Abstract
Background
Malnutrition and inflammation are prevalent in hemodialysis (HD) patients and linked to cognitive impairment (CI). Naples prognostic score (NPS) is a comprehensive measure of patients’ inflammation and nutritional status. This study is to evaluate the association of NPS and CI risk in HD patients.
Methods
Two thousand seven hundred twenty-five HD patients were recruited and NPS score obtained based on albumin, total cholesterol (TC), lymphocytes, neutrophils, and monocytes. Cognitive function was assessed with Mini-Mental State Examination score (MMSE). Logistic regression models, interactive analyses were conducted.
Results
Among 2725 HD patients, the prevalence of CI is 33.8%, the mean MMSE score was 26.9 ± 3.9. After adjusting clinical confounders, NPS showed a positive associated with CI both as a continuous variable (OR = 1.120, 95% CI 1.029–1.221, p = 0.009) and as a categorical variable (OR = 1.552, 95%CI: 1.146–2.110, p = 0.005). The analysis revealed a negative correlation between NPS and MMSE scores, observed both as a continuous variable (β = –0.178, 95% CI −0.321 to −0.035, p = 0.015) and as a categorical variable. Higher NPS was significantly associated with increased dementia risk (adjusted OR = 1.153, 95% CI 1.035–1.286, p = 0.010). Among CI patients, the proportion of males was higher than that of females. Subgroup analysis showed that the effect of NPS on CI was more pronounced in individuals under 65 years, without diabetes and cerebrovascular disease (CVD). Except for males, low education level, non-CVD, and HD frequency less than three times per week, the association between NPS and dementia was more significant.
Conclusions
NPS was associated with cognitive impairment in HD patients with a positive dose-response effect.
Keywords: Naples prognostic score (NPS), cognitive impairment (CI), hemodialysis, dementia
Introduction
Research indicates that cognitive impairment (CI) is a common and notable health concern among hemodialysis (HD) patients, with research indicating that 60%–90% of end-stage kidney disease (ESKD) patients on hemodialysis experience cognitive impairment [1]. Patients undergoing hemodialysis who experience cognitive impairment face substantial challenges related to their disease, such as increased risk of hospitalization, higher mortality rates, elevated healthcare costs, and diminished quality of life [2]. Despite progress in diagnostic and treatment technologies, a considerable number of individuals undergoing hemodialysis continue to experience cognitive impairment. The causes of CI in hemodialysis patients are multifactorial, including factors related to uremia and the dialysis process, such as uremic toxins, anemia, rapid fluid shifts, and cerebral hypoperfusion. Individuals undergoing hemodialysis often experience systemic inflammation and concurrent protein-energy malnutrition [2], which is known to play a significant role in the development of cognitive impairment. The inflammatory and malnutrition status of these patients can lead to complications related to hemodialysis, ultimately resulting in cognitive impairment and increased mortality. Currently, there are no effective pharmacological treatments specifically targeting cognitive impairment in hemodialysis patients. Therefore, identifying modifiable risk factors for cognitive impairment in this population is of great importance.
The Naples Prognostic Score (NPS) is an innovative scoring system that integrates various inflammation and nutritional biomarkers, including total cholesterol, lymphocyte to monocyte ratio (LMR), neutrophil to lymphocyte ratio (NLR), and serum albumin. Studies have shown that NPS can predict acute renal failure post-myocardial infarction, malnutrition in hypertensive individuals, and postoperative complications in diverticulitis patients [3,4]. A study on kidney transplant recipients, those with a creatinine reduction ratio below 30% demonstrated a significantly prevalence of NPS 3–4, reduced LMR levels, and increased neutrophil and NLR levels [5]. An empirical study on patients with resected Cholangiocarcinoma revealed a direct association between preoperative malnutrition and the NPS, with the NPS independently indicating prognosis [6]. The NPS is a valuable tool for assessing prognostic outcomes across various cancers. However, there is a lack of research on the association between NPS and CI in HD patients. Given that NPS is a comprehensive indicator of nutrition and inflammation, understanding its impact on cognitive dysfunction is of great significance. Previous studies have shown as association between malnutrition, inflammation, and cognitive decline in the general population. Individuals undergoing HD, who often experience elevated inflammation and malnutrition, are at a higher risk of cognitive impairment [7]. Fluctuations in inflammation and nutritional status in HD patients may affect prognostic indicators for cognitive impairment over time. Despite this potential relationship, limited research has explored the link between changes in common nutritional and inflammatory conditions and cognitive impairment.
Given the high prevalence of inflammation and malnutrition among HD patients with cognitive dysfunction, a plausible association between these comorbidities is suggested. However, research on this specific cohort is scarce. Thus, this investigation aimed to assess the associations between nutritional status and inflammation with cognitive impairment in HD patients.
Methods and materials
Study design and participants
This multicenter, cross-sectional study enrolled patients receiving maintenance hemodialysis at twenty-two dialysis centers in Guizhou Province, China between 1 June 2021 and 30 September 2021. Inclusion criteria included individuals with end-stage kidney disease (ESKD) undergoing regular bicarbonate-based dialysis (twice or thrice weekly) for a minimum of three months, aged 18 years or older, and with complete biochemical, anthropometric, and questionnaire data.
Exclusion criteria comprised individuals with: (1) mental illness, severe aphasia, or critical illness impeding questionnaire completion; (2) extreme weakness resulting in a life expectancy of less than six months; (3) patients with severe liver failure, lung disease, or other conditions significantly affecting cognitive function; (4) dependence on psychotropic drugs or alcohol; (5) significant limb defects, deformities, or presence of metal stents hindering bioelectrical impedance analysis.
The study followed the Declaration of Helsinki and obtained approval for informed consent from the Institutional Review Board of Guizhou Provincial People’s Hospital (Approval Number: (2020)208). All participants provided written informed consent.
