Lifestyle Factors and Indices of Kidney Function in the Framingham Heart Study

Abstract

Background and objectives

Lifestyle characteristics are modifiable factors that could be targeted as part of chronic kidney disease (CKD) prevention. We sought to determine the association of lifestyle characteristics with incident estimated glomerular filtration rate (eGFR)<60mL/min/1.73m² and rapid eGFR decline in older adults in the United States.

Methods

Prospective cohort study of Framingham Offspring participants with baseline eGFR<60mL/min/1.73m² (n=1802) who attended the seventh (1998–2001; baseline) and eighth (2005–2008; follow-up) examinations (mean age=59years, 54.8% women). Predictors included measures of diet quality, physical activity, alcohol intake, and current smoking status assessed during baseline. Outcomes were based on creatinine-based eGFR at baseline and follow-up and included incident eGFR<60mL/min/1.73m² (at follow-up) and rapid eGFR decline (annual eGFR decrease≥3mL/min/1.73m²).

Results

Over 6.6 years average follow-up 9.5% (n=171) of participants developed incident eGFR<60. A trend was observed across quartiles of diet quality, with higher levels of diet quality associated with a decreased odds ratio [OR] of incident eGFR<60 (p-trend=0.045). Higher diet quality was associated with decreased odds of rapid eGFR decline (p-trend=0.03) and was attenuated with additional adjustment (p-trend=0.07). In sensitivity analysis for rapid eGFR decline using a secondary definition (annual eGFR decrease≥3 and incident eGFR<60), diet associations remained significant with additional adjustment (p-trend=0.04). No associations were observed with physical activity, smoking status, or alcohol intake with incident eGFR<60 or rapid eGFR decline (all p>0.19).

Conclusions

Higher diet quality may be associated with a decreased risk of incident eGFR<60mL/min/1.73m², and rapid eGFR decline. Whether adherence to a healthy diet can prevent reduction in kidney function warrants further study.

Introduction

Chronic kidney disease (CKD) affects over 26 million adults in the United States and its incidence is increasing.[1] CKD is a risk factor for cardiovascular disease morbidity and mortality [2–4] and several established cardiovascular risk factors are also risk factors for CKD, including diabetes, hypertension, and obesity.[5] Obesity is of particular interest in relation to CKD. In addition to being an established risk factor for the development of cardiovascular and metabolic diseases,[6,7] obesity is a common condition in the United States [8] that can be modified through lifestyle interventions.[9] A growing body of evidence supports the role of both generalized and central obesity in the development of CKD [5,10–15] and microalbuminuria.[16–20] Several modifiable lifestyle factors, including physical activity, smoking, and diet factors, are associated with the development of obesity and obesity-related disease and may also be associated with the development of CKD, potentially through obesity-related pathways. Potential mechanisms relating lifestyle factors and ectopic fat accumulation are also of interest, as we have previously observed that high renal sinus fat is independently associated with hypertension and chronic kidney disease.[21]

Several studies have investigated the association of lifestyle factors and the development of kidney disease or renal function decline; the results for some lifestyle factors are inconsistent and few have investigated the impact of overall dietary patterns.[5,22–36] Thus, the primary aim of our study was to evaluate the association of potentially modifiable lifestyle factors with the development of incident estimated glomerular filtration rate (eGFR)<60mL/min/1.73m², and rapid renal function decline in the Framingham Offspring cohort. Further, we sought to determine if such associations persist after adjustment for body mass index as a measure of overall adiposity, and the presence of other cardiometabolic conditions associated with incident eGFR<60. Through this analysis, we aim to identify lifestyle factors that could be targeted as part of an approach to prevent kidney function decline.

