subjects but had the lowest increased likelihood compared with control subjects of all age groups.

Table 4.

Proportional hazards regression models of users.

Age at IBD Diagnosis	Group		PPI Users		HR (95% CI)
	Case	Control	Case	Control
All	4508	36 831	1420	5297	2.54 (2.39-2.71)
<18 y	552	5066	179	261	7.83 (6.36-9.63)
18-39 y	2094	17 146	540	1932	2.55 (2.30-2.82)
40-64 y	1446	11 636	530	2312	2.22 (2.01-2.45)
65+ y	416	2983	171	792	1.88 (1.57-2.25)
Drug-free at diagnosis and using PPI 1 y postdiagnosis
All	4058	36 129	1008	4895	1.94 (1.8-2.08)
<18 y	449	4975	83	249	3.72 (2.86-4.85)
18-39 y	1942	16 886	409	1823	2.04 (1.82-2.28)
40-64 y	1308	11 390	401	2132	1.82 (1.62-2.03)
65+ y	359	2878	115	691	1.45 (1.17-1.79)
Drug-free at diagnosis and using PPI 2 y postdiagnosis
All	3820	34 909	874	4502	1.84 (1.70-1.99)
<18 y	422	4751	74	233	3.49 (2.64-4.60)
18-39 y	1839	16 456	350	1697	1.88 (1.67-2.12)
40-64 y	1226	10 997	347	1966	1.71 (1.52-1.93)
65+ y	333	2705	103	606	1.52 (1.21-1.91)
Drug-free at diagnosis and using 5 y postdiagnosis
All	3078	30 115	534	3349	1.53 (1.39-1.69)
<18 y	340	4086	37	193	2.10 (1.45-3.03)
18-39 y	1514	14 441	226	1290	1.62 (1.40-1.88)
40-64 y	972	9339	214	1476	1.44 (1.24-1.67)
65+ y	252	2249	57	390	1.31 (0.96-1.79)

Values are n, unless otherwise indicated.

Abbreviations: CD, Crohn’s disease; CI, confidence interval; IBD, inflammatory bowel disease; PPI, proton pump inhibitor; UC, ulcerative colitis.

There were 4407 persons with IBD who were available to be included in a nested case-control study of persons with IBD who either used PPIs pre–IBD diagnosis (n = 1068) or did not. Table 5 shows the OR of having any of the outcomes that identify more aggressive disease. There was no difference in these outcomes whether or not persons with IBD were PPI users.

Table 5.

Outcomes suggesting aggressive disease among PPI users vs nonusers

Characteristic	HR (95% CI)
Any anti-TNF prescription after IBD diagnosis
IBD	1.09 (0.93-1.27)
CD	1.14 (0.94-1.38)
UC	1.01 (0.78-1.31)
At least 2 anti-TNF prescriptions after IBD diagnosis
IBD	1.12 (0.96-1.32)
CD	1.18 (0.97-1.43)
UC	1.04 (0.80-1.35)
Hospitalized for IBD
IBD	1.05 (0.90-1.22)
CD	1.06 (0.87-1.30)
UC	1.03 (0.82-1.31)
IBD-related surgery
IBD	1.02 (0.84-1.24)
CD	0.98 (0.77-1.25)
UC	1.11 (0.80-1.54)
At least 1 Entyvio prescription post–IBD diagnosis
IBD	0.97 (0.51-1.84)
CD	0.62 (0.20-1.88)
UC	1.27 (0.58-2.79)

Abbreviations: CD, Crohn’s disease; CI, confidence interval; HR, hazard ratio; IBD, inflammatory bowel disease; PPI, proton pump inhibitor; TNF, tumor necrosis factor; UC, ulcerative colitis.

PPI use is generally more common in females for cases and control subjects (Supplementary Table 1) and has been increasing in more recent years (Supplementary Table 2)

Discussion

In this retrospective analysis, we found that PPIs were 3 times more likely to be prescribed to persons with IBD in the 3 years prior to their IBD diagnosis than matched control subjects. There was an abrupt rise in PPI use within 6 months of the diagnosis of IBD, and this increase compared with control subjects persisted for at least 3 years postdiagnosis. Interestingly, persons with CD were more likely to have been PPI users prediagnosis than persons with UC. Children with IBD were more likely to be prescribed IBD prediagnosis than persons over 18 years of age, and persons diagnosed with IBD prior to 18 years of age were more likely to be using PPIs post–IBD diagnosis.

