The inflammation–nutrition axis and survival outcomes: Role of the neutrophil percentage-to-albumin ratio in patients with sleep disorders

Abstract

Background

The neutrophil percentage-to-albumin ratio has shown prognostic significance across several chronic diseases; however, its relevance within the context of sleep disorders has not been investigated.

Methods

This study utilized data from the National Health and Nutrition Examination Survey for the years 2005–2014 to explore the connection between the neutrophil percentage-to-albumin ratio and all-cause mortality in people with sleep disorders. Mortality risk was assessed using Cox regression, and the analyses included restricted cubic spline, Kaplan–Meier survival, subgroup, and time-dependent receiver operating characteristic curve analyses.

Results

A positive correlation was identified between the neutrophil percentage-to-albumin ratio and mortality risk among individuals with sleep disorders (hazard ratio = 1.09, 95% confidence interval: 1.05, 1.14). A comparison between the highest (Q4) and lowest (Q1) quartiles revealed that patients in Q4 had an 86% higher mortality risk (hazard ratio = 1.86, 95% confidence interval: 1.31, 2.65). Subgroup analysis further confirmed consistency across different demographic and clinical strata (all p for interaction > 0.05). Furthermore, the neutrophil percentage-to-albumin ratio demonstrated strong predictive performance for early mortality risk in patients with sleep disorders (1-year area under the curve = 0.751, 95% confidence interval: 0.693, 0.808).

Conclusions

Among patients diagnosed with sleep disorders, higher levels of NPAR are independently linked to an elevated risk of all-cause mortality.

Graphical abstract.

In U.S. adults with sleep disorders, a higher neutrophil percentage-to-albumin ratio (NPAR) is significantly associated with increased all-cause mortality. This suggests NPAR could serve as a practical and accessible biomarker for identifying high-risk individuals in this population.

Introduction

Sleep is a fundamental physiological necessity for humans, critical for maintaining diverse biological and cognitive functions. ¹ Sleep disorders, which include conditions such as insomnia, hypersomnia, sleep-related breathing disorders, and restless legs syndrome, disrupt normal sleep patterns and impair sleep quality, duration, and continuity.^2,3 These disorders can result in cognitive impairment, mood disturbances, and behavioral abnormalities while also compromising immune function, thereby influencing the progression and outcomes of various diseases, including cardiovascular diseases, cognitive decline, and cancer development.^4–7 As a global health issue, sleep disorders affect populations not only in high-income and developed countries but also in low-income regions and across different racial groups.^8,9 In the United States, approximately 30% of adults experience sleep disorders, contributing to significant health burdens, including an estimated 38,000 cardiovascular deaths annually and economic losses surpassing US$66 billion annually, considering both direct and indirect costs.^10,11 Although evidence linking sleep disorders to all-cause mortality continues to accumulate, the nature of this relationship remains debated. Previous studies have shown a U-shaped pattern between sleep duration and long-term mortality, where both too little and too much sleep are associated with an increased mortality risk. ¹² Furthermore, insomnia, difficulty initiating sleep, and early morning awakenings have been identified as risk factors for increased all-cause mortality. ¹³ Therefore, investigating potential mortality risk factors in patients with sleep disorders is critical.

The neutrophil percentage-to-albumin ratio (NPAR) is an emerging inflammatory marker that is based on the proportion of neutrophils and albumin levels, offering a cost-effective, straightforward, and easily accessible measurement. ¹⁴ Distinct from other inflammatory markers, the NPAR uniquely integrates the assessment of inflammatory status and nutritional condition, potentially offering a comprehensive indicator of overall health. Neutrophils, as key mediators of the innate immune response, reflect systemic inflammation, while albumin, a major plasma protein, serves as a marker of nutritional status and synthetic liver function.^15,16 By combining these two components, the NPAR captures both the inflammatory burden and nutritional–metabolic state, making it a versatile biomarker for evaluating disease severity and prognosis. ¹⁷ Although extensive research has established a strong link between NPAR and the prognosis of chronic diseases,^18–22 including cancer, ²³ its potential impact on mortality in individuals with sleep disorders remains insufficiently explored.

The present study aimed to determine the association of the NPAR with all-cause mortality in a cohort of individuals with sleep disorders, based on data from the National Health and Nutrition Examination Survey (NHANES), 2005–2014.