Assessment of covariates
Standard questionnaires were used to gather demographic information, such as age, sex, ethnicity, household income, educational level (low:<9th grade; high:≥9th grade), physical activity levels, smoking status (yes or no), alcohol status (yes or no), medical history, and presence of any diseases during the interviews. Hypertension was defined as systolic blood pressure (SBP) ≥140mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, or self-reported, or a medical record of responding diagnosis or medication (yes or no). Diabetes mellitus was diagnosed as fasting plasma glucose ≥126 mg/dL (7 mmol/L) or random blood glucose ≥200 mg/dL (11.1 mmol/L) or HbA1c ≥6.5%, or self-reported, or a medical record of responding to diagnosis or medication (yes or no). The following information on HD therapy was also recorded: prior to hemodialysis, anthropometric measurements including height, waist circumference, weight, systolic blood pressure, diastolic blood pressure, and hip circumference were obtained by two trained nephrologists. Additionally, data on dialysis frequency (twice/thrice per week), hemodialysis vintages, dialysis modality, dialysis dehydration amount, and complications during dialysis, such as hypotension and hypoglycemia were recorded.
Biochemical measurements
For laboratory variables, all participants provided venous blood samples after fasting for 8–10 h and were collected before the initiation of hemodialysis therapy. The following parameters were evaluated: leukocyte, hemoglobin, platelet, neutrophil, lymphocyte, monocyte, parathyroid hormone, albumin, serum uric acid, urea, total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, and other biochemical indicators.
Cognitive function evaluation
Cognitive function was assessed using the Mini-Mental State Examination (MMSE) questionnaire by the trained personnel of the research group. The total score of the evaluation scale is 30 points. It assesses cognitive function in 5 components: orientation (5 points for temporal orientation, 5 points for spatial orientation), memory (3 points for immediate recall, 3 points for delayed recall), serial subtraction (5 points), language ability (2 points for naming, 3 points for oral command comprehension, 1 point each for repetition, reading, and writing), and visuospatial ability (1 point) in order. A score of 30–27 points means no cognitive dysfunction. A score <27 on the MMSE can be diagnosed as CI, dementia diagnosis was evidence of cognitive deficit from objective assessment (MMSE score ≤23 or neuropsychological domain scores 2 SDs below the age and education–adjusted means), and evidence of social or occupational function impairment (dependency in ≥1 basic activities of daily living or Clinical Dementia Rating ≥1) [2,8]. All patients completed the MMSE at baseline and follow-up.
Calculation of Naples prognostic score
The NPS was calculated based on serum albumin, TC, NLR, and LMR. According to previous reports, serum albumin ≥40 g/L, TC >180 mg/dL, NLR <2.96, or LMR >4.44 was scored as 0, while serum albumin <40 g/L, TC ≤180 mg/dL, NLR ≥2.96, or LMR ≤4.44 was scored as 1 [4]. NPS is the sum of the scores for each of the four factors. The calculation of the NPS involves adding up the scores assigned to the parameters mentioned earlier. Patients were divided into three groups based on NPS: group 0 (score of 0 or 1), group 1(score of 2 or 3), and group 2 (score of 4).
Statistical analysis
Categorical data were expressed as numbers and percentages. Normally distributed continuous data were represented by mean and standard deviation while skewed distributions were described as medians and interquartile ranges. As there is no unified grouping standard for NPS, we divided patients into three groups based on increasing Naples scores (groups 0–2): group 0 with a score of 0 or 1, group 1 with a score of 2 or 3, group 2 with a score of 4.
Univariate and multivariate logistic regression analyses were performed to investigate the correlation between NPS and the prevalence of CI. Additionally, univariate and multivariate linear regression analyses were conducted to examine the relationship between MMSE scores and NPS. Odds ratios (ORs), unstandardized coefficients (β), and confidence intervals (CIs) summarized, respectively. Model 1 was adjusted for education. In model 2, we adjusted for model 1 and age, sex, BMI, Hypertension, alcohol, diabetes mellitus, smoking, SBP1, DBP1, and CVD; in model 3, we adjusted for model 2 and HD vintages and dialysis frequency; Model 4, adjusted for model 3 and uric acid, creatinine, potassium, phosphorus, and high-density lipoprotein (HDL). A subgroup analysis was conducted to explore the potential effect modification by sex (female or male), age (<65 or ≥65 years), education level (low or high), HD frequency (<3times or ≥3 times), smoking history (yes or no), history of hypertension (yes or no), diabetes mellitus history (yes or no), history of diabetes (yes or no), and CVD (yes or no).
All statistical analyses were performed by IBM SPSS Statistics (version 27.0, Chicago, USA) the R open-source software (version 4.3.3.), a two-sided P-value < 0.05 was considered statistically different.
Results
Characteristics of study participants
As illustrated in Figure 1, 3902 hemodialysis patients were enrolled in the study, after applying the exclusion criteria, the final analysis included 2725 HD patients, with a mean age of 54.44 ± 14.8 years, and 1668 (61.2%) were male. The median dialysis vintage was 54.0 (33.0, 87.0) months. The average number of NPS components was 2.8 ± 1.0. The mean MMSE score was 26.87 ± 3.9, the prevalence of CI (MMSE < 27) was 33.8%, and the prevalence of dementia (MMSE ≤ 23) was 16.8%. The patients were stratified into two groups based on their cognitive status: the normal cognition group (n = 1803, 66.2%) and the CI group (n = 922, 33.8%).
Figure 1.
Flowchart of the study. HD: hemodialysis; NPS: Naples prognostic score; CI: cognitive impairment.
