OBJECTIVE- To analyze smell function in diabetic patients and healthy
control subjects, with reference to individual characteristics and maj
or complications of the disease. RESEARCH DESIGN AND METHODS- A cross-
sectional study of smell recognition in 68 diabetic patients and 30 co
ntrol subjects without known cause of smell impairment was conducted.
Smell was studied using a kit of flavors that patients were asked to r
ecognize, leading to a smell recognition score ranging from 0 (no reco
gnition) to 20 (perfect recognition). Demographic and clinical data, a
nd the electrogustometric threshold were recorded in both groups. Meta
bolic control and degenerative complications were studied in diabetic
patients. RESULTS- The SRS was significantly lower in diabetic patient
s than in control subjects (12.4 +/- 0.5 vs. 15.1 +/- 0.5, P < 0.001).
By use of univariate and multivariate analyses, this difference could
not be explained by individual factors such as age, sex, body mass in
dex, blood pressure, or tobacco or alcohol consumption but was related
to diabetes. SRS correlated with EGT (r = 0.70, P < 0.001). in the di
abetic group, SRS was associated with age (r = 0.29, P < 0.05), durati
on of diabetes (r = 0.27, P < 0.05), microalbuminuria (12.2 +/- 0.8 vs
. 14.4 +/- 0.9, P < 0.05), peripheral neuropathy (10.3 +/- 1.0 vs. 14.
1 +/- 0.9, P < 0.01), and EGT (r = 0.39; P < 0.001). By use of multiva
riate analysis, the strongest association was found with EGT (R2 = 0.1
5). CONCLUSIONS- These results suggest that smell recognition is impai
red in patients with diabetes mellitus. Smell dysfunction is associate
d with age and degenerative complications of diabetes, suggesting a de
generative mechanism related to diabetes.