Methods

Study population

The NHANES utilizes a sophisticated sampling framework to guarantee national representativeness among various demographic and regional groups in the U.S. The standardized examination protocol integrates comprehensive physical measurements, detailed health and nutritional assessments, and systematic biological sample collection. This robust, multidimensional data collection system has been extensively utilized in public health research, with complete methodological details having been previously published in established scientific literature. ²⁴

This study analyzed data from five consecutive 2-year cycles (2005–2014) of the NHANES, comprising a total of 50,965 participants. Based on predefined inclusion and exclusion criteria, 2395 participants with a diagnosis of sleep disorders were initially identified. Exclusions were applied to 80 participants aged <20 years, 206 participants with missing NPAR data, and 393 participants with incomplete or outlier data for other covariates. Following the application of exclusion criteria, a total of 1716 participants remained eligible for the final analysis, as outlined in the selection flowchart (Figure 1).

Figure 1.

Flowchart depicting participant selection.

Outcome variables

We identified participants with a sleep disorder based on the NHANES variable SLQ050, which documents a history of diagnosis by a physician or other healthcare professional. This method has been previously validated and demonstrates high reliability in capturing sleep disorder history. ²⁵ Mortality status was assessed using the NHANES Linked Mortality File, updated until 31 December 2019 and verified using the National Death Index.

Exposure variables

Blood specimens were obtained from all participants via venipuncture. Following standardized processing, these samples were stored at −20°C and subsequently shipped to a certified laboratory for analysis. Neutrophil percentage data were obtained from the NHANES complete blood count (CBC) dataset, while albumin levels were sourced from the biochemistry profile. Finally, the NPAR was calculated using the formula: NPAR = (neutrophil percentage (%) × 100) / albumin (g/dL). ²⁶

Definitions of covariates

Consistent with prior research, potential confounders associated with sleep disorders were identified and adjusted for in the analysis. Factors considered included sex, age, race, marital status, educational level, poverty-to-income ratio (PIR), body mass index (BMI), smoking status and alcohol consumption, and a history of hypertension and diabetes. Additional covariates included serum alanine aminotransferase (ALT) and serum creatinine (SCr) levels. Information on smoking status, alcohol consumption, and history of hypertension and diabetes was obtained from on self-reported responses to their corresponding standardized questionnaires (SMQ020, ALQ101, B0, and DIQ010).

Statistical analyses

Participant characteristics were analyzed according to demographic features based on the NPAR quartiles. Appropriate statistical adjustments were made using the NHANES-provided sample weights (WTMEC2YR), stratification variables (SDMVSTRA), and clustering variables (SDMVPSU). Mean ± SD values were used to represent continuous variables, while categorical variables were presented as frequencies (percentages). Statistical significance of between-group differences was determined using weighted rank-sum tests for continuous variables and weighted chi-square tests for categorical variables.

The statistical approach involved several methods to evaluate the link between the NPAR and all-cause mortality. For the primary analysis, Cox regression models were developed, treating the NPAR as both continuous and quartile-based categorical variable (Q1 reference). This was performed across three adjustment levels: (a) crude (Model 1); (b) partially adjusted for demographics (Model 2); and (c) fully adjusted (Model 3), yielding hazard ratios (HRs) and 95% confidence intervals (CIs). To explore the dose–response relationship, restricted cubic spline (RCS) analysis was conducted. For survival visualization, Kaplan–Meier curves were generated. Finally, to assess the robustness of the association, we performed stratified analyses across a comprehensive list of subgroups, including those defined by sex, age, race, marital status, educational level, PIR, BMI, smoking status, alcohol consumption, and histories of diabetes and hypertension. Additionally, time-dependent ROC analysis was performed, and corresponding area under the curve (AUC) curves were generated over time to assess the NPAR's prognostic capability.

R Studio software was utilized for all statistical procedures, and statistical significance was defined as a two-sided p-value <0.05.

Results

Baseline information

The analytical cohort included 1716 participants, which, after weighting, represented an estimated 14,642,888 U.S. adults with sleep disorders. The mean age of the cohort was 51.0 ± 0.37 years; 53.2% were men, and 76.2% were of non-Hispanic White ethnicity. The overall all-cause mortality rate for the entire cohort was 14.1%. As shown in Table 1, which stratifies baseline data by NPAR quartiles, significant differences were observed between the highest (Q4) and lowest (Q1) quartiles. Specifically, participants in the Q4 group were older and more likely to be of non-Hispanic White ethnicity; furthermore, a higher proportion of them were overweight. This group also exhibited higher rates of smoking, alcohol consumption, diabetes history, and hypertension history, and consequently experienced significantly greater all-cause mortality compared with the Q1 group.

