• Open Access

Correspondence

Authors


Dear Sir,

Piek and colleagues1 present interesting data on the treatment and prognosis in a cohort of dogs with a diagnosis of immune-mediated haemolytic anemia and draw some interesting conclusions about prognostic factors. We believe, however, that their 6 month survival statistics should be viewed with caution. They report that the last date of contact with the owner occurred after a median of 46 days (approximately 6 weeks), yet extrapolate using the Kaplan-Meier method to estimate a 6 month survival of 72.6% (95% CI: 64.9–81.3%). The cases lost to follow-up are numerous. A total of 52/149 (35%) cases were shown to be censored on the Kaplan-Meier curve presented (Fig. 1); furthermore, 23 cases were also reported as lost to follow-up in the first 14 days following diagnosis, yet only 3 censor marks are evident in Figure 1 in these first 2 weeks. It would seem that at least 72/149 (48%) cases were censored from survival analysis, which substantially impacts confidence in the survival estimates derived. There was no reported attempt to determine the true status (alive or dead) of the cases at a point 6 months from presentation, from which a genuine survival percentage could be given. It appears tthat cases may have relapsed and potentially died or have been euthanized after their second follow-up visit yet have been censored from the survival data. We would be grateful if you would invite the authors to further clarify how follow-up data were sought, to summarise the vital status of the study cohort at 6 months, and to comment on the validity of extrapolations of survival statistics when working with censored data.

Yours sincerely,

Ancillary