The neutrophil-percentage-to-albumin ratio (NPAR) shows a nonlinear correlation with all-cause mortality in patients with anemia

Abstract

The neutrophil percentage-to-albumin ratio (NPAR) is an emerging indicator of inflammation that has been associated with the prognosis of various diseases, including hypertension, diabetes, and cardiovascular disease. NPAR is calculated by multiplying the neutrophil percentage by 100 and dividing it by the albumin value. This study aimed to evaluate the predictive value of NPAR for all-cause mortality in anemic patients. We employed Kaplan–Meier analysis, multiple regression models, restricted cubic splines (RCS), and threshold effect analysis to explore these associations. This study included 3,258 anemic individuals. The relationship between NPAR and all-cause mortality was evaluated using Kaplan–Meier analysis and multiple regression models. RCS curves were employed to assess potential nonlinear relationships, while threshold effect analysis was used to identify breakpoints in the association between NPAR and mortality risk. Subgroup analyses were conducted to examine variations in this relationship across different population strata. The study population included 3,258 anemic patients, with a mean follow-up period of 94.3 months. During follow-up, 904 participants (27.75%) died. Kaplan–Meier analysis revealed that individuals in the highest NPAR quartile (Q4) exhibited significantly higher cumulative mortality compared with those in the lowest quartile (Q1) (p < 0.001). Multiple regression analysis showed that, in model 3, each unit increase in NPAR among individuals in the highest quartile was associated with a 51.8% increased risk of all-cause mortality. RCS and threshold effect analyses revealed a nonlinear relationship between NPAR and mortality, with a breakpoint at NPAR = 15.119. Below this threshold, no statistical association was observed, while above the threshold, NPAR was positively associated with increased mortality risk. Subgroup analysis revealed a significant positive correlation between anemia in individuals under 60 years of age, with diabetes mellitus and a body mass index ≥ 30 kg/m², and all-cause mortality. Our findings indicate that elevated NPAR is independently associated with increased all-cause mortality in anemic individuals. The nonlinear relationship and identified breakpoint suggest that NPAR may serve as a valuable prognostic biomarker for mortality risk in this population. Further research should focus on the clinical utility of NPAR in guiding therapeutic interventions for anemia.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10238-026-02119-1.

Introduction

Anemia is characterized by a lower-than-normal hemoglobin concentration in the blood, impairing oxygen transport and resulting in various clinical symptoms and pathophysiological changes [1]. Anemia in children can affect cognitive and motor development, while in adults, it is associated with increased risks of infection, heart failure, and mortality [2–5]. According to the World Health Organization (WHO), anemia is defined as hemoglobin levels < 12.0 g/dL for women and < 13.0 g/dL for men [6]. Common symptoms in adults include fatigue, reduced activity levels, and shortness of breath [7, 8]. In severe cases, anemia can precipitate heart failure or death [9–14]. These findings underscore the need for improved strategies to prevent anemia, alleviate symptoms, and enhance patient outcomes.

Neutrophils, as first responders during inflammation, release mediators like LTB4, TNF-α, and IL-1β, which exacerbate vascular permeability and oxidative stress [15, 16]. Albumin, the most abundant plasma protein, provides antioxidant and immunomodulatory functions [17, 18]. Inflammatory responses inhibit albumin synthesis and distribution, compounding oxidative stress [19–21]. Elevated NPAR, reflecting increased neutrophil counts and/or decreased albumin levels, thus serves as a marker of heightened inflammatory states [19, 22]. Inflammation is intricately linked to anemia, as it disrupts iron homeostasis, inhibits red blood cell production, and shortens red cell lifespan [23, 24].

It is suggested that the percentage of neutrophils and the albumin ratio may serve as novel laboratory indicators for the identification of inflammation [25, 26]. Although traditional inflammatory markers such as C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) are widely utilized within clinical practice, the percentage of neutrophils and albumin values are more readily obtainable and less expensive to test [27]. Elevated NPAR has been linked to increased mortality in conditions such as chronic obstructive pulmonary disease (COPD), diabetes, and heart failure [28–30]. However, its prognostic value in anemia-related mortality remains unexplored. This study seeks to address this gap by investigating NPAR as a predictor of all-cause mortality in anemic individuals.

