Low Protein Intake Associated with Higher Mortality Risk

Glomerular filtration rate (GFR) is a key index of kidney function. Protein intake is a major dietary modulator of GFR and standard practice calls for low protein intake to slow the progression of chronic kidney disease (CKD). Results of studies on low protein intake in kidney disease have been inconsistent regarding decline in kidney function and/or mortality over time.

Massimo Cirillo, MD, and colleagues recently conducted an observational, longitudinal cohort study to examine the relationships of low protein intake with mortality and kidney function decline over time in the population sample of the Gubbio Study, a population-based cohort study in the city of Gubbio, Italy. Results of the analysis were reported in the Journal of Renal Nutrition [2018;28(4):235-244].

The current analysis utilized data from the 1988-1992 examination (baseline), data from the 2001-2007 examination (follow-up), and data on mortality from baseline to December 31, 2007. The main outcome measures were mortality and estimated GFR (eGFR) decline, defined as eGFR change from baseline to follow-up ≤ mean –1 standard deviation (Z-score ≤ –1).

The target cohort included 4679 adults in the Gubbio cohort, with adult age at baseline (range, 18-97 years). Of those, 4307 had complete baseline data and were included in the final study cohort. Because urine excretion of urea nitrogen (UUN) differed by 16.5% between sexes, low UUN was separately defined in men and women.

Compared with individuals with non-low UUN, those with low UUN were older, had lower plasma urea nitrogen, lower urine sodium, lower urine potassium, lower urine sodium/potassium ratio, higher alcohol intake, lower physical activity, lower eGFR, higher prevalence of reduced eGFR, lower body mass index, lower obesity prevalence, and higher underweight prevalence. Compared with those with non-low UUN, those with low UUN also had significantly different levels of plasma urea nitrogen, urine sodium/potassium ratio, alcohol intake, obesity prevalence, and underweight prevalence 6 years prior to baseline and at follow-up.

During the observation period of 15.9 years, there were 871 deaths (annual mortality rate=1.26%); of those, 417 were due to cardiovascular disease, 276 from neoplastic disease, and 178 from other diseases (annual rate=0.61%, 0.40%, and 0.26%, respectively).

In unadjusted analyses, there was an association between low UUN and higher mortality in men (low UUN and non-low UUN: number of deaths=127 and 340; annual rate=2.22% and 1.38%; hazard ratio [HR], 1.62; 95% confidence interval [CI], 1.32-1.99; P<.001) and women (number of deaths 118 and 286; annual rate=1.61% and 0.92%; HR, 1.75; 95% CI, 1.41-2.17; P<.001). Following adjustment for sex and other covariates, the association was consistent in the entire cohort and across all subgroups. In the small subgroup of those with baseline eGFR <60 mL/min/1.73 m2 (n=249), findings were consistent with those of the overall cohort, but not significant (number of deaths, 67; HR, 1.18; 95% CI, 0.85-1.63; P=.327).

When analyzed separately from baseline to year 5 of follow-up, from year 5 to year 10 of follow-up, and from year 10 of follow-up on, the association of low UUN with higher mortality in the overall cohort was significant or borderline significant.

In unadjusted analyses, there was an association between low UUN and higher mortality from cardiovascular disease (annual rate, 0.95% in men and 0.53% in women; HR, 1.81; 95% CI, 1.47-2.23; P<.001) and from neoplastic disease (annual rate, 0.62% and 0.35%; HR, 1.87; 95% CI, 1.37-2.30; P<.001). There was no association between low UUN and mortality from other disease. In adjusted analyses, competing-risks regression indicated an independent association of low UUN with mortality from neoplastic disease but not with mortality from cardiovascular disease or from other disease.

The follow-up examination was completed by 2845 baseline examinees (43.9% men). Of those participants, decline in eGFR was seen in 454 individuals (16.0% with Z-score < of –1, i.e., with eGFR change equal to or more negative than –20 mL/min/1.73 m2). In unadjusted analyses, there was an association between low UUN and lower incidence of eGFR decline in men (low UUN and non-low UUN; n=219 and 1030; incidence=10.0% and 15.7%; OR, 0.60; 95% CI, 0.37-0.96; P=.033) and women (n-292 and 1304; incidence=9.9% and 18.5%; OR, 0.49; 95% CI, 0.32-0.73; P<.001). In analysis of men and women combined with adjustment for sex and other covariates, the association of low UUN and eGFR decline was not significant.

When analyzed by UUN quintile, the trend between UUN and decline in eGFR was significantly positive due to high prevalence of eGFR decline in UUN quintile 5. Findings were similar in analyses by quintile of estimated protein intake per kg of ideal weight.

Limitations cited by the researchers included the lack of nutritional data at follow-up and the use of UUN as a marker of low protein that implied a slight underestimate of protein intake.

“Briefly, this observational, longitudinal, population-based cohort study reported that, in the general population, a urinary marker of low protein intake predicted higher mortality, in particular from neoplastic disease. In the subgroup with reduced kidney function, low protein intake predicted higher mortality likewise in the general population but also a lower incidence of kidney function decline over time. Study results suggest the need of further investigation on the safety of long-term restriction of dietary protein below currently recommended allowance of 0.8 g/day/kg of ideal weight,” the researchers said.

Takeaway Points

  1. There are few data on the relationship of low protein intake with mortality and kidney function decline over time. Researchers conducted an observational, longitudinal cohort study using data from the Gubbio Study conducted in Gubbio, Italy.
  2. In the overall cohort, low protein intake, as measured by overnight urine urea nitrogen (UUN) excretion, predicted higher mortality in the general study population.
  3. In a subgroup with reduced kidney function, low protein intake predicted higher mortality but also a lower incidence of decline in kidney function.