Conflicting data exist regarding the effects of statin therapy on the prevalence of inflammatory bowel diseases. We aimed to examine the association of statin therapy with diagnoses of inflammatory bowel diseases and noninfectious gastroenteritis. This is a retrospective study using data of a military health care system from October 1, 2003, to March 1, 2012. Based on medication fills during fiscal year 2005, patients were divided into: (1) statin users (received at least 90-day supply of statin) and (2) nonusers (never received a statin). A propensity score–matched cohort of statin users and nonusers was created using 80 variables. Primary analysis examined the risks of being diagnosed with inflammatory bowel diseases and noninfectious gastroenteritis between statin users and nonusers in the propensity score–matched cohort. Secondary analyses examined the risk of outcomes in the whole cohort and in patients with no comorbidities according to Charlson Comorbidity Index. Of 43,438 patients meeting study criteria (13,626 statin users and 29,812 nonusers), we propensity score matched 6,342 statin users with 6,342 nonusers. For our primary analysis, 93 statin users and 92 nonusers were diagnosed with inflammatory bowel diseases (odds ratio = 1.01, 95% confidence interval = 0.76 to 1.35), and 632 statin users and 619 nonusers were diagnosed of noninfectious gastroenteritis (odds ratio = 1.02, 95% confidence interval = 0.91 to 1.15). In conclusion, the risks of inflammatory bowel diseases and noninfectious gastroenteritis among statin users and nonusers are similar after adjusting for other potential confounding factors.
Statins (3-hydroxy-3-methyl-glutaryl-CoA reductase inhibitors) exhibit anti-inflammatory and antifibrotic effects, which might be beneficial in inflammatory bowel diseases (IBD). In a model of experimentally induced colitis, simvastatin ameliorated inflammation and resulted in a dose-dependent decrease in the level of a fibrosis-related growth factor. In another experimental model, simvastatin inhibited proinflammatory gene expression in intestinal epithelial cells and attenuated dextran sulfate sodium-induced colitis. In an observational study, patients with IBD on statins had lesser use of oral steroids and a reduced incidence of surgery for IBD. Additionally, in an open-label study of 10 patients with Crohn’s disease, the addition of atorvastatin 80 mg for 23 weeks resulted in significant decrease in inflammatory markers. In contrast, case reports have associated fatal colitis with stain use. Additionally, some investigators argued that statins might cause mitochondrial dysfunction and proinflammatory effects, which may induce autoimmune diseases. The objective of this study was to examine the association of statin use on the prevalence of IBD and noninfectious gastroenteritis (NI-GE) in a cohort of patients who were followed longitudinally in a military health care system, where patients had similar access and availability of health care.
Methods
This study was approved by the Institutional Review Boards of Brooke Army Medical Center and VA North Texas Healthcare System. This is a retrospective cohort study including all patients aged 30 to 85 years enrolled in the San Antonio Military Multi-Market Area as Tricare Prime or Plus. We retrieved the medical encounters and their associated medication fill histories, diagnoses, and procedure codes data, from October 1, 2003, to March 1, 2012, using the Military Health System Management Analysis and Reporting Tool (M2). The reliability and reproducibility of M2 Data are well documented. Each patient in M2 is uniquely identified by an identification code that is consistent throughout the data. These data encompass the full spectrum of all clinical and administrative data integrated with eligibility and enrollment data including outpatient electronic medical records, inpatient electronic medical records, medical benefits claims data, laboratory data performed within military facilities, and pharmacy data regardless of pharmacy location or affiliation.
Our study duration was divided into 2 periods: baseline period and follow-up period. Baseline period included the period from October 1, 2003, to September 30, 2005, and was used for description of baseline characteristics. The follow-up period included the period from October 1, 2005, to March 1, 2012, and was used for examining outcomes of patients. All patients were enrolled in the system throughout the study period; hence, we have no missing data.
Inclusion criteria were detailed in a previous publication ; our cohort included all patients aged 30 to 85 years who were (1) enrolled in Tricare Prime or Plus in the San Antonio Multi-Market Area; (2) had at least 1 medical encounter during the baseline period and 1 during the follow-up period; and (3) had at least 1 prescription medication during the baseline period.
