Featured Articles

QuarterWatch™ Latest Edition: Cancer Risks with Biological Products for Psoriasis; Sudden Hearing Loss with Tadalafil and Sildenafil

ISMP’s QuarterWatch™ provides an overview of drug safety issues reflected in adverse drug events reported to the US Food and Drug Administration (FDA) in the third quarter of 2015 (the latest data available for research). Based on these data and the previous 3 quarters of data, this report identifies:

  • Reports of cancer associated with products used to treat psoriasis, particularly potent immunosuppressants
  • New evidence that drugs for erectile dysfunction and pulmonary hypertension can cause sudden hearing loss

Reporting totals

In the third quarter of 2015, FDA received 332,226 adverse drug event reports, a 31% increase over the second quarter in 2015. The large increase in reports consisted almost entirely of non-serious events from drug manufacturers. In fact, reports of serious events in the US declined by 2% from 78,854 reports in the second quarter to 77,117 reports in the third quarter. Among the 190,911 non-serious events in the third quarter of 2015, most (80%) were from consumers who submitted reports to manufacturers, stating that a drug was ineffective, or that they had experienced nausea, fatigue, or a headache. 

Reports of cancer with psoriasis drugs

Psoriasis is a chronic skin disorder that affects an estimated 7.5 million people in the US. Because psoriasis has an autoimmune component, the more serious cases are often treated with powerful immunosuppressants, which can increase the risk of cancer. To evaluate this adverse event, we examined reports of cancer and benign tumors among the 38,952 cases of all types of adverse events with psoriasis medications reported in the 12 months ending with September 30, 2015. The psoriasis medications were grouped primarily by the mechanism of action. We excluded cases for drugs sometimes used to treat psoriasis where the diagnosis or indication was for some other medical condition, or was not specifically stated. Apremilast (OTEZLA), a new psoriasis medication without known immunosuppressant properties, was used as the reference drug for comparison.

We identified 1,315 cases (3.4%) indicating a form of neoplasm, mostly malignant, among all types of adverse events with psoriasis medications (see Table 1 in the PDF newsletter). Non-melanoma skin cancers formed the largest group, followed by breast and prostate cancers. The strongest association with cancer was with ustekinumab (STELARA), the only human monoclonal antibody used to treat psoriasis that binds to and interferes with interleukin (IL)-12 and IL-23, as well as other proinflammatory proteins. We found that the odds of cancer being reported when taking ustekinumab were 15 times higher than with the reference drug, apremilast. IL-12 itself has an anti-tumor effect; therefore, blocking it might permit cancer growth. In animal testing, mice that were engineered to have no IL-12 or IL-23 developed bigger skin cancers more rapidly when exposed to ultraviolet radiation than did mice with normal IL-12 and IL-23. 

A strong association with cancer was also observed for biological products that block the tumor necrosis factor (anti-TNF medications), including: etanercept, inFLIXimab, adalimumab, certolizumab, and golimumab. The odds of cancer being reported when taking an anti-TNF drug were 5 times higher than with the reference drug, apremilast (see Table 1 in the PDF). We saw no signal for secukinumab (COSENTYX), a drug that inhibits IL-17A, but our sample of cases was too small for an evaluation.

Biological products that target different elements of the immune system are being used for a growing spectrum of medical disorders, including psoriasis, rheumatoid arthritis, multiple sclerosis, Crohn’s disease, ankylosing spondylitis, and lupus. The evaluation of cancer risk with these biologics is scientifically challenging using any of the accepted methods. Animal carcinogenicity, when risk is found, typically triggers a debate about whether the findings apply to humans. Registries and long-term clinical trials usually lack a meaningful comparison group and have high dropout rates. However, the relative scarcity of definitive findings about the cancer risks of drugs should not be construed as evidence that the risks are minor or not important. This report, by analyzing a large group of cases in the same patient population using standard event definitions and meaningful comparators, provides the first comparative assessment of cancer risks for these drugs.  

Erectile dysfunction drugs and sudden hearing loss

We found new evidence that tadalafil (CIALIS, ADCIRCA) and sildenafil (VIAGRA, REVATIO) can cause sudden loss of hearing in one or both ears. These drugs inhibit phosphodiesterase 5 (PDE5) and are established treatments for both erectile dysfunction and pulmonary hypertension. Tadalafil is also approved for benign prostatic hyperplasia. Early reports of isolated cases of hearing loss began to appear in 2007, and FDA required a cautiously worded warning, specifically noting that impaired hearing was common in older men, with the qualification that a causal relationship had not been proven. In some studies, the prevalence of impaired hearing in older persons was around 17%.

To assess this risk, we selected all cases for any of the PDE5 drugs (n=10,174) in the most recent 12 months ending with September 30, 2015, and compared them to a group of reports that were otherwise similar in type, report source, and completeness (n=40,705). The outcomes we studied were sudden hearing loss or any preferred term including the word deafness. Reports of sudden hearing loss were rare—we found 214 reports among our 50,879 study cases—but the odds of a hearing loss report among the PDE5 drugs were 21.5 times higher than otherwise similar comparators (see Table 2 in the PDF). Tadalafil had a disproportionately large effect on the results, both in the number of cases and size of the effect—the odds of a hearing loss report were about 30 times higher than with otherwise similar comparators. There were not enough cases to evaluate vardenafil (LEVITRA).

While these results establish a strong association between the PDE5 drugs and reports of hearing loss (especially for tadalafil), adverse drug event reports do not provide an estimate of how frequently they might be occurring in a broad patient population. These data also lacked information about how much exposure to the drug over time had preceded the onset of hearing loss, and follow up to assess whether the hearing loss resolved when treatment was discontinued.  

Two previous investigations also found a relationship between PDE5 drugs and sudden hearing loss. In 2007, FDA assessed 29 cases of sudden hearing loss, including 10 cases where the event occurred after the first dose. In 2011, a British team evaluated 47 cases worldwide for all three drugs and found 88% of the cases occurred in only one ear, with most occurring within 24 hours of ingesting the drug for intermittent use. FDA also identified 5 cases among 660 patients in a clinical trial of continuous, rather than intermittent, sildenafil use for pulmonary arterial hypertension.

Our analysis of sudden hearing loss and the PDE5 inhibitor drugs confirms and extends the evidence that these drugs are a causal agent. It is noteworthy that a medically significant side effect of a drug taken each year by millions of people remained uncertain more than 25 years after the approval of the first drug in this class. Earlier efforts to pinpoint this side effect were complicated by the problem that hearing loss in older men was very common, aside from drug-related causes. Our analysis avoided this pitfall by focusing on cases that had sudden onset, or were extensive enough to justify the term deafness.

Download the full QuarterWatch™ report below.