Materials and Methods

Study Sample

The original Framingham Heart Study cohort was established in 1948 and consisted of 5209 women and men from Framingham, Massachusetts. In 1971, the Framingham Offspring Cohort was established and consisted of 5124 women and men who were children or the spouses of children from the original cohort. For the incident eGFR<60 sample, offspring participants who attended the seventh (1998–2001) and eighth (2005–08) examination cycles were eligible for the present analysis (n=2869). Participants with missing serum creatinine measurements at either exam (n=165), missing dietary data (n=722), missing data on baseline hypertension, diabetes, or proteinuria (n=18), or with eGFR<60 at baseline (n=162) were excluded, leaving a study sample of 1802 Offspring participants. There were no differences in development of eGFR<60 at follow-up or lifestyle characteristics (dietary intake, physical activity, smoking, or alcohol intake) at baseline. However, participants excluded due to missing dietary data were more likely to have hypertension and diabetes, suggesting that these participants were likely to be less healthy. For rapid eGFR decline, participants with baseline eGFR<60 were not excluded, for a sample size of 1964 participants. The study was approved by the institutional review boards of the Boston University Medical Center. All subjects provided written informed consent.

Exposure Assessment

Dietary intakes were assessed using the Harvard semi-quantitative food frequency questionnaire (FFQ)[37]. The FFQ consists of a list of foods with a standard serving size and a selection of 9 frequency categories ranging from never or <1 serving/month to ≥6 servings/day. Participants were asked to report their frequency of consumption of each food item during the past year. Participants could also add up to three additional foods that are important components of their diets but are not listed on the questionnaire. Information on nutrient supplement use was also obtained by the FFQ. Dietary information was judged as unreliable and excluded from further study if reported energy intakes were <600 kcal/d or >4000 kcal/d (women) or >4200 kcal/d (men) or if >12 food items were left blank. Out of the N=3539 participants in the FHS Offspring Study at baseline, N=414 (11.7%) had no FFQ data and N=95 (2.7%) had unreliable diet data.

Diet quality was assessed using the Dietary Guidelines Adherence Index (DGAI), a previously validated index that assesses adherence to key dietary recommendations in the 2005 Dietary Guidelines for Americans (DGA), as previously described.[38] In brief, the DGAI is reported on a 0–20 scale based on adherence to 20 energy-specific food and nutrient recommendations,[38] with higher DGAI scores representing higher diet quality. As diet information was collected during the seventh examination cycle for offspring participants (through the FFQ), prior to the release of these guidelines, the DGAI is a measure of overall consistency of baseline diet with these recommendations in the 2005 DGA. Physical activity was assessed using a physical activity index (PAI). The PAI that takes into account the average number of hours/day at varying levels of activity intensity based on self-report, with higher scores representing higher levels of usual physical activity.[39] Alcohol intake was assessed using a physician-administered questionnaire and categorized into three intake groups: none (0 drinks/week), low-to-moderate (1–7 drinks/week in women and 1–14 drinks/week in men), and high intake (>7 drinks/week in women and >14 drinks/week in men).[40] Current smoking was defined as smoking ≥1 cigarette/day in the past year.

Outcome Assessment

Serum creatinine was measured using the modified Jaffe method from fasting blood samples collected during the seventh and eighth exams (inter-assay coefficient of variation [CV] of 2.8% and intra-assay CV of 4.0%). Serum creatinine measures can vary widely across different laboratories. Therefore, we calibrated serum creatinine values through a two-step process: 1) calibration of National Health and Nutrition Examination Survey III (NHANES III) creatinine values to the Cleveland Clinic Laboratory resulting in a correction factor of 0.23 mg/dl, and 2) alignment of mean serum creatinine values from the Framingham Offspring Study by sex-specific age groups (20–39, 40–59, 60–69, and ≥70 years) with the corresponding corrected NHANES III age- and sex-specific means.[41] The CKD-EPI Equation was used to estimate glomerular filtration rate (eGFR).[42] Incident eGFR<60 was defined as the presence of an eGFR<60mL/min/1.73m² at the eighth examination (follow-up) among participants with eGFR≥60mL/min/1.73m² at the seventh examination (baseline). The average annual rate of change in eGFR during follow-up was determined by dividing the difference in eGFR by the number of years of follow-up between the seventh and eighth examinations. Rapid eGFR decline was defined as an annual decrease in eGFR of at least 3 mL/min/1.73m² per year. In sensitivity analyses, a secondary rapid eGFR decline outcome was defined by a combination of 1) an annual decrease in eGFR of at least 3 mL/min/1.73m² per year and 2) incident eGFR<60mL/min/1.73m² at follow-up.

