Objective: Although psychiatric patients have a shortened life expectancy due to increased coronary heart disease (CHD), early identification of high-risk patients and targeted prevention for reduction of low-density lipoprotein (LDL) cholesterol are suboptimal in clinical care. We aimed to compare the accuracy of a single LDL-cholesterol intervention threshold of > 130 mg/dL (recently proposed for psychiatric patients) with that of the more complex LDL-cholesterol targets defined by the National Cholesterol Education Panel (NCEP). The study was performed in patients receiving second-generation antipsychotics (SGAs), a medication class associated with CHD risk.
Method: Three hundred fifty-six psychiatric patients receiving SGAs underwent standard LDL-cholesterol target assessments upon admission to the hospital between August 1, 2004, and March 1, 2005. The expert consensus-recommended > 130-mg/dL LDL-cholesterol threshold was used to determine false-negative results among patients with above-target NCEP-defined LDL cholesterol and false-positive results in the group with below-target NCEP-defined LDL cholesterol.
Results: The > 130-mg/dL threshold misclassified 15 (14.9%) of 101 high-risk patients and 31 (12.2%) of 255 low-risk patients (mean +/- SD 10-year CHD risk: 23.1% +/- 12.2% and 2.1% +/- 2.2%, respectively). Results were similar in the 171 schizophrenia patients. Misclassified patients with above-target LDL cholesterol were more likely than correctly identified patients to have diabetes (p = .0002), greater 10-year CHD risk (p = .0006), higher age (p = .0008), metabolic syndrome (p = .0018), and past CHD events (p = .0025). No distinguishing factors for false-positive cases could be identified.
Conclusions: The > 130-mg/dL LDL-cholesterol intervention threshold operated poorly in our psychiatric population. To avoid substandard care, NCEP-defined LDL-cholesterol targets should be used for the routine detection of psychiatric patients treated with antipsychotics who require interventions to decrease CHD risk.