Blood cell ratio flags hidden cardiopulmonary risk in high-altitude polycythemia

Patients living at high altitude can develop dangerously high hemoglobin levels, yet the severity of their symptoms often bears little relationship to how elevated their blood counts actually are. A new cross-sectional study of 198 patients with high altitude polycythemia (HAPC) suggests that the immune system may hold the key to understanding why some of these patients become severely breathless while others with identical hemoglobin levels remain relatively comfortable. The research, published in Frontiers in Physiology, identified four distinct immune-based subgroups among HAPC patients and found that a simple blood ratio could identify the subgroup carrying the heaviest cardiopulmonary symptom burden.

The study, conducted at West China Hospital in Chengdu, enrolled adult patients admitted between September 2014 and September 2025 who met strict diagnostic criteria for HAPC: hemoglobin of at least 21 grams per deciliter for men or 19 grams per deciliter for women, permanent residence above 2,500 meters, and a first-time hospitalization. Patients with polycythemia vera, acute infections, established thrombotic disease, or pre-existing respiratory or cardiovascular conditions were excluded, a design choice intended to isolate the immunological features of HAPC itself rather than those of unrelated diseases.

Rather than relying on any single blood measurement, the researchers applied an unsupervised machine learning approach. They fed standardized absolute counts of five peripheral white blood cell subsets—neutrophils, lymphocytes, monocytes, eosinophils, and basophils—into a K-means clustering algorithm. The optimal number of groups was determined using the elbow method and silhouette analysis, and the resulting four-cluster structure was cross-checked with hierarchical clustering and tested for stability through bootstrap resampling, which produced Jaccard indices between 0.718 and 0.807, indicating reasonably stable assignments.

A striking feature of the resulting classification is what it did not depend on. Baseline demographics, county-level estimated residential altitude, and hemoglobin levels were statistically comparable across all four clusters. The overall cohort had a median hemoglobin of 219 grams per liter, a median age of 47.25 years, and a median body mass index of 26.96 kilograms per square meter, with the vast majority of patients being men of Tibetan ethnicity. None of these variables distinguished one immune phenotype from another, suggesting that the clustering captured biological variation that conventional clinical measures miss entirely.

One cluster in particular stood out. Cluster 2, comprising 62 patients, showed the highest absolute lymphocyte count and consequently the lowest monocyte-to-lymphocyte ratio (MLR) of any group. Despite having quiescent systemic inflammatory markers—including the lowest D-dimer values in the cohort—these patients reported the highest incidence of breathlessness and palpitations, at 32.3 percent. By contrast, Cluster 1, which displayed the highest neutrophil and monocyte counts along with peak levels of C-reactive protein and D-dimer, reported no breathlessness or palpitations at all. Cluster 3 appeared as a relatively quiescent baseline phenotype, while Cluster 4 showed elevated eosinophils and basophils along with the highest overall symptom scores and headache incidence.

To test whether the MLR could independently predict membership in the highