QuarterWatch™ (2014 Quarter 2 data) Zolpidem and Canagliflozin Safety Signals, Thousands of Incomplete Rosiglitazone Reports
A new edition of ISMP’s QuarterWatch™ studies adverse events associated with a longstanding hypnotic taken by more than 5 million people to treat insomnia, zolpidem (AMBIEN and others), and a relatively new antidiabetic agent, canagliflozin (INVOKANA). The report examines why unsafe use of zolpidem accounts for more emergency department (ED) visits than any other psychoactive drug, and whether enough is known about canagliflozin to assure its benefits outweigh its risks. The report also identifies a flood of low quality, incomplete manufacturer reports of heart attack and stroke associated with rosiglitazone (AVANDIA), providing another example of why the US Food and Drug Administration (FDA) needs to modernize its primary postmarket surveillance reporting program.
This edition of QuarterWatch™ includes analysis of domestic fatal, disabling, or other serious events reported in the second quarter in 2014, during which FDA received 75,643 reports meeting our criteria. Cases included 7,071 deaths, 1,596 cases indicating sustained harm or disability, and 16,717 cases of injury severe enough to require hospitalization.
Zolpidem and Unsafe Use
In 2014, the Centers for Disease Control and Prevention (CDC) reported the results of a survey of ED visits resulting from adverse drug events with psychiatric medications. The investigators were surprised to discover that zolpidem ranked first, accounting for 11.5% of all psychiatric medication related ED visits among all adults, and 21% of visits by patients 65 years or older. The study estimated that zolpidem accounted for 10,212 annual ED visits, substantially more than the second most frequently involved psychiatric medication, QUEtiapine, with 6,900 ED visits.
Given these findings, we examined 12 months of QuarterWatch™ data ending with the second quarter in 2014. We identified 1,030 serious adverse events in which zolpidem was the primary or secondary suspect drug. Frequently reported adverse events included suicidal and self-injurious behavior (n=159); unaware sleep activities such as “sleep-driving,” “sleep-walking,” making and eating food, or having sex (n=62); amnesia (n=40); hallucinations (n=31); memory impairment (n=25); falls (n=35); car accidents (n=32); and impaired driving ability (n=18).
We also independently analyzed data from a new data source to provide a broader perspective on how safely zolpidem is used. The new data source was an annual healthcare survey from 2012, the Medical Expenditure Panel Survey (MEPS), available from the Agency for Healthcare Research and Quality for research use. The key problem with zolpidem, the QuarterWatch™ analysis concluded, is a broad pattern of unsafe use that is contrary to FDA and manufacturer recommendations:
- Long-term use. While zolpidem is recommended for short-term use, 68% of patients were sustained users (defined as 3 or more prescriptions/refills) with a mean of 229 days’ supply for this group.
- Higher doses. FDA now recommends a lower starting dose of 5 mg (or 6.25 mg, extended-release) for women and the elderly. In 2012 only 5% of women and 10% of people 65 years or older were prescribed the lower dose. More recent prescription data showed use of the lower dose was increasing.
- Concomitant psychoactive medications. While zolpidem prescribing information warns against concomitant use of other central nervous system (CNS) depressants, 22% of zolpidem patients with sustained use were also sustained users of opioids. In addition, 23% of sustained users were also taking other drugs active on the same GABA (gamma-aminobutyric acid) receptors (e.g., benzodiazepines, pregabalin, gabapentin).
The QuarterWatch™ adverse event report data also demonstrated that the primary suspect drugs taken with zolpidem were similar to those raising safety concerns above: opioids, including HYDROcodone, morphine, traMADol, oxyCODONE, and fentaNYL; and benzodiazepines, including ALPRAZolam, diazepam, and clonazePAM. Neither the QuarterWatch™ nor the MEPS data provided information about three other potential concerns with zolpidem: illicit or inappropriate use; concomitant alcohol use; and failing to take the drug at least 8 hours prior to next-day activities requiring mental alertness.
Conclusion. The analysis of reported events and typical usage patterns associated with zolpidem illustrates a different dimension of drug safety. Apart from the question of whether a medication might cause adverse effects is whether the pattern of clinical use is consistent with minimizing its known risks. Our findings that most zolpidem use does not follow safety recommendations should be of concern. It illustrates that, to minimize harm from drug treatment requires not only accurate information about adverse effects, but also adherence to recommended prescribing and administration practices.
Unanswered Questions about Canagliflozin
Canagliflozin, marketed in the US by Janssen Pharmaceuticals and approved in 2013, is the first in a new class of oral diabetes agents. The drug blocks a key transport protein in the kidney filtration process—sodium-glucose cotransporter 2 (SGLT2)–thereby causing some of the circulating glucose to be excreted in the urine rather than returned to the circulating blood. In the QuarterWatch™ data we identified reports indicating serious injuries involving kidney function and hypersensitivity reactions. While most could have been reasonably anticipated, given the mechanism of action and preapproval clinical trial data, these data also underscore the importance of unanswered questions about the new drug.
