Primary prevention lipid-lowering drug therapy.

To the Editor: We appreciate the editorial comments provided by Drs Gotto and Kuller1 on our article published in the January 15, 2002, issue of Circulation,2 but we wish to respond to several issues raised. The methodology used to project population size is standard for the Third National Health and Nutrition Examination Survey (NHANES III) and is described thoroughly in the publication by the National Center for Health Statistics, which we reference3 in our article. We did not feel it necessary to reiterate that method. We assume that the phrase “significance of the data” is a reference to statistical significance testing which remains a controversial issue in some epidemiological analyses.4 Given the large sample size inherent to these types of studies, even small differences can be declared statistically significant with very low probability values. We therefore felt that a list of highly significant probability values would not add to the interpretation of results. With regard to under-reporting of coronary heart disease, as Gotto and Kuller admit, the estimates in the survey reflect the misdiagnosis of coronary heart disease in women and elderly in actual practice. Because these same criteria will be used to determine eligibility for treatment in the clinician’s office, we maintain that the same level of misclassification will occur in the office as in the NHANES III survey. Thus, our estimates represent those who would be so classified (as primary) in actual practice. Admittedly, “other forms of cardiovascular disease (CVD),” such as peripheral artery disease and symptomatic carotid disease, would be detectable in clinical practice. These will not constitute a large percentage of the total misclassified. Lastly, our estimates for primary prevention treatment–eligible patients under Adult Treatment Panel II correspond closely to those reported by Hoerger et al.5 That study reported 15.7 million patients aged 20 years as treatment eligible. Our study reported a similar number (15.1 million) for those aged 20 to 79 years. We disagree with the conclusion that “more women than men aged 20 to 29 years are eligible for primary prevention.” In fact, the observed difference is actually very small (78 735) relative to the total eligible population in that age group (2.4 million). We agree that global risk assessment is preferable to the narrow application of the Adult Treatment Panel III algorithm. However, it is unclear how the clinician will combine the algorithm’s quantitative output with the list of global risk factors (Table 2 of the editorial1) to decide treatment. That subjective judgment could not be reliably incorporated into our analyses. The fact remains that 70% of the eligible patients can be selected for treatment solely on the basis of having a low-density lipoprotein cholesterol level 160 mg/dL.