Materials and methods

Study population

This study included 101,316 individuals between 1999 and 2018. Among them, 46,235 individuals under the age of 20 years were excluded from the study, as were 6,299 individuals with missing data on neutrophil percentage and albumin values, and 44,004 individuals with missing hemoglobin data but who were not anemic. Subsequently, a total of 1520 individuals with missing data on covariates were excluded, including 640 individuals with no history of alcohol consumption, 1 individual with no history of hypertension, 44 individuals with a history of smoking, 372 individuals with a history of diabetes mellitus, 5 pregnant women, 7 individuals with missing education levels, 30 individuals with missing marital status, 127 individuals with missing body mass index, 333 individuals with missing household income and poverty rates, and 1 individual with missing data on death. A total of 3258 anemic patients were included in the final analysis (Fig. 1). This study analyzed NHANES data spanning 1999–2018, comprising 101,316 individuals. After excluding participants aged < 20 years, those with missing neutrophil or albumin values, and non-anemic individuals, 3,258 anemic patients were included in the final analysis. Details of the selection process are shown in Fig. 1.

Fig. 1.

Selection process of the participant

Exposure and outcomes

In this study, NPAR is the exposure variable. The NPAR formula is calculated by multiplying the neutrophil percentage by 100 and dividing it by the albumin value. The albumin value is obtained from the standard biochemical profile of laboratory data, and the neutrophil percentage is derived from the complete blood count and five-part differential white blood cell count. The study endpoint is defined as all-cause mortality, or mortality from any cause. For each participant, the follow-up duration was calculated from enrollment until either death or December 31, 2019, whichever came later. This date is the latest available in the NDI database (https://www.cdc.gov/nchs/datalinkage/mortality-public.htm). Mortality data were extracted from the 1999–2018 NHANES mortality files (https://www.cdc.gov/nchs/nhanes/index.htm). In this study, NPAR served as the exposure variable. The NPAR formula is calculated as the neutrophil percentage multiplied by 100 and then divided by the albumin value. Albumin values were obtained from the standard biochemical profile in laboratory data, while neutrophil percentages were derived from complete blood counts and five-part differential white blood cell counts. The study endpoint was defined as all-cause mortality, or mortality from any cause. Follow-up time was calculated for each participant from enrollment until death or December 31, 2019 (the latest data available in the NDI database). Mortality data were extracted from the 1999–2018 NHANES mortality files.

Covariates

Variables studied included age, gender (including both men and women), and race (including Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other). Marital status is categorized as married or cohabiting with a partner, and single (never married, divorced, separated, or widowed). Educational level is divided into three categories: less than high school, high school, and more than high school. Household income and poverty rate (PIR) were classified into three groups: [1] < 1.5; [2] 1.5–3.5; and [3] ≥ 3.5. Body mass index (BMI) was generally classified into the following three groups: [1] BMI < 25 kg/m²; [2] BMI 25–30 kg/m²; and [3] BMI ≥ 30 kg/m² [31]. The five categories of drinking status are as follows: [1] ex-drinkers, who used to drink but do not drink now; [2] never drinkers, who have not drunk any alcohol in their lifetime; [3] light drinkers, who consume one to two glasses of alcohol per day; [4] moderate drinkers, who consume no more than four glasses of alcohol per day for men and no more than three glasses of alcohol per day for women; and [5] heavy drinkers, who consume more than four glasses of alcohol per day for men and more than three glasses of alcohol per day for women. The smoking status of individuals is categorized into three distinct groupings: [1] never-smokers, defined as those who have smoked less than 100 cigarettes in their lifetime; [2] ex-smokers, characterized by having smoked more than 100 cigarettes in their lifetime, but are no longer active smokers; and [3] current smokers, comprising individuals who have smoked more than 100 cigarettes in their lifetime, and who currently smoke either regularly or daily [32]. Hypertension is diagnosed if any of the following conditions are met: Firstly, a physician diagnoses hypertension. Secondly, antihypertensive medication. Thirdly, the average systolic blood pressure must be ≥ 140 mmHg, and/or the average diastolic blood pressure must be ≥ 90 mmHg. The diagnosis of diabetes is made if any of the following conditions are met: [1] a doctor’s diagnosis of diabetes; [2] a glycated hemoglobin (HbA1c) level ≥ 6.5%; [3] a fasting blood glucose level ≥ 126 mg/dl; [4] current use of diabetes medication or insulin. The present study included covariates related to anemia, including laboratory parameters for white blood cells (WBC), red blood cells (RBC), platelet count (PLT), mean corpuscular hemoglobin concentration (MCHC), mean corpuscular hemoglobin (MCH), and mean corpuscular volume (MCV), Hemoglobin (HGB). These data are available from the CDC National Health Survey (https://www.cdc.gov/nchs/nhanes).

