The incidence of both asymptomatic and symptomatic colonic diverticulosis is increasing.1 However, only 10–20% of patients with diverticulosis develop diverticulitis.2 Factors such as physical inactivity, constipation, obesity, smoking, and the use of nonsteroidal anti-inflammatory drugs (NSAIDs) have been associated with an increased risk of diverticular disease.3
Proton pump inhibitors (PPIs) are primary agents for the treatment of gastroduodenal ulcers and gatroesophageal reflux disease and are effective in preventing NSAID/aspirin-associated peptic ulcer and ulcer bleeding.4 PPIs are among the most frequently prescribed classes of medications worldwide because of their high rate of efficacy and low toxicity.5 However, PPIs have recently been reported to produce adverse effects.6–8 PPIs can interact with targets other than the gastric H+/K+ ATPase. Colonic epithelial cells, which are involved in maintaining local electrolyte balance, express proton pumps in the colon.9 PPIs also influence microbial growth by inhibiting the colonic H+–K+ ATPase.10
Bacterial colonization by exogenous enteric microbes is inhibited by host defense mechanisms such as gastric acid, host gut microflora, local gut immunity, intestinal motility, intestinal secretion, and the epithelial barrier.11 PPI use4,12 and colonic diverticulitis13 are both associated with bacterial enteric infection and translocation. The relationship between PPI use and diverticulitis is not well documented. Therefore, we assessed the association between PPI use and colonic diverticulitis by conducting a nested case-control study based on the National Health Insurance Research Database (NHIRD) in Taiwan.
2.1. Ethics statement
All data that could be used to identify patients were encrypted in the National Health Insurance (NHI) files used in our study. The confidentiality of the data presented in our study, which was approved by the Taiwan's National Health Research Institute (NHRI), adheres to the regulations of the Bureau of National Health Insurance (BNHI), Taiwan. This study was also approved by the Institutional Review Board (IRB) of Taipei City Hospital (IRB No. TCHIRB-1021103-E). Written consent was waived by the IRB.
2.2. Data source
This nationwide cohort study was based on patient data obtained from the NHIRD, which is managed by the Taiwan NHRI. The NHIRD contains healthcare data for 99% of Taiwan's population (approximately 23 million).14 The sample files of NHIRD contained comprehensive enrollment and drug-use information for a randomly selected sample of 1,000,000 NHI beneficiaries, representing approximately 5% of all enrollees in the year 2000. We used the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) to define diseases.
2.3. Identification of cases and control
We conducted a population-based nested case-control study. Diverticulitis patients were identified from the NHIRD by using inpatient discharge records based on ICD-9-CM codes (562.11 and 562.13) following computed tomography, magnetic resonance imaging, colonoscopy, and barium radiological examination between January 1, 2000 and December 31, 2010. Patients under 20 years of age and patients with prior colonectomy, celiac disease, and inflammatory bowel disease were excluded. We also excluded patients diagnosed with colonic cancer between January 1, 1996 and the index date. Control group patients did not have the code ICD-9-CM: (562.xx) in their inpatient records or in the ambulatory-care claims. Four control patients were selected using data from inpatient medical databases to match each newly recorded colonic diverticulitis patient by conducting random sampling stratified by age, sex, NSAID use, and laxative use in the same observational period. Fig. 1 illustrates a flowchart of the patient selection process.
2.4. Determination of medication exposure for PPIs
2.4.1. Exposure assessment
Information on the prescribed drugs was extracted from the NHRI prescription database, and the defined daily doses (DDDs) recommended by the World Health Organization were used to quantify the use of PPIs.15 The cumulative DDD was estimated as the sum of the dispensed DDD of any PPI (omeprazole, lansoprazole, pantoprazole, rabeprazole, or esomeprazole) with the final dose taken within the specified period prior to the index date. The gathered data comprised the date of prescription, daily dosage, and the number of days of drug use.
2.4.2. Drug exposure data
We defined current users as patients exposed to PPIs for 28 days prior to the claimed date of hospitalization for acute diverticulitis. We defined previous users as patients who experienced PPI exposure between January 1, 2000 and the claimed date of hospitalization for colonic diverticulitis, excluding the current users. Patients with no prescription for any PPIs at any time between January 1, 2000 and the claimed date of hospitalization for colonic diverticulitis were defined as nonusers.