Table 1.

Baseline characteristics of participants, stratified by NPAR quartiles.

Characteristics	Overall, N = 14,642,888 Overall, n = 1716	Q1 n = 429	Q2 n = 429	Q3 n = 429	Q4 n = 429	p-value
Sex						0.183
Male	53.2%	56.3%	55.6%	53.2%	47.4%
Female	46.8%	43.7%	44.4%	46.8%	52.6%
Age group (years)						<0.001
<60	71.4%	79.8%	74.1%	68.9%	63.0%
≥60	28.6%	20.2%	25.9%	31.1%	37.0%
Ethnicity						<0.001
Non-Hispanic White	76.2%	69.5%	78.0%	79.1%	77.5%
Non-Hispanic Black	9.7%	15.2%	8.3%	8.0%	7.8%
Others	14.1%	15.3%	13.7%	12.9%	14.7%
Marital status						0.228
Married	58.6%	60.1%	55.7%	62.4%	56.0%
Unmarried and others	41.4%	39.9%	44.3%	37.6%	44.0%
Educational level						0.097
Less than high school	14.1%	14.1%	13.4%	13.8%	15.2%
High school or GED	24.2%	18.6%	24.6%	23.6%	29.9%
Above high school	61.7%	67.2%	62.0%	62.6%	54.9%
PIR						0.222
<1.5	29.0%	32.2%	25.5%	25.7%	33.8%
1.5–3.5	29.1%	29.0%	30.6%	28.6%	28.2%
>3.5	41.8%	38.8%	43.9%	45.7%	38.0%
BMI						<0.001
<25	16.1%	22.8%	18.0%	13.9%	9.7%
25–30	24.8%	30.3%	27.0%	21.0%	21.0%
>30	59.2%	46.9%	55.0%	65.1%	69.3%
Smoking status						0.010
Yes	56.7%	53.0%	54.9%	53.3%	66.2%
No	43.3%	47.0%	45.1%	46.7%	33.8%
Alcohol consumption						0.019
Yes	76.8%	79.8%	81.2%	72.8%	73.3%
No	23.2%	20.2%	18.8%	27.2%	26.7%
History of diabetes						<0.001
Yes	20.5%	13.5%	16.1%	24.8%	27.8%
No	79.5%	86.5%	83.9%	75.2%	72.2%
History of hypertension						0.003
Yes	51.8%	44.0%	50.3%	51.9%	61.3%
No	48.2%	56.0%	49.7%	48.1%	38.7%
ALT	27.9 ± 33.93	29.6 ± 17.55	27.4 ± 16.28	26.7 ± 18.49	28.0 ± 62.52	0.001
SCr	83.8 ± 51.23	80.7 ± 19.92	83.0 ± 35.38	81.5 ± 34.53	90.2 ± 89.26	0.571
All-cause mortality						<0.001
Yes	14.1%	7.7%	9.2%	14.7%	25.4%
No	85.9%	92.3%	90.8%	85.3%	74.6%

Continuous variables are presented as mean ± SD, and categorical variables are presented as percentages.

PIR: poverty index ratio; BMI: body mass index; ALT: serum alanine aminotransferase; SCr: serum creatinine; NPAR: neutrophil percentage-to-albumin ratio; Q: quartile; GED: General Educational Development.

Associations of the NPAR with all-cause mortality

During a mean follow-up duration of 107 months, 295 all-cause deaths were observed. The Cox regression analysis, detailed in Table 2, consistently demonstrated that higher NPARs were associated with increased all-cause mortality. This positive relationship held true whether the NPAR was treated as a continuous variable (fully adjusted HR = 1.09, 95% CI: 1.05, 1.14) or as a categorical variable. In the categorical analysis, participants in the highest NPAR quartile (Q4) had a significantly elevated risk of death than those in the lowest quartile (Q1) (HR = 1.86, 95% CI: 1.31, 2.65). A strong dose–response relationship was also evident, with a p for trend across quartiles of < 0.001. The unadjusted (Model 1) and partially adjusted (Model 2) models for continuous NPAR yielded similar positive findings (HR = 1.15 and HR = 1.12, respectively).

Table 2.

Association between the NPAR and all-cause mortality.

