Statistics from Altmetric.com
We were interested to read the report by Ransford and Langman of their analysis of yellow card reports of suspected adverse drugs reactions for sulphasalazine and mesalazine (
). These reports, submitted to the Committee on Safety of Medicines, may provide useful flags to signal unrecognised hazards of drugs. However, as adverse reactions are not always recognised or reported to the regulatory authorities by physicians, these reports usually underestimate the frequency of any adverse drug reaction. Of greater importance, underreporting is usually not random but selective, which may introduce serious bias when comparing different drugs.1 Various examples have been described previously of drugs that showed substantive differences in reporting rates, which were not substantiated after further research.1 For this reason, it is recommended that, once there is a signal for a suspected adverse drug reaction, other sources of data are investigated.2
We recently initiated a study to quantify the risk of renal toxicity in patients taking aminosalicylate (5-ASA) drugs in the UK. The General Practice Research Database (GPRD) was used for this study, with data collected as part of routine medical practice. The GPRD has previously been demonstrated to be a representative sample of the general population of England and Wales, and the completeness and validity of the GPRD recording of medically significant events is well established. Its data have been used frequently to quantify the risk of adverse drug reactions.3 Our study population included almost 40 000 patients. We found that the overall incidence of renal damage (which included interstitial nephritis) was rare in patients taking 5-ASA drugs, but was increased relative to control patients (table 1). The risk of renal toxicity in patients taking mesalazine and sulphasalazine was comparable. Interestingly, we found that the risk of renal disease was related to indicators of severity of inflammatory bowel disease and to concomitant disease and drug treatment. A recent report also suggested that the kidney can be an extraintestinal target in Crohn’s disease.4 We presented the results of this study at the recent British Society of Gastroenterology meeting.
Our findings also highlight the substantive underreporting of the data used by Ransford et al (table 1). Given the selected and incomplete nature of the reports of suspected adverse drugs reactions, one needs to establish whether physicians reported cases of interstitial nephritis equally for users of different 5-ASA drugs. The authors did not provide any data for the comparability of the users of the various 5-ASA drugs in the UK.
In conclusion, while we agree that renal function should be evaluated and monitored in patients taking 5-ASAs, the results of our large epidemiological study show no difference in renal toxicity between mesalazine and sulphasalazine and that confounding factors can also significantly affect the overall risk. A statistical analysis of suspected adverse drug reaction reports may generate signals but does not provide conclusive evidence of differences in safety between drugs.
We do of course recognise that making deductions from examination of spontaneous adverse reaction reports poses problems from the incomplete nature of the data, and the limited knowledge of biases. Thus we say “spontaneous reporting....cannot be used to determine true rates of reaction”. We also speculate that reporting rates of interstitial nephritis with mesalazine may be high “because a specific warning of possible renal toxicity had been issued”.
The comparatively equal values quoted by Logan and van Staa of 1.2 (mesalazine) and 1.7 (sulphasalazine) cases of interstitial nephritis per 1000 patient years are very different from the 29 cases reported spontaneously on yellow cards for mesalazine with none for sulphasalazine in the time period assessed (there are a total of 47 for mesalazine and two for sulphasalazine, a fairly large difference).
Being aware of the problems of judging true rates of reaction from spontaneous reports, and knowing that there was (as for interstitial nephritis) a relative paucity of reports for pancreatitis with sulphasalazine, we have recently analysed data from the General Practice Research Database (GPRD) on prior drug exposure in cases of acute pancreatitis. This clearly shows raised odds ratios for mesalazine, but not for sulphasalazine, and with the odds ratio for mesalazine being particularly high in those with first exposure in the prior three months.1 The finding is consonant with the spontaneous adverse drug reaction data presented by us.
Rates from GPRD for interstitial nephritis, as presented by Logan and van Staa per 1000 patient years, are difficult to relate to individual patient exposures. It would be valuable to have such information. Given that sulphasalazine is the older drug, one would expect longer exposure in each such taker (particularly if it was for inflammatory bowel disease). It would also be valuable to know if the cases of renal damage in sulphasalazine takers identified by Logan and van Staa were in patients with inflammatory bowel disease rather than in those with rheumatoid disease, where confounding by use of other agents, notably penicillamine and gold, and by complicating renal amyloid, would need to be borne in mind. Differences between our findings may be resolved in due course by current surveillance studies being conducted by the British Society of Gastroenterology.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.