As shown in Table 1, the mean age of the overall population was 54.44 ± 14.8 years, 1668 (61.2%) were male. Among CI patients, the proportion of males was higher than that of females; compared to patients with normal cognition, those with CI were older, lower DBP levels, higher prevalence rates of hypertension, lower uric acid levels, and lower hemoglobin, albumin, urea nitrogen, and creatinine levels. The mean MMSE score was 26.9 ± 3.9. The mean scores for the subgroups were as follows: Orientation, 9.5 ± 1.2; Registration, 2.8 ± 0.5; Attention and Calculation, 3.9 ± 1.5; Recall, 2.3 ± 0.9; and Language, 8.3 ± 1.3.
Table 1.
Baseline characteristics of hemodialysis patients according to cognitive function.
| Variables | All (n = 2725) | Normal cognition (n = 1803) | Cognitive impairment (n = 922) | P |
|---|---|---|---|---|
| Male, (n,%) | 1668 (61.2%) | 1166 (64.7%) | 502 (54.4%) | <0.001 |
| Age (years) | 54.4 ± 14.8 | 52.1 ± 14.4 | 59.1 ± 14.5 | <0.001 |
| High education level (n,%) | 161 (5.9%) | 113 (6.3%) | 48 (5.2%) | 0.266 |
| Dialysis frequency (<thrice/week) (n,%) | 439 (16.1%) | 298 (16.5%) | 141 (15.3%) | 0.407 |
| Dialysis vintage (months) | 54.0 [33.0,87.0] | 55.0 [35.0,90.0] | 52.0 [31.0,84.0] | 0.007 |
| Alcohol history (n,%) | 195 (7.2%) | 118 (6.5%) | 77 (8.4%) | 0.083 |
| Smoking history (n,%) | 1041 (38.2%) | 675 (37.4%) | 366 (39.7%) | 0.251 |
| History of hypertension (n,%) | 2228 (81.8%) | 1434 (79.5%) | 794 (86.1%) | <0.001 |
| Diabetes mellitus (n,%) | 933 (34.2%) | 568 (31.5%) | 365 (39.6%) | <0.001 |
| Heart failure (n,%) | 675 (24.8%) | 440 (24.4%) | 235 (25.5%) | 0.535 |
| CVD (n,%) | 238 (8.7%) | 145 (8.0%) | 93 (10.1%) | 0.074 |
| Pre-dialysis SBP (mmHg) | 136.6 ± 18.4 | 136.7 ± 18.5 | 136.3 ± 18.4 | 0.646 |
| Pre-dialysis DBP (mmHg) | 78.2 ± 14.0 | 78.8 ± 13.8 | 76.9 ± 14.3 | 0.001 |
| BMI (kg/m2) | 22.9 ± 3.6 | 22.9 ± 3.6 | 23.0 ± 3.6 | 0.624 |
| Leukocyte (x109/L) | 6.4 ± 2.1 | 6.5 ± 2.2 | 6.3 ± 2.1 | 0.046 |
| Hemoglobin (g/L) | 109.5 ± 21.0 | 110.1 ± 20.6 | 108.3 ± 21.7 | 0.034 |
| Platelet (x109/L) | 177.8 ± 63.2 | 178.6 ± 63.7 | 176.3 ± 62.2 | 0.380 |
| Neutrophils (x109/L) | 4.4 ± 1.8 | 4.5 ± 1.9 | 4.4 ± 1.8 | 0.153 |
| Lymphocyte (x109/L) | 1.3 ± 0.7 | 1.3 ± 0.8 | 1.2 ± 0.6 | 0.013 |
| Monocyte (x109/L) | 0.5 ± 0.3 | 0.5 ± 0.4 | 0.5 ± 0.3 | 0.410 |
| Parathyroid hormone (pg/mL) | 316.9 [169.8,573.5] | 322.9 [171.6,593.7] | 305.1 [165.5,531.5] | 0.055 |
| Albumin (g/L) | 40.0 ± 4.4 | 40.3 ± 4.2 | 39.3 ± 4.6 | <0.001 |
| Serum uric acid (mmol/L) | 427.3 ± 173.0 | 436.0 ± 190.5 | 410.3 ± 130.8 | <0.001 |
| Urea (mmol/L) | 20.8 ± 14.3 | 21.3 ± 14.5 | 19.8 ± 13.9 | 0.009 |
| Serum creatinine (μmol/l) | 888.0 [683.0,1103.0] | 932.0 [732.0,1135.2] | 808.6 [584.0,1016.6] | <0.001 |
| Potassium (mmol/L) | 4.7 ± 0.8 | 4.8 ± 0.8 | 4.6 ± 0.8 | <0.001 |
| Total protein (mmol/L) | 68.8 ± 6.4 | 69.0 ± 6.3 | 68.3 ± 6.5 | 0.006 |
| Sodium (mmol/L) | 139.0 ± 3.7 | 138.9 ± 3.7 | 139.1 ± 3.6 | 0.177 |
| Calcium (mmol/L) | 2.2 ± 0.3 | 2.2 ± 0.3 | 2.2 ± 0.3 | 0.115 |
| Total cholesterol (mmol/L) | 4.0 ± 1.0 | 4.0 ± 1.0 | 4.0 ± 1.1 | 0.449 |
| HDL (mmol/L) | 1.2 ± 0.4 | 1.1 ± 0.4 | 1.2 ± 0.4 | 0.005 |
| Phosphorus (mmol/L) | 1.8 ± 0.6 | 1.8 ± 0.6 | 1.7 ± 0.6 | <0.001 |
| Triglyceride (mmol/L) | 1.9 ± 1.5 | 2.0 ± 1.5 | 1.9 ± 1.5 | 0.285 |
| LDL (mmol/L) | 2.2 ± 0.8 | 2.2 ± 0.8 | 2.2 ± 0.9 | 0.913 |
| NPS (continous) | 3.0 [2.0,4.0] | 3.0 [2.0,3.0] | 3.0 [2.0,4.0] | 0.001 |
| MMES | 26.9 ± 3.9 | 29.1 ± 1 | 22.5 ± 3.8 | <0.001 |
| Orientation | 9.5 ± 1.2 | 9.9 ± 0.4 | 8.72 ± 1.80 | <0.001 |
| Registration | 2.8 ± 0.5 | 3.0 ± 0.2 | 2.6 ± 0.8 | <0.001 |
| Attention and calculation | 3.9 ± 1.5 | 4.7 ± 0.6 | 2.3 ± 1.5 | <0.001 |
| Recall | 2.3 ± 0.9 | 2.7 ± 0.6 | 1.7 ± 1.0 | <0.001 |
| Language | 8.3 ± 1.3 | 8.8 ± 0.6 | 7.2 ± 1.7 | <0.001 |
According to the stratification of the NPS, a higher NPS score is associated with an increased incidence of CI, and the clinical and biochemical characteristics as shown in the Supplementary Table S1.