	Model 1			Model 2			Model 3
Characteristic	HR	95% CI	p-value	HR	95% CI	p-value	HR	95% CI	p-value
NPAR (continuous)	1.15	1.11, 1.20	<0.001	1.12	1.08, 1.17	<0.001	1.09	1.05, 1.14	<0.001
NPAR
Q1	Ref	Ref		Ref	Ref		Ref	Ref
Q2	1.14	0.77, 1.68	0.516	1.07	0.72, 1.59	0.743	1.03	0.69, 1.53	0.888
Q3	1.55	1.07, 2.23	0.019	1.24	0.85, 1.79	0.265	1.17	0.80, 1.71	0.408
Q4	2.89	2.07, 4.04	<0.001	2.28	1.62, 3.21	<0.001	1.86	1.31, 2.65	<0.001
p for trend			<0.001			<0.001			<0.001

Model 1: Unadjusted

Model 2: Adjusted for sex, age, and race

Model 3: Adjusted for sex, age, race, marital status, educational level, PIR, BMI, smoking status, alcohol consumption, history of diabetes, history of hypertension, ALT, and SCr

NPAR: neutrophil percentage-to-albumin ratio; Q: quartile; HR: hazard ratio; CI: confidence interval; PIR: poverty-to-income ratio; BMI: body mass index; ALT: alanine aminotransferase; SCr: serum creatinine

Further analyses provided additional support for these findings. First, the adjusted RCS analysis revealed a significant and positive linear association between the NPAR and all-cause mortality, with the test for non-linearity being non-significant (p-nonlinear = 0.122; Figure 2). Second, Kaplan–Meier survival curves indicated that participants in the highest NPAR quartile (Q4) had substantially lower long-term survival rates than those in the lowest quartile (Q1) (log-rank p < 0.001; Figure 3). Finally, subgroup analyses confirmed the consistency of this association across various demographic and clinical strata as no significant effect modification was detected (all p for interaction > 0.05; Figure 4). The predictive performance of NPAR for mortality over time was evaluated by performing time-dependent ROC analysis (Figure 5). The resulting AUC values were 0.751 (95% CI: 0.693, 0.808) at 1 year, 0.703 (95% CI: 0.656, 0.750) at 3 years, and 0.630 (95% CI: 0.586, 0.673) at 5 years.

Figure 2.

RCS analysis of the association between the NPAR and all-cause mortality.

RCS: restricted cubic spline; NPAR: neutrophil percentage-to-albumin ratio.

Figure 3.

Kaplan–Meier survival curves for all-cause mortality according to NPAR quartiles.

NPAR: neutrophil percentage-to-albumin ratio.

Figure 4.

Subgroup analysis of the association of the NPAR with all-cause mortality.

NPAR: neutrophil percentage-to-albumin ratio.

Figure 5.

(a) Time-dependent ROC curves and (b) time-dependent AUC values of the NPAR for predicting all-cause mortality.

NPAR: neutrophil percentage-to-albumin ratio; ROC: receiver operating characteristic; AUC: area under the curve.

Discussion

Among the 1716 patients with sleep disorders analyzed, the NPAR demonstrated a clear positive linear relationship with all-cause mortality. Survival analysis demonstrated that the highest NPAR group had significantly lower survival rates than the other groups. Subgroup analysis further confirmed that stratification variables did not significantly modify this association, supporting its robustness. Additionally, time-dependent ROC analysis indicated strong predictive performance of the NPAR for short-term mortality risk. These results underscore the potential of the NPAR as an accessible and practical biomarker for prognostic evaluation in sleep disorder patients, offering new perspectives on survival risk assessment in this group.

Previous studies have established that patients with sleep disorders often present with chronic low-grade inflammation. In an analysis of the NHANES data from 22,599 participants, You et al. demonstrated that sleep disorders were associated with elevated levels of blood cell inflammatory biomarkers, including neutrophils (odds ratio (OR) = 1.032, 95% CI: 1.005, 1.059). ²⁷ Kadier et al. assessed various inflammatory markers in relation to sleep-related disorders and found that the systemic immune–inflammation index (SII) exhibited the most robust correlation. ²⁸ In the study conducted by Piber ', a significant association was identified between self-reported sleep disorders and the activation of inflammatory pathways. ²⁹ In contrast, our study employed sleep disorder diagnoses assessed by physicians or other healthcare professionals rather than relying on self-reported data. This professionally evaluated diagnostic approach improves the objectivity and reliability of the findings, enabling a more accurate investigation of the relationship between the NPAR and mortality in patients with sleep disorders. Additionally, sleep disorders may lead to malnutrition or metabolic abnormalities, resulting in reduced albumin levels. ³⁰ In their study involving 9973 U.S. adults, Li and Guo demonstrated that shorter sleep duration is significantly associated with lower albumin levels (β = −1.00, 95% CI: −1.26, −0.74). ³¹

