Renal function changes and seasonal temperature in patients undergoing cardiac surgery

Some observations in humans and other mammalians suggest that serum creatinine (SC) and estimated glomerular filtration rate (eGFR) may change during the warm season. The objective of this study is to determine if temperature-dependent seasonal changes in levels of SC and eGFR are detectable in cardiac surgery patients, with associated changes in postoperative acute kidney injury (AKI) incidence. This is a single-center retrospective study based on the institutional database of cardiac surgery in the period 2000-2012. Sixteen-thousand and twenty-three consecutive adult patients undergoing cardiac surgery comprised the study population. Baseline and postoperative SC and eGFR values, and AKI rate according to the month when surgery was performed were measured. The month-related changes SC and eGFR, and AKI rate, were assessed in crude and adjusted models, and their association with the correspondent meteorological data registered at the time of surgery was tested. Patients operated in the six warmest months (May through October) had a significant (p<0.001) higher value of baseline SC (1.17 +/- 0.7 mg/dL) versus the six coldest months (1.12 +/- 0.6 mg/dL), and a significantly (p = 0.031) higher value of peak postoperative SC (1.31 +/- 0.85 mg/dL) versus the 6 coldest months (1.28 +/- 0.89), with maximum values between July and August. A similar behaviour was found for eGFR. After adjustment for other confounders, the AKI rate was not significantly different in the warmest months, even if a trend towards a higher rate in August was observed (odds ratio 1.287, 95% confidence interval 0.96-1.74, p = 0.097). Baseline (p<0.001) and peak postoperative (p = 0.0054) serum creatinine levels were significantly higher for increasing mean ambient temperature. Humidity and wind speed were negatively associated with pre- and postoperative eGFR. In conclusion, patients operated during the warmest season, have higher levels of SC and lower levels of eGFR, without a correspondent increase in the AKI rate. Different hypotheses underlying this pattern are generated by this study, including a dehydration status, concomitant anemia, and a higher transfusion rate.