The RxView Blog

Real world data, analytics, and insights for healthcare decision makers

Advera Health Analytics, Inc.

What happens when a Safety Signal contradicts Clinical Trial data?

August 4, 2017

Last month, FDA released their quarterly watch list of ongoing drug safety concerns. My colleague Jim wrote a blog post on some of the major talking points around the update, not on the drugs or the risks themselves, but general pharmacovigilance themes. Our post this week focuses on one of the risks that was specifically discussed by FDA on the SGLT-2 class of diabetes drugs. The SGLT-2's had triggered a safety signal for nephrolithiasis (aka kidney stones). Evidex RxSignal analysis had also predicted this safety signal for most of these SGLT-2 drugs, with many of the signals being triggered more than a year prior to this alert.

However, unlike the 12 other safety issues in this same alert, the FDA quickly determined that despite the safety signal shown in the postmarketing data, “no action is necessary at this time, based on available information”. In other words, the FDA does not believe that kidney stones should be disclosed as a risk to the labels of these medications.

Why did the FDA make this determination so quickly? What information did the FDA use?

According to the FDA’s Guidance for Industry on the Adverse Reaction section, the following factors are used to determine whether a specific AE seen from the spontaneous reporting should be added to a drug’s label:

(1) seriousness of the event

(2) number of reports

(3) strength of causal relationship to the drug.

Looking at the seriousness factor, kidney stones are indeed on the Important Medical Event seriousness list. However, the outcomes for were not particularly serious despite the serious nature of kidney stones. Only 17% of the Jardiance kidney stone cases involved hospitalization, with 29% for Invokana and 41% for Farxiga. None of the cases for any of this class of drugs resulted in death or any long-term injury. An example of what this looks like on the Evidex signal validation report for Jardiance can be seen here.


To establish causation, the third factor in the FDA’s guidance, the FDA provides additional factors of what should be considered: (1) the frequency of reporting, (2) whether the adverse event rate for the drug exceeds the placebo rate, (3) the extent of dose-response, (4) the extent to which the adverse event is consistent with the pharmacology of the drug, (5) the timing of the event relative to the time of drug exposure, (6) existence of challenge and dechallenge experience, and (7) whether the adverse event is known to be caused by related drugs.

The second factor is an assessment of pooled clinical trial data to determine whether reporting rates are higher with the drug in question versus controls. The clinical trial details for the SGLT-2 drugs did not confirm any of the kidney stone signals. Evidex has pooled curated data from thousands of trials to provide background rates for both treatment and control groups for any adverse event signal being validated. In the case of the SGLT-2 drugs, Evidex analyzed 84 trials covering 35,000 patients to provide the expected rates of kidney stones for each of these drugs.

One of the three drugs, Farxiga, actually showed zero cases of kidney stones across all 33 of the trials with 2 patients (0.04%) experiencing them in the control groups of those trials. Pooled results for Jardiance showed a higher rate of kidney stones than their control groups (0.12% vs. 0.07%) but those results were not statistically significant (OR 1.19, p=0.781). These results can be viewed easily in our Jardiance Signal Validation report:


In one of our latest enhancements, Evidex also quickly summarizes the individual trial data for safety signals to determine consistency of results across different trials. The individual trial data can also highlight any specific characteristics of the trials showing a safety signal. In the case of Jardiance, kidney stones only showed up in trials where patients were exposed to Jardiance for at least a year:


The Evidex signal validation reports can also be useful in analyzing the other factors that the FDA suggests in determining causation. This includes a clean list of confirmed medical literature references with links to the PubMed sources for references. In the case of the SGLT-2 drugs, there were no literature references confirming causation of kidney stones.

A list of all relevant case reports in the Evidex report signal allows subscribers to dig into further detail on dose response factors, time to onset of adverse event, dechallenge and rechallenge data, and a list of other drugs taken which are known causes of the safety issue.

When drug safety data come together, signal validation is easier and more intuitive. Learn more by reading our comprehensive guide to using drug safety data. 


Topics: FDA, Clinical Evidence, Evidex

New Call-to-action

Subscribe to Email Updates