PPIs are acid suppressants used to reduce gastric acid secretions, thereby preventing and treating acid peptic disorders like gastroesophageal reflux and peptic ulcer disease. Over the last decade, the use of PPIs has increased. A retrospective analysis from a Danish Prescription Registry of 1 617 614 Danish adults found that from 2002 to 2014, the prevalence of PPI use increased 4-fold to 7.4%.¹² The Danish study also showed a strong age dependency, whereby 20% of adults 80 years of age or older used PPIs.¹² Similarly, in Manitoba, we found that over the last 15 years, PPI use for the general population has increased from 1.5% to 7.0%, while PPI use for persons with IBD has increased to 17%. Additionally, age was an important predictor of PPI use, as persons older than 65 years of age had a significantly higher use of PPIs.

Others have also shown an increased use of PPIs prior to IBD diagnosis. In a pooled analysis of the Nurses’ Health Study, PPI use showed a positive association with IBD compared with non-PPI users (HR, 1.42, 95% CI, 1.22-165). Use of PPIs prior to a diagnosis of IBD suggests that either IBD-related symptoms are being mistakenly treated with a PPI or that persons with IBD develop increased acid reflux–related symptoms. This may suggest that persons with IBD may develop impaired forward motility of the small bowel, thereby causing delayed gastric emptying and, in turn, gastroesophageal reflux. The increased PPI use prediagnosis that is accentuated in CD may be due to a higher likelihood of impaired small bowel motility. However, it is also possible that the increased use of PPIs in persons with IBD before diagnosis is secondary to doctors indiscriminately using PPIs to treat gastrointestinal symptoms, including abdominal pain, whether or not a definite acid peptic disorder is being diagnosed. Similarly, a case-control study found that PPIs were associated with an increased risk of pediatric IBD.¹³ In fact, our data suggest that nearly 11% of persons who were diagnosed under 18 years of age were PPI users within 1 to 3 years prior to their IBD diagnosis. A nested case-control study found that the OR for the association of at least 1 PPI or histamine H₂ blocker prescription and a diagnosis of IBD was 3.6 (95% CI, 1.1-11.7) for PPIs and 1.6 (95% CI, 0.7-3.7) for histamine H₂ blockers. However, for some it may be that PPI use prior to IBD diagnosis is treating acid-peptic symptoms, and for others it may be that PPI treatment is for gastrointestinal symptoms that in fact may not warrant a PPI. While the increased use of PPIs post–IBD diagnosis may reflect either of the previous 2 possibilities, it is also possible that the PPI prescription is simply being continued unnecessarily postdiagnosis. Persons using PPIs pre–IBD diagnosis should have the ongoing use of PPIs postdiagnosis reviewed to determine the appropriateness of their ongoing use.

Regardless as to why PPIs are prescribed at a higher rate pre–IBD diagnosis than in control subjects, it is possible that their use enhances the likelihood of an IBD diagnosis by their role in altering the gut microbiota.¹⁴ The acidic pH is the first line of defense against enteric pathogens.⁶ Loss of this first line by PPI use eliminating gastric acid might alter the downstream gut microbiota. This may be how PPIs can be a factor in contributing to inflammation and the development of mucosal lesions in IBD.¹⁵ A more altered microbiota has been found in CD compared with UC.¹⁵ The exact impact of PPI use on the gut microbiota in CD or UC has not been undertaken. Mechanistically, we propose that PPIs can change the gut microbiome.¹⁴ A double-blind randomized clinical trial comparing daily pantoprazole with placebo over a median of 3 years with >53 000 patients found no association with any diseases (including C. difficile infection), except for a higher risk of enteric infections (transmitted by ingestion of infected food or drink).¹⁶ In this trial, IBD was not identified as an adverse outcome after 3 years of PPI use. Nevertheless, a higher risk of enteric infections has been suggested to be an underlying mechanism in the paradigm of development of IBD, as it can impact the gut microbiome.⁸ Furthermore, it has been suggested that PPIs may result in a decreased abundance of Faecalibacterium and increased pathogenic species (eg, Escherichia coli, C. difficile), thereby increasing the risk of IBD.^17–21 We and others have shown that other factors may impact the gut microbiome²¹ and are a risk factor for development of IBD.^22–24