Additional Covariate Assessment

Body mass index (BMI, kg/m²) was calculated using weight and height measurements collected during the baseline clinic exam. Diabetes was defined as baseline plasma fasting glucose ≥126 mg/dL or use of oral hypoglycemic agents or insulin. The presence of dipstick proteinuria at baseline was defined as a measurement of trace protein or higher in a spot urine sample.

Statistical Analyses

Incident eGFR<60 and rapid eGFR decline were each modeled as functions of individual lifestyle factors using logistic regression in primary and sensitivity analyses. Models were initially adjusted for age, sex, and baseline eGFR and included diet quality (DGAI, expressed as indicator variables for quartiles), physical activity (PAI, expressed as indicator variables for quartiles), current smoking status, and alcohol intake (indicator variables for low-to-moderate and high alcohol intake, with no intake as the reference group, Model 1). Models were further adjusted for baseline BMI, hypertension, diabetes, and dipstick proteinuria (Model 2). We tested for linear trends across diet quality and physical activity quartiles by performing the models described above with diet quality and physical activity expressed as 4-level ordinal variables. Statistical analyses were performed using SAS Version 9.2 (SAS Institute, http://www.sas.com/).

Results

Lifestyle Factors and Incident eGFR<60mL/min/1.73m²

Overall, 54.8% of our study sample participants were women with an average baseline age of 59 years; additional baseline characteristics are presented in Table 1. Participants were overweight on average (mean BMI of 28.0 kg/m²), with a low to moderate level of daily physical activity and a diet quality consistent with about half of the diet recommendations reflected in the DGAI. The prevalence of current smoking was 11.9% and the majority of participants consumed either low-to-moderate levels or no alcohol.

Table 1.

Baseline characteristics of the study sample of Offspring cohort participants who had eGFR≥60mL/min/1.73m² at baseline by development of incident eGFR<60mL/min/1.73m² at follow-up.

Variable	Participants Free of Baseline CKD (n=1802)	Participants with eGFR≥60 at Follow-up (n=1631)	Participants with eGFR<60 at Follow-up (n=171)
Age (Years)	59 ± 9	59 ± 8	66 ± 7
Sex (N, % female)	987 (54.8)	894 (54.8)	93 (54.4)
BMI (kg/m2)	28.0 ± 5.3	28.0 ± 5.4	28.0 ± 4.5
Dietary Guidelines Adherence Index*	9.3 ± 2.8 [9.0 (7.0–11.3)]	9.3 ± 2.8 [9.0 (7.3–11.3)]	9.3 ± 2.7 [9.0 (7.0–11.5)]
Quartile 1 (N=451)- lowest quality	5.8 ± 1.0 [6.0 (5.0–6.5)]	5.7 ± 1.0 [6.0 (5.0–6.5)]	6.1 ± 1.0 [6.5 (5.8–6.8)]
Quartile 2 (N=459)	8.2 ± 0.6 [8.3 (7.8–8.8)]	8.2 ± 0.6 [8.3 (7.8–8.8)]	8.2 ± 0.6 [8.1 (7.8–8.8)]
Quartile 3 (N=442)	10.3 ± 0.6 [10.3 (9.8–10.8)]	10.3 ± 0.6 [10.3 (9.8–10.8)]	10.2 ± 0.6 [10.0 (9.5–10.8)]
Quartile 4 (N=450)- highest quality	12.9 ± 1.2 [12.8 (12.0–13.5)]	12.9 ± 1.2 [12.5 (12.0–13.5)]	12.9 ± 1.1 [12.8 (12.0–13.5)]
Physical Activity Index*	37.1 ± 8.0 [36.5 (33.3–41.0)]	37.0 ± 8.0 [36.4 (33.2–40.7)]	37.9 ± 8.4 [37.6 (34.0–42.5)]
Quartile 1 (N=447)- lowest activity	28.5 ± 7.9 [31.1 (29.2–32.2)]	28.6 ± 7.8 [31.1 (29.3–32.2)]	27.1 ± 9.0 [30.3 (28.1–31.4)]
Quartile 2 (N=454)	34.9 ± 0.9 [35.0 (34.1–35.7)]	34.9 ± 0.9 [35.0 (34.1–35.8)]	34.8 ± 0.8 [34.9 (34.1–35.4)]
Quartile 3 (N=450)	38.6 ± 1.3 [38.4 (37.5–39.7)]	38.6 ± 1.3 [38.4 (37.5–39.7)]	38.7 ± 1.4 [38.8 (37.5–40.1)]
Quartile 4 (N=451)- highest activity	46.3 ± 5.4 [44.6 (42.4–48.5)]	46.3 ± 5.5 [44.5 (42.4–48.5)]	46.4 ± 4.6 [45.3 (42.8–49.0)]
Current Smoking (N, %)	215 (11.9)	200 (12.3)	15 (8.8)
Alcohol Intake (N, %) †
None	536 (29.7)	472 (28.9)	64 (37.4)
Low-to-moderate	945 (52.4)	864 (53.0)	81 (47.4)
High	321 (17.8)	295 (18.1)	26 (15.2)
Hypertension (N, %)	692 (38.4)	589 (36.1)	103 (60.2)
Diabetes (N, %)	138 (7.7)	111 (6.8)	27 (15.8)
Dipstick proteinuria (N, %)	227 (12.6)	193 (11.8)	34 (19.9)

