Pulmonary function in diabetes: a metaanalysis.

Abstract

Research into the association between diabetes and pulmonary function has resulted in inconsistent outcomes among studies. We performed a metaanalysis to clarify this association. From a systematic search of the literature, we included 40 studies describing pulmonary function data of 3,182 patients with diabetes and 27,080 control subjects. Associations were summarized pooling the mean difference (MD) (standard error) between patients with diabetes and control subjects of all studies for key lung function parameters. For all studies, the pooled MD for FEV(1), FVC, and diffusion of the lungs for carbon monoxide were -5.1 (95% CI, -6.4 to -3.7; P < .001), -6.3 (95% CI, -8.0 to -4.7; P < .001), and -7.2 (95% CI, -10.0 to -4.4; P < .001) % predicted, respectively, and for FEV(1)/FVC 0.1% (95% CI, -0.8 to 1.0; P = .78). Metaregression analyses showed that between-study heterogeneity was not explained by BMI, smoking, diabetes duration, or glycated hemoglobin (all P > .05). Diabetes is associated with a modest, albeit statistically significant, impaired pulmonary function in a restrictive pattern. Since our results apply to the diabetic subpopulation free from overt pulmonary disease, it would next be interesting to investigate the potential clinical implications in those patients with diabetes who carry a pulmonary diagnosis, such as COPD or asthma.