Jackson Heart Study Post Hoc Analysis: Protein Intake and Long-Term Decline in Kidney Function

The estimated prevalence of chronic kidney disease (CKD) in the United States is 14% to 16% of the adult population. Patients with CKD commonly progress to end-stage renal disease (ESRD), requiring dialysis therapy or transplantation; CKD is associated with substantial morbidity and mortality. Diabetes and hypertension are major contributors to kidney disease; other modifiable risk factors, including diet, may be associated with the development of incident CKD and decline in kidney function.

Rakesh Malhotra, MD, MPH, and colleagues recently conducted an observational cohort study designed to test the hypothesis that there would be an association between higher protein intake and decline in kidney function; the researchers also hypothesized that patients with diabetes would experience greater declines in estimated glomerular filtration rate (eGFR) over a follow-up of 8 years. The researchers utilized data from participants in the Jackson Heart Study to evaluate decline in eGFR. Study results were reported in the Journal of Renal Nutrition [2018;28(4):245-250].

The Jackson Heart Study enrolled 5301 noninstitutionalized African American participants from three counties in Jackson, Mississippi between 2000 and 2004. For the current analysis, participants without dietary data were excluded (n=164), as were those without serum creatinine measurements at baseline or at follow-up (n=1521). An additional 230 were excluded due to implausible energy intakes, 34 who had no data on alcohol intake, smoking, body mass index (BMI), or diabetes status, and 154 with eGFR <60 mL/min/1.73 m2 at baseline. The final analysis included 3165 participants.

Of the final cohort, 65% were women, 54% were nonalcohol drinkers, 90% were nonsmokers, 57% had a history of hypertension, and 81% had no history of diabetes. Median (25th, 75th percentile) percent energy intake from protein was 14.3 (12.4, 16.4) for total protein, 9.6 (7.6, 11.8) for animal protein, and 4.4 (3.7, 5.2) for plant protein. There was a modest correlation between animal protein and plant protein (r=–0.27; P<.001).

Among participants with diabetes, total protein intake at baseline was significantly higher (P<.0001) compared with participants without diabetes (15.5% vs 14.0% of energy). Among participants with hypertension, total protein intake was also higher compared with those without hypertension (14.4% vs 14.2% of energy; P=.01); this difference was driven by differences in plant protein intake (4.5% vs 4.3%; animal intake was the same in each group, 9.6%).

There were no significant differences in age across categories of protein intake. Differences in percent decline in eGFR were nonsignificant across quintiles of percent energy from protein in unadjusted analysis.

With regard to decline in eGFR, there was significant interaction between protein intake and diabetes (protein ✕ diabetes interaction=0.03). Across all quintiles of protein intake, following adjustment for age, sex, smoking status, alcohol use, BMI, systolic blood pressure, diastolic blood pressure, diabetes status, years between creatinine measurements, total energy intake, and percent of energy from saturated fat, polyunsaturated fat, trans fat, and carbohydrate, the percentage decline in eGFR over an 8-year period was higher in participants with diabetes compared with those without diabetes. There was also a U-shaped dose-response relationship, with higher degree of decline in eGFR among participants with diabetes for the lowest two (–15.9 and –16.7 mL/min/1.73m2) and highest quintiles (–19.1 and –20.0 mL/min/1.73 m2) of protein intake, compared with those with middle quintile (–12.0 mL/min/1.73 m2) category.

In a separate analysis stratified by glycosylated hemoglobin (HbA1c), there was an association between increasing protein intake to greater than ~15% of energy and greater decline in eGFR, particularly for those with HbA1c >8.0%. There was no association between decline in eGFR in those with HbA1c <6.5%.

There were some limitations to the study, according to the researchers, including results of the food-frequency questionnaire (FFQ) being self-reported, thus prone to misreporting and measurement error; assessing FFQ only once, with no updating of dietary intake data during the study period; and the inability to exclude residual confounding secondary to measurement error and unknown or unmeasured confounders that may influence the relationship between protein intake and kidney decline.

“In conclusion, our results show that among African Americans with diabetes, higher protein intake, expressed as percent of daily total energy intake, was positively associated with greater decline in eGFR in analyses that accounted for risk factors for kidney disease. Further studies, including clinical trials in patients with diabetes, are warranted to corroborate these findings and to elucidate the underlying mechanism,” the researchers said.

Takeaway Points

  1. Researchers conducted an analysis of data from the Jackson Heart Study to determine whether there was an association between higher protein intake and decline in kidney function; the analysis also examined whether the decline varied by diabetes status.
  2. Total protein intake was significantly higher among participants with diabetes than among those without diabetes at baseline.
  3. There was a significant interaction between protein intake and diabetes with regard to decline in estimated glomerular filtration rate over an 8-year period. There was also a U-shaped dose-response relationship, with higher degree of decline in eGFR among those with diabetes for the two lowest and highest quintiles of protein intake compared with those in the middle quintile.