We aimed to review occupational burnout predictors, considering their type, effect size and role (protective versus harmful), and the overall evidence of their importance. MEDLINE, PsycINFO, and Embase were searched from January 1990 to August 2018 for longitudinal studies examining any predictor of occupational burnout among workers. We arranged predictors in four families and 13 subfamilies of homogenous constructs. The plots of z-scores per predictor type enabled graphical discrimination of the effects. The vote-counting and binomial test enabled discrimination of the effect direction. The size of the effect was estimated using Cohen's formula. The risk of bias and the overall evidence were assessed using the MEVORECH and GRADE methods, respectively. Eighty-five studies examining 261 predictors were included. We found a moderate quality of evidence for the harmful effects of the job demands subfamily (six predictors), and negative job attitudes, with effect sizes from small to medium. We also found a moderate quality of evidence for the protective effect of adaptive coping (small effect sizes) and leisure (small to medium effect sizes). Preventive interventions for occupational burnout might benefit from intervening on the established predictors regarding reducing job demands and negative job attitudes and promoting adaptive coping and leisure.