Reader Comments

Post a new comment on this article

Publisher's Note: Correction for “Long-term survival in a large cohort of patients with venous thrombosis: incidence and predictors. L.E. Flinterman, A. Van Hylckama Vlieg, S.C. Cannegieter, F.R. Rosendaal. PLoS Med. 2012 Jan;9(1):e1001155”

Posted by plosmedicine on 07 Jan 2013 at 16:20 GMT

Shortly after the publication of this paper a coding error was discovered in the database. Vital status was retrieved from the municipal registries, and for all participants from one city who did not die or emigrate during follow-up an end of follow-up date of 2004 instead of 2007 was entered. This implies there were three additional years of follow-up in 315 patients with thrombosis and 491 control subjects. This amounted to 3.7% of the total follow-up reported in the paper.

After correction of the database, all estimates were recalculated. Since the error only concerned survivors, this led to minor corrections to several of the previously reported figures, but not in all, and none of the conclusions changed. The changes in the estimates are:

The overall duration of follow-up changed from 26515 years to 27270 years for the patients with venous thrombosis (2.8% change). The duration of follow-up for the controls changed from 28433 to 29729 patient years (4.6% change). Median duration of follow-up was still 5.5 years for patients and 4.4 years for controls. The overall mortality rate changed from 22.7 (CI95 21.0-24.6) to 22.0 (CI95 20.3-23.9) per 1000 person years for the patients and from 4.7 (CI95 4.0-5.6) to 4.5 (CI95 3.8-5.4) per 1000 person years for the controls. For individuals without malignancies and other comorbidities the hazard ratio of death (patients with thrombosis versus controls) remained 2.5 (CI95 1.9-3.4), for those with comorbidities the HR increased slightly from 1.2 (CI95 0.9-1.7) to 1.4 (CI95 1.0-2.0).The recalculated numbers for the SMRs and Hazard ratios from the tables are shown in the link below.


No competing interests declared.