Table 2.

Association of Cannabis Use with Acute Kidney Injury

Cannabis use versus negative for all substance use	Unadjusted/Model 1, OR (95% CI)	p	Adjusted (IPTW weights), OR (95% CI)	p
Main analysis	0.87 (0.53–1.42)	0.6	0.85 (0.38–1.87)	0.7
Sensitivity analyses
Double robust estimation			0.79 (0.34–1.83)	0.6
Adjusted (winsorized IPTW weights)			0.81 (0.46–1.43)	0.5
DRE (winsorized IPTW weights)			0.79 (0.34–1.83)	0.6
Incremental analysis (weights not included)
Model 2			0.84 (0.51–1.39)	0.5
Model 3			0.91 (0.55–1.51)	0.7
Model 4			0.93 (0.55–1.57)	0.8
Model 5			0.83 (0.48–1.41)	0.5

IPTW weights are calculated from the variables shown in Table 1. Doubly robust estimation: “doubly robust” means the model includes all the variables used to calculate IPTW weights and weights too. Winsorized weights: Weights adjusted at 90 percentiles, respectively, for each group. Model 1: Unadjusted analysis. Models 2–5 presented here are incrementally adjusted, as follows: Model 2: Model 1+demographics (sex, race, and age). Model 3: Model 2+comorbidities (chronic pulmonary disease, liver disease, diabetes, hyperlipidemia, hypertension, and post-traumatic stress disorder)+access type+smoking. Model 4: Model 3+use of medications (analgesics, psychiatric drugs, antimicrobials, antiretrovirals, cardiovascular medications, chemotherapeutic inhibitors, diuretics, proton pump inhibitors, H2 receptor blockers, warfarin, anticoagulants, and benzodiazepines). Model 5: Model 4+vital signs (mean systolic blood pressure, mean DBP, mean BMI, and mean pain score)+baseline eGFR (eGFR intercept). DRE, doubly robust estimation; IPTW, inverse probability of treatment weighting.