More than 90 percent of patients who stop and restart statin therapy tolerated the drug long term. The results, published in the April 2 issue of the Annals of Internal Medicine, used natural language processing software and an EHR to study patients commonly seen in clinical practice.
Huabing Zhang, MD, of Brigham and Women's Hospital in Boston, and colleagues conducted the retrospective cohort study to find reasons for discontinuation of statins and the role of statin-related clinical events or symptoms. In clinical trials, the rate of statin-related events ranged between 5 and 10 percent, according to the researchers, but the trials likely enrolled patients who were younger with fewer comorbidities than patients seen in routine practice.
To capture adverse events in a real-world setting, they used two technologies: an EHR system that served practices affiliated with Brigham and Women's Hospital and Massachusetts General Hospital, also in Boston, and an algorithm that processed physicians’ narrative electronic notes. The software was validated for statins and had a sensitivity of at least 86.5 percent and had a specificity of at least 91.9 percent.
The researchers identified 107,835 patients who had at least one statin prescription given between 2000 and 2008. Of that total, 53.1 percent discontinued statins at least once and 17.4 percent had a documented statin-related event. The most common reason for discontinuation was “no longer necessary,” a default reason for medical discontinuation in the EHR.
Myalgia or myopathy topped the list for documented statin-related events. More than half of patients with a statin-related event were re-prescribed a statin, and more than 90 percent of those patients ultimately tolerated the drug 12 months later. Only 13.2 percent subsequently had a documented second statin-related event.
Zhang et al found that almost one in five of patients in their study discontinued statin therapy for at least 12 months. “The rate of reported statin-related events to statins was nearly 18 percent, substantially higher than the 5 percent to 10 percent rate usually described in randomized, placebo-controlled clinical trials,” they wrote. “This finding is consistent with previously published observational studies.”
Patient selection, including the possible exclusion of sicker patients, patients with a history of statin-related events or those considered at high risk of statin-related events, may contribute to the different findings. The authors noted it is unlikely that a randomized clinical trial could be designed to study such patients, who in clinical practice may still be prescribed statins to reduce the risk of cardiovascular disease events.
They recommended physicians challenge patients who discontinue statins to try again. They also suggested that technologies similar to the software and EHR in their study may be used to monitor adverse drug events retrospectively or prospectively.
The researchers acknowledged that the accuracy of the algorithm “although high, was not perfect.” Although the study was large, it may not be generalizable to other settings and missing laboratory data may have led to an underestimation of the frequency of statin-related events.