The secret of immortal time bias in epidemiologic studies.

In the March 2007 issue of JASN, Hemmelgarn et al.1 reported a 50% reduction in the risk for all-cause mortality for patients who had chronic kidney disease (CKD) and attended multidisciplinary care (MDC) clinics compared with those who received usual care. Their survival curves showed a clear divergence in rates of death between the two groups in the first 6 mo of follow-up. We suggest that it is less plausible from a biologic perspective that use of MDC clinics immediately reduces the short-term risk for death. Rather, much of the early observed effect may be due to survivor treatment selection bias, also known as immortal time bias. Here we consider this issue. In the Hemmelgarn study, a retrospective cohort of 187 clinic patients who were exposed to a MDC clinic were matched to 187 non-MDC clinic control patients to examine the association between MDC and survival.1 Control subjects were chosen on the basis of propensity matching, whereby individuals in the control group had a similar likelihood of being referred to a MDC clinic as those in the MDC clinic group. Figure 1 shows a schematic of how patients with CKD entered the cohort. All patients were required to have an outpatient serum creatinine test performed between July 1 and December 31, 2001. Patients in the MDC clinic group were also required to have attended a MDC clinic between July 1, 2001, and December 31, 2002. The primary analysis for the study was the association between MDC clinic visits and survival, modeled using a Cox regression analysis. Survival time was measured starting from each patient’s serum creatinine test date. In other words, the date of each patient’s serum creatinine represented the date they entered the cohort, or time 0. Patients were followed until the end of the assessment (December 31, 2004) or death, whichever came first. A difference in survival between the two groups was illustrated using Kaplan-Meier survival curves (Figure 2). In this analysis, censoring occurred only at the end of assessment; therefore, the curves essentially represent the proportion of patients who were still alive at each time during follow-up. The curves were step-like in shape, and a dip in the curve occurred when a patient in that group died.2 As can be seen from Figure 2, the curves diverged almost immediately, with the non-MDC clinic curve dipping below the MDC clinic curve, signifying an increased death rate for the non-MDC control group. The difference in the proportion of individuals alive between the two groups steadily increased until about 1.5 yrs, after which point the rate of decline was similar between the groups. The difference in curves was tested using a log-rank test and found to be highly significant (P 0.008). The Cox model yielded a risk reduction of 50% with 95% confidence limits ranging from 29 to 65%. Is this result biologically plausible? From a mechanistic perspective, we suggest that it is less plausible that attending MDC clinics confers an immediate survival advantage over regular care for elderly patients with CKD. These clinics concentrate on better education, lifestyle modification, and medical management over that provided in routine care. Although better efforts at smoking cessation, weight management, dietary protein restriction, glycemic control, renin-angiotensin blockade, BP lowering, and statin use all could improve survival in this high-risk population, practical experience suggests that such a benefit would likely take longer to manifest.3– 6 It is also improbable that better potassium control explains the large early survival benefit. Much of the early observed beneficial effect may be due to survivor treatment selection bias,7 more recently described as immortal time bias.8 First noted in 1885,9 the bias explains the suggestion that Popes seem to live longer than artists10 or Oscar winners longer than nonwinners.11 In general, such individu-