Population attributable fraction and its relation to time and expected glomerular filtration rate
Yoojin Lee, Sung Kweon Cho
Ajou University, College Of Medicine, Suwon-Si
Potential risk factors for hyperuricemia have been known as body mass index (BMI), hypertension, type 2 diabetes, metabolic syndrome, and kidney function. Gender and body mass index are potential moderators. However, little research investigates how much these risk factors attribute to hyperuricemia. Using the longitudinal nhanes cohort from 1998 to 2018, we investigated those risk factors' population attributable fraction (PAF).
First, we examined the difference in the hyperuricemia prevalence from the first to the last collection cycle by ethnicity. Incidence of hyperuricemia increases in african american and hispanic populations (16.29% to 19.81%, 11.90% to 14.96%, respectively), not in caucasian and the other ethnic groups. We then estimated PAF using the punaf package in stata, whether the estimated PAF was predicted by time or degrees of kidney function (expected glomerular filtration rate (EGFR)) using a simple linear regression model. BMI, hypertension, diabetes, smoking, metabolic syndrome are the factors associated with PAF of hyperuricemia over time, whereas EGFR and gender are not. We also found that PAF of BMI and smoking decreased in the other ethnic group. BMI, hypertension, diabetes, smoking, and metabolic syndrome are associated with PAF of hyperuricemia throughout the EGFR level, whereas gender is not. Ethnic-specific findings were also observed in the EGFR model (hypertension for caucasian, diabetes for hispanic and caucasian groups).
Our results showed the unique trend of PAF of hyperuricemia and increase of hyperuricemia prevalence. Ethnic-specific risk factors for hyperuricemia needs to be investigated further. We will apply this approach in korean cohorts - knhanes and koges.