Is Corneal Arcus Independently Associated with Incident Cardiovascular Disease in Asians?

The more common a disease is, the greater the difference between its OR and RR estimates. Because of this drawback, calculation of more intuitive risk estimates such as RRs and related statistics (eg, attributable risk, number needed to harm) may be preferred.4 Cardiovascular disease (CVD) is unlikely to be a rare event in the elderly—hence the possibility of ORs' overestimation of the risk factors' effect size. [...]the addition of several risk factors can mean substantial increase in the overall risk. [...]the combined effect of some factors may not be simply additive but rather synergistic.5 The use of multivariable logistic regression—a statistical analysis that simultaneously models several factors in an attempt to predict 1 binary multifactorial outcome—takes into account such pooled influence and also controls for potential confounding that may exist. Yet, findings of systematic reviews and meta-analyses are surely used to make clinical decisions. Besides the statistical significance and the magnitude of effect size, the range of confidence intervals, the method of evidence synthesis, the internal and external validity of the study, and the characteristics of the disease and target population, among other factors, are important determinants of clinical significance.