Abstract
Introduction: There is increasing public health concern about the consequences that modern diet, which is rich in processed and red meat, may have on the risk of chronic non-communicable diseases. Chronic kidney disease (CKD) is among the most widespread common chronic diseases in industrialized countries, where it represents a major public health issue. CKD affects nearly 10% of the world’s population, with globally increasing prevalence and limited treatment options [1]. It represents a high burden for the health system and is associated with high disease-attributable mortality and multi-systemic complications [2]. Elevated protein intake may damage the kidney, leading to increased glomerular pressure and glomerular hyperfiltration: the consequent damage to the glomerular structure may lead to CKD or accelerate its progression [3]. However, cross-sectional and longitudinal analyses of association between protein intake and kidney function provide different and often contrasting perspectives [4]. While data from the general population are scarce, they may help elucidate the protein intake - kidney function relationship in a real-world scenario.
Objectives: Aim of our study was to investigate the relationship between total daily protein intake (TDPI) and kidney function and the CKD in a general population sample.
Methods: The analysis was carried out in the Cooperative Health Research In South Tyrol (CHRIS) study, an ongoing population-based study being conducted in Val Venosta (South Tyrol, Italy) [5]. Participants completed a Food Frequency Questionnaire (FFQ) developed by the Global Allergy and Asthma European Network (GA2LEN) study [6], which included local foods of the South Tyrolean region. The frequency of food consumption was transformed into grams per day for each food item and converted into nutrient estimates such as the TDPI, using the UK’s Food Composition Table [7]. Kidney function was quantified by the glomerular filtration rate estimated from serum creatinine (eGFR) using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [8]. CKD was defined as eGFR<60 ml/min/1.73m2. The relationship between eGFR and TDPI was assessed by fitting linear mixed regression models adjusting for age, sex, and other potential confounders, while logistic regression was used to asses the association between CKD and TDPI.
Results: Complete data for analyses was available for 5883 out of 10,518 participants (CHRIS study data release no. 3.0, June 7th, 2018), of whom 55% were females. The median age was 46 years old (interquartile range, IQR = 33, 58). The mean eGFR was 93.27 ml/min/1.73m2 (standard deviation, SD = 15.89). One-hundred thirty-three participants (2.26%, 95% confidence interval, CI = 1.90, 2.67) were affected by CKD. The median calculated TDPI was 83.17 g (IQR = 66.58, 103.52). The effect of TDPI on eGFR levels was of –0.003 ml/min/1.73 m2 per one unit increase of TDPI (standard error, SE = 0.005; p-value = 0.48). Results did not change when additionally accounting for the body mass index (BMI; effect = -0.003, SE = 0.005, p-value = 0.53). The odds ratio (OR) of CKD per one unit increase of TDPI was 1.00 (95% CI = 0.99, 1.01; p-value = 0.77) and remained unchanged when accounting for BMI (OR = 1.00, 95% CI = 0.99, 1.01; p-value = 0.79).
Conclusions: Our results did not support an association between TDPI and kidney function or CKD. This might be explained by CKD being typically silent in its earliest stages, for which reason, people might not modify their diet early enough to prevent progression of the kidney damage. Another aspect is the temporal allocation of our measurements: while eGFR chronically declines over time, along with people ageing, we could only assess dietary habits relatively to 12 months prior to the participation, which may prevent from observing a long-term effect of TDPI on kidney function. The main strength of our study is the assessment of kidney function and dietary habits in a large sample from the general population. Among the limitations, total protein intake was estimated based on an FFQ and not objectively quantified. In addition, eGFR was approximated by serum creatinine and CKD was estimated based on a single eGFR measurement, whilst at least two measures at a distance of at least 3 months should be considered for a proper diagnosis. In conclusion, longitudinal studies that can allocate dietary habits earlier in time as compared to the CKD onset are warranted. Additionally, for chronic conditions that remain silent for a long time, it would be important to include the information about clinical diagnosis or individuals’ awareness in the models, for a better interpretation of possible cause-effect relationship.