Continuous characteristics are presented as mean ± standard deviation.

^*Median and interquartile range [median (IQR)] are also provided for Dietary Guidelines Adherence Index and Physical Activity Index

^†Alcohol Intake defined as: none = 0 drinks/week; low-to-moderate = 1–7 drinks/week in women and 1–14 drinks/week in men; high = >7 drinks/week in women and >14 drinks/week in men

During follow-up (mean of 6.6 years), 9.5% (n=171) participants developed incident eGFR<60. After accounting for age, sex, baseline eGFR, physical activity, alcohol intake, and current smoking, we observed a significant trend across diet quality quartiles suggesting that higher diet quality associated with a decreased odds of incident eGFR<60 (p-trend=0.045, Table 2). These results were similar after further adjustment for BMI, hypertension, diabetes, and dipstick proteinuria (Table 2). In contrast, we did not observe associations with physical activity, smoking status, or alcohol intake with incident eGFR<60 (All p>0.39, Table 2).

Table 2.

Odds ratios and 95% confidence intervals of incident eGFR<60mL/min/1.73m² and estimated using the CKD-EPI equation, by lifestyle exposures.

	Model 1†		Model 2*
	OR (95% CI)	p	OR (95% CI)	p
Dietary Guidelines Adherence Index
Quartile 1 (Q1) – lowest quality	1.0 (ref)	-	1.0 (ref)
Q2	0.80 (0.49, 1.30)	0.37	0.77 (0.47, 1.27)	0.30
Q3	0.55 (0.33, 0.92)	0.02	0.52 (0.31, 0.89)	0.02
Q4 – highest quality	0.64 (0.39, 1.07)	0.09	0.63 (0.38, 1.07)	0.09
p-trend		0.045		0.045
Physical Activity Index
Q1 – lowest activity	1.0 (ref)	-	1.0 (ref)
Q2	1.12 (0.67, 1.90)	0.66	1.16 (0.69, 1.97)	0.58
Q3	1.10 (0.66, 1.86)	0.71	1.07 (0.63, 1.81)	0.80
Q4 – highest activity	1.23 (0.74, 2.04)	0.42	1.19 (0.71, 1.99)	0.51
p-trend		0.46		0.60
Current Smoker
No	1.0 (ref)	-	1.0 (ref)
Yes	1.28 (0.68, 2.39)	0.44	1.25 (0.66, 2.35)	0.49
Alcohol Consumption
None	1.0 (ref)	-	1.0 (ref)
Low-to-moderate	0.84 (0.57, 1.24)	0.39	0.94 (0.63, 1.40)	0.75
High	0.83 (0.49, 1.40)	0.48	0.84 (0.49, 1.43)	0.51

