As physicians with a long-time interest in the pharmacoprophylaxis of post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP), we greatly enjoyed reading the excellent meta-analysis performed by Elmunzer et al (Gut 2008;57:1262–7) based on four randomised clinical trials (RCTs). The analysis demonstrated a pooled relative risk (RR) for PEP of 0.36 (95% confidence interval (CI) 0.22 to 0.60) after prophylactic rectal administration of non-steroidal anti-inflammatory drugs (NSAIDs), with a relative risk (RR) of 0.30 (95% CI 0.01 to 0.76) for moderate and severe pancreatitis. The authors concluded that rectal administration of NSAIDs is effective in preventing PEP. While these results have significant clinical implications, several issues remain unsolved. For example, do risk factors influence the prophylactic effect?

Therefore, we performed a complementary meta-analysis based on the methodology and the source articles identical to those used by Elmunzer et al. The data extracted from the four RCTs were stratified by risk factors of patients (high-risk vs low-risk population)1 and the severity of PEP (mild vs moderate to severe).2 It was shown that administration of NSAIDs was associated with decreased incidence of PEP in patients with low (RR = 0.29, 95% CI 0.12 to 0.71, p = 0.006) and high risks (RR = 0.40, 95% CI 0.23 to 0.72, p = 0.002). Moreover, administration of NSAIDs significantly reduced incidence of mild PEP (RR = 0.49, 95% CI 0.28 to 0.84), p = 0.009), but marginally reduced incidence of moderate to severe PEP (RR = 0.13, 95% CI 0.02 to 0.99, p = 0.050), which was probably due to the limited number of cases with moderate to severe PEP (n = 7, all in the placebo group) (tables 1 and 2).

There was no death or severe adverse events recorded among the 912 patients treated with NSAIDs or placebo in the four RCTs, indicating that prophylactic rectal NSAIDs were safe.

Therefore, both the main meta-analysis by Elmunzer et al and our complementary meta-analysis showed beneficial effect of prophylactic rectal NSAIDs in preventing PEP. However, it must emphasised that there were several limitations of the both meta-analyses including small sample sizes (for both subjects and studies), inconsistent definition of PEP, and less representative populations. First, only four trials met the criteria for meta-analysis, two of which included fewer than 200 patients. Second, the PEP diagnostic criteria were not consistent throughout the four RCTs. Whereas two adopted the generally accepted criteria, ie, serum amylase levels 3-fold the normal upper limits, the other two adopted 4-fold. This may lead to over-diagnosis of PEP, and increased incidence of PEP. Theoretically, the data of these four RCTs using different PEP criteria should not have been pooled together for the meta-analysis. Third, the four RCTs were carried out in only three countries, ie, Iran (n = 542), Mexico (n = 150) and Scotland (n = 220). Thus, the results of the RCTs need to be reproduced in other populations. Finally, it was noticed that only two NSAIDs were used, indomethacin in two studies and diclofenac in the other two. Whether there was a difference in prophylactic effect between the two drugs and whether other NSAIDs possess the same effect needs further investigation.

In conclusion, rectal administration of NSAIDs is effective in the prevention of PEP in patients with low or high risk factors. However, multi-centre clinical trials with large sample sizes and consistent criteria for PEP are required to confirm the beneficial effect.