The link between elevated NPAR and mortality risk has been documented across numerous populations. To illustrate, research involving peritoneal dialysis patients showed a 51% higher risk of all-cause mortality for every 1-unit increase in the NPAR (HR = 1.51, 95% CI: 1.14, 1.98), demonstrating its strong predictive capacity. ³² A Shanghai-based study involving 1141 older atrial fibrillation patients aged ≥80 years found that NPAR levels were significantly correlated with 28-day all-cause mortality, demonstrating excellent predictive power for all-cause mortality (AUC = 0.81, 95% CI: 0.77, 0.85). ³³ In our study, the NPAR also exhibited strong predictive performance for short-term mortality risk in patients with sleep disorders.

The link between the NPAR and mortality risk in sleep disorder patients may be driven by interconnected mechanisms involving both immune hyperactivation and nutritional–metabolic impairment. First, sleep disorders profoundly disrupt the hypothalamic–pituitary–adrenal (HPA) axis and induce sympathetic nervous system (SNS) overactivation, resulting in the elevated production of proinflammatory cytokines. ³⁴ This neuroendocrine–immune shift triggers Toll-like receptors (TLRs) and chemokine receptors on neutrophils, boosting their migration, tissue infiltration, and degranulation functions. Recent experimental and clinical cohort studies have confirmed that sleep deprivation directly activates the sympathetic–immune axis, leading to enhanced neutrophil trafficking and altered innate immune responses.^35,36 Once activated, these neutrophils release massive amounts of reactive oxygen species (ROS), myeloperoxidase (MPO), and neutrophil extracellular traps (NETs). These factors directly damage vascular endothelial cells, accelerate atherosclerotic plaque formation, and promote coagulation abnormalities, thereby substantially increasing the risk of fatal cardiovascular events. ³⁷

Conversely, chronic sleep fragmentation and the resulting systemic inflammation significantly impair metabolic homeostasis, directly contributing to the decrease in serum albumin levels. Although sleep disturbances are known to alter feeding behaviors and induce malnutrition, the chronic low-grade inflammation typical of sleep disorders—characterized by sustained elevations in the levels of cytokines such as interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α)—acts as a primary driver of hypoalbuminemia. Accumulating evidence has indicated that these proinflammatory cytokines directly suppress hepatic albumin synthesis as part of an acute-phase response shift.^38,39 Decreased albumin levels, a key molecule for maintaining colloid osmotic pressure, can subsequently lead to tissue edema, impaired organ perfusion, and induction of acute kidney injury or heart failure, thereby elevating mortality risk. ⁴⁰ In the specific pathological context of sleep disorders, hypoxia, circadian rhythm disruption, and autonomic dysfunction create a vicious cycle that further amplifies the pathological effects of the NPAR. The findings of our study not only deepen our understanding of the mechanisms linking sleep disorders to mortality but also provide a theoretical foundation for targeted interventions focusing on the inflammation–nutrition axis.

This study has several strengths. The chief strength is the analysis of a large, nationally representative sample from the NHANES dataset, with rigorous adjustments for its complex survey design and weighting. Second, the study rigorously adjusted for multiple potential confounders, ensuring the robustness of the findings. Furthermore, subgroup analyses across various groups supported the reliability of the results. However, it is important to recognize certain study limitations. The median follow-up period was <10 years; a longer follow-up duration may offer more reliable conclusions. Certain other limitations should also be acknowledged. The calculation of the NPAR based on measurements made at a single baseline time point does not account for its potential fluctuations over time. The absence of data on important potential confounders, such as medication use, may have introduced residual bias and affected the true magnitude of the NPAR–mortality association. Finally, because the analysis was confined to U.S. adults, the applicability of the findings to populations in other countries with differing genetic, dietary, and lifestyle characteristics may be limited. Future studies employing more recent datasets and a broader range of intervention indicators are needed to further refine these findings.

Conclusion

This study offers robust evidence of a positive link between the NPAR and all-cause mortality risk in U.S. individuals with sleep disorders. As an affordable, efficient, and clinically viable biomarker, the NPAR could assist in identifying high-risk individuals and shaping tailored intervention approaches. Additional prospective research is required to confirm these results.