Thirteen studies (n=522 patients, range eight to 94 patients) were included in the review: four RCTs, one CCT and eight prospective cohort studies.
All four RCTs had explicit eligibility criteria. Three were open-label rather than double-blinded. There was no loss to follow-up in three studies and 15.4% loss to follow-up in the fourth study, which did not use intention to treat. The CCT had explicit eligibility criteria, was non-randomised and open-label and had no reported loss to follow-up. Four of the eight prospective cohorts had explicit eligibility criteria. All were open label. No loss to follow-up was reported in six studies; for the other studies loss to follow-up was 3.3% and 40.2%
Proteinuria: Three studies (n=69 patients, range 17 to 30 patients) measured proteinuria. They included patients with nephrotic range proteinuria and used caloric restriction to try to reduce weight loss and produced a pooled mean difference in overt proteinuria of -1.66 g/day (95% CI -2.63 to -0.69 g/day, I2=59.5%), which suggested a statistically significant association between weight loss and proteinuria levels for this patient group.
Microalbuminuria: Eight trials (n=391 patients, range eight to 94 patients) produced a statistically significant pooled mean difference of -13.87mg/day (95% CI -17.12 to -10.61, I2=50%) that favoured weight loss interventions
Creatinine clearance: Seven trials (n=203 patients, range eight to 61 patients) produced a statistically significant pooled mean difference of -23.73 mL/minute (95% CI -36.12 to -11.35, I2=0%) that favoured weight loss interventions.
A number of meta-regressions and subgroup analyses were presented.