Modeling

The NHANES data have undergone a complex stratified, multi-stage sampling design; if it is not weighted, this may lead to biased results. Consequently, the implementation of weighting methodologies is imperative to enhance the generalisability of the study sample. The sampling weight is divided into two parts, and the calculation method is as follows: Specifically, the 4-year weight WTMEC4YR x 2/10 is utilized for the period 1999–2002, while the 2-year weight WTMEC2YR/10 is employed for the subsequent period from 2003–2018 [33]. Ultimately, the weights described above are amalgamated through the implementation of a meticulous calculation approach for NHANES composite survey cycle weights (https://wwwn.cdc.gov/nchs/nhanes/tutorials/Weighting.aspx), thereby yielding the aggregate weight.

Continuous variables that follow a normal distribution are represented by the mean and standard deviation; continuous variables with non-normal distributions are represented by the median (Q1, Q3). Categorical variables are represented by percentages. The participants’ baseline characteristics are represented by quartiles from NPAR. The Kaplan-Meier method was employed to ascertain the association between variations in NPAR levels and all-cause mortality in anemic patients. A multiple regression model was employed to evaluate the relationship between NPAR levels and all-cause mortality in anemic patients. To eliminate the potential impact of confounding factors, three models have been developed: Model 1 is not adjusted for covariates, and Model 2 is adjusted for the variables of gender, age, and race. The third model is adjusted for all potential confounding factors, including gender, age, race, education level, marital status, PIR, smoking history, drinking history, and medical history of hypertension, diabetes, white blood cells, platelets, and HGB. The present study employed restricted cubic spline analysis to explore the nonlinear relationship between NPAR and all-cause mortality in anemic patients. We assessed multicollinearity among the covariates and found that several hematological indicators related to red blood cells, such as RBC, MCV, MCH, MCHC, and HGB, exhibited VIF values over 10. This indicates high collinearity (see Supplementary Table 1). HGB is a key indicator for diagnosing and treating anemia and is irreplaceable clinically. Therefore, in Model 3, we excluded RBC, MCV, MCH, and MCHC while retaining HGB. After removing these four hematological indicators, HGB’s VIF value decreased from 22.96 to 1.31, and the VIF values of all covariates fell below 10. This suggests that the multicollinearity within the model has been resolved. Furthermore, we proceeded to utilize threshold effect analysis to ascertain the breakpoint of the nonlinear relationship between NPAR and all-cause mortality in anemic individuals based on Model 3. Stratified analysis and interaction tests were performed to explore whether there was a statistically significant association between NPAR and all-cause mortality in anemic individuals in different subgroups.

Statistical analysis

All statistical analyses were performed using the following software: Empower Stats 4.0, Free Statistics 1.9.2, and R Studio 4.4.2. A p-value of less than 0.05 was considered statistically significant.

Results

Baseline characteristics of participants

Among the 3,258 participants, 37.02% were male and 62.98% were female. The mean age was 56.199 ± 18.315 years, and the median NPAR was 14.526. The demographic composition of the study sample is shown in Table 1. Participants were grouped by NPAR quartiles: Q1 (below 12.543), Q2 (12.543–14.525), Q3 (14.526–16.424), and Q4 (above 16.424). Compared to individuals in the lowest NPAR quartile (Q1), those in the highest quartile (Q4) were more likely to exhibit the following characteristics: [1] male; [2] older age; [3] non-Hispanic white; [4] less than a high school education; [5] married or cohabiting with a partner; [6] moderate income; [7] higher body mass index; [8] history of diabetes; [9] history of hypertension; [10] prior smoking; [11] prior alcohol consumption; [12] higher white blood cell count; [13] higher platelet count; and [14] lower red blood cell count. No statistically significant differences in mean red blood cell hemoglobin concentration were observed between different NPAR groups.