Post–IBD diagnosis, use of PPIs may further impact the disease course, either by making therapy less effective or through triggering disease relapse by microbiome alteration. A study by Lu et al²⁵ analyzed 5 randomized controlled trials and found that persons with IBD who were taking PPIs were less likely to achieve week 30 remission while on infliximab therapy (OR, 0.45; P < .001). Furthermore, adjusting for confounders by multivariable analysis and propensity score–matched analysis found that the week 30 remission rates were 30% and 49% in patients with and without PPI therapy, respectively (P < .001).²⁵ A case-control study using the Veterans Affairs Database found a modestly increased risk of IBD-related hospitalization and surgeries within 30 days of using a PPI in patients with UC (adjusted incidence density ratio, 1.18, 95% CI, 1.03-1.34) and CD (adjusted incidence density ratio, 1.12, 95% CI, 1.02-1.22).²⁶ Conversely, it has been suggested that PPIs (specifically omeprazole) might improve the clinical outcome of IBD and decrease the risk of pouchitis in UC.^27,28

While PPI use was increased prior to and post–IBD diagnosis, PPI users did not have a different rate of more aggressive disease than nonusers as measured by use of anti-tumor necrosis factor therapy, IBD-related hospitalizations, or IBD-related surgeries.

A strength of our study is that our data are with a well-identified cohort of persons with IBD and their matched control subjects. There are, however, some limitations. We assessed PPI prescriptions that were dispensed, but we could not monitor for adherence. Furthermore, because we could not be certain as to whether persons dispensed PPIs were using 1 per day or twice daily, we thus could not assess a dose-response relationship between PPI use and being subsequently diagnosed with IBD. Finally, we could not identify over-the-counter use of PPIs, but it is likely that their use should be proportionally similar among the cases and the matched control subjects. Further, over-the-counter use of PPIs in Canada is limited, and the majority of people obtain prescriptions for PPI use.

Conclusions

In summary, our retrospective analysis showed an increase in PPI use over the past 20 years among all Manitobans. There was a highly significant increase in the use of PPIs for 5 years prior to a diagnosis of IBD compared with control subjects. This may reflect an increase in acid peptic diseases in persons with IBD. Alternatively, it may be due to a prescription of PPIs for gastrointestinal symptoms that are generated by IBD, or the erroneous use of PPIs for gastrointestinal symptoms that in fact are not impacted on by PPIs. Regardless of the purpose for a PPI prescription, it is possible that PPI use enhances IBD progression by altering the gut microbiome. Future studies are needed to delineate whether the antecedent use of PPIs increases the risk for IBD, or alternatively whether persons with IBD have an increased risk of acid peptic diseases preceding their diagnoses. In this regard, we could not do a comparative study with use of histamine H₂ blockers, as the widespread over-the-counter use of these agents in Canada for many years would limit any assessment of their use by prescription data analysis. It is possible that PPIs are simply prescribed for symptoms that are IBD related, or perhaps PPI use is indiscriminate and should be reduced. Similarly, the ongoing use of PPIs post–IBD diagnosis needs to be explored, as to whether the indication for ongoing PPI use is appropriate.

Acknowledgments

The authors acknowledge the Manitoba Centre for Health Policy for use of the Manitoba Population Research Data Repository. The results and conclusions presented are those of the authors and no official endorsement by the Manitoba Centre for Health Policy, Manitoba Health, or other data providers is intended or should be inferred.

Author Contributions

Study concept and design (N.S., Z.N., C.N.B.). Data acquisition (Z.N., C.N.B.). Data analysis (N.S., ZS, H.S., S.S., C.N.B.). Manuscript drafting (N.S.). Final editing and approval of the manuscript (N.S., Z.N., H.S., S.S., C.N.B.).

Funding

This study was unfunded.

Conflicts of Interest

C.N.B. is supported by the Bingham Chair in Gastroenterology; has served on the advisory board for AbbVie Canada, Amgen Canada, Bristol Myers Squibb Canada, JAMP Pharmaceuticals, Roche Canada, Janssen Canada, Sandoz Canada, Takeda Canada, and Pfizer Canada; has served as a consultant for Mylan Pharmaceuticals and Takeda; has received educational grants from AbbVie Canada, Pfizer Canada, Takeda Canada, and Janssen Canada; has served on the speakers panel for AbbVie Canada, Janssen Canada, Medtronic Canada, and Takeda Canada; and has received research funding from AbbVie Canada and Pfizer Canada. N.S. has served on the advisory board or as a consultant for Pendopharm, Ferring Canada, Amgen Canada, Roche Canada, Sandoz Canada, Takeda Canada, Bristol-Myers Squibb Canada, and Guardant Health, Inc. S.R.S. has consulted for Takeda Canada. The other authors have nothing to declare.

Data Availability

The data underlying this article are available in the article and in its online supplementary material.