Table 6.
Cox proportional hazards regression analysis to identify the association between insulin use and adverse outcomes in DM cohort
| HR (95% CI) | P-value | |
|---|---|---|
| Unadjusted model | ||
| Non-insulin-treated DM | 1.00 (reference) | |
| Insulin-treated DM | 3.01 (1.66–5.44) | P < 0.001 |
| Model 1: adjusted for age, sex, BMI* and NYHA function class | ||
| Non-insulin-treated DM | 1.00 (reference) | |
| Insulin-treated DM | 2.94 (1.62–5.33) | P < 0.001 |
| Model 2: adjusted for model 1 combined with etiology of HF | ||
| Non-insulin-treated DM | 1.00 (reference) | |
| Insulin-treated DM | 2.93 (1.61–5.34) | P < 0.001 |
| Model 3: adjusted for model 2 combined with NT-proBNP§, eGFR, and use of sacubitril/valsartan | ||
| Non-insulin-treated DM | 1.00 (reference) | |
| Insulin-treated DM | 2.85 (1.40–5.51) | P = 0.007 |
| Model 4: adjusted for model 3 combined with HbA1C and DM duration | ||
| Non-insulin-treated DM | 1.00 (reference) | |
| Insulin-treated DM | 2.16 (1.08–4.59) | P = 0.030 |
* BMI is considered as binary variable (BMI ≥ 25 kg/m2) when including in the Cox model
§ NT-proBNP is log-transformed before being included in the Cox model
Abbreviations as in Table 5