The results of this study showed a consistently increased risk for upper GI adverse event associated with nsNSAIDs among cirrhotic patients. Use of celebrex only showed a slight increase in bleeding risk and the association did not reach statistical significance. Concomitant use of proton pump inhibitors and histamin-2 receptor antagonists tended to decrease the upper GI toxicity associated with nsNSAIDs and celecoxib.
Because most NSAIDs are largely metabolized by hepatic cytochrome P450 enzyme and heavily protein bound, increased serum levels of NSAIDs can be anticipated in cirrhotic patients. Dysregulated focal factors such as prostaglandin, nitric oxide and other cytokines (tumour necrosis factor α, epidermal growth factor) resulted in decreased gastric mucosal defense in patients with portal hypertension. Meanwhile, NSAIDs could also cause clinically important bleeding, either from gastroesophageal varices or portal hypertensive gastropathy due to local mucosa injury and inhibition of platelet function. The severity of liver disease and the nature and size of the varices are also important long term risk factors for variceal bleeding.
Few published studies examined the NSAIDs GI toxicity in cirrhotic patients. de Ledinghen et al. reported a three to five times risk for first variceal bleeding in cirrhosis patients using aspirin, with or without NSAIDs.[24,25] Several case series of variceal and non-variceal (ulcers and erosions) bleeding among patients with cirrhosis also observed a high prevalence of NSAIDs use before the episode of GI adverse event.[1,2,26,27] However, no conclusions on exclusively non-aspirin NSAID use could be drawn due to small sample size and difficulty in control selection.
This study demonstrated that the risk associated with nsNSAIDs use for variceal bleeding was similar to that for non-variceal bleeding. The risk associated with celecoxib use in cirrhotic patients was similar to the traditional nsNSAIDs, although the CIs were wide due to few patients took celecoxib in our study population. This finding was in contrast to the evidence from clinical trials which revealed a favourable effect of celecoxib among mostly arthritis patients without severe liver diseases. There was also a positive association for acetaminophen use but not for use of anti-glaucoma medication. Acetaminophen may not cause upper GI events, and rather it can be considered as a marker of deterioration in general health condition or increased medicalization that occurring just before hospitalization of upper GI adverse events.
Our risk estimates for NSAIDs-induced upper GI toxicity in cirrhotic patients were smaller as compared with nearly three- to four-fold increases in risk reported by previous studies including general population or arthritis patients.[15,28,29] There might be several reasons. Due to the potentially high risk for an upper GI adverse event in cirrhosis population, physicians might be more careful in prescribing these anti-inflammatory drugs with lower dose and shorter duration. Physicians' sensitivity to early GI symptoms and treating patients early might decrease the occurrence of severe GI event that needs hospitalization. We observed that concomitant use of proton pump inhibitors and histamine-2 receptor antagonists seemed to lower the risk of both variceal and non-variceal upper GI events which suggested a potential role of anti-ulcer drugs in protecting the esophagogastric mucosa among cirrhotic patients with bleeding varices.
There were several limitations in our study. Firstly, we identified cirrhotic patients with upper GI events by ICD-9-CM codes. Some patients with variceal events might be misdiagnosed or incorrectly coded as having non-variceal events. Secondly, information related to variceal and non-variceal bleeding, such as severity of liver disease, hepatic venous pressure gradient, size of varices, cigarette smoking and alcohol consumption were lacking in our database. However, these unmeasured factors could be largely controlled by case-crossover design. Thirdly, we could not evaluate whether effects of NSAIDs and acid-suppressive agents varied among different site of varices (oesophageal, cardiac, fundic). Forthly, information of over-the counter NSAIDs use was not covered in our datasets. The undifferential misclassification of the drug exposure would bias the result towards the null. Fifthly, we could not exclude the possibility that the multivariable analysis may over-control some of the intermediate variables such as concomitant use of proton-pump inhibitors or histamine-2 receptor antagonists that were actually effects caused by NSAIDs-related GI toxicity. Finally, as this study only assessed risk for short-term use of NSAIDs in cirrhotic patients, our findings would probably not generalize to the risk associated with long-term use.
In conclusion, our study found that nsNSAIDs caused a two-fold increased risk of variceal and non-variceal upper GI events among cirrhotic patients. Use of celecoxib in cirrhotic patients could possibly pose a risk similar to that of traditional NSAIDs. Use of NSAIDs, selective or non-selective, should be cautious in patients with cirrhosis.
Conflicts of interest
This study was supported by Taiwan Department of Health grant DOH098-TD-D-113-098016.
Liver International. 2012;32(5):859-866. © 2012 Blackwell Publishing