2.5. Assessment of cofounders and covariates
2.5.1. Comorbid illness
Conditions that required inpatient care or three or more ambulatory-care visits between January 1, 1996 and the index date of the study were defined as comorbidities. The comorbidities identified in our cohort and their corresponding ICD-9-CM diagnosis codes are listed as follows: coronary artery disease (ICD-9-CM 410–414), cerebral vascular disease (ICD-9-CM 430–438), dyslipidemia (ICD: 272.xx), chronic obstructive pulmonary disease (ICD-9-CM 490–492, 494, and 496), diabetes mellitus (DM; ICD-9-CM 250), and peptic ulcers (ICD-9-CM 531, 532, and 533).
2.5.2. Use of other medications
Patients were defined as users of NSAIDs, laxatives, aspirin, cyclooxygenase-2-specific inhibitors, steroids, and antidiarrhea drugs based on whether they had consumed the prescribed medication within 28 days prior to the claimed date of hospitalization for colonic diverticulitis.
2.6. Statistical analysis
For comparing the proportions, chi-square statistics was used. A conditional logistic regression model was used to estimate the relative magnitude in relation to the use of PPIs. Exposure was defined as receiving PPIs between January 1, 2000 and the index date. Patients were categorized into one of the three PPIs exposure categories for analysis: nonusers, previous users, and current users. In the analysis, participants were categorized into PPI use below the median (<42 or <55 DDD) and PPI use equal or above the median (≥42 or ≥55 DDD) over an exposure period of 90 and 180 days, respectively, prior to the claimed date of hospitalization for colonic diverticulitis. Patients with no exposure were used as the reference for calculating odds ratios (ORs) and their 95% confidence intervals (CIs). Models with and without adjusted variables were presented. All statistical analyses were performed using the SAS System for Windows, version 9.2 (SAS Institute, Cary, NC, USA).
We enrolled 690 patients with acute diverticulitis, as well as 2760 patients who comprised the control group, based on NHIRD records from 2000 to 2010. Table 1 displays the demographic data including age, sex, comorbidities, and medications. Patients with colonic diverticulitis had a considerably low rate of dyslipidemia, but increased rates of peptic ulcer, aspirin use, cyclooxygenase-2 inhibitor use, and antidiarrhea drug use.
Compared with PPI nonusers, the crude ORs were 1.12 (95% CI = 0.69–1.84) and 1.34 (95% CI = 1.08–1.67) for current users and previous users, respectively. However, after adjusting for possible confounders (including sex, age, comorbidities, and medication), the adjusted ORs were 0.90 (95% CI = 0.48–1.72) and 1.01 (95% CI = 0.84–1.43) for the current users and previous users, respectively (Table 2). When PPI use was categorized by cumulative dosage, the crude ORs were 0.99 (95% CI = 0.60–1.62) and 1.10 (95% CI = 0.73–1.67) for the groups with cumulative PPI use of <42 and <55 DDDs, respectively, and 1.67 (95% CI = 1.04–2.68) and 1.40 (95% CI = 0.91–2.17) for the groups with cumulative PPI use of ≥42 and ≥55 DDDs over an exposure period of 90 and 180 days, respectively, prior to the claimed date of hospitalization for acute diverticulitis, compared with PPI nonusers. However, after adjusting for possible confounders (including sex, age, comorbidities, and medication), the adjusted ORs were 0.79 (95% CI = 0.44–1.42) and 0.93 (95% CI = 0.56–1.53) for the groups with cumulative PPI use of <42 and <55 DDDs, respectively, and 1.29 (95% CI = 0.70–2.36) and 1.02 (95% CI = 0.59–1.76) for the groups with cumulative PPI use of ≥42 and ≥55 DDDs over an exposure period of 90 and 180, days, respectively, prior to the claimed date of hospitalization for colonic diverticulitis (Table 2). In addition, values of 1.28 (95% CI = 1.01–1.61) and 1.27 (95% CI = 1.01–1.60) for DM patients, 1.30 (95% CI = 1.04–1.61) and 1.31 (95% CI = 1.05–1.63) for peptic ulcer patients, and 1.47 (95% CI = 1.22–1.77) and 1.47 (95% CI = 1.22–1.78) for antidiarrhea medication users, were recorded over an exposure period of 90 and 180 days, respectively, prior to the claimed date of hospitalization for colonic diverticulitis (Table 3).
This clinical epidemiological study is the first of its kind that focuses exclusively on inpatients with colonic diverticulitis who receive PPIs. The nationwide population-based, nested case-control study did not reveal that use of PPI significantly increased the risk of colon diverticulitis after adjustment for possible confounding factors. Factors such as constipation and NSAIDs have been associated with an increased risk of colonic diverticulitis.3 Therefore, four control patients were selected to match colonic diverticulitis case stratified by laxative drugs use and NSAID use.