We excluded the following groups: (1) patients who started on statins at the end of baseline period (after September 30, 2005), thus allowing us to have 2 treatment groups of statin users and nonusers with equal periods of follow-up; (2) patients who received statins for <90 days during the study period; and (3) patients who were diagnosed with trauma or burn throughout the study period because these patients’ outcomes were likely affected by the extent of their trauma or burn rather than statin use. Diagnoses with trauma or burn were based on the International Classification of Diseases, Ninth Revision, Clinical Modification [ ICD-9-CM ] codes as defined by the Agency for Healthcare Research and Quality Clinical Classifications Software (AHRQ-CCS), category 240, and previous publications.
Patients who met the study criteria were divided into 2 treatment groups: (1) statin users: patients who were on statin medication for a cumulative period of at least 90 days including the period from October 1, 2004, to September 30, 2005, and (2) nonusers: patients who never received statin from October 1, 2003, to March 1, 2012.
We identified ICD-9 codes consistent with outcomes of interest during the follow-up period. Our outcome diagnosis groups were defined as the followings: (1) IBD outcome group: this group included all ICD-9-CM codes of regional arteritis and ulcerative colitis as defined by AHRQ-CCS, category 144 (5550, 5551, 5552, 5559, 556, 5560, 5561, 5562, 5563, 5564, 5565, 5566, 5568, and 5569); and (2) NI-GE outcome group: this group included ICD-9-CM codes of noninfectious gastroenteritis as defined by the AHRQ-CCS, category 154 (55841, 55842, and 5589).
Patients’ comorbidities were characterized using the Charlson Comorbidity Index (CCI), Deyo’s method, and several diagnostic categories of AHRQ-CCS. A logistic regression model was created to develop the propensity score and test the balance of covariates. Matching with a caliper of 0.001 was then performed. We included 80 candidate variables in the propensity score that we believed would either be potentially associated with using statins or the outcomes of interest. These variables included patient age, gender, 17 comorbidities that constituted all components of CCI using Deyo’s method, and total CCI ; the presence of the following disease categories as defined by AHRQ-CCS : hypertension, hypertension with complications, chest pain, acute myocardial infarction, dysrhythmia, valvular heart disease, coronary artery disease, heart disease not otherwise specified, cardiovascular conduction disorder, congestive heart failure, cardiomyopathy, vascular aneurysm, arterial thromboembolism, peripheral vascular disease, cardiac arrest/ventricular fibrillation, cerebrovascular disease, acute cerebrovascular disease, asthma, chronic obstructive lung disease, smoking, cor pulmonale, respiratory failure, diabetes mellitus, diabetes mellitus with complications, gastrointestinal ulcers, gastritis/duodenitis, gastrointestinal hemorrhage, acute kidney injury, chronic kidney disease, rheumatoid arthritis/systemic lupus, obesity, psychosis, metastatic neoplasm, osteoarthritis, pathologic fracture, alcohol abuse/dependence, illicit drug use, rehabilitation services, suicide, number of both inpatient and outpatient medical encounters during the baseline period as a measure for health care utilization, number of encounters for immunization as a surrogate marker for health conscious behavior, and use of 20 classes of medications.
We compared the risk of our outcome groups (IBD and NI-GE) between statin users and nonusers in the propensity score–matched cohort. We also determined the risk of outcomes in statin users and nonusers in the following groups: (1) All-patients cohort: all patients who met inclusion and exclusion criteria. Using logistic regression models, we examined the risk of each outcome as a dependent variable, treatment group as an independent variable, and propensity score as a covariate. (2) No-Charlson comorbidity cohort: patients from the “All-patients cohort” with any positive element of their CCI were excluded. Using a logistic regression model, as described earlier, we examined the risk of outcomes while adjusting for propensity score.
Baseline characteristics of statin users and nonusers were compared using appropriate 2-way tests (chi-square for categorical variables and Student’s t for continuous variables). Continuous variables were summarized by the mean and SD. For the propensity score–matched analyses, we presented the outcomes analyzed using the chi-square tests and odds ratios using conditional logistic regression analysis. For secondary analyses, multivariable logistic regression analysis was used to examine the risks of outcomes; each outcome measure was independently examined as a dependent variable, and statin utilization as a predictor variable, while adjusting for propensity score. Comparisons were considered to be statistically significant with a 2-tailed p value of ≤0.05. Statistical analyses were performed using STATA 12 (Stata Statistical Software: Release 12, 2011; StataCorp., College Station, Texas) and SPSS Statistical Software, version 21 (SPSS Statistics for Windows; IBM Corp., Armonk, New York).
