Mukherjee M, Dastur D K, Soneji S L
Thrombosis Research Institute, Chelsea, London, UK.
J Assoc Physicians India. 1996 Jan;44(1):12-8.
A reliable method for identification of the subset of population predisposed to coronary heart disease (CHD) would aid a targetted implementation of intervention strategies. To this end, a mathematical formula was developed based on stepwise linear discriminant analysis. Age, body mass index, the number of associated coronary risk factors and a large number of biochemical markers were analysed by computerised discriminant analysis on a test sample of 203 subjects. Unstandardised canonical discriminant coefficients of statistically significant independent variables were used to derive the total discriminant score or the 'risk score'. The 'low-risk' persons not in need of immediate preventive measures of CHD could be distinguished from the 'high-risk' individuals with an almost 90% correctness. As compared with the existing methods such as clinical evaluation and cardiac stress test, the risk scores derived by the new method, and based chiefly on blood markers besides clinical and anthropometric variables, appeared to correctly predict the future coronary episodes in members of the test sample selected at random. The risk scores were also tested on a new sample of 50 subjects; while low scores were not associated with CHD, high scores in some patients were associated with myocardial ischemia. It appears that the preventive measures of CHD may be directed at people who have no clinical manifestations of CHD, but whose risk scores are greater than 0.1. On the other hand, if the score is less than -1.0, immediate preventive measures may not be necessary. If the score is between -1.0 and 0.1 (borderline), no immediate action may be taken but the score may be determined after six months, and action taken accordingly.
一种可靠的识别易患冠心病（CHD）人群亚组的方法将有助于有针对性地实施干预策略。为此，基于逐步线性判别分析开发了一个数学公式。通过对203名受试者的测试样本进行计算机判别分析，分析了年龄、体重指数、相关冠心病危险因素的数量以及大量生化标志物。使用具有统计学意义的独立变量的非标准化典型判别系数来得出总判别分数或“风险分数”。可以以近90%的正确率区分出不需要立即采取冠心病预防措施的“低风险”人群和“高风险”个体。与临床评估和心脏应激测试等现有方法相比，新方法得出的风险分数主要基于血液标志物以及临床和人体测量变量，似乎能够正确预测随机选择的测试样本成员未来的冠心病发作情况。还在50名受试者的新样本上对风险分数进行了测试；虽然低分与冠心病无关，但一些患者的高分与心肌缺血有关。看来冠心病的预防措施可能针对那些没有冠心病临床表现但风险分数大于0.1的人。另一方面，如果分数小于 -1.0，则可能不需要立即采取预防措施。如果分数在 -1.0 和0.1之间（临界值），则可能不立即采取行动，但可以在六个月后确定分数，并据此采取行动。
