Background
Chronic kidney disease (CKD) is a growing health problem with an estimated prevalence of 10.8–16% in major developed and developing countries [
1,
2]. Individuals with CKD are at risk of progressive kidney failure, cardiovascular events, and death [
3,
4].
Nonmicrobial inflammation contributes to CKD progression and fibrosis [
5]. The neutrophil count reflects inflammation, while the lymphocyte count indicates the status of general stress and nutrition. The neutrophil-to-lymphocyte ratio (NLR) in CKD patients provides information on the inflammation status [
6]. It is suggested that NLR is a complementary prognostic marker for evaluating the cardiovascular risk in CKD3-5 patients [
6]. Studies demonstrated that an increase in neutrophil count coupled with a reduction in lymphocyte counts predicts mortality in hemodialysis patients [
7] and peritoneal dialysis patient [
8]. Also, NLR indicates the rate of stage 4 chronic kidney disease progressing to dialysis [
9]. Nevertheless, studies of NLR for its prognosis potential towards ESRD, CVD and all-cause mortality in patients with stage 1–4 CKD other than stage 5 CKD are rare. Since NLR is readily derived from complete blood count tests, its potential as a predictor should be investigated in a large cohort of patients with stage 1–4 CKD.
The objective of this study was to evaluate whether NLR, a simple marker of chronic systemic inflammation, predicts the progression of ESRD, CVD and all-cause mortality among CKD1–4 patients in the Chinese Cohort Study of Chronic Kidney Disease (C-STRIDE), the first national prospective CKD cohort of the Chinese population.
Materials and methods
The design, methods, and baseline characteristics of the Chinese Cohort Study of Chronic Kidney Disease (C-STRIDE) study population has been published; this is a prospective investigation involving 39 clinical centers in 28 cities of 22 provinces in China [
10]. The design of C-STRIDE has been described in detail elsewhere [
11]. A total of 3358 participants were included until December 31, 2017 after exclusion of patients with missing value of key demographic variables (including blood routine and blood lipid) or loss of follow-up. Of these 3358 participants, 938 patients have NLR and other relevant clinical variables and were used in the current analysis. The C-STRIDE study was approved by the ethics committee of Peking University First Hospital and was in adherence with the Declaration of Helsinki. All participants have consented this study. During the study visit, all C-STRIDE study data were collected by trained staff according to the study protocol [
12]. Data were obtained by questionnaires, anthropometric measurements, collection of blood and urine specimens. The albumin/creatinine ratio (ACR) stage was categorized according to an analysis of a spot urine sample: A1 (normoalbuminuria), ACR < 30 mg/g creatinine; A2 (microalbuminuria), 30 ≤ ACR < 300 mg/g creatinine; or A3 (macroalbuminuria), ACR ≥ 300 mg/g creatinine. GFR was estimated from serum creatinine measurements and demographic characteristics by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [
1,
13]. Smoking was defined as currently smoking or had ever smoked. Alcohol consumption was defined as a habitual drinker (drink once a day or more). Hypertension was defined as a systolic blood pressure (BP) ≥ 140 mmHg and/or a diastolic BP ≥ 90 mmHg, or a self-reported history of hypertension. Levels of hemoglobin, NLR, fasting blood glucose, serum triglycerides, total cholesterol, LDL-cholesterol, HDL cholesterol, uric acid, serum phosphate, serum calcium, hs-CRP, iPTH, HbA1c, were documented. Plain lateral abdominal x-ray film showed abdominal aortic calcification. CKD-MBD was defined as a triad of interrelated abnormalities of serum biochemistry (serum phosphate levels > 1.49 mmol/L or < 0.87 mmol/L; serum calcium levels > 2.57 mmol/L or < 2.1 mmol/L, and iPTH > 70 pg/mL or < 35 