Results
A total of 43,438 patients met study inclusion and exclusion criteria (13,626 statin users and 29,812 nonusers). The mean (SD) cumulative duration of statin use was 4.65 (1.82) years. Statin users received prescriptions for various forms and dosages of the drugs during the study period. Overall, the statins prescribed were simvastatin (73.5%), atorvastatin calcium (17.4%), pravastatin sodium (7.0%), rosuvastatin calcium (1.7%), and fluvastatin sodium or lovastatin (0.2%).
The propensity score–matched cohort included 12,684 patients (6,342 statin users and 6,342 nonusers). There were no statistically significant differences in baseline characteristics between the 2 groups after matching ( Table 1 ). Of statin users, 5,276 (83.2%) continued to take their statins for ≥2 years and 3,887 (61.3%) for 4 years. The No-Charlson comorbidity cohort constituted 30,127 patients: 5,761 statin users and 24,366 nonusers. Statin users were older, had a higher proportion of men, obesity, and smokers, had more frequent use of most medication classes, and had higher health care utilization ( Table 2 ).
| Variable | Statin users | P value | |
|---|---|---|---|
| NO (n = 6342) | YES (n = 6342) | ||
| Age (year), mean ± SD | 56.0 ± 12.0 | 55.7 ± 12.4 | 0.12 |
| Men | 3486 (55.0%) | 3418 (53.9%) | 0.23 |
| Charlson Comorbidity Index components during baseline period ∗ | |||
| Chronic obstructive lung disease | 771 (12.2%) | 780 (12.3%) | 0.81 |
| Diabetes mellitus | 544 (8.6%) | 581 (9.2%) | 0.26 |
| Diabetes mellitus with complications | 449 (7.1%) | 454 (7.2%) | 0.86 |
| Malignancy | 365 (5.8%) | 384 (6.1%) | 0.48 |
| Cerebrovascular disease | 176 (2.8%) | 185 (2.9%) | 0.63 |
| Rheumatologic diseases | 142 (2.2%) | 136 (2.1%) | 0.72 |
| Peripheral vascular disease | 116 (1.8%) | 144 (2.3%) | 0.08 |
| Congestive heart failure | 108 (1.7%) | 117 (1.8%) | 0.55 |
| Kidney disease | 77 (1.2%) | 83 (1.3%) | 0.63 |
| Peptic ulcer disease | 53 (0.8%) | 57 (0.9%) | 0.70 |
| Acute myocardial infarction | 37 (0.6%) | 48 (0.8%) | 0.23 |
| Metastatic neoplasm | 31 (0.5%) | 25 (0.4%) | 0.43 |
| Mild liver disease | 28 (0.4%) | 28 (0.4%) | 1.00 |
| Dementia | 20 (0.3%) | 21 (0.3%) | 0.88 |
| Hemiplegia/paraplegia | 11 (0.2%) | 14 (0.2%) | 0.56 |
| HIV | 7 (0.1%) | 9 (0.1%) | 0.80 |
| Liver disease (moderate/severe) | 4 (0.1%) | 5 (0.1%) | 0.75 |
| Charlson Comorbidity Index total score: mean (SD) | 0.64 ± 1.23 | 0.66 ± 1.25 | 0.29 |
| Obesity † | 993 (15.7%) | 960 (15.1%) | 0.43 |
| Smoker | 534 (8.4%) | 509 (8.0%) | 0.44 |
| Alcohol abuse/dependence ‡ | 83 (1.3%) | 78 (1.2%) | 0.69 |
| Illicit drug use ‡ | 24 (0.4%) | 22 (0.3%) | 0.77 |
| Hypertension ‡ | 3766 (59.4%) | 3707 (58.5%) | 0.30 |
| Hypertension with complications ‡ | 176 (2.8%) | 181 (2.9%) | 0.79 |
| Valvular heart disease ‡ | 402 (6.3%) | 403 (6.4%) | 0.97 |
| Asthma ‡ | 375 (5.9%) | 366 (5.8%) | 0.73 |
| Coronary artery disease ‡ | 277 (4.4%) | 314 (5.0%) | 0.12 |
| Gastritis/duodenitis | 216 (3.4%) | 215 (3.4%) | 1.00 |