Table 1.

Baseline characteristics of study subjects expressed as NPAR quartiles

	Q1(<12.543)	Q2(12.543–14.525)	Q3(14.526–16.424)	Q4(>16.424)	P- value
	N = 814	N = 815	N = 814	N = 815
Age(years)	54(39,69)	52(40,70)	58(42,75)	64(46,76.5)	< 0.001
WBC(1000 cells/uL)	5.65(4.6,6.9)	6.3(5.3,7.7)	6.7(5.6,8.1)	7.6(6.3,9.2)	< 0.001
RBC(million cells/uL)	4.09(3.75,4.44)	4.09(3.82,4.39)	4.095(3.8,4.44)	3.99(3.69,4.35)	< 0.001
MCV(fl.)	86(78.15,91.4)	85.6(77.9,91)	84.7(77.6,91.25)	86(80.05,92.1)	< 0.001
MCH (pg)	28.4(25.3,30.8)	28.5(25.4,30.8)	28.2(25.125,30.8)	28.7(26.05,31)	0.016
MCHC (g/dL)	32.9(32.1,33.8)	33.2(32.4,34)	33.1(32.2,33.9)	33.2(32.3,34)	0.475
PLT(1000 cells/uL)	246.5(198,304)	255(210,312)	253(210,315)	260(208,333)	< 0.001
Gender (%)					< 0.001
Male Female	31.267	23.968	31.517	36.015
	68.733	76.032	68.483	63.985
Race/ethnicity (%)					< 0.001
Mexican American Other Hispanic Non-Hispanic White Non-Hispanic Black Other Race	8.333	7.76	8.017	6.87
	5.014	7.475	6.09	4.417
	36.044	50.818	52.987	58.219
	42.701	27.359	26.547	23.984
	7.908	6.589	6.359	6.51
Education (%)					0.045
Less than high school High school More than high school	22.199	20.105	20.641	26.36
	22.909	24.59	23.743	23.921
	54.892	55.305	55.615	49.719
Marital status (%)					< 0.001
married or cohabiting with a partner Single	52.485	58.112	63.816	54.937
	47.515	41.888	36.184	45.063
BMI (%)					< 0.001
< 25 25–30 ≥ 30	39.857	31.928	32.092	26.002
	29.426	29.727	25.716	24.123
	30.717	38.345	42.192	49.876
PIR (%)					< 0.001
<1.5 1.5–3.5 ≥3.5	36.17	32.584	29.292	35.716
	34.404	32.153	35.115	38.752
	29.427	35.262	35.593	25.532
Smoking (%)					< 0.001
Never Former Now	62.579	63.632	65.844	55.568
	23.503	23.968	24.218	30.931
	13.918	12.399	9.939	13.5
Drinking (%)					< 0.001
Never Former Mild Moderate Heavy	20.354	19.495	15.771	15.903
	17.804	16.818	17.755	29.278
	34.149	31.507	36.931	32.459
	12.122	18.344	16.351	11.057
	15.571	13.836	13.193	11.302
Hypertension history(%)					< 0.001
Yes No	45.701	45.012	48.257	64.358
	54.299	54.988	51.743	35.642
Diabetes(%)					< 0.001
Yes Prediabetes No	16.352	20.478	26.601	33.523
	4.945	3.766	5.209	8.739
	78.704	75.756	68.19	57.737

Q1-Q4 are grouped according to NPAR quartiles. Continuous variables are expressed as mean ± standard deviation, and p-values were calculated using a weighted linear regression model. Percentages indicate categorical variables, and p-values were calculated using a weighted chi-square test. WBC: white blood cells; RBC: red blood cells; PLT: platelet count; MCHC: mean corpuscular hemoglobin concentration; MCH: mean corpuscular hemoglobin; MCV: mean corpuscular volume; PIR: household income-to-poverty ratio; BMI: body mass index

