Albuminuria is not associated with bacteriuria in patients with type 2 diabetes mellitus: a meta-analysis of observational studies

Abstract

Background/aims: It has been recommended that urinary albumin be measured in sterile urine for the proper diagnosis of diabetic nephropathy. However, the association between bacteriuria and urinary albumin is controversial.

Methods: A systematic review and meta-analysis was performed to investigate the association of albuminuria and bacteriuria in patients with diabetes. Medline and Embase were searched (beginning in 1950 until November 2010). Data were extracted independently by two investigators. The pooled OR estimates were calculated using the random effects model.

Results: We identified 305 studies in the database searches. A total of seven studies were included, providing data from 1,552 patients (mean age 56.4 years). The OR of bacteriuria for the presence of micro- and/or macroalbuminuria was 1.60 (95% CI: 0.97-2.66, I(2) = 66.6%) as compared to patients without bacteriuria. Funnel plots and the Egger regression test suggested no significant asymmetry in the analysis (p = 0.21). In a sensitivity analysis including the five studies (1,197 participants) that evaluated microalbuminuria as the outcome, the OR of bacteriuria for microalbuminuria was 1.22 (95% CI: 0.68-2.19).

Conclusion: In conclusion, no association was identified between albuminuria and bacteriuria considering the current literature. Further prospective studies of a large diabetic population are needed to clarify such an association.