^†Adjusted for other listed lifestyle factors, age, sex, and baseline eGFR

^*Adjusted for other listed lifestyle factors, age, sex, baseline eGFR, BMI, hypertension, diabetes, and dipstick proteinuria

Lifestyle Factors and Rapid eGFR Decline

During follow-up, 12.1% (n=238 of 1964) of participants experienced rapid eGFR decline (primary definition of eGFR decline≥3mL/min/1.73m² per year). Similar to incident eGFR<60, a significant trend was observed across diet quality quartiles (p-trend=0.03), and the highest diet quality quartile was associated with a decreased odds of rapid eGFR decline after accounting for age, sex, baseline eGFR and the other lifestyle factors (Table 3, Quartile 4 vs. Quartile 1: OR 0.65, 95% CI 0.43 to 0.98, p=0.04). The associations of diet quality with rapid eGFR decline were attenuated with additional adjustment for BMI, hypertension, diabetes, and dipstick proteinuria (p-trend=0.07). However, in sensitivity analyses when the analysis was repeated with a stricter definition of rapid eGFR decline (eGFR decline≥3mL/min/1.73m² per year and eGFR<60mL/min/1.73m² at follow-up; n=80 cases/1722 controls), the inverse association between increasing diet quality and decreasing odds of rapid eGFR decline remained significant and was robust to adjustment (p-trend=0.02 and p-trend=0.04; Supplemental Table 1). Physical activity, current smoking, and alcohol consumption were not significantly associated with rapid eGFR decline in primary (Table 3) or sensitivity analyses (Supplemental Table 1).

Table 3.

Odds ratios and 95% confidence intervals of rapid estimated glomerular filtration rate (eGFR) decline≥3mL/min/1.73m² per year (n=1964), defined using eGFR estimated using the CKD-EPI equation, by lifestyle exposures.

	Model 1†		Model 2*
	OR (95% CI)	p	OR (95% CI)	p
Dietary Guidelines Adherence Index
Quartile 1 (Q1) – lowest quality	1.0 (ref)		1.0 (ref)
Q2	0.82 (0.56, 1.20)	0.31	0.83 (0.56, 1.22)	0.34
Q3	0.70 (0.47, 1.04)	0.08	0.73 (0.49, 1.10)	0.13
Q4 – highest quality	0.65 (0.43, 0.98)	0.04	0.69 (0.45, 1.05)	0.08
p-trend		0.03		0.07
Physical Activity Index
Q1 – lowest activity	1.0 (ref)		1.0 (ref)
Q2	1.26 (0.85, 1.85)	0.25	1.30 (0.88, 1.93)	0.19
Q3	1.23 (0.84, 1.81)	0.29	1.23 (0.83, 1.82)	0.31
Q4 – highest activity	0.88 (0.58, 1.32)	0.53	0.93 (0.61, 1.42)	0.73
p-trend		0.57		0.72
Current Smoker
No	1.0 (ref)		1.0 (ref)
Yes	1.21 (0.80, 1.82)	0.37	1.19 (0.78, 1.81)	0.42
Alcohol Consumption
None	1.0 (ref)		1.0 (ref)
Low-to-moderate	0.87 (0.64, 1.20)	0.39	0.99 (0.71, 1.37)	0.94
High	1.01 (0.68, 1.50)	0.98	1.12 (0.74, 1.68)	0.60

^†Adjusted for other listed lifestyle factors, age, sex, and baseline eGFR

^*Adjusted for other listed lifestyle factors, age, sex, baseline eGFR, BMI, hypertension, diabetes, and dipstick proteinuria

Discussion

Among Framingham Offspring participants, our findings suggest that higher diet quality is associated with a decreased risk of developing incident eGFR<60mL/min/1.73m² and rapid eGFR decline after a mean of 6.6 years of follow-up. These findings were not materially different after accounting for BMI, diabetes, hypertension, and proteinuria, suggesting that the decreased risk is not fully attributable to the potential association of higher diet quality with obesity and cardio-metabolic diseases. In contrast, other lifestyle characteristics, including physical activity, smoking, and alcohol intake, were not associated with developing incident eGFR<60 or rapid eGFR decline over the same time period.