Current long-term PPI therapy has been associated with the induction of severe hypomagnesemia, iron and vitamin B12 deficiency, community-acquired pneumonia, hip fracture, and interstitial nephritis.16 Elevation in gastric pH levels may exert deleterious effects on the gastrointestinal host defense including delayed gastric emptying, increased bacterial translocation, decreased gastric mucus viscosity, and changes in the normal microbial flora.17 The pathogenesis of diverticulitis involves an altering of the gut flora13 and a stasis or obstruction in the narrow-necked pseudodiverticulum leading to bacterial overgrowth.18 Suppressing gastric acidity may therefore result in an increased load of pathogenic microbes in the gastrointestinal tract. However, our current study addresses the effect of PPIs on colonic diverticulitis and it remains a null association. PPIs have been implicated in the disruption of the ecology of the gut and changes in bacterial growth.19 Janarthanan et al20 conducted a meta-analysis and suggested that PPIs increase the incidence of Clostridium difficile infection. The case-control study conducted by Garcia Rodriguez et al21 revealed an increased incidence of Salmonella and Campylobacter jejuni gastroenteritis in patients taking PPIs. People with comorbidities (especially elderly patients) are highly likely to be prescribed acid-suppressive therapy.20 The role of PPI dosage and duration of use in immunocompetent and immunocompromised patients who have contracted varying levels of acute and recurrent diverticulitis requires investigation. Further study to evaluate whether use of PPI increased the risk of colon diverticulitis is warranted, especially in subgroup patients with risks or immune-compromised patients.
The clinical process of colonic diverticulitis varies in accordance with the extent of the disease's progress. Because only 10–20% of patients with diverticulosis develop diverticulitis, proving that these patients have recurrent asymptomatic diverticulitis might be difficult. Therefore, we adopted a strict definition of the control group, in which a patient cannot exhibit ICD-9-CM code 562.xx in their inpatient records or in the ambulatory care claims between 2000 and 2010. To be included in the study group, patients were identified using ICD-9-CM codes following computed tomography, magnetic resonance imaging, colonoscopy, and barium radiological imaging (excluding 765 cases without image examination). Although imaging reports are unavailable in the NHIRD, the Bureau of the NHI randomly samples the claims data from every hospital and reviews charts to verify diagnostic validity. ICD-9-CM coding was strictly audited for the purpose of reimbursement.
Our observations are consistent with those of previous reports regarding the role of DM in diverticulitis, but no such association was observed for diverticulitis in steroid users. This study showed that DM is an independent risk factors of colonic diverticulitis, which is in agreement with Floch's22 study. However, Humes et al23 revealed that perforated diverticular disease is a severe surgical emergency, and oral corticosteroids (OR = 2.74; 95% CI = 1.63–4.61) are strongly associated with an increased risk of perforation. By contrast, our data suggested that steroid use did not increase the risk of acute diverticulitis, which is dissimilar to the result reported by Humes et al.23 The primary discrepancy is the enrolled patient criteria. Further study is required to elucidate the role of corticosteroids, including duration and dose in immune-competent and immunocompromised patients, diagnosed with varying extents of diverticulitis.
This study had several limitations. First, factors such as patients’ lifestyles, dietary habits (especially fiber intake), body mass indices, and smoking were not available for adjustment.24–26 Second, certain data, such as the location of diverticulitis, are unavailable in the NHIRD. Left-sided diverticular disease is most common in Western societies, whereas right-sided diverticular disease is more prevalent in Asia. In addition, colonic diverticular disease is not as common in Asian societies as in Western societies, so the results cannot be applied to western societies. Third, the consumption of PPIs could not be confirmed in this study. However, the Bureau of the NHI has formed various audit committees that randomly sample claims data from every hospital and review charts on a regular basis to verify the diagnostic validity and quality of care. Fourth, the causality of this phenomenon has not been established by any randomized placebo-controlled clinical trials. Determining causality is difficult because neither withholding PPI therapy from those who require it, nor administering PPIs to those who do not require them are ethically feasible.20 Therefore, we conducted a nested case-control study by using data from the Taiwan NHIRD. In Taiwan's NHI program, the use of PPIs is strictly limited to treating patients with reflex esophagitis or peptic ulcer disease for 4 months after gastroscopic confirmation.27 Therefore, we analyzed the cumulative dose effect of PPIs on acute diverticulitis for an exposure period of 90 and 180 days prior to the claimed date of hospitalization for colonic diverticulitis. Fifth, hospitalization is indicated if the patient is unable to tolerate oral intake or if the patient has acute diverticulitis. Establishing a standard definition is crucial; therefore, we analyzed only the cases requiring hospitalization for colonic diverticulitis by using data from inpatient medical databases. The actual prevalence rate of diverticulosis is difficult to determine because most patients are asymptomatic. Therefore, we conducted the current study by using the nested case-control method instead of the cohort study method. Finally, use of over-the-counter oral PPI medications was not included in our database.
