Letter by Lai and Chan Regarding Article, “Bleeding Risk in Very Old Patients on Vitamin K Antagonist Treatment: Results of a Prospective Collaborative Study on Elderly Patients Followed by Italian Centres for Anticoagulation”
To the Editor:
We read with interest the article by Poli et al1 regarding their recent study, from which the authors concluded that the observed rate of major bleeding in patients who began vitamin K antagonist at ≥80 years of age was acceptably low and that age in itself should not be considered a contraindication to vitamin K antagonist treatment. The authors modeled hemorrhagic risk using a competing-risk regression analysis, according to Fine and Gray,2 and considered nonhemorrhagic death as a competing risk. The authors also pointed out that the rate of bleeding incidents was higher among patients taking vitamin K antagonist for venous thromboembolism than among those taking vitamin K antagonist for atrial fibrillation (hazard ratio 1.4; 95% confidence interval, 1.1–1.8). However, as revealed by the only Figure in the article, the 2 cumulative incidence curves interlaced for most of the follow-up period but diverged after about 4000 days of follow-up. In our opinion, it seems likely that the proportional hazards assumption for the regression model was violated. Furthermore, the observed difference between groups after 4000 days might have been confounded by the relatively small sample size (or risk set) after longer follow-up.
There are several methodological approaches to deal with the violation of the proportional hazards assumption. One of the operable strategies is to perform a time-dependent Cox regression model with partition of the time axis into 2 intervals (time <4000 days and time ≥4000 days) and estimate regression coefficients (and hazard ratios) for these 2 time periods separately.3 On the basis of the cumulative incidence curves presented by the authors, it seems more appropriate to estimate different hazard ratios in 2 time periods rather than to obtain 1 pooled estimate throughout the long follow-up period.
Nevertheless, the major clinical implication of this article is that the results provide evidence and support for aggressive anticoagulant therapy in geriatric patients suffering from venous thromboembolism or atrial fibrillation.
Chao-Lun Lai, MD, PhD
Department of Internal Medicine
National Taiwan University Hospital Hsin-Chu Branch
Wan-Khey Chan, MD
Far Eastern Memorial Hospital
New Taipei City, Taiwan
- © 2012 American Heart Association, Inc.
- Poli D,
- Antonucci E,
- Testa S,
- Tosetto A,
- Ageno W,
- Palareti G