Association between NPS and cognitive impairment
Patients with NPS = 4 exhibited a higher prevalence of cognitive impairment compared to those with NPS = 2–3 and NPS = 0–1 (39.4% vs. 32.3% and 29.4%, respectively), as shown in Figure 2. The multivariate adjusted ORs and 95% CIs for the prevalence of cognitive impairment based on the continuous NPS value and increasing Naples scores (groups 0–2) are presented in Table 2 and Supplementary Table S2. When regarded as a continuous variable, NPS (OR = 1.120, 95% CI 1.029–1.221, p = 0.009) had a higher risk of CI occurrence. The patients were divided into three groups based on increasing Naples scores (groups 0–2). After adjusting Model 4, compared with those NPS of 0–1 as a reference, HD patients with an NPS of 4 (OR = 1.552, 95%CI: 1.146–2.110, p = 0.005) showed a higher risk of CI. While, no significant differences were found in the 2–3 group.
Figure 2.
Prevalence of cognitive impairment in HD patients stratified by NPS.
Table 2.
Logistic regression analysis of cognitive impairment prevalence according to NPS.
| Variables | Model 1 |
Model 2 |
Mode 3 |
Model 4 |
||||
|---|---|---|---|---|---|---|---|---|
| OR (95%CI) | P-value | OR (95%CI) | P-value | OR (95%CI) | P-value | OR (95%CI) | P-value | |
| Categorical variable | ||||||||
| 0–1 | Reference | Reference | Reference | Reference | ||||
| 2–3 | 1.149 (0.885–1.501) | 0.302 | 1.100 (0.838–1.452) | 0.497 | 1.109 (0.845–1.466) | 0.459 | 1.159 (0.880–1.537) | 0.298 |
| 4 | 1.573 (1.184–2.100) | 0.002 | 1.466 (1.090–1.981) | 0.012 | 1.472 (1.094–1.990) | 0.011 | 1.552 (1.146–2.110) | 0.005 |
| Continuous variable | ||||||||
| NPS | 1.131 (1.044–1.226) | 0.003 | 1.105 (1.018–1.202) | 0.019 | 1.105 (1.017–1.202) | 0.019 | 1.120 (1.029–1.221) | 0.009 |
Note: p < 0.05 was considered statistically significant. Abbreviations: OR: odds ratio; CI: confidence interval; Model 1, education; Model 2, model 1 + age, sex, BMI, Hypertension, alcohol, diabetes mellitus, smoke, SBP1, DBP1, CVD; Model 3, model 2 + HD vintages, Dialysis frequency; Model 4, model 3 + uric acid, creatinine, potassium, phosphorus, and high density lipoprotein (HDL).
Table 3 shows the relationship between the NPS and MMSE scores. When regarded as a continuous variable, after adjusting for Model 4, NPS (β = -0.178, 95% CI −0.321 to −0.035, p = 0.015) was significantly associated with lower MMSE scores. Patients were further grouped by increasing Naples scores (groups 0–2), Compared to those NPS 0–1 score group, the MMSE score of those with 4 score group is 0.681 lower (p = 0.008).
Table 3.
Association of NPS with MMSE score, using linear regression analysis among hemodialysis patients.
| Variables | Model 1 |
Model 2 |
Mode 3 |
Model 4 |
||||
|---|---|---|---|---|---|---|---|---|
| β (95%CI) | P-value | β (95%CI) | P-value | β (95%CI) | P-value | OR (95%CI) | P-value | |
| Categorical variables | ||||||||
| 0–1 | Reference | Reference | Reference | Reference | ||||
| 2–3 | −0.249 (−0.721,0.223) | 0.301 | −0.123 (−0.577,0.330) | 0.594 | −0.137 (−0.590,0.317) | 0.555 | −0.210 (−0.663,0.244) | 0.365 |
| 4 | −0.834 (−1.355,−0.313) | 0.002 | −0.603 (−1.098,−0.094) | 0.019 | −0.607 (−1.110,−0.106) | 0.018 | −0.681 (−1.186,−0.176) | 0.008 |
| Continuous variable | ||||||||
| NPS | −0.231 (−0.378,-0.084) | 0.002 | −0.158 (−0.299,-0.017) | 0.029 | −0.159 (−0.300,−0.018) | 0.028 | −0.178 (−0.321,−0.035) | 0.015 |
Note: p < 0.05 was considered statistically significant. Abbreviations: β Unstandardized coefficient; CI: confidence interval；MMSE Mini-mental state examination. Model 1, education; Model 2, model 1 + age, sex, BMI, Hypertension, alcohol, diabetes mellitus, smoke, SBP1, DBP1, CVD; Model 3, model 2 + education, HD vintages, Dialysis frequency; Model 4, model 3 + uric acid, creatinine, potassium, phosphorus, high density lipoprotein (HDL).