In conclusion, the study showed that use of PPIs did not increase the risk of colon diverticulitis. Prospective studies are necessary to confirm this association in the general population and explore whether this risk can be reversed by avoiding unnecessary PPI use and administering the optimal PPI DDD.
We thank the Taiwan Ministry of Education for supporting this work through their “Aim for the Top University Plan.” In addition, we thank the National Health Insurance Bureau and the NHRI for making the NHIRD available to us for this study. We also thank the anonymous reviewers and editors for their comments.
1. Delvaux M. Diverticular disease of the colon in Europe: epidemiology, impact on citizen health and prevention. Aliment Pharmacol Ther
2. Almy TP, Howell DA. Medical progress. Diverticular disease of the colon. N Engl J Med
3. Stollman NH, Raskin JB. Diverticular disease of the colon. J Clin Gastroenterol
4. Luo JC. Gastroprotective strategy in aspirin users. J Chin Med Assoc
5. Forgacs I, Loganayagam A. Overprescribing proton pump inhibitors
6. Leontiadis GI, Yuan Y, Howden CW. The interaction between proton pump inhibitors
and clopidogrel and upper gastrointestinal bleeding. Gastrointest Endosc Clin N Am
7. Chen J, Yuan YC, Leontiadis GI, Howden CW. Recent safety concerns with proton pump inhibitors
. J Clin Gastroenterol
8. Moayyedi P, Leontiadis GI. The risks of PPI therapy. Nat Rev Gastroenterol Hepatol
9. Rechkemmer G, Frizzell RA, Halm DR. Active potassium transport across guinea-pig distal colon: action of secretagogues. J Physiol (Lond)
10. Vesper BJ, Jawdi A, Altman KW, Haines GK 3rd, Tao L, Radosevich JA. The effect of proton pump inhibitors
on the human microbiota. Curr Drug Metab
11. DuPont AW, DuPont HL. The intestinal microbiota and chronic disorders of the gut. Nat Rev Gastroenterol Hepatol
12. Canani RB, Terrin G. Gastric acidity inhibitors and the risk of intestinal infections. Curr Opin Gastroenterol
13. Szojda MM, Cuesta MA, Mulder CM, Felt-Bersma RJ. Management of diverticulitis. Aliment Pharmacol Ther
14. Cheng SH, Chiang TL. The effect of universal health insurance on health care utilization in Taiwan. Results from a natural experiment. JAMA
15. WHO Collaborating Center for Drugs Statistics Methodology: ATC Index with DDDs 2003.
Oslo: WHO; 2003.
16. Ali T, Roberts DN, Tierney WM. Long-term safety concerns with proton pump inhibitors
. Am J Med
17. Dial MS. Proton pump inhibitor use and enteric infections. Am J Gastroenterol
18. Jacobs DO. Clinical practice. Diverticulitis. N Engl J Med
19. Choung RS, Ruff KC, Malhotra A, Herrick L, Locke GR 3rd, Harmsen WS, et al. Clinical predictors of small intestinal bacterial overgrowth by duodenal aspirate culture. Aliment Pharmacol Ther
20. Janarthanan S, Ditah I, Adler DG, Ehrinpreis MN. Clostridium difficile
-associated diarrhea and proton pump inhibitor therapy: a meta-analysis. Am J Gastroenterol
21. Garcia Rodriguez LA, Ruigomez A, Panes J. Use of acid-suppressing drugs and the risk of bacterial gastroenteritis. Clin Gastroenterol Hepatol
22. Floch CL. Diagnosis and management of acute diverticulitis. J Clin Gastroenterol
23. Humes DJ, Fleming KM, Spiller RC, West J. Concurrent drug use and the risk of perforated colonic diverticular disease: a population-based case-control study. Gut
24. Strate LL, Modi R, Cohen E, Spiegel BM. Diverticular disease as a chronic illness: evolving epidemiologic and clinical insights. Am J Gastroenterol
25. Bose KP, Khorshidi I, Southern WN, Brandt LJ. The impact of ethnicity and obesity on the course of colonic diverticulitis. J Clin Gastroenterol
26. Humes DJ, Spiller RC. The pathogenesis and management of acute colonic diverticulitis. Aliment Pharmacol Ther
27. Luo JC, Leu HB, Hou MC, Huang CC, Lin HC, Lee FY, et al. Incidence of bleeding from gastroduodenal ulcers in patients with end-stage renal disease receiving hemodialysis. CMAJ