Purpose: To examine whether there is an association between diabetes mellitus (DM) and keratoconus. Design: A retrospective comparison of the proportion of keratoconus patients with DM versus the proportion of overall patients with DM,and a retrospective cross-sectional study of a cohort consisting of all diabetic keratoconus patients and randomly selected keratoconus patients without DM at a single center. Participants: Patients seen at theWilmer Eye Institute from January 1,1995,through March 18,2004. Methods: Review of billing data and clinic charts. Eligibility criteria for the cross-sectional study were 1 or more clinic visits,lack of other eye pathologic features (excluding cataract and diabetic retinopathy),and absence of bilateral penetrating keratoplasty (PK) at presentation. Application of novel keratoconus severity index was based on best-corrected visual acuity (BCVA) in the better eye at last visit and defined as: grade 1 (least severe),spectacle wear with BCVA of 20/40 or better; grade 2 (intermediate),spectacle wear with BCVA worse than 20/40 or rigid gas permeable lens wear; grade 3 (most severe),PK. Main Outcome Measures: Prevalence of DM in keratoconus patients and those without keratoconus,odds ratio of having DM on a diagnosis of more severe keratoconus,and prevalence of DM in keratoconus patients and those without keratoconus who underwent corneal transplantation. Results: There was no difference in the prevalence of DM in keratoconus patients and those without keratoconus,and there was no difference in the prevalence of DM in keratoconus patients and those without keratoconus undergoing PK. However,our results suggest a negative association between DM and severity of keratoconus (P=0.03,Fisher exact test). The odds of being in the most severe group as opposed to the least severe group were lower in DM patients than in those without DM(P=0.01; odds ratio OR=0.20; 95% confidence interval CI,0.05-0.70). Compared with those without DM,DM patients also had lower odds of being in the intermediate group than in the least severe group (P=0.02; OR=0.25; 95% CI,0.08-0.80). After adjustment for age,gender,and race,these differences remained statistically significant. Conclusions: We found that DM is not associated with a diagnosis of keratoconus,but having DM decreases the odds of having more severe keratoconus. Design: A retrospective comparison of the proportion of keratoconus patients with DM versus the proportion of overall patients with DM, and a retrospective cross-sectional study of a cohort consisting of all diabetic keratoconus patients and randomly selected keratoconus patients without DM at a single center. Participants: Patients seen at the Wilmer Eye Institute from January 1, 1995, through March 18, 2004. Methods: Review of billing data and clinic charts. Eligibility criteria for the cross-sectional study were 1 or more clinic visits, lack of other eye pathologic features (excluding cataract and diabetic retinopathy), and absence of bilateral penetrating keratoplasty (PK) at presentation. Application of novel keratoconus severity index was based on best- corrected visual acuity (BCVA) in the better eye at last visit and defined as: grade 1 (least severe), spectacle wear with BCVA of 20/40 or grade 2 (intermediate), spectacle wear with BCVA worse than 20/40 or rigid gas permeable lens wear; grade 3 (most severe), PK. Main Outcome Measures: Prevalence of DM in keratoconus patients and those without keratoconus, odds ratio of having DM on a diagnosis of more severe keratoconus, and prevalence of DM in keratoconus patients and those without keratoconus who underwent corneal transplantation. Results: There was no difference in the prevalence of DM in keratoconus patients and those without keratoconus, and there was no However, our results suggest a negative association between DM and severity of keratoconus (P = 0.03, Fisher exact test). The odds of being in the most severe group as opposed to the least severe group were lower in DM patients than in those without DM (P = 0.01; odds ratio OR = 0.20; 95% confidence interval CI, 0.05-0.70)lso had lower odds of being in the intermediate group than in the least severe group (P = 0.02; OR = 0.25; 95% CI, 0.08-0.80). After adjustment for age, gender, and race, these differences remained statistically significant. Conclusions: We found that DM is not associated with a diagnosis of keratoconus, but having DM decreases the odds of having more severe keratoconus.