This study further explored the relationship between NPS and the prevalence of dementia (Table 4; Supplementary Table S3). The finding indicated that, when regarded as a continuous variable, NPS had higher risk of dementia with multivariate-adjusted ORs of 1.153 (95% CI 1.035 − 1.286, p = 0.010). When divided into three groups based on increasing Naples scores (groups 0–2). After adjusting Model 4, compared with those NPS of 0–1 as a reference, HD patients with NPS of 4 (OR = 1.538, 95%CI: 1.055–2.276, p = 0.028) showed a higher risk of CI. While, no significant differences were found in the 2–3 group.
Table 4.
Logistic regression analysis of dementia prevalence according to the NPS.
| Variables | Model 1 |
Model 2 |
Mode 3 |
Model 4 |
||||
|---|---|---|---|---|---|---|---|---|
| OR (95%CI) | P-value | OR (95%CI) | P-value | OR (95%CI) | P-value | OR (95%CI) | P-value | |
| Categorical variables | ||||||||
| 0–1 | Reference | Reference | Reference | Reference | ||||
| 2–3 | 1.101 (0.789–1.566) | 0.583 | 1.055 (0.747–1.517) | 0.764 | 1.066 (0.754–1.533) | 0.725 | 1.111 (0.784–1.603) | 0.563 |
| 4 | 1.579 (1.105–2.293) | 0.014 | 1.468 (1.014–2.158) | 0.046 | 1.477 (1.019–2.173) | 0.043 | 1.538 (1.055–2.276) | 0.028 |
| Continuous variable | ||||||||
| NPS | 1.162 (1.049–1.290) | 0.004 | 0.130 (1.025–1.268) | 0.016 | 1.140 (1.026–1.270) | 0.016 | 1.153 (1.035–1.286) | 0.010 |
Note: p < 0.05 was considered statistically significant. Abbreviations: OR: odds ratio; CI: confidence interval; Model 1, education; Model 2, model 1 + age, sex, BMI, Hypertension, alcohol, diabetes mellitus, smoke, SBP1, DBP1, CVD; Model 3, model 2 + education, HD vintages, Dialysis frequency; Model 4, model 3 + uric acid, creatinine, potassium, phosphorus, high density lipoprotein (HDL).
Subgroup analyses of association between NPS and the prevalence of CI
As show in Figure 3(A) a stronger association between NPS and CI prevalence was found among patients aged < 65 years (OR 1.210, 1.095–1.348, p < 0.001), low educational level (OR 1.129, 1.034–1.234, p < 0.001), under 3 times HD frequency per week (OR 1.110, 1.01–1.22, p = 0.026), hypertension history (OR 1.156, 1.053–1.271, p < 0.001), no-diabetes mellitus history (OR 1.125, 1.013–1.251, p = 0.03), no-CVD history (OR 1.158, 1.058–1.268, p < 0.001). Significant interaction effect of sex, age, education, CVD on the CI was observed (all P for interaction < 0.001), respectively.
Figure 3.
Subgroup analyses of association between NPS and cognitive impairment (A), and subgroup analyses of association between NPS and dementia (B). Model was adjusted for age, sex, BMI, hypertension, alcohol, diabetes mellitus, smoke, SBP1, DBP1, CVD, education, HD vintages, dialysis frequency, uric acid, creatinine, potassium, phosphorus, high density lipoprotein (HDL). OR: odds ratio, CI: confidence interval, HD hemodialysis, CVD: cerebral vascular disease.
Subgroup analyses of association between NPS and the prevalence of dementia
Figure 3(B) shows a stronger association between NPS and dementia prevalence was observed among male patients (OR 1.174, 1.009–1.371, p < 0.04), aged <65 years (OR 1.252, 1.091–1.442, p < 0.002), low educational level (OR 1.149, 1.029–1.285, p < 0.014), under 3 times HD frequency per week (OR 1.518, 1.111–2111, p = 0.011), hypertension history (OR 1.201, 1.066–1.355, p < 0.003), diabetes mellitus history (OR 1.242, 1.033–1.503, p = 0.023), no-CVD history (OR 1.186,1.058–1.332, p < 0.004). Significant interaction effect of sex, education, HD frequency, CVD on the dementia was observed (P for interaction 0.004, 0.007, 0.024, <0.001, respectively).
Discussion
This study represents the initial investigation into the horizontal relationship between NPS and CI prevalence in hemodialysis patients. Our results suggest that higher NPS scores were associated with lower MMSE scores and an increased prevalence of CI even after adjusting for Model4. Sex, age, education, CVD had an interactive role in the association between NPS and CI prevalence. Sex, education level, HD frequency, and CVD interacted in the association between NPS and dementia prevalence. Previous researches have indicated that malnutrition and inflammation leads to cognitive impairment in HD patients [2], with each component independently associated with CI. NPS as a comprehensive indicator of inflammation and nutritional status has been utilized as a prognostic marker in various cancers and acute kidney disease [4]. However, this comprehensive scoring system has not been examined in CI, we propose to investigate the relationship between NPS and CI in hemodialysis patients.
Malnutrition is highly prevalent among HD patients, with studies demonstrating a clear association between malnutrition and CI [9]. The causes of malnutrition are complex and can be attributed to non-iatrogenic and iatrogenic factors. Serum albumin levels are a sensitive indicator for identifying HD patients at risk for malnutrition [10]. Albumin serves as a marker for malnutrition and also reflects various non-nutritional factors that frequently occur among HD patients, including inflammation, anemia, dialysate losses, and hydration status [11]. Serum albumin has the ability to withstand both exogenous and endogenous oxidants, and a reduction in its levels may lead to an imbalance in oxidation and antioxidation processes, potentially resulting in cognitive impairment [12]. However, albumin concentrations can be affected by age, liver function, comorbidities, inflammation and alterations in body fluid volume, particularly in HD patients [13]. Several authors have proposed adding plasma cholesterol levels as a means to enhance the assessment of nutritional status [14].
