Reply to letter to the editor: hypoglycemia and treatment with opioids

We thank Brož et al. for their interest in our study aiming to determine if drug-induced hypoglycemia could be a class effect for opioids [1] and for their useful comments.

The French PharmacoVigilance Database (FPVD) contains a narrative description of cases provided by the declarant, which we examined for this study. The description of the report’s narrative in pharmacovigilance study is a complement to disproportionality analysis that may identify potential mechanisms underlying the associations found between drugs and adverse drug reactions. We are glad that they shared our interest in finding such mechanisms to further investigate the link between hypoglycemia and opioids. As they accurately mentioned, 45 of the 93 tramadol-associated hypoglycemia cases included in our descriptive study in the FPVD were related to diabetic patients. Indeed, hypoglycemia in these patients could have been favored by the antihyperglycemic treatment they received. We agree with them that sleepiness, one of the common side effect of opioids, combined to the use of those antihyperglycemic treatments could have been one of the factors leading to the onset of hypoglycemia in those patients. However, not all the reports mentioned sleepiness and the prescription of opioids appeared to be the trigger factor causing the hypoglycemia even in diabetic patients, regardless of the underlying mechanism.

In pharmacovigilance databases, narrative descriptions are not subject to validation, and their content does not have to fulfill completeness criteria before being submitted. Therefore, narratives contain missing data. Although we agree with them that it would have been very interesting to collect the frequency of hypoglycemia and their severity in the period before and after the opioid initiation to make a comparison or to collect the exact definition of hypoglycemia and a relevant mean of the value of hypoglycemia used in the report, we regret that the narratives did not provide such information. They may have read in our methods that we only included cases with at least a C2 chronology score according to the French drug reaction assessment method to reduce the risk of bias. A C2 chronology means that the glycemia improved after the opioid was stopped (except if the patient died), which is in favor of the imputability of the opioid. The short time to onset (usually a few hours to a few days) of hypoglycemia after the opioid initiation was also in favor of the imputability of those treatments.

In the absence of a clear definition of hypoglycemia and a relevant mean of the value of hypoglycemia, we are aware that some cases could have been falsely interpreted as indicating hypoglycemia by the declarant. Still, they may also have read in our methods that cases with other obvious etiologies (i.e. overdose of insulin or sulfonamide, recent initiation of a beta-blocker, etc.) were excluded from our study in the FPVD.

We agree with them that in both studies cited in our article to support the connection between tramadol and hypoglycemia, the definition of hypoglycemia used is disputable and that a somewhat limited number of cases could have falsely been interpreted as indicating hypoglycemia [2,3]. However, we think the safety signal of hypoglycemia with tramadol they highlight must be acknowledged and investigated.

To conclude, the limitation raised in their letter about our work in the FPVD is indubitably exact, but we believe they do not impact the conclusion of our study, supported by the disproportionality analysis in Vigibase® (https://www.who-umc.org/vigibase/vigibase/) and by literature data. Those limitations are inherent to the use of pharmacovigilance database, where data are incomplete. We meet their views that only a prospective study aiming to evaluate the risk of hypoglycemia associated with opioids could reduce the impact of confounding biases. We sincerely hope such a study could be undertaken in the near future.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.