Canagliflozin had a successful launch, with 426,859 outpatient prescriptions dispensed in the second quarter of 2014, according to IMS Health. In the same period we identified 457 serious adverse event reports. This was a higher total of adverse event reports than observed for 92% of the drugs we regularly monitor. Most patients experiencing reported events were also taking other diabetes drugs, notably insulin (22.7%) and metFORMIN (23.6%). Examining these cases we identified 6 signals:
- Renal failure or impairment. A number of cases (n=54) reported acute or unspecified renal failure, although some included details about dehydration or hypotension. (During clinical trials, estimated glomerular filtration rates were reduced by small amounts during short-term use, with unknown effects over longer treatment periods. Higher dose animal studies showed long-term kidney damage, kidney and testicular cancers, and bone abnormalities.)
- Fluid and electrolyte imbalances. In a number of cases (n=54), the diuretic effect of canagliflozin was suspected to cause serious adverse effects, including dehydration (n=36) and renal failure (n=13). These were severe events, with 54% requiring hospitalization.
- Urinary tract infections. During clinical trials, the rates of urinary tract infections were high: 14% of women and 3.9% of men with near-normal kidney function developed fungal infections at rates 4-6 times higher than a comparison group of patients. Thus, the cases (n=50) reported to FDA illustrate that only a small fraction of probable events are being reported.
- Kidney stones. This adverse outcome (n=11) was not mentioned in FDA safety reviews prior to approval or in the prescribing information.
- Weight loss. While the mean weight loss seen in clinical studies (2-3 kg) may benefit a largely obese patient population, these adverse event cases (n=52) indicated weight loss could be large enough to contribute to serious adverse events.
- Hypersensitivity. With these reported events (n=50), some were emergency cases of angioedema; others were adverse effects on the skin, ranging from rash (n=11) and urticaria (n=8), to more severe skin exfoliation (n=2).
We shared these results with Janssen, and the company responded by noting that the drug already contains warnings or alerts about these adverse effects, except kidney stones about which the company said there was insufficient evidence to assess a causal role.
Conclusion. The history of oral diabetes medications shows that most agents have yet to demonstrate long-term benefits against clinical endpoints, and some have caused increased harm. SGLT2 inhibitors are gaining rapid acceptance in clinical practice despite important unanswered questions about both risks and benefits. While Janssen notes that the drug has now been tested in more than 10,000 patients, the data are still of insufficient duration to establish whether the drug has a measurable clinical benefit on the complications of type 2 diabetes. The current data are also insufficient to address unanswered questions raised about whether long-term use might result in a steady decline in kidney function, increased risk of bone fractures, or more cardiovascular events. By contrast, we observed clear evidence of harm to some patients in terms of an array of renal adverse effects and hypersensitivity reactions.
Incomplete Rosiglitazone Reports
The prior issue of QuarterWatch™ examined growing problems in the FDA Adverse Event Reporting System (FAERS) and the steps needed to modernize it. Among the issues are outdated regulations and guidances, flawed inspection and enforcement programs, and low quality adverse event reports from drug manufacturers. These problems were again illustrated in the second quarter of 2014 when GlaxoSmithKline (GSK) submitted 20,632 incomplete adverse event reports for rosiglitazone, 99% during April 2014.
When a typical drug that QuarterWatch™ monitors accounts for a median of 7 adverse event reports per quarter, and when 160 reports in a quarter places the drug in the upper 5% of all drugs, discovering more than 20,000 adverse event reports for rosiglitazone immediately attracted our attention. These adverse events were primarily myocardial infarction (n=12,159) and stroke (n=5,169). But the report quality was among the poorest seen for a large group of reports. The elementary essentials of patient age, gender, and event date were missing in all of these case reports, which had other deficiencies as well.
GSK told us that the reports stemmed from a 2012 legal settlement in federal district court. Other published reports said the company paid more than $2 billion to an estimated 40,000 rosiglitazone patients who alleged the drug caused heart attacks, strokes, and other cardiovascular problems. Two years later, the company discovered records of the 20,632 cases from the settlement. Although GSK provided FDA with a spreadsheet containing the limited data it received from the plaintiffs’ lawyers, FDA required GSK to file separate event reports for each case, despite the lack of relevant detail. GSK told us the reports will be updated with additional information, beginning with 2014 quarter 3 data.
Conclusion. The incomplete rosiglitazone adverse event cases provide additional evidence as to why FDA needs to modernize its reporting system and requirements. The reporting regulations and guidelines predate the age of digital marketing by manufacturers that involves many forms of contact with consumers and health professionals. FDA has compounded the problem with compliance and inspection actions that allow reporting of incomplete data about patient health problems without determining whether the drug was suspected of causing them. The modern digital age in which manufacturers know more about patients who take their products ought to produce richer safety data from a broad population often excluded from clinical trials. Instead, the agency receives increasing numbers of incomplete reports of little value in assessing possible harm from drugs.
The full QuarterWatch™ report can be viewed here.