Patients with cognitive impairment (CI) exhibit significantly lower levels of TC, LDL, and HDL compared to those with normal cognitive function [15]. Increasing evidence suggests that insufficient cholesterol in brain cells may contribute to Alzheimer’s disease (AD) pathology, as neuronal morphology, function, and synaptic transmission depend on adequate cholesterol for maintaining cell membrane integrity and regulating signaling molecules [16].
Chronic inflammation is prevalent among HD patients. Numerous studies have demonstrated that the detrimental effect of elevated inflammatory status on the cognitive abilities [17], particularly in CKD and HD patients. NLR has been identified as being in closely associated with cognitive ability and various mental states. Yu et al. found higher neutrophil counts and lower lymphocyte counts in the mild cognitive impairment group compared to those with normal cognitive function in their research [18]. A study on older individuals in Northeast China found a positive correlation between the NLR and the presence of Alzheimer’s disease (AD) and MCI. The NLR was also found to be useful in distinguishing normal cognitive function from AD or MCI [19]. In a study focusing on older patients with esophageal cancer, a significant correlation was initially observed between postoperative NLR levels and MMSE scores [20]. Fang et al. demonstrated that a high NLR was linked to poor visual memory and visuospatial performance [21]. Recent findings indicated that a lower lymphocyte count could predict ApoE “4-related cognitive decline in Parkinson’s disease [22] and was associated with an increased risk of subsequent Parkinson’s disease diagnosis [23], and associated with an increased risk of subsequent Parkinson’s disease diagnosis. Monocytes play a critical role in the innate immune response by serving as the primary defense against pathogens. Research indicates that elevated levels of peripheral soluble CD163 and CD14 are commonly linked to cognitive impairment [24]. In HIV-uninfected individuals, increased levels of intermediate monocytes may potentially serve as a biomarker for distinguishing between individuals with depressive symptoms and those without such symptoms [25].
The precise pathophysiological mechanism connecting NLR and cognitive impairment remains uncertain. Neutrophils play a role in combating pathogens through phagocytosis and the ingestion of cellular debris, releasing inflammatory factors from granules and vesicles upon activation [26]. In the pathophysiology of neurodegenerative diseases, inflammation is a crucial factor. Persistent inflammation can cause vascular dysfunction, leading to compromised cerebral blood flow and oxygenation. Prolonged inflammation may lead to adverse outcomes such as neuronal dysfunction, synaptic degeneration, and ultimately neuronal death, which can impact cognitive functions [27]. However, activated neutrophils have been shown to not only activate microglia and release neurotoxic substances, but also disrupt the blood-brain barrier through the promotion of reactive oxygen species (ROS) release. Additionally, activated neutrophils have been found to stimulate T lymphocytes by enhancing antigen presentation, ultimately resulting in neuroinflammation and neuronal damage [16,28]. Studies have shown that lymphocyte infiltration into the ischemic brain persists for a minimum of 72 days following the onset of a stroke, indicating a substantial role of lymphocytes in post-stroke cognitive impairment. This suggests that therapeutic strategies aimed at depleting CD4+ T cells may be a promising approach for treatment [29].
According to the Standard Operating Procedure for HD patients, the Subjective Global Assessment (SGA) and Malnutrition-Inflammation Score (MIS) methodologies are routinely employed to assess the nutritional and inflammatory statuses of hemodialysis patients. However, the SGA scoring heavily relies on subjective perceptions and reports from patients and their families, which can be influenced by psychological factors and does not evaluate the inflammatory status. In contrast, the MIS scoring requires more data and involves more intricate calculation methods, necessitating specialized equipment and personnel, which may increase medical costs. The Naples prognostic score, a composite index of inflammatory and nutritional markers, has been previously investigated for its prognostic value in cancer patients [6]. The NPS incorporates three leukocyte subtypes, as well as albumin and cholesterol, suggesting that the combined effects may be more advantageous than those of the individual components. Our results indicate that NPS is associated with a higher risk of CI occurrence, whether used as a continuous or categorical variable. This adaptability allows researchers to choose the most suitable approach for utilizing this metric, thereby enhancing the effectiveness of prediction and intervention strategies. The constituent components of the NPS are commonly employed in clinical settings and are relatively simple to calculate. Early computation of the score could potentially aid in identifying individuals who might benefit from preventive interventions. In our study, among CI patients, there is a higher proportion of males compared to females. Subgroup analysis revealed that the impact of NPS on CI was more significant in individuals under 65 years of age, with lower educational levels, without diabetes and Cardiovascular Disease (CVD) among Hemodialysis (HD) patients. These findings suggest potential associations with gender-specific physiological, psychological, or social factors that may contribute to suboptimal nutrition or inflammation in the absence of timely recognition and intervention. Individuals under 65 years may exhibit relatively better physical health and be more sensitive to physiological and psychological responses to nutrition or inflammation. The negative impact of NPS on cognitive function may be more significant. Patients with lower education may have limited cognitive reserves and information processing abilities, potentially exacerbating the negative effects of NPS. Diabetes mellitus is known to influence cognitive function. The findings of our study indicate that, among individuals without diabetes, NPS exert a more significant impact on cognitive function. This underscores the critical importance of managing NPS in non-diabetic populations. Additionally, the presence of cerebrovascular diseases can further impair cognitive function. In patients without these comorbidities, NPS may play a more prominent role in affecting cognitive function. Our study revealed a significant association between NPS and dementia in patients who receive hemodialysis less than three times per week, except for males, those with a low education level, and those without cerebrovascular disease. These findings offer a new perspective on the relationship between NPS and cognitive impairment in hemodialysis patients. Furthermore, prior research has established a correlation between DBP and the risk of dementia [30]. Our findings align with these studies, suggesting that low DBP may reflect impaired vascular constriction ability, resulting in inadequate cerebral perfusion. Particularly in dialysis patients, factors such as concurrent anemia exacerbate cerebral ischemia and hypoxia during hemodialysis, consequently elevating the risk of dementia.
