Non-steroidal Anti-inflammatory Drugs Use and Risk of Upper Gastrointestinal Adverse Events in Cirrhotic Patients

Yen-Chieh Lee; Chia-Hsuin Chang; Jou-Wei Lin; Hsi-Chieh Chen; Min-Shung Lin; Mei-Shu Lai


Liver International. 2012;32(5):859-866. 

In This Article


The study protocol was approved by the National Taiwan University Hospital Research Ethics Committee. We conducted a retrospective case-crossover study to investigate the association between NSAIDs use and upper GI adverse outcome among cirrhotic patients. We have used this design to examine NSAIDs lower and upper GI safety in the general population.[14,15] Instead of using matched controls which are difficult to find in the evaluation of NSAID toxicity in cirrhotic patients, the case-crossover design uses past experience of the case as case's own control. The major advantage of this design is that stable risk factors such as cigarette smoking, alcohol consumption, or GI distress symptoms, despite they are not recorded in the claims database, will not confound the analysis results.[16,17]

Data Source

The Taiwan National Health Insurance database includes complete outpatient visits, hospital admissions, prescriptions, disease and vital status for 99% of the population (about 23 millions) in Taiwan. We established the longitudinal medical history of each beneficiary by linking several computerized administrative, claims datasets and National Death Registry through the civil identification number unique to each beneficiary and date of birth.

Study Population

Firstly, we searched the Taiwan National Health Insurance Database for the source population to identify all hospitalized patients aged ≥ 20 years with a primary or secondary diagnosis for an upper GI adverse event in 2006 based on International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes. The diagnostic codes for upper GI adverse events included gastric, duodenal, peptic and gastrojejunal ulcer; gastritis and duodenitis; oesophageal ulcer and haemorrhage; gastrointestinal haemorrhage; and oesophageal and gastric variceal bleeding (Table S1). For those who had ≥ 2 hospitalizations for GI adverse events in 2006 only the first event was included and the date of first hospitalization was defined as index date. Patients were classified as having cirrhosis and included in the analysis if they had at least one hospital admission with a diagnostic code of cirrhosis (ICD-9-CM codes 571.2, 571.5, 571.6) or two or more outpatient visits with a cirrhosis diagnostic code before index hospitalization. Previous validation study using hospital administrative database reported a positive predictive value of 90% using this definition.[18] Patients with concomitant diagnoses of trauma (ICD-9-CM code 800–804, 850–854, V57), and those with a history of cancer (ICD-9-CM code 140–208) or coagulopathy (ICD-9-CM code 286) before hospitalization were excluded. We also excluded those with records of hospitalization within 120 days before index date due to the dosage or duration of NSAIDs use in hospitalization was difficult to ascertain and their upper GI events may be precipitated by other underlying diseases.

Data on Drug Exposure and Comorbidities

The main exposures of interest in this study were most commonly used selective and non-selective NSAIDs in Taiwan in 2006. Based on the drug utilization profile, celecoxib (the only coxib that was on the market), indomethacin, sulindac, diclofenac, acemetacin, ketorolac, piroxicam, meloxicam, ibuprofen, naproxen, ketoprofen, flurbiprofen, tiaprofenic acid and mefenamic acid were included. We retrieved data from the outpatient pharmacy prescription database for the information of type of drug prescribed (according to the anatomic therapeutic chemical ATC classification system), dosage, route of administration (oral, parenteral), date of prescription, days supply and total number of drug pills dispensed. Mean daily dose of NSAIDs exposure was determined by multiplying the number of pills dispensed by the dose prescribed divided by the recorded days supply and was presented as the number of defined daily doses (DDD), the typical maintenance dose required when the drug is used for its main indication in an adult.[19] Other concomitant drug use with potential risk modification effect for GI adverse events comprised histamine-2 receptor antagonists(ATC A02BA), proton pump inhibitors (A02BC), sucralfate (A02BX02), selective serotonin reuptake inhibitors (N06AB), other antidepressants (N06AX exclude N06AX01 N06AX02), nitrates (C01DA), calcium channel blockers (C08), beta blockers(C07A),vitamin K antagonists (B01AA), non-aspirin antiplatelet agents (B01AC except B01AC04 and B01AC06), low-dose aspirin (B01AC06), clopidogrel (B01AC04) and systemic corticosteroids (H02AB, H02B). The following information regarding patients' characteristics and comorbidities at admission were also collected, which included age, gender, cause of cirrhosis (viral hepatitis B and C, alcoholic liver disease), prior hospitalization for cirrhosis-related complications (varices, ascites, spontaneous bacterial peritonitis, encephalopathy and hepatorenal syndrome) and other chronic conditions (myocardial infarction, stroke, diabetes mellitus, hypertension, ischaemic heart disease, atrial fibrillation, congestive heart failure, chronic renal disease, chronic liver disease, chronic lung disease, rheumatoid arthritis, osteoarthritis and peptic ulcer disease) (Table S1).

Statistical Analysis

In the case-crossover design, each patient represented a matched set of data for time-varying exposures between the case and control period. For each patient, the odds of medication use during the case period were compared with the odds of medication use during the control periods. We defined case period as 1–30 days prior to index date and control period as 31–60 days prior to index date. Persons were considered current users of an NSAID during the start date and end date of a prescription.

We used conditional logistic regression to calculate crude odds ratios (ORs) and their 95% confidence intervals (CIs) by comparing current use of selective NSAID, nsNSAIDs overall and individual NSAID, with nonuse as the reference group. Patients were further classified as variceal (oesophageal and gastric) or non-variceal (oesophageal, gastric, duodenal and peptic ulcer and bleeding, gastritis and duodenitis) events. Separate analyses were conducted for different routes of NSAIDs use (oral or parenteral). Meanwhile, the relationship between acetaminophen (ATC code N02BE01, N02BE51) and anti-glaucoma medication (S01E) use and upper GI adverse events risk was also investigated as a comparison. To evaluate a possible frequency-response relationship, we estimated the mean daily dosage of NSAID use during the case and control period among users and classified them as frequent (≥0.5 DDD/day) or intermittent (<0.5 DDD/day) users in the analysis and calculated risk estimates for different frequency of use. In the multivariable analysis, we calculated adjusted OR simultaneously controlled for use of other individual NSAIDs and potential time-varying confounding variables including discordant use of systematic corticosteroids, aspirin, selective serotonin reuptake inhibitors, other anti-depressants, calcium channel blockers, nitrates, beta-blockers, proton-pump inhibitors, histamine-2 receptor antagonists and sucralfate between case and control period. In the sensitivity analysis, we estimated the risk of upper GI adverse event associated with NSAIDs use with different definition in control period (61–90 and 91–120 days prior to the index date).

Furthermore, subgroup analysis was performed to evaluate potential effect modification by items of stratification chosen a priori. The cases were separated according to (i) age (<65 and ≥ 65 years), (ii) sex (male, female), (iii) prior GI event and (iv) osteoarthritis. We also examined whether the risk was modified by concomitant use of aspirin, systematic corticosteroids, nitrates, proton pump inhibitors and histamine-2 receptor antagonists. A formal test of interaction was performed for each of all subgroups to examine if the difference in effect size between two subgroups was statistically significant. Two-sided P value < 0.05 was considered to be statistically significant. All statistical analyses were performed with sas 9.2 (SAS Institute, Cary, NC, USA).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.