| Rheumatoid arthritis/systemic lupus ‡ | 172 (2.7%) | 156 (2.5%) | 0.37 |
| Peripheral vascular disease ‡ | 153 (2.4%) | 169 (2.7%) | 0.37 |
| Cardiovascular conduction disorder ‡ | 130 (2.0%) | 137 (2.2%) | 0.71 |
| Respiratory failure ‡ | 30 (0.5%) | 32 (0.5%) | 0.80 |
| Pathologic fracture ‡ | 25 (0.4%) | 28 (0.4%) | 0.68 |
| Suicide ‡ | 4 (0.1%) | 3 (0.00%) | 0.73 |
| Number of outpatient visits during baseline period: mean ± SD | 31.67 ± 36.76 | 31.81 ±40.63 | 0.84 |
| Number of inpatient admissions during baseline period: mean ± SD | 0.25 ± 0.75 | 0.26 ± 0.77 | 0.75 |
| Number of encounters for immunization during baseline period: mean ± SD | 0.48 ± 1.60 | 0.49 ± 3.71 | 0.75 |
| Rehabilitation care ‡ | 1227 (19.3%) | 1205 (19.0%) | 0.62 |
| Medications: | |||
| Non-statin lipid lowering drugs | 373 (5.9%) | 391 (6.2%) | 0.52 |
| NSAID | 3729 (58.8%) | 3702 (58.4%) | 0.64 |
| ACE/ARB | 2137 (33.7%) | 2141 (33.8%) | 0.96 |
| Proton pump inhibitor | 2009 (31.7%) | 2030 (32.0%) | 0.70 |
| Aspirin | 1835 (28.9%) | 1890 (29.8%) | 0.28 |
| Diuretic | 1740 (27.4%) | 1718 (27.1%) | 0.68 |
| Sedatives | 1221 (19.3%) | 1243 (19.6%) | 0.62 |
| Beta-blocker | 1099 (17.3%) | 1123 (17.7%) | 0.58 |
| SSRI | 1059 (16.7%) | 1067 (16.8%) | 0.87 |
| Calcium channel blocker | 987 (15.6%) | 1001 (15.8%) | 0.75 |
| Bisphosphonate | 573 (9.0%) | 536 (8.5%) | 0.26 |
| Cytochrome p450 | 386 (6.1%) | 414 (6.5%) | 0.32 |
| Systemic corticosteroid | 257 (4.1%) | 249 (3.9%) | 0.75 |
| Oral hypoglycemic | 218 (3.4%) | 254 (4.0%) | 0.10 |
| Antipsychotic | 85 (1.3%) | 84 (1.3%) | 1.00 |
| Mean LDL-C in baseline period in mg/dL: mean ± SD ‡ | 107 ± 27 | 118 ± 34 | < 0.001 |
| Mean HDL-C in baseline period in mg/dL: mean ± SD § | 59 ± 19 | 55 ± 15 | < 0.001 |
| Mean Triglycerides in baseline period in mg/dL: mean ± SD § | 129 ± 91 | 150 ± 88 | < 0.001 |
| Mean LDL-C in follow-up period in mg/dL: mean ± SD § | 108 ± 26 | 110 ± 32 | < 0.001 |
| Mean HDL-C in follow-up period in mg/dL: mean ± SD § | 57 ± 18 | 53 ± 14 | < 0.001 |
| Mean Triglycerides in follow-up period in mg/dL: mean ± SD § | 120 ± 73 | 142 ± 77 | < 0.001 |
∗ Diagnosis is based on ICD-9-CM codes as identified in Deyo method for applying the Charlson Comorbidity Index.
† Diagnosis is based on selected ICD-9-CM diagnosis codes from category 56 of the Agency of Health Research and Quality-Clinical Classification Software (other nutritional; endocrine; and metabolic disorders) related to overweight, obesity and hyperalimentation (codes: 2780, 27800, 27801, 27802, 27803, 2781, 2788, and 7831).
‡ As defined by the Agency for Health Research and Quality-Clinical Classifications Software.
§ Lipid measurements represent the mean value for each patient throughout the baseline or follow-up periods; these laboratory measurements were available only for those who had their laboratory tests performed at the military treatment facilities under study (3,879 statin users and 3,320 nonusers). Lipid parameters were not included in propensity score matching.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
