Background: In older adults, kidney function declines with age. People with advanced kidney diseases may have poor olfaction. However, it is unclear whether poor olfaction is a marker for declining renal function or future risk of chronic kidney disease (CKD). We therefore investigated olfaction in relation to kidney function and risk of CKD.

Methods: These secondary data analyses were limited to participants of the year 3 clinical visit of the Health Aging and Body Composition Study. The analytic sample size varied between 1427 to 2531, depending on participant eligibility and data availability for each analysis. Olfaction was tested using the Brief Smell Identification Test (B-SIT), defined as anosmia (score≤6), hyposmia (7-8), moderate (9-10), and good function (10-11) at baseline. We estimated glomerular filter rate (eGFR) at baseline and seven years later using the CKD-EPI creatinine-cystatin C equation, and defined incident CKD as eGFR<60 ml/min/1.73m2 and eGFR decline ≥1 ml/min/1.73m2/year. Further, we identified CKD hospitalization events from hospitalization and death records. We used inverse probability weighting and weighted multivariable regressions to account for censoring in the prospective analyses and used absolute risk regression to account for competing risk of death.

Results: At baseline, compared to participants with good olfaction, the multivariable-adjusted mean eGFR was 3.00 ml/min/1.73m2 lower (95% confidence interval (CI): -5.25, -0.75) for those with anosmia and 1.87 lower (95% CI: -3.94, 0.21) for those with hyposmia with a P for linear trend < 0.001. Those with anosmia at baseline was had a significantly lower eGFR seven years later (-5.31, 95% CI: -8.58, -2.04, P for trend = 0.002), but the association was attenuated after further accounting for baseline eGFR (-2.37, 95%CI: -4.91, 0.16, P for linear trend = 0.147). Olfactory function was not associated with incident CKD or CKD hospitalization.

Conclusion: In older adults > age 70 years, poor olfaction is associated with lower kidney function, but not future CKD risk. These associations should be further investigated in relatively younger population.