While recent studies have reported on the role of specific nutrient intake and the development of incident eGFR<60, [24–26,36] limited data are available addressing the association of overall diet quality or dietary patterns and renal function.[22,23] Adherence to a Mediterranean dietary pattern and creatinine clearance were assessed in a cross-sectional sample of individuals aged 18–89 in Greece, with higher adherence associated with a higher creatinine clearance.[23] In a sample of ~3000 women from the Nurses’ Health Study, an overall diet consistent with a Western diet pattern was associated with an increased risk of rapid eGFR decline whereas a diet consistent with a Dietary Approach to Hypertension (DASH) diet pattern was associated with a decreased risk of rapid eGFR decline.[22] Our findings extend the current literature to suggest that beneficial dietary associations observed for other diseases include diabetes and cardiovascular disease [43,44] may also extend to incident eGFR<60 and rapid eGFR decline.

Our results indicate that physical activity levels are not associated with incident eGFR<60 in our study sample selected from Framingham Offspring participants. In the Australian Diabetes, Obesity and Lifestyle (AusDiab) Study, when compared to participants reporting sufficient leisure-time physical activity (≥150 minutes/week), participants who were inactive or insufficiently active (>0 but <150 minutes/week) were not at an increased risk of developing de novo low eGFR.[27] In relation to rapid eGFR decline, no consistent trend was observed with higher levels of physical activity. In contrast, among older adults in the Cardiovascular Health Study, a higher physical activity score (based on leisure-time activity and walk pace) was associated with a decreased risk of rapid eGFR decline over 7 years follow-up.[28] Similarly, physical activity during non-working hours was associated with an increase in eGFR when compared to physical inactivity after 7 years of follow-up among women in the Tromsø study.[29] Our disparate findings may be attributable to differences in exposure assessment (leisure-time physical activity vs. overall physical activity) or study population demographics; for example, the CHS study sample participants were on average older than the participants in the present analysis.

The association of alcohol intake and changes in renal function has been investigated previously in prospective analyses, with conflicting results. Moderate alcohol consumption was [22]not associated with declining renal function over 11 years follow-up in the Nurses’ Health Study [30] and alcohol consumption was not associated with rapid eGFR decline in the CHS,[31] whereas results from the Beaver Dam Study suggest that heavy alcohol intake (≥4 drinks/day) is associated with an elevated risk of developing CKD over 5 years.[32] Protective associations have also been reported in other study populations. Among men in the AusDiab Study, a protective association for de novo low eGFR was observed for moderate to heavy alcohol consumption when compared to light consumption; among current drinkers, a protective association was also observed for consuming ≥30 grams/day when compared to <10 grams/day.[33] High alcohol consumption among men in the Tromsø study was also associated with an increase in eGFR over 7 years of follow-up when compared to alcohol abstention.[29] A protective association for moderate alcohol consumption and CKD has also been observed over 10-years follow-up in a community-based sample from Japan [34] (< 20 grams/day) as well over 14 years follow-up in the Physicians’ Health Study, with consumption of ≥7 drinks/week associated with a decreased risk of reduced eGFR.[35] We did not observe a significant association between alcohol intake, incident eGFR<60, or rapid eGFR decline, although our results were suggestive of a protective effect of low-to-moderate or high alcohol intake when compared to non-drinkers.

Our current smoking results are in contrast to our previous report in the Framingham Offspring cohort, in which we observed that current smoking is a risk factor for incident eGFR<60after 18.5 years of follow-up,[5] whereas in the present analysis we observed that current smoking at baseline was not associated with incident eGFR<60 or rapid eGFR decline. Our disparate results may be due to differences in baseline study samples and length of follow-up. Specifically, our prior study consisted of individuals with a mean age of 42 years and a smoking prevalence of 33%,[5] whereas the present study consists of surviving participants nearly two decades later (and therefore two decades older) who by then had a lower smoking prevalence of 11.9%. Additionally, the follow-up period was shorter in the present analysis (6.6 years vs. 18.5 years), and this may not be long enough to capture the effect of smoking on renal function decline.