Association of NPAR with all-cause mortality in individuals with anemia

Because the endpoint of this study was death, a Kaplan-Meier analysis was performed to examine the effect of NPAR on all-cause mortality in anemic individuals over time. After a mean follow-up period of 94.337 months (range: 0–248 months), 904 out of 3,258 participants (27.75%) died. (See Fig. 2.) In the Kaplan-Meier analysis, which was stratified by NPAR quartiles and adjusted for all covariates, individuals in the highest NPAR quartile (Q4) exhibited the highest all-cause mortality during the follow-up period, compared to individuals in the other quartiles. Meanwhile, individuals in the lowest NPAR quartile (Q1) exhibited the lowest all-cause mortality, with significant differences in all-cause mortality rates observed among the four groups (p < 0.001).

Fig. 2.

Adjusted variables such as gender, race, age, education level, marital status, household income to poverty ratio, body mass index, hypertension, smoking, drinking, diabetes, WBC, HGB, PLT

In Model 1, the highest quartile (Q4) of NPAR was associated with a 2.460-fold increased risk of all-cause mortality (HR: 2.460; 95% CI: 2.004–3.020; p < 0.001) when the lowest quartile (Q1) of NPAR was used as the reference value(See Table 2). Model 2 incorporated demographic variables, including gender, age, and ethnicity. Compared to Q1, each additional unit of NPAR in Q4 individuals was associated with a 1.597-fold increased risk of all-cause mortality (HR: 1.597; 95% CI: 1.297–1.967; p < 0.001). Model 3 included all covariates: gender, race, age, education level, marital status, poverty ratio, body mass index, hypertension, smoking status, alcohol consumption, diabetes status, white blood cell count, platelet count, and hemoglobin. The findings indicated that with each additional unit of NPAR in Q4, there was a 1.518-fold increase in all-cause mortality risk (HR: 1.518; 95% CI: 1.253–1.840; p < 0.001).

Table 2.

Multiple regression model of NPAR and all-cause mortality in individuals with anemia

Exposure	Model 1	Model 2	Model 3
	HR (95% CI)P value	HR (95% CI)P value	HR (95% CI)P value
All-cause mortality
NPAR	1.136(1.107, 1.166) < 0.001	1.069(1.040, 1.098) < 0.001	1.051(1.025, 1.077) < 0.001
NPAR category
Q1 Q2 Q3 Q4	Reference	Reference	Reference 0.969(0.782, 1.200)0.77 0.985(0.803, 1.207)0.881 1.518(1.253, 1.840) < 0.001
	1.057(0.832, 1.345)0.649	1.033(0.815, 1.309)0.787
	1.255(0.981, 1.604)0.07	0.913( 0.729, 1.144)0.429
	2.460( 2.004, 3.020) < 0.001	1.597( 1.297, 1.967) < 0.001
Trend test	< 0.001	< 0.001	< 0.001

Model 1 was not adjusted for covariates. Model 2 was adjusted for gender, age, and race. Model 3 was adjusted for gender, race, age, education level, marital status, PIR, BMI, hypertension, smoking, drinking, diabetes, white blood cells, HGB and platelets

Compared to Model 1, the all-cause mortality risk for the highest quartile (Q4) of NPAR decreased from 2.460 to 1.597 after adjusting for variables in Model 2. This suggests that one of these variables is a risk factor for NPAR and that adjusting for it reduces both NPAR and mortality risk. However, after adjusting for all covariates in Model 3, the all-cause mortality risk in the highest quartile decreased from 1.597 to 1.518. Following the adjustment of all covariates in Model 3, the all-cause mortality rate in the highest quartile (Q4) of NPAR decreased from 1.597 to 1.518. The findings of this investigation suggest that specific variables present in laboratory tests may function as risk factors for NPAR. The adjustment for these variables has the effect of reducing NPAR levels and consequently lowering the risk of all-cause mortality. All three models yielded p-values < 0.001. Although the all-cause mortality risk in the highest quartile fluctuated across these models, the finding that higher NPAR correlates with increased all-cause mortality risk remained consistent. Our results confirm that NPAR is a highly reliable biological marker for predicting all-cause mortality in individuals with anemia.