pg/mL); or calcification of lateral abdominal x-ray film. Patients were considered to have diabetes mellitus if they had a fasting glucose ≥ 7.0 mmol/L; an HbA1c ≥ 6.5%; took insulin or other anti-diabetic medications; or reported a history of diabetes. CVD history was defined as the past occurrence of a myocardial infarction, admittance into a hospital for congestive heart failure, or severe cardiac arrhythmia incidents (resuscitated cardiac arrest, ventricular fibrillation, sustained ventricular tachycardia, paroxysmal ventricular tachycardia, atrial fibrillation or flutter, severe bradycardia, or heart block). Hyperlipidemia was defined as total cholesterol level ≥ 5.7 mmol/L or an LDL-cholesterol level ≥ 3.6 mmol/L. Hyperuricemia was defined as a serum concentration of uric acid ≥ 420 μmol/L for men and ≥ 360 μmol/L for women. The abnormal of hsCRP was ≥ 3 mg/dL [
14]. The kidney disease outcomes evaluated in this paper are three facets: (1) ESRD that was defined as progression to hemodialysis, peritoneal dialysis or renal transplantation; (2) CVD risk that was evaluated based on onset of CVD events (myocardial infarction, heart failure, arrhythmia and cerebrovascular disease and peripheral arterial disease) and (3) all-cause mortality (deaths). Doctors at the clinical centers were requested to submit related clinical data to Renal Institute of Peking University. All events were adjudicated by an independent committee consisting of specialists.
Statistical analyses
Demographic information and other baseline variables were described using mean ± SD or median and interquartile range for continuous variables and frequency and proportion for categorical variables. Chi-square or Fisher’s exact test was used for comparison of categorical variables between groups, while Mann–Whitney U test or unpaired Student’s t test was used for continuous variables. We identified the cutoff point for NLR level using maximally selected log-rank statistics. Prevalence of events according to NLR categorical variables and estimate survival time for each category are calculated with Kaplan–Meier Survival and time-to-event analysis of outcomes were performed using Cox proportional hazards model, including adjustment for potential confounding factors. Covariates for the models were selected based on prior knowledge about the factors that could be potential confounders of the associations of NLR with ESRD. The potential confounders including age (continuous), gender (male vs. female), smoking (yes vs. no), drinking more than once a day (yes vs. no), clinical characteristics [hypertension (yes vs. no), CKD-MBD (yes vs. no), diabetes (yes vs. no), CVD (yes vs. no), hyperlipidaemia (yes vs. no), hyperuricemia (yes vs. no)], hemoglobin (continuous), eGFR (continuous), ACR categories (A1, A2, A3). A p value < 0.05 was regarded as statistically significant. Proportional hazards assumptions were verified by testing the interaction with time using the likelihood ratio test, which yielded non-significant p values. The results are presented as hazard ratios (HRs) with 95% confidence intervals (CIs). p values for trend were given by treating dichotomy NLR as a continuous variable. We hypothesized that the effect of NLR might be modified by the CKD stage. The interaction terms were generated between NLR and each of the eGFR levels (CKD1-2 vs. all others, CKD3 vs. all others, CKD4 vs. all others). A stratified analysis in the prediction of ESRD events was conducted by the stages of CKD. A series of sequential models were fit to evaluate the effect of adding certain sets of covariates. All statistical analyses were performed using SPSS 24.0 for Windows (SPSS Inc., Chicago, IL, USA).
Discussion
The current study suggests that baseline NLR is associated with an increased risk of ESRD in pre-dialysis patients with stage 4 CKD; the association was independent of traditional risk factors of CKD including ACR and eGFR.
