Abstract
Background: This cohort study was conducted to devise an approximation formula for predicting the glomerular filtration rate (GFR) after 1 year using annual medical examination data from the general population. Methods: Consecutive annual medical examination data were obtained for 41,337 inhabitants. Machine learning with the random forest method was used to assess the importance of each clinical parameter in terms of its association with estimated GFR (eGFR) after 1 year. An approximation formula was developed by multiple linear regression analysis based on the four most important clinical parameters. The relationship between the GFR after 1 year approximated by our formula and the eGFR after 1 year was analyzed using Pearson's correlation coefficient. Results: The following approximation formula was obtained by multiple linear regression analysis: approximate GFR after 1 year (mL/min/1.73 m(2)) = -0.054 × age + 0.162 × hemoglobin - 0.085 × uric acid + 0.849 × eGFR + 11.5. The approximate GFR after 1 year was significantly and strongly correlated with the eGFR at that time (r = 0.884; p < 0.001). Conclusions: An approximation formula including age, hemoglobin, uric acid, and eGFR may be useful for predicting GFR after 1 year among members of the general population.