As NPAR is a continuous variable, restricted cubic splines can model the nonlinear relationship between continuous variables (NPAR) and the outcome (all-cause mortality). This process facilitates the identification of inflection points. Restricted cubic splines have been demonstrated to plot risk curves according to the trajectory of the data, thus circumventing the fitting bias that is generated by linear models. Furthermore, restricted cubic splines demonstrate a high degree of sensitivity to the number and location of nodes, resulting in smoother curves that are more suitable for clinical indicators. Using restricted cubic splines within model 3, we determined the association between NPAR and all-cause mortality, revealing a nonlinear relationship in anemic individuals (see Fig. 3). Specifically, we found that NPAR was positively associated with all-cause mortality in anemic individuals to the right of the breakpoint. Subsequently, we conducted a threshold effect analysis on model 3 (Table 3). The results revealed a nonlinear relationship between NPAR and all-cause mortality, with a breakpoint at 15.119. To the left of the breakpoint, NPAR showed no statistically significant association with all-cause mortality in individuals with anemia. However, to the right of the breakpoint, NPAR was positively correlated with all-cause mortality in anemic individuals(HR: 1.137; 95%CI: 1.102–1.172; p < 0.001). The log-likelihood ratio test yielded p < 0.001.

Fig. 3.

The solid red line shows the cumulative risk of death from any cause for patients with anemia, and the two dashed blue lines indicate the 95% confidence interval

Table 3.

NPAR has a non-linear relationship with the all-cause mortality rate of anemic individuals

Fitting by the standard linear model	Adjusted HR (95% CI), p Value
Linear Effect	1.066(1.044,1.088) < 0.001
Fitting with a two-piecewise linear model
Breakpoint(K)	15.119
NPAR < 15.119	0.976(0.939,1.013)0.2
NPAR > 15.119	1.137(1.102,1.172) < 0.001
The effect difference between 2 and 1	1.165(1.100,1.234) < 0.001
Log likelihood ratio	< 0.001

The adjustment variables are the same as in Model 3

Subgroup analysis

Subgroup analysis identified significant associations between NPAR and mortality in participants aged < 60 years, those with diabetes and those with BMI ≥ 30 kg/m² (Fig. 4). Subgroups were stratified by age, sex, PIR, and BMI, smoking, alcohol consumption and hypertension, and diabetes. They were analyzed using Model 3. Age was categorized as < 60 or ≥ 60 years. Interaction analyses revealed that, after adjusting for all possible confounding variables, there were statistically significant differences in the association between NPAR and all-cause mortality in anemic individuals across age groups (less than 60 years, more than 60 years), diabetes status (diabetes, prediabetes, no diabetes), and BMI categories (less than 25 kg/m², 25–30 kg/m², more than 30 kg/m²) (P < 0.05). Specifically, the presence of anemia was found, in this cohort, to be significantly positively associated with all-cause mortality in individuals aged < 60 years, those with diabetes, and those with a BMI ≥ 30 kg/m². After the analysis, it was determined that no statistically significant interactions were observed across subgroups stratified by sex, household income relative to the poverty line, smoking status, alcohol consumption, or history of hypertension.

Fig. 4.

