In this community-based longitudinal study of adults, the risks related to a faster decline in eGFR were different among age groups. In relation to hypertension, proteinuria, and smoking status, faster declines in eGFR were observed in the older group. Other risk factors including anemia and diabetes mellitus were related to a similar rate of eGFR decline among the age groups.
Age is one of the most fundamental considerations for providing personalized care. The risk factors for kidney dysfunction have been evaluated in cohort studies or interventional trials [5, 22]. For example, a study of a general Japanese population showed that the relationship between blood pressure and GFR slopes varied across age groups . However, age-adjusted risks, not age-stratified risks, have usually been evaluated and thus reported risk factors are limited. To achieve more personalized care, we conducted an age-stratified risk assessment of comprehensive risk factors for loss of kidney function.
Clinical significance and study implications
In this study, higher systolic blood pressure was related to a faster loss of kidney function in older people. This result was consistent with those of previous studies reporting that the absence of hypertension was related to a slower progression of kidney disease in older people [23, 24]. A possible mechanism for this association might be due to glomerular hypertension by hyalinization of afferent arterioles  that may cause loss of autoregulation of glomerular blood flow . Because active treatment of hypertension is a risk factor for kidney failure , a study of renal outcome with respect to hypertension treatment in older people is required to evaluate the balanced target of blood pressure treatment in older people.
Current smoking is a known risk factor for the loss of kidney function in middle-aged or older people . It was interesting that this relationship with smoking was observed even in the participants aged ≥ 80 years. Current smoking was not associated with reduced kidney function in young people, but this result should be interpreted carefully because smoking may transiently increase eGFR.
In this study, higher BMI was related to a slightly faster loss of kidney function in older people. However, no age-dependent trend was observed. Moreover, the effect sizes related to higher BMI were small in each age group. Considering that most participants had BMI < 30 kg/m2, this result was consistent with that in other studies that reported that a BMI of 25 to 30 kg/m2 was not related to a faster loss of kidney function in any age group .
Due to the large sample size, this study might have detected a small difference in the GFR slope. However, most of the significantly different risk factors observed in this study had a difference of -0.10 mL/min/1.73 m2 from the mean slope. Considering that the mean slope was 0.39, the presence of multiple factors could lead to a clinically significant loss of kidney function. Furthermore, risk factors that increased in association with older age might become more important over time, which is shown in Fig. 2.
Strengths and limitations
More than 50,000 Japanese people with ages 40 to 80 years or older participated in this study. The participants may not be different from the typical Japanese population. The mean eGFR slope in the study population was 0.39 mL/min/1.73 m2 per year, which was almost the same as that reported in a nationwide study in Japan . Furthermore, the wide range of characteristics may contribute to the generalizability of the results.
Another point of strength of this study was the inclusion of > 8,000 participants aged ≥ 80 years. Old-age populations are usually not stratified because of either their limited number or they are not the population of interest. People aged ≥ 80 years are increasing worldwide . The results of this study provide basic information about the risk factors of loss of kidney function in older people. Interventional trials including this age group are warranted to validate the results of this present study.
This study has several limitations. First, drug classes for the treatment of hypertension or diabetes mellitus were not considered in the analysis. Some medications, including angiotensin II receptor blockers or sodium-glucose cotransporter-2 inhibitors, may affect kidney function . Second, we did not investigate the underlying diseases of kidney dysfunction that might affect the progression of the disease. Third, smoking history was not included in the questionnaire. Because the questionnaire asked about the current smoking status at the time of examination, the risks related to smoking in older people may have been underestimated. Fourth, although a median follow-up period of 4.0 years may be short due to the slow progression of renal impairment in the general population, studies of kidney function have found that annual changes in eGFR over a 3-year follow-up period can be used as a surrogate outcome for assessing risk of ESKD even in those with preserved kidney function [33, 34].