In this cross-sectional study, we assessed the association between renal function and plasma NfL and P-tau181 levels, which yielded three main findings. First, eGFR was negatively associated with plasma NfL and P-tau181. After adjusting for demographic characteristics and comorbid diseases, eGFR remained significantly associated with plasma NfL, but not with P-tau181. Second, a significant interaction between age and eGFR was found on the impact of plasma NfL. Third, in participants ≥ 70 years and with eGFR < 60 mL/min/1.73 m2, the correlation between eGFR and plasma NfL was particularly noteworthy.
Blood-based biomarkers are promising in the diagnosis and prognosis of AD [27]. However, biological and technical factors that might negatively affect diagnostic accuracy must be characterized before widespread application[28, 29]. Recent studies showed that renal function might potentially impact blood biomarkers. Elevation of plasma Aβ40, Aβ42, total-tau, NfL, and P-tau181 was demonstrated in patients with CKD [15, 30, 31]. Liu et al. reported that serum Aβ levels were higher in CKD patients, and dialysis CKD patients had lower Aβ levels than non-dialysis ones [30]. Syrjanen et al. found that CKD was associated with higher plasma Aβ40, Aβ42, total-tau, and NfL levels[11]. An investigation by Adam et al. revealed that less than 10% of CKD patients were aware of their disease[32] because CKD remained asymptomatic until the late stage[33]. In previous studies, individuals with renal dysfunction might be underestimated because most diagnoses were based on self-report or medical history. Therefore, we calculated eGFR to reduce reporting bias and enable reliable analysis in a scalable manner.
NfL is released into the cerebrospinal fluid (CSF) and blood upon axon injury [8]. Although being a sensitive indicator for axon damage and neuron death [34], plasma NfL was reported to increase under various conditions other than neurodegenerative diseases. Blood NfL concentrations were reported to be correlated with various factors including BMI, age, diabetes, hypertension, dyslipidemia, and renal function [10, 12, 16, 35, 36]. Consistent with these findings, our study found that eGFR was negatively correlated with plasma NfL concentration. There might be some underlying reasons. Since the kidney is responsible for the clearance of most proteins[37], a low eGFR may reduce the clearance of proteins in the blood, and increase the circulating levels for biomarkers, including NfL. This was proposed as the main explanation in prior studies [17, 38]. Meanwhile, elevated NfL levels were found to be associated with cardiovascular diseases, which were often comorbid with CKD [39]. But the association in our study remained significant after adjustment for cardiovascular diseases, i.e., hypertension and diabetes, which suggested other possibilities. Notably, neurological complications were prevalent in patients with renal dysfunction[40]. Accumulation of toxic compounds, including various uraemic compounds, homocysteine, etc, could induce glial dysfunction and neuronal apoptosis, and eventually cause NfL elevation in the blood [41, 42].
Plasma P-tau181 had been validated as a superb marker for AD [7, 28, 43]. However, only a few studies had explored the impact of comorbidities on P-tau181. Mielke et al. reported that plasma P-tau level was linked to CKD and other comorbidities[15]. Berry et al. observed patients with cirrhosis and found serum creatinine was negatively associated with P-tau181 [16]. Stocker et al. found that kidney function was negatively associated with P-tau181 [36]. In the current study, although P-tau181 was elevated in the low eGFR group, the difference was not statistically significant after adjusting for confounding factors. The inconsistent result might be related to age, gender, renal function composition, and discrepant calculation for eGFR. Prior studies had demonstrated that plasma P-tau181 specifically reflected AD pathology, and was less related to other pathological conditions, including comorbid diseases [27]. Another possibility was that P-tau181 might be cleared through means other than the kidney, which, however, remained to be elucidated. Albeit preliminary, these findings suggested that P-tau181 was a relatively robust measure that was resilient to biological confounding factors.
Age was one of the most profound risk factors and the main driving force for neurodegeneration. In our study, among different demographic and comorbidity characteristics, age was the only modifier on the impact of eGFR on plasma NfL. The eGFR decreased along with aging. Hypertension and diabetes which mainly caused kidney dysfunction were also increasingly prevalent in older adults. Moreover, under chronic stress during aging, cellular senescence was activated with the secretion of various pro-inflammatory molecules, leading to the increment in peripheral markers. Collectively, all the aforementioned mechanisms might contribute to the alteration of plasma NfL and confuse the picture for precise diagnosis. In this study, the correlation of renal function with NfL concentration was significantly remarkable, especially in individuals ≥ 70 years and with eGFR < 60 mL/min/1.73 m2. Special attention should be paid to this group when setting up reference ranges of blood-based NfL.
Several advantages of our research should be emphasized. Comprehensive laboratory measures of serum creatinine and cystatin C were assayed and used for the CKD-EPI equation to calculate eGFR, which was superior to other estimations [44, 45]. We used eGFR, rather than the medical history of CKD or other self-reported kidney dysfunction to objectively analyze the relationship between renal function and plasma biomarkers. Additionally, the current study included a large sample of dementia-free individuals from a population-based cohort, which well represented a general older population. Furthermore, establishing the reference ranges of blood biomarkers for the population with diverse ethnicities is warranted for clinical practice. Our findings complemented prior work with valuable data from the Chinese population.
Our findings shall be cautiously interpreted with the following limitations. Firstly, given the cross-sectional study design, we cannot infer a causal relationship between decreased renal function and elevated plasma NfL levels. Secondly, individuals with advanced renal failure were not able to participate in the current study. Most of the participants (83%) had normal or fair renal function (eGFR > 60mL/min/1.73m2), which might underestimate the impact of renal function on plasma biomarkers. However, the eGFR distribution of this study was in accordance with the worldwide CKD prevalence in the general population[46]. Thirdly, only essential demographic and comorbid factors were included in our analysis. There were probably other potential confounders related to renal function and blood biomarkers that might impact their associations. Finally, the study participants were from an urban community in a developed metropolis. Whether our findings could represent the population living in other areas remained to be elucidated.