On average, patients in the trials with the highest mortality rates -- versus those in trials with the lowest rates -- were older (64.2 versus 59), had a longer diabetes duration (15.2 versus 7.2 years), and had a higher systolic blood pressure (145 versus 136 mm Hg), serum creatinine (1.8 versus 0.9 mg/dL), and proteinuria prevalence (100% versus 9%).
Both the average LDL cholesterol level and prevalence of current smoking decreased across the increasing mortality categories.
"While the lower prevalence of smoking is probably related to the inclusion criteria, reports from cohorts of chronic kidney disease and heart failure demonstrated that lower cholesterol correlates with higher mortality -- what is described as 'reverse epidemiology' in other populations of patients with advanced disease," the authors wrote.
Selection for hypertension was common across trials in all four of the mortality categories, ranging from 69% to 97% of patients. Hypertension was not a strong predictor of mortality, but rates were lowest in primary prevention trials and those in which hypertension was the only additional inclusion criterion.
Mortality rates were higher in trials selecting for prior cardiovascular disease compared with those for which cardiovascular disease was simply permitted but not required (2.0 to 2.5 versus 0.9 to 2.0 per 100 patient-years).
The highest rates, however, occurred in the three trials that required the presence of chronic kidney disease -- defined as either elevated serum creatinine or eGFR less than 60 mL/min/1.73 m2 with or without the presence of proteinuria.
The authors acknowledged that the study was limited by the lack of individual patient data, of information on the effects of medications and therapeutic interventions during the trials, and of information on the effects of the randomized treatment arms and their relationships with mortality.