Coding Muddies Stats on Statin Side Effect - Rhabdomyolysis

This was a poor article, I had to look up the definition of  Rhabdomyolysis myself. Be careful out there.
http://www.medpagetoday.com/Cardiology/Dyslipidemia/32210
 definition first, from NIH,
Rhabdomyolysis is the breakdown of muscle fibers that leads to the release of muscle fiber contents (myoglobin) into the bloodstream. Myoglobin is harmful to the kidney and often causes kidney damage.

Symptoms

• Abnormal urine color (dark, red, or cola colored)
• Decreased urine production
• General weakness
• Muscle stiffness or aching (myalgia)
• Muscle tenderness
• Weakness of the affected muscles

Other symptoms that may occur with this disease:

• Fatigue
• Joint pain
• Seizures
• Weight gain (unintentional)

 The article:
Relying on use of the standard diagnostic code for rhabdomyolysis, a serious but rare side effect of statin therapy, to gauge incidence of the condition may lead to both under- and overestimates, researchers said.
Analysis of patient records maintained by the Group Health Cooperative over a 5-year period indicated that 292 charts of statin-treated patients had the ICD-9 code for rhabdomyolysis, but only 22 were validated by elevated creatine kinase levels and muscle symptoms, according to James S. Floyd, MD, of the University of Washington in Seattle.
Another seven cases of probable rhabdomyolysis were identified for which the ICD-9 code was not used, the researchers reported in a research letter appearing in the April 18 issue of the Journal of the American Medical Association.
In short, an ICD-9 code for the condition had a positive predictive value of just 7.5% (95% CI 5.0% to 11.1%), Floyd and colleagues wrote, while its sensitivity was 76% (95% CI 58% to 88%).
The researchers also found that the inaccuracy was not spread evenly across different types of statin therapy. In particular, it led to an underestimate of the actual rhabdomyolysis risk associated with high-dose simvastatin (Zocor), which was the subject of an FDA warning last year.
A higher proportion of the validated rhabdomyolysis cases were in patients taking simvastatin at 80 mg/day or more than among those with the ICD-9 code, Floyd and colleagues found.
As a result, the incidence rate ratio for validated cases in patients on high-dose simvastatin, relative to those taking doses of 20 to 39 mg/day, was 12.2 (95% CI 3.6 to 52.3), whereas use of ICD-9 codes produced an incidence rate ratio of only 1.77 (95% CI 1.05 to 2.88).
The researchers' analysis also showed a trend toward an increase in risk associated with more moderate simvastatin doses -- an incidence rate of 14.1 per 100,000 person-years (95% CI 6.1 to 27.8) at doses of 40 to 79 mg/day, whereas for all other statin drugs combined -- mostly lovastatin (Mevacor) or atorvastatin (Lipitor) -- the rate was 5.2 per 100,000 person-years (95% CI 1.9 to 11.2).
"The use of administrative data alone in studies of adverse drug reactions with multiple causes may fail to detect actionable and clinically important harms," Floyd and colleagues concluded.
To validate a case of rhabdomyolysis, the researchers looked for chart records indicating muscle symptoms with a peak creatine kinase level of 10 times the upper limit of normal or higher.
When peak creatine kinase fell between 5 and 10 times the upper limit of normal, they classed the event as myopathy. The rate of myopathy also appeared to be elevated in patients on high-dose simvastatin versus those taking other statins or lower simvastatin doses.
Floyd and colleagues noted that they did not adjust for potential confounding factors and also may have missed some cases of statin-associated rhabdomyolysis.
In searching for rhabdomyolysis cases that lacked the ICD-9 code, the researchers looked for clear records of the defining symptoms and also for the words "rhabdo" and "statin" appearing near each other anywhere in the record.