The adjustment variables are the same as in Model 3

Discussion

Our study investigated the association between the novel inflammatory marker NPAR and all-cause mortality in anemic individuals. The Results indicated that, after adjusting for confounding factors, higher NPAR levels were positively correlated with all-cause mortality in anemic individuals. We observed differing associations between NPAR and all-cause mortality in anemic individuals on either side of the cutoff point (NPAR = 15.119). When NPAR > 15.119, a positive correlation with all-cause mortality was found, whereas no statistically significant association was observed when NPAR < 15.119. We found that NPAR was significantly positively associated with all-cause mortality in anemic individuals, particularly in those younger than 60 years old, those with diabetes, and those with a body mass index (BMI) ≥ 30 kg/m². Our study identified a non-linear association between NPAR and all-cause mortality in anemia patients, with a breakpoint at 15.119. This suggests that NPAR could be used as a biomarker to predict mortality risk in anemia patients, potentially helping clinicians to develop personalized treatment strategies. However, the extrapolation and clinical applicability of this breakpoint require validation using external data. An NPAR value exceeding 15.119 indicates elevated neutrophil counts or low albumin levels, suggesting the patient is in a state of severe inflammation or malnutrition. This breakpoint helps clinicians initially screen high-risk anemia patients. When the NPAR value exceeds 15.119, a comprehensive patient evaluation may be necessary to rule out severe infection or cachexia and identify potential causative factors. This breakpoint helps clinicians initially screen high-risk anemia patients. When the NPAR value exceeds 15.119, a comprehensive patient evaluation may be necessary to rule out severe infection or cachexia and identify potential causative factors. From a clinical perspective, the neutrophil-to-albumin ratio test is simple and easy to perform, aiding in the rapid preliminary identification of high-risk anemia patients and thereby improving diagnostic efficiency. Routine NPAR monitoring from the time of hospitalization assists clinicians in the timely identification of patients with elevated NPAR (> 15.119). Reducing NPAR levels below 15.119 through treatment adjustments may indicate effective inflammation control and serve as a marker for a favourable prognosis. Conversely, if NPAR levels remain above 15.119 after treatment, it may signal a poor prognosis. Such anemic patients require enhanced monitoring and comprehensive clinical examinations to promptly determine if medical intervention or treatment adjustments are necessary. It is evident from extant reports that this phenomenon also occurs in other disease populations. When NPAR > 148.56, there is a significant increase in all-cause mortality in patients with hypertension-associated arthritis [34]. When NPAR > 14.8, the risk of all-cause mortality increases in patients with diabetic nephropathy [35]. Consequently, the breakpoint of 15.119 in this study’s NPAR is considered to possess both high generalizability and reliability.

According to existing research, anemia is associated with certain inflammatory markers. Elevated levels of the neutrophil-to-lymphocyte ratio (NLR), systemic inflammatory index (SII), and platelet-to-lymphocyte ratio (PLR) increase the likelihood of developing anemia [36–38]. According to existing research, anemia is associated with certain inflammatory markers. Elevated levels of the neutrophil-to-lymphocyte ratio (NLR), systemic inflammatory index (SII), and platelet-to-lymphocyte ratio (PLR) increase the likelihood of developing anemia [39]. Reduced albumin levels weaken the body’s antioxidant capacity, intensifying inflammatory responses and tissue damage [40]. The mechanisms by which inflammation causes anemia primarily involve the following points: when the body experiences inflammation, it enters a state of stress, which triggers the activation of immune cells and the release of inflammatory cytokines such as IL-6、IL-1、and TNF-α [41]. IL-6 activates the STAT3 signalling pathway in the liver, thereby upregulating ferritin expression [42]. Ferritin has been shown to bind to ferroportin (FP1), inducing its degradation and consequently resulting in iron being stored in macrophages. This process impedes the absorption of iron in the duodenum, thereby reducing the amount of iron available as a raw material for the production of red blood cells and leading to anemia [43, 44]. In addition, IL-1 and TNF have been demonstrated to inhibit the production of EPO by renal epithelial cells, thereby reducing red blood cell production and leading to the development of anemia [45]. It has been established that the occurrence of inflammation is accompanied by the deposition of autoantibodies and complement on red blood cells. This, in turn, results in enhanced phagocytosis of red blood cells by macrophages and increased red blood cell destruction. The consequence of these processes is the onset of anemia [41].

Compared to traditional inflammatory markers, the advantages offered by neutrophil percentage and albumin levels include convenient acquisition and lower cost. Based on existing research, these markers may be able to predict the prognosis of related diseases [46]. This study’s findings suggest that modulating the inflammatory response could improve the prognosis of patients with anemia [47, 48]. The indicators representing inflammation, such as SII and PLR, are nonlinearly related to anemia, which is consistent with the results of this study. Furthermore, NPAR is nonlinearly related to the all-cause mortality of individuals suffering from anemia [37, 38]. The results of this study suggest that elevated NPAR values in anemic patients may indicate a poorer prognosis. These results demonstrate the potential of NPAR to predict favourable outcomes in patients with anemia.

