Examining Bias in Studies of Statin Treatment and Survival in Patients With Cancer

Abstract

Importance Cancer patients who use statins appear to have a substantially better survival than non-users in observational studies. However, this inverse association between statin use and mortality in cancer patients may be due to selection bias and immortal time bias. Objective We used observational data to emulate a randomized trial of statin initiation that is free of selection bias and immortal time bias. Design We used data on 17,372 cancer patients from the SEER-Medicare 2007-2009 database with complete follow-up until 2011. Individuals were duplicated, with each replicate assigned to either the strategy “statin initiation within 6 months of diagnosis” or “no statin initiation”. Replicates were censored when they stopped following their assigned strategy, and the potential selection bias was adjusted for via inverse-probability weighting. Hazard ratios (HR), cumulative incidences, and risk differences were calculated for all-cause mortality and cancer-specific mortality. We then compared our estimates with those obtained using the same analytic approaches as in previous observational studies. Setting The SEER-Medicare data is a linkage between 17 American cancer registries and claims files from Medicare and Medicaid in 12 states. Participants We included individuals with a new diagnosis of colorectal, breast, prostate, and bladder cancer who had not been prescribed statins for at least 6 months before cancer. Exposure Statin initiation within 6 months from cancer diagnosis. Main outcome measure Cancer-specific and all-cause mortality. Results The adjusted HR (95% CI) comparing statin initiation vs. no initiation was 1.00 (0.88-1.15) for cancer-specific mortality and 1.07 (0.93-1.21) for overall mortality. Cumulative incidence curves for each group were almost overlapping. On the contrary, the methods used by prior studies resulted in a strong inverse association between statin use and mortality. Conclusion and relevance After using methods that are not susceptible to selection bias from prevalent users and to immortal time bias, initiation of statins within 6 months of cancer diagnosis did not appear to improve 3-year cancer-specific or overall survival.

Association between metformin use after surgery for colorectal cancer and oncological outcomes: A nationwide register‐based study

“…These findings are not consistent when compared to patients not diagnosed with diabetes and the results seen for metformin in the meta‐analyses of observational studies have not been replicated in meta‐analyses of randomized controlled trials . The association between metformin and improved cancer risk found in observational studies may be due to selection bias and immortal time bias and a recent large retrospective study not suffering from the mentioned biases were not able to show any association between metformin use and cancer risk . To emulate a randomized controlled trial and to avoid immortal time bias, a study design excluding prevalent users and only investigating post‐diagnostic use as a time‐dependent covariate were chosen.…”

“…Patients who were diagnosed with diabetes and redeemed prescriptions for antidiabetic medication in the year preceding surgery were excluded in the main analysis. This was done to emulate a randomized trial study design, and only examine incident patients with diabetes and naïve users of antidiabetic medication. Patients who died within 180 days after surgery and patients with metastases at time of surgery or within 180 days after surgery were excluded, as it was not relevant to investigate recurrence in these patients.…”

“Resuscitation time bias”—A unique challenge for observational cardiac arrest research

“…studies of statin exposure [6,7]) and has been described in the critical care setting [8,9]. The concept of immortal time bias is illustrated in Fig.…”

How to estimate the effect of treatment duration on survival outcomes using observational data

“…Suppose we want to estimate the effect of statins on the mortality of patients with cancer using a healthcare database 1. A direct comparison of long term users, short term users, and non-users would be biased because long term users have, by definition, survived for a long time.…”