Determinants Of Acute Kidney Injury Risk During Simulated Occupational Heat Stress

PURPOSE: Preclinical models and lab- and field-based studies in humans indicate that occupational heat stress (OHS) is associated with AKI risk. Hyperthermia and dehydration are equally important in determining AKI risk in a lab model with a fixed level of OHS. The relative importance of hyperthermia and dehydration in development of AKI risk across a range of OHS conditions is unknown. We tested the hypothesis that hyperthermia and dehydration are equally important in AKI risk during simulated OHS that varied by heat exposure (wet bulb globe temperature, WBGT), rate of metabolic heat production (Hprod), and work-rest ratios. METHODS: 60 healthy adults (29 females, age: 28 ± 6 y, BMI: 24.9 ± 3.0 kg/m2) completed 1-5 OHS simulations for 2- (n = 12) or 4- hours (n = 122) and differed by WBGT (24 °C, 26 °C, 28 °C, 30 °C, 36 °C), work-rest ratio (15-, 30-, 45-, or 60- min work per hour), and Hprod (230 W, 338 W, 400 W, 430 W, 530 W). Sport drinks were available for drinking ad libitum. Core temperature (Tcore) was measured and percent changes in nude body weight (%ΔBW) provided an index of hydration. Urine was collected pre-, post- and 1 h post- OHS. AKI risk was quantified as the peak change in the product of specific gravity corrected urine insulin-like growth factor-binding protein 7 and tissue inhibitor of metalloproteinase 2 (Δ[IGFBP7·TIMP-2]). Data were analyzed using a two component zero-adjusted (log normalized) regression model with Δ[IGFBP7·TIMP-2] values <0.3 (ng/mL)2/1000 interpreted as no AKI risk. Initial predictors were age, BMI, gender, and normalized peak Tcore and %ΔBW. RESULTS: Peak Tcore, %ΔBW, and Δ[IGFBP7·TIMP-2] was: 38.1 ± 0.5 °C, -0.4 ± 1.0%, 3.9 ± 14.3 (ng/mL)2/1000. Age (p = 0.702), BMI (p = 0.862) and gender (p = 0.345) were not significant predictors in the initial model. A lower peak Tcore (β = -2.9, p = 0.004) and smaller reductions in %ΔBW (β = 5.8, p = 0.004) predicted the odds of no AKI risk. Higher peak Tcore (β = 3.7, p < 0.001) and greater reductions in %ΔBW (β = -4.5, p < 0.001) predicted Δ[IGFBP7·TIMP-2] when ≥0.3 (ng/mL)2/1000. The interaction terms for peak Tcore and %ΔBW were not significant for the odds of no AKI risk (p = 0.163) and mean AKI risk (p = 0.486). CONCLUSIONS: Tcore and %ΔBW are independent predictors AKI risk during simulated OHS, with normalized %ΔBW being the stronger predictor. Supported by: R01OH011528