We assessed diet quality using the DGAI: a summary score that takes into account eleven food-specific intake recommendations as well as nine “healthy choice” recommendations in the 2005 Dietary Guidelines for Americans.[38] Higher diet quality assessed by DGAI indicates dietary intake consistent with recommended intakes of fruits and vegetables, meat and legumes, dairy, whole grains and fiber, sodium, alcohol, and discretionary energy intake based on dietary patterns defined by estimated energy requirements. Additional penalties for overconsumption of energy dense foods (high in fats and sugars) are also included in the score. Higher scores reflect better adherence to recommended levels of intake and greater variety of intake within each food group. DGA recommendations for foods higher in proteins may result in protein intakes that exceed levels indicated by the KDOQI nutrition guidelines for individuals with mild kidney disease; however, sources of protein encouraged are similar to the DASH diet [22] – vegetable, low-fat dairy, whole grains, nuts, legumes, fish and poultry – and have not typically been associated with decline in kidney function. Higher DGAI scores have been previously been positively associated with intake of healthy food items such as fruits and vegetables and negatively associated with intake of fats and other energy-dense foods.[38] Additionally, higher DGAI scores are associated with lower levels of central adiposity,[45] and decreased insulin resistance.[46] Our findings suggest that diet quality may be a modifiable kidney function risk factor among older adults that could potentially be targeted as an approach to reduce the incidence of eGFR<60 or rapid eGFR decline. Of note, trends persisted after accounting for baseline BMI, diabetes, hypertension, and proteinuria, suggesting that potential mechanistic pathways underlying our observed association may be independent of the impact of diet quality on obesity and these cardio-metabolic conditions. We were unable to compare the associations of diet quality and kidney function between this study and our previous study examining lifestyle factors in FHS [5] because dietary data was unavailable prior to the seventh examination. Future prospective research in a randomized setting is necessary to determine whether interventions promoting adherence to US dietary guidelines are feasible and effective in improving renal outcomes.

Strengths of the present analysis include its prospective design in a community-based sample with well-measured clinical covariates. There are several potential limitations that warrant mention. While we did observe wide range of DGAI scores observed, few participants had high DGAI scores thus suggesting the association of higher dietary quality and lower risk of CKD may have been even stronger among those closely adhering to the US dietary recommendations. Incident eGFR<60 and rapid eGFR decline were determined based on single serum creatinine measurements at baseline and end of follow-up, leading to potential outcome misclassification. However, any non-differential misclassification would tend to attenuate the true association. In addition, the inverse trend in associations observed between dietary quality and rapid eGFR decline were robust to a secondary (and more strict) definition of rapid eGFR decline in sensitivity analysis. Furthermore, we used a spot urine dipstick test to classify proteinuria status in place of measuring urinary albumin-to-creatinine ratio. In addition, we observed significant differences for hypertension and diabetes status between participants included and excluded from the analysis based on missing dietary intake data. This suggests that the participants excluded based dietary data were likely to be less healthy and excluding these individuals from the analysis likely attenuated the true association between lifestyle factors and incident eGFR<60. Finally, our study sample is comprised predominately of older participants of European ancestry, and the generalizability of our results to younger populations or other race or ethnic groups may be limited.

In conclusion, our findings suggest that higher diet quality is a modifiable lifestyle factor that may be associated with a decreased risk of developing incident eGFR<60mL/min/1.73m² and rapid eGFR decline. Whether improving diet quality based on the US Dietary Guidelines for Americans (or its individual food components) may lead to improved renal outcomes among older adults in the US population of European and non-European ancestral background is still in question. Our study highlights the importance of detailed analyses of lifestyle factors and measures of kidney function, and also suggests investigation of individual foods and food groups with kidney function in for future study.