Algorithms to predict statin side effects published
MedWire News: Researchers have developed algorithms that predict patients’ risk for myopathy, acute renal failure, and cataract while they are taking statins.
Julia Hippisley-Cox and Carol Coupland (Division of Primary Care, Nottingham, UK) also developed an algorithm to predict moderate or severe liver dysfunction, but this performed poorly in validation tests.
The researchers derived their algorithms from data on 225,922 new statin users and 1,778,770 nonusers contained in the QResearch database. They validated the algorithms in an additional 118,372 new statin users and 877, 812 nonusers from the QResearch database and in 282,056 new users and 1,923,840 nonusers from the THIN database.
The algorithm for moderate-to-severe myopathy comprised ethnicity, corticosteroid use, and the comorbidities diabetes, hypertension, chronic liver disease, and hypothyroidism. The algorithm explained 42.15% of variation in the probability of developing myopathy in women and 35.98% of variation in men.
The risk for renal failure was increased by Townsend score, smoking status, corticosteroid use, and the comorbidities diabetes, congestive heart failure, hypertension, and chronic kidney disease. The algorithm based on these variables explained 59.62% of variation in women and 59.68% of variation in men.
Cataract risk was increased by non-White ethnicity, smoking, corticosteroid use, and the comorbidities cardiovascular disease, diabetes, rheumatoid arthritis, and atrial fibrillation. The resulting algorithm explained 59.14% of variation in women and 59.42% in men.
The receiver operating characteristic curve statistics for these three algorithms ranged from 0.739 to 0.878, indicating that the algorithms could distinguish between patients who would and would not develop the side effect with at least 73.9% accuracy.
In contrast, the algorithm for liver dysfunction, based on ethnicity, smoking habit, corticosteroid use, and various comorbidities, explained just 15.55% and 10.83% of variation in women and men, respectively. The receiver operating characteristic curve statistics were just 0.646 for women and 0.612 for men, indicating that the algorithm performed little better than chance.
“While the algorithms have specifically been designed to help inform treatment choices on risks and benefits of statins, they can also be applied to the general population,” say Hippisley-Cox and Coupland in the journal Heart.
They conclude: “Three of the algorithms could be used to identify patients at high risk of these clinical outcomes so that patients can be monitored more closely.
“Also, the algorithms can be used within a consultation to assess the balance of risks and benefits at the start of statin treatment.”
MedWire (www.medwire-news.md) is an independent clinical news service provided by Current Medicine Group, a trading division of Springer Healthcare Limited. © Springer Healthcare Ltd; 2010
By Eleanor McDermid