Estimated GFR and Incidence of Major Surgery: A Population-Based Cohort Study

RATIONALE & OBJECTIVE Kidney disease is associated with an increased risk of postoperative morbidity and mortality. However, the incidence of major surgery on a population level is unknown. We aimed to determine the incidence of major surgery by level of kidney function. STUDY DESIGN Retrospective cohort study with entry from January 1 2008 until December 31 2009, and outcome surveillance from January 1 2010 to December 31 2016. SETTING & PARTICIPANTS Population-based study using administrative health data from Alberta, Canada; adults with an outpatient serum creatinine measurement or on chronic dialysis formed the study cohort. EXPOSURE Participants were categorized into six eGFR categories (all in mL/ min/1.73m2), ≥ 60 (G1-2), 45-59 (G3a), 30-44 (G3b), 15-29 (G4), and <15 with (G5D) and without dialysis (G5ND). eGFR was examined as a time-varying exposure based on means of measures within three-month ascertainment periods throughout the study period. OUTCOME Major surgery defined as surgery requiring admission to hospital for at least 24 hours. ANALYTICAL APPROACH Incidence rates (IRs) for overall major surgery were estimated using quasi-Poisson regression, and adjusted for age, sex, income, location of residence, albuminuria, and Charlson comorbidities. Age and sex-stratified IRs of 13 surgery subtypes were also estimated. RESULTS 1,455,512 cohort participants were followed for a median 7.0 years (IQR: 5.3), over which time 241,989 (16.6%) underwent a major surgery. Age and sex modified the relationship between eGFR and incidence of surgery. Males less than 65 years of age receiving chronic dialysis experienced the highest rates of major surgery, with an adjusted IR (95% Confidence Interval) of 243.8 per 1,000 person-years (179.8, 330.6). There was a consistent trend of increasing surgery rates at lower eGFR levels for most subtypes of surgery. LIMITATIONS Outpatient preoperative serum creatinine measurement was necessary for inclusion and outpatient surgical procedures were not included. CONCLUSIONS People with reduced eGFR have a significantly higher incidence of major surgery compared to those with normal eGFR, and age and sex modify this increased risk. This study informs our understanding of how surgical burden changes with differing levels of kidney function.