This study presents new insights into the association between NPS and cognitive impairment in HD patients. Nevertheless, it is crucial to recognize some limitations. Firstly, NPS combines inflammatory and nutritional indicators, facilitating the identification of high-risk patients and informing personalized treatment strategies. Despite its demonstrated predictive value in a range of diseases, further large-scale clinical studies are necessary to substantiate its applicability across diverse patient populations. Secondly, all participants in this study were recruited from a specific region in Southwestern China, indicating a regional constraint in generalizing the findings, at the same time, the absence of patient follow-up precludes the possibility of conducting longitudinal data analysis. Thirdly, the study did not explore other potential markers such as C-reactive protein, interleukins, erythrocyte sedimentation rate, and oral medications. Lastly, the diagnosis of cognitive impairment relied solely on the Mini-Mental State Examination (MMSE), which may lack the diagnostic accuracy provided by comprehensive neuropsychological test batteries. The MMSE performance can be influenced by linguistic factors, potentially leading to false-positive outcomes in individuals speaking dialects. To address this issue, the study deployed a consistent team of trained research physicians proficient in local languages to conduct clinical assessments using face-to-face interviews and physical evaluations. In the future, we intend to undertake prospective investigations and explore the combined prediction of NPS with other biomarkers such as C-reactive protein, interleukins, and more. This study has been previously posted as a preprint on Research Square [31].
Conclusion
In summary, we used a prediction model based on NPS and evaluated the prevalence of CI. We found that in patients on HD, higher baseline NPS was associated with the occurrence of CI. This study provides a justification for the assessment of malnutrition and inflammation using NPS as part of the cognitive function assessment. NPS provides a comprehensive evaluation of the nutritional and inflammatory status of HD patients, and we look forward to conducting prospective studies to further validate our findings.
Supplementary Material
Funding Statement
This study was supported by Science and Technology Cooperation Foundation of Guizhou Province (ZK [2021]381). Technology Department (QKHCG2023-ZD010).
Authors’ contributions
Yan Ran: investigation, data curation, methodology, validation, writing–original-draft, writing–review & editing. Yuqi Yang: investigation, methodology, conceptualization. Yanzhe Peng, Jingjing Da, Zuping Qian, Jing Yuan: investigation, conceptualization. Yan Zha: project administration, conceptualization, supervision. All authors have read and approved the final version of the manuscript.
Disclosure statement
The authors declare no competing interests.
Ethics approval and consent to participate
The study followed the Declaration of Helsinki and received informed consent approval from the Institutional Review Board of Guizhou Provincial People’s Hospital (Approval Number: (2020)208). All participants provided written informed consent.
Data availability statement
All data used in this study are publicly available. To assess the data, please contact the corresponding author.
References
- 1.van Zwieten A, Wong G, Ruospo M, et al. Prevalence and patterns of cognitive impairment in adult hemodialysis patients: the COGNITIVE-HD study. Nephrol Dial Transplant. 2018;33(7):1197–1206. doi: 10.1093/ndt/gfx314. [DOI] [PubMed] [Google Scholar]
- 2.Yang Y, Da J, Li Q, et al. The impact of malnutrition, inflammation on cognitive impairment in hemodialysis patients: a multicenter study. Kidney Blood Press Res. 2022;47(12):711–721. doi: 10.1159/000527453. [DOI] [PubMed] [Google Scholar]
- 3.Russell B, Zager Y, Mullin G, et al. Naples prognostic score to predict postoperative complications after colectomy for diverticulitis. Am Surg. 2023;89(5):1598–1604. doi: 10.1177/00031348211069803. [DOI] [PubMed] [Google Scholar]
- 4.Karakoyun S, Cagdas M, Celik AI, et al. Predictive value of the Naples Prognostic score for acute kidney injury in ST-elevation myocardial infarction patients undergoing primary percutaneous coronary intervention. Angiology. 2024;75(6):576–584. doi: 10.1177/00033197231161922. [DOI] [PubMed] [Google Scholar]
- 5.Aytaç İ, Güven Aytaç B, Kilci O, et al. Naples prognostic score for graft functions after renal transplantation: a retrospective analysis. Ann Transplant. 2023;28:e942007. doi: 10.12659/AOT.942007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Xu B, Zhu J, Wang R, et al. Clinical implications of Naples prognostic score for patients with resected cholangiocarcinoma: a real-world experience. J Inflamm Res. 2024;17:655–667. doi: 10.2147/JIR.S446735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhong C, Bu X, Xu T, et al. Serum matrix metalloproteinase-9 and cognitive impairment after acute ischemic stroke. J Am Heart Assoc. 2018;7(1):e007776. doi: 10.1161/JAHA.117.007776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ng TP, Feng L, Nyunt MS, et al. Metabolic syndrome and the risk of mild cognitive impairment and progression to dementia: follow-up of the Singapore longitudinal ageing study cohort. JAMA Neurol. 2016;73(4):456–463. doi: 10.1001/jamaneurol.2015.4899. [DOI] [PubMed] [Google Scholar]
- 9.Rotondi S, Tartaglione L, Pasquali M, et al. Association between cognitive impairment and malnutrition in hemodialysis patients: two sides of the same coin. Nutrients. 2023;15(4):813. doi: 10.3390/nu15040813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zhang K, Gao J, Chen J, et al. MICS, an easily ignored contributor to arterial calcification in CKD patients. Am J Physiol Renal Physiol. 2016;311(4):F663–F670. doi: 10.1152/ajprenal.00189.2016. [DOI] [PubMed] [Google Scholar]
