Skip to main content
. 2016 May 25;25(10):1196–1203. doi: 10.1002/pds.4030

Table 3.

Hazard ratio of lactic acidosis for metformin by CLD/CKD

Variables Main cohort (before matching)* Main cohort adjusted by propensity score (20 strata) Propensity score‐matched cohort
Number of LA cases 30 30 10
Non metformin users 23 23 5
Metformin users 7 7 5
Hazard ratio
Non metformin users Ref Ref Ref
Metformin users 0.92 [0.33–2.55] 0.89 [0.32–2.49] 0.90 [0.26–3.11]
Age ≥ 75 y 1.67 [0.79–3.53] 1.36 [0.61–3.04] 2.02 [0.41–9.09]
Female 1.03 [0.49–2.17] 1.10 [0.52–2.30] 1.05 [0.29–3.77]
Duration of metformin use before cohort entry (day) 1.00 [1.00–1.01] 1.00 [1.00–1.01] 1.01 [1.00–1.01]
Insulin 1.64 [0.76–3.51] 2.85 [0.31–26.39]
Hypertension 1.26 [0.54–2.89] 1.03 [0.21–5.10]
Ischemic heart disease 1.11 [0.42–2.89] 0.96 [0.83–1.10]
Heart failure 2.37 [0.98–5.75] 2.39 [1.00–5.71]
Complications of diabetes 1.73 [0.78–3.84] 1.94 [0.90–4.19] 2.58 [0.67–10.01]
Dyslipidemia 0.48 [0.06–3.81]
Cerebrovascular disease 1.65 [0.20–13.73]
Gastric ulcer 2.05 [0.86–4.93] 1.64 [0.20–13.79]
Chronic liver disease 1.44 [0.65–3.17] 1.42 [0.65–3.13] 3.35 [0.97–11.58]
Liver Cirrhosis
Chronic kidney disease 7.33 [3.17–16.96] 7.11 [3.08–16.43] 23.12 [4.90–109]
Renal failure
*

The covariates which caused at least a 10% shift in the risk estimate for the univariate analysis were adjusted in the final model. The COX model for each analysis was constructed separately.

We categorized the continuous propensity score into 20 strata of 5% each for the distribution of scores. The covariates which caused at least a 10% shift in the risk estimate after given by stratum of propensity score were adjusted in the final model.

Variables in propensity score matched cohort were adjusted for stratum of propensity score only, which were estimated using the same variables.