Abstract
Purpose :
While studies have suggested relationships between neurotoxic heavy metals (HM) and diabetes (DB), they did not correlate such findings with complications like diabetic retinopathy (DR). We performed a retrospective, observational cohort study to assess the relationship between HM and DR.
Methods :
We extracted all patient data from CDC’s NHANES with DB and ophthalmic parameters, corresponding to NHANES 2005/6 and 2007/8 cycles. We recorded socio-demographic factors, DB-related factors, HM measures and DR severity. To correct for cycle variations in testing, HM levels were transformed into cycle-specific quartiles. We performed statistical tests (Mann-Whitney U, Pearson Chi2, Spearman’s Rho Coefficient (p), Welch’s T-test, multivariable logistic regression), in SPSS, v26. Outcome measures included: serum and urine levels of HMs compared to the degree of DR, past medical history and social history.
Results :
884 eligible patients were classified into 4 groups based on their DR severity: no DR (n=565, 46.5% men), mild DR (n=203, 56.7% men), moderate DR (n=85, 56.3% men), and proliferative DR (n=31, 35.5% men). The average age in each group was respectively 63.1±10.6, 65.1±11.0, 61.4±10.2 and 63.4±7.4 years. Mean serum HM levels for lead, cadmium, total mercury and inorganic mercury did not differ significantly in the presence or absence of DR (p>0.05, Welch’s t-tests). All mean urinary metal levels were not significantly different in the presence or absence of DR, except for thallium (UThal) which had significant lower levels in DR (p<0.013). Adjustments for cycle variations confirm the independent association of DR and UThal, and the absence of correlation with other blood and urinary HM levels (Figure 1). UThal was the most significant correlation with a Spearman p of -0.214 (p=0.0003). We then analyzed the population to see which other parameter was associated with UThal levels. UThal was significantly associated with eGFR (p=0.002), hemoglobin (p=0.040), and insulin therapy (p=0.010). Multivariate logistic regression with 3 models are presented in Table 1. UThal was still independently associated with DR in all models (p=0.011-0.042).
Conclusions :
UThal is significantly associated to DR severity in the NHANES database. Additional validation for geographic factors are warranted. If validated, these findings suggest a possible role for thallium, a neurotoxic HM, in DR.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.