- 11.Friedman AN, Fadem SZ.. Reassessment of albumin as a nutritional marker in kidney disease. J Am Soc Nephrol. 2010;21(2):223–230. doi: 10.1681/ASN.2009020213. [DOI] [PubMed] [Google Scholar]
- 12.Wang L, Wang F, Liu J, et al. Inverse relationship between baseline serum albumin levels and risk of mild cognitive impairment in elderly: a seven-year retrospective cohort study. Tohoku J Exp Med. 2018;246(1):51–57. doi: 10.1620/tjem.246.51. [DOI] [PubMed] [Google Scholar]
- 13.Tokunaga R, Sakamoto Y, Nakagawa S, et al. CONUT: a novel independent predictive score for colorectal cancer patients undergoing potentially curative resection. Int J Colorectal Dis. 2017;32(1):99–106. doi: 10.1007/s00384-016-2668-5. [DOI] [PubMed] [Google Scholar]
- 14.Toyokawa T, Kubo N, Tamura T, et al. The pretreatment controlling nutritional status (CONUT) score is an independent prognostic factor in patients with respectable thoracic esophageal squamous cell carcinoma: results from a retrospective study. BMC Cancer. 2016;16(1):722. doi: 10.1186/s12885-016-2696-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dietschy JM, Turley SD.. Thematic review series: brain Lipids. Cholesterol metabolism in the central nervous system during early development and in the mature animal. J Lipid Res. 2004;45(8):1375–1397. doi: 10.1194/jlr.R400004-JLR200. [DOI] [PubMed] [Google Scholar]
- 16.Ozturk HM, Ogan N, Erdogan M, et al. The association between total cholesterol and cognitive impairment in chronic obstructive pulmonary disease patients. Prostaglandins Other Lipid Mediat. 2023;164:106697. doi: 10.1016/j.prostaglandins.2022.106697. [DOI] [PubMed] [Google Scholar]
- 17.Bauer IE, Pascoe MC, Wollenhaupt-Aguiar B, et al. Inflammatory mediators of cognitive impairment in bipolar disorder. J Psychiatr Res. 2014;56:18–27. doi: 10.1016/j.jpsychires.2014.04.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yu ZW, Wang Y, Li X, et al. Association between the neutrophil to lymphocyte ratio and mild cognitive impairment in patients with type 2 diabetes. Aging Clin Exp Res. 2023;35(6):1339–1345. doi: 10.1007/s40520-023-02420-z. [DOI] [PubMed] [Google Scholar]
- 19.Dong X, Nao J, Shi J, et al. Predictive value of routine peripheral blood biomarkers in Alzheimer’s disease. Front Aging Neurosci. 2019;11:332. doi: 10.3389/fnagi.2019.00332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhao J, Dai T, Ding L, et al. Correlation between neutrophil/lymphocyte ratio, platelet/lymphocyte ratio and postoperative cognitive dysfunction in elderly patients with esophageal cancer. Medicine (Baltimore). 2023;102(10):e33233. doi: 10.1097/MD.0000000000033233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fang Y, Doyle MF, Alosco ML, et al. Cross-sectional association between blood cell phenotypes, cognitive function, and brain imaging measures in the community-based Framingham heart study. J Alzheimers Dis. 2022;87(3):1291–1305. doi: 10.3233/JAD-215533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tsukita K, Sakamaki-Tsukita H, Takahashi R.. Lower circulating lymphocyte count predicts ApoE ε4-related cognitive decline in Parkinson’s disease. Mov Disord. 2021;36(12):2969–2971. doi: 10.1002/mds.28799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Jensen MP, Jacobs BM, Dobson R, et al. Lower lymphocyte count is associated with increased Risk of Parkinson’s disease. Ann Neurol. 2021;89(4):803–812. doi: 10.1002/ana.26034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Pulliam L, Gascon R, Stubblebine M, et al. Unique monocyte subset in patients with AIDS dementia. Lancet. 1997;349(9053):692–695. doi: 10.1016/S0140-6736(96)10178-1. [DOI] [PubMed] [Google Scholar]
- 25.Lynall ME, Turner L, Bhatti J, et al. Peripheral blood cell-stratified subgroups of inflamed depression. Biol Psychiatry. 2020;88(2):185–196. doi: 10.1016/j.biopsych.2019.11.017. [DOI] [PubMed] [Google Scholar]
- 26.Satoh A, Imai SI, Guarente L.. The brain, sirtuins, and ageing. Nat Rev Neurosci. 2017;18(6):362–374. doi: 10.1038/nrn.2017.42. [DOI] [PubMed] [Google Scholar]
- 27.Kure CE, Rosenfeldt FL, Scholey AB, et al. Relationships among cognitive function and cerebral blood flow, oxidative stress, and inflammation in older heart failure patients. J Card Fail. 2016;22(7):548–559. doi: 10.1016/j.cardfail.2016.03.006. [DOI] [PubMed] [Google Scholar]
- 28.Sayed A, Bahbah EI, Kamel S, et al. The neutrophil-to-lymphocyte ratio in Alzheimer’s disease: current understanding and potential applications. J Neuroimmunol. 2020;349:577398. doi: 10.1016/j.jneuroim.2020.577398. [DOI] [PubMed] [Google Scholar]
- 29.Weitbrecht L, Berchtold D, Zhang T, et al. CD4(+) T cells promote delayed B cell responses in the ischemic brain after experimental stroke. Brain Behav Immun. 2021;91:601–614. doi: 10.1016/j.bbi.2020.09.029. [DOI] [PubMed] [Google Scholar]
- 30.Ou Y-N, Tan C-C, Shen X-N, et al. Blood pressure and risks of cognitive impairment and dementia: a systematic review and meta-analysis of 209 prospective studies. Hypertension. 2020;76(1):217–225. doi: 10.1161/HYPERTENSIONAHA.120.14993. [DOI] [PubMed] [Google Scholar]
- 31.Ran Y, Yang Y, Peng Y, et al. The impact of Naples prognostic score on cognitive impairment in hemodialysis patients: a multicenter study. Research Square. [Preprint]. 2024; Available from doi: 10.21203/rs.3.rs-4773830/v1. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data used in this study are publicly available. To assess the data, please contact the corresponding author.