Finally, the Cox proportional hazards model did not detect a significant covariate after adjusting for sex, age, total number of injections, and major concomitant diseases such as hypertension, diabetes mellitus, dyslipidemia, cerebrovascular accident, and lung disease through model 1 to model 4.