(162) Evaluation of The Case-Time-Control Design to Examine the Association Between Anticancer Drug Use and Acute Kidney Injury

Friday, August 26, 2022

8:00 AM – 6:00 PM CEST

Location: Hall C

Publication Number: 493

Presenting Author(s)

Philip Munch, MD
– Department of Clinical Epidemiology, Aarhus University, Aarhus N, Midtjylland, Denmark

Co-Author(s)

Simon K. Jensen, MD, PhD
– Epidemiologist, Department of Clinical Epidemiology, Department of Clinical Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark, Department of Clinical Epidemiology, Aarhus University, Aarhus, Midtjylland, Denmark
Uffe Heide-Jørgensen
– Aarhus N, Midtjylland, Denmark
Mette Norgaard, MD, PhD
– Professor, MD, PhD, Department of Clinical Epidemiology, Aarhus University, Aarhus N, Midtjylland, Denmark
Henrik Birn
– Department of Clinical Medicine, Aarhus University, Denmark, Denmark
Christian Christiansen, MD,PhD
– Professor, Consultant, Department of Clinical Epidemiology, Aarhus University and Aarhus University Hospital, Aarhus, Denmark, Aarhus N, Midtjylland, Denmark

Background: Studies suggest that use of anticancer drugs (ADs), including platinum-based agents is associated with acute kidney injury (AKI); however, observational studies aiming at estimating the effect of ADs on the risk of AKI are challenged by a lack of active comparator drugs with the same indication and by confounding by indication. The case-time-control design minimizes the time-stable confounding by indication and the effect of time trends of drug use.

Objectives: To evaluate the use of case-time-control (CTC) design to mitigate confounding by indication when examining the association between AD use and AKI in colorectal cancer (CRC) patients.

Methods: Using population-based registries, we identified patients diagnosed with CRC in Denmark between 2010 and 2018 with recorded creatinine measurements. We defined AKI according to the Kidney Disease Improving Global Outcomes (KDIGO) creatinine criteria. To investigate the transient impact of ADs on the risk of AKI events within two years after cancer diagnosis, we compared the incidence of AKI after AD administration (any type and oxaliplatin) using a CTC design, with the periods 1-30 and 61-90 days before AKI serving as the hazard and comparison period, respectively. For each case, we identified 5 controls matched on stage and year of cancer diagnosis. We excluded AKI events in the first 90 days after cancer diagnosis to ensure relevant hazard and comparison periods. In a sensitivity analysis, this period was increased to 180 days. A priori, we expected a transient increased incidence of AKI after AD administration, i.e., the odds ratios (ORs) from the CTC model would be greater than one.

Results: We identified 19,880 CRC patients with data on creatinine measurements and cancer stage. Of them, 8,870 (44.6%) received ADs and 1,857 (9.3%) developed AKI within 90 days to two years after the cancer diagnosis. The CTC model yielded an OR for AKI of 0.59 (95% confidence interval (CI): 0.47—0.75) for AD overall and 0.78 (95% CI: 0.55—1.10) for oxaliplatin in the period 90 days to two years after cancer diagnosis. The ORs for AKI in the period 180 days to two years after cancer diagnosis were 0.98 (95% CI: 0.72—1.33) and 1.34 (95% CI: 0.81—2.20), respectively.

Conclusions: The results from the CTC model varied substantially depending on whether AKI events were included from day 90 or day 180. The results from the period 90 days to two years after cancer diagnosis showed a transient unexpected lower risk of AKI after AD administration, indicating that the design of the study is susceptible to time-varying confounding. The variation of the results together with the risk of time-varying confounding complicate the interpretation of the results and indicate that the CTC design may not be suitable in this setting.