Sir,
Kim et al.
1 examined the usefulness of the Clinical Dementia Rating (CDR) for predicting Alzheimer's disease (AD) dementia progression in 59 elderly subjects with mild cognitive impairment (MCI). By logistic regression analysis, odds ratio [95% confidence interval (CI)] of CDR Orientation subscale for AD dementia progression was 12.3 (3.5-43.8). The authors recommended CDR Orientation subscale score for predicting AD dementia progression in elderly subjects with MCI. I have some concerns about their study.
First, Xue et al.
2 investigated risk factors for progression to AD and death in 437 elderly subjects with MCI. They adopted Cox model and adjusted hazard ratios (95% CIs) of female, older age, reading occasionally, current smoking, light-moderate alcohol drinking, cerebrovascular disease, hyperlipidemia and diabetes for transition from MCI to AD were 1.82 (1.20-2.77), 3.09 (1.81-5.25), 1.79 (1.11-2.89), 1.74 (1.15-2.65), 2.24 (1.42-3.53), 2.70 (1.68-4.34), 1.87 (1.16-3.02), and 1.81 (1.18-2.77), respectively. These significant factors should be considered as contributors of AD dementia progression in elderly subjects with MCI.
Second, Kim et al.
1 handled a limited number of patients with MCI. From 59 subjects, 22 subjects progressed from MCI to AD after the 2 years of follow-up. I suppose that the wide range of 95% CI in odds ratio would reflect unstable estimation and summing-up of samples are recommend.
3
Finally, a simple and easy clinical measure is important for predicting AD dementia progression in elderly subjects with MCI. But statistical significance by logistic regression analysis and discriminant ability differs, and combination of several predictors should be considered for the progression.