In CKD patients, declines in glomerular filtration rate are associated with increases in the risk of CVD and rapid progression of CKD to end-stage renal disease and mortality [
15,
16]. CKD is a chronic inflammatory condition and remains a substantial economic burden on the patient, caregiver and society [
17]. Inflammation is one of the most important initiators of progressive tubule-interstitial fibrosis, which usually culminates in ESRD [
18]. Several studies have reported that inflammation plays a role in the decline of kidney function [
19]. Anti-inflammatory treatment in tubulointerstitial fibrosis of CKD may have renal protective effects [
20‐
23]. An alternative marker of systemic inflammation, c-reactive protein (CRP) has been shown to be a superior prognostic marker if it is persistently elevated [
24,
25]. HsCRP was found to be a predictor of mortality and ESRD in pre-dialysis patients with chronic kidney disease [
26]. HsCRP, which is synthesized mainly in the liver, only marginally increased even in patients with active infection [
24]. Furthermore, it is non-specific and may not fully capture all forms of inflammation [
27]. For example, the Multi-Ethnic Study of Atherosclerosis (MESA) demonstrated obesity being independently associated with subclinical atherosclerosis irrespective of hsCRP, with no additive effect when elevated hsCRP was present [
28]. What is more, hsCRP is influenced by other factors such as body mass index, weight loss, smoking, active alcohol consumption and diabetes [
29]. Similar in our study, we observed that hsCRP was not correlated with the outcome in CKD stage 1–4 patients. hsCRP is thus likely not a good inflammation indicator in CKD patients; other inflammation biomarkers should be investigated to stratify patients with stage 1–4 CKD who are at risk of poor prognosis and to monitor treatment effects.
Studies with relatively small-scale have recently been conducted on whether NLR may be a predictor of CKD progression. The balance between the inflammatory and immune response is also reflected by NLR [
7], which shows chronic low-grade inflammation. In addition to a variety of well-known risk factors for CKD progression, this study was aimed to evaluate the relationship between NLR and CKD progression. We used NLR as a surrogate marker of systemic inflammation. NLR is composed of two different complementary immune pathways [
30], and is less likely to be influenced by various physiological conditions such as dehydration. Because some inflammatory cytokines, such as hsCRP, interleukin-6 and TNF-α are limited to research for nonconventional tests, NLR would be more easily available for clinical practice [
31]. NLR displays a prognostic value for proteinuria [
32,
33], mortality and RRT [
34]. Nevertheless, recent evidence indicates NLR was not an independent predictor of CKD progression in CKD stage 2–4 patients [
35]. In our study of a large chinese CKD population for relationship between NLR, ESRD, CVD, or all-cause mortality, our longitudinal study revealed NLR as an independent risk factor of ESRD only in patients with stage 4 CKD after adjusting for classic risk factors of CKD including ACR and eGFR. These patients have severely damaged renal function prior to dialysis [
36]; chronic low-grade inflammation will likely promote disease progression towards ESRD in stage 4 CKD patients. In this regard, NLR may not be an independent predictor of CKD progression in patients with CKD at early stages [
35]. In renal replacement patients, including hemodialysis and peritoneal dialysis, evidence suggests that NLR predicts CVD and all-cause mortality [
7]. Nevertheless, in our study, this relationship could not be demonstrated in pre-dialysis patients. It remains possible that NLR predicts CVD and all-cause mortality only in stage 5 CKD patients.
Our study has several limitations. First, although most well-established risk factors of CKD progression were included in our multivariable regression models, the possibility of residual confounding still exists. Secondly, despite using a large patient cohort organized by multiple centers, risk factor exposures, CVD risk and deaths in some regions were affected by data shortages. Thirdly, the NLR only captured in 938 of 3358 patients, selection bias could not be excluded. Forthly, the cohort has a relatively short duration of follow-up and a limited number of cardiovascular events and death [
37], which limited our power to investigate the association between the levels of NLR and CVD and all-cause mortality.
Authors’ contributions
QJY searched the literature; QJY and JWW analysed the data, interpreted the results and drafted the manuscript. KH revised the manuscript. HX conceived and design the study, organised and supervised the study, interpreted the results, and revised the manuscript. LXZ and MHZ obtained funding. Other members collected and analysed the data. HX is the guarantor and takes full responsibility for the work as a whole, including the study design, access to data, and the decision to submit and publish the manuscript. All authors read and approved the final manuscript.