Our study demonstrated the predictive value of NPAR for anemia prognosis. Using Kaplan-Meier analysis, multivariable Cox proportional hazards models, restricted cubic splines, threshold effect analysis, stratified analysis, and interaction tests, we demonstrated a significant positive correlation between elevated NPAR levels and all-cause mortality. However, this study has limitations. The RCS curve may undergo subtle directional changes at nodes, and such minor alterations can cause shifts in the curve’s inflection points. Consequently, RCS predictions for inflection points exhibit a certain degree of deviation. At both ends of the RCS curve, due to sparse sample sizes, the 95% confidence intervals are wide, the variance is high, and uncertainty is significant.

The NHANES data are subject to complex stratified, multistage sampling designs and weighting procedures in order to ensure the representation of the U.S. population. However, it should be noted that dietary habits, ethnic variations, cultural customs, and healthcare environments differ significantly across countries worldwide. It is important to note that laboratory testing results may vary due to differences in instrumentation and analytical methods. It is important to note that these factors have the capacity to introduce measurement biases in albumin and neutrophil levels, which can result in fluctuations in NPAR concentrations.

In the present study, NPAR cannot be treated as a time-dependent covariate due to the current research conditions. Consequently, it is not possible to predict the all-cause mortality rate of anemic patients by establishing a time-dependent Cox model. This approach does not fully capture the dynamic relationship between NPAR changes over time and the all-cause mortality rate in anemic patients.

Despite utilizing the extensive NHANES database for analysis, the cross-sectional design of the study precludes the direct determination of a causal relationship between NPAR and all-cause mortality in anemic patients. Residual confounders not included in the adjustment, such as genetic factors, medication history, and other chronic diseases, may influence both NPAR levels and all-cause mortality in anemic patients. However, following the exclusion of 358 follow-up patients from the initial two years, a multivariable Cox proportional hazards model was utilized for the re-analysis of the data (see Supplementary Table 2). The findings indicated that, in model 3, in comparison with individuals in quartile Q1, those in quartile Q4 demonstrated a 1.401-fold increase in all-cause mortality (p = 0.003). This finding indicates that, after ruling out reverse causality through sensitivity analysis, the association between NPAR and all-cause mortality remains consistent among anemic patients. If NPAR is validated in additional cohorts, it could be a useful biomarker for clinicians to use in identifying risk stratification among anemic patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Peng Anqi, Li Jizhe(both authors contributed equally to this study); Methodology: Peng Anqi, Li Jizhe; Software: Peng Anqi, Li Jizhe; Validation: Peng Anqi, Li Jizhe; Formal analysis: Peng Anqi, Li Jizhe; Visualization: Peng Anqi, Li Jizhe; Investigation: Huang Jun, Yang Yuqiong; Resources: Huang Jun, Yang Yuqiong; Data curation: Huang Jun, Yang Yuqiong; Writing – original draft: Peng Anqi, Li Jizhe, Huang Jun, Yang Yuqiong; Project administration: Jiang Yizhi, Huang Dongping; Supervision: Jiang Yizhi, Huang Dongping; Funding acquisition: Jiang Yizhi, Huang Dongping; Writing – review & editing: Jiang Yizhi, Huang Dongping.

Funding

This research was funded by the Natural Science Foundation of Anhui Province (2023AH040255) and the National Natural Science Foundation of China (82200146).

Data availability

The datasets supporting the conclusions of this article are available in the NHANES (https://www.cdc.gov/nchs/nhanes).

Declarations

Consent for publication

Not an application.

Competing interests

The authors declare no competing interests.

Ethics

The authors confirm that all experiments were performed in accordance with the Declaration of Helsinki.

Ethical approval statement

This study is an analysis derived from publicly available NHANES data, which was acquired with the approval of the Ethics Review Board of the National Center for Health Statistics (https://www.cdc.gov/nchs/nhanes/irba98.htm). Given that the NHANES public use data does not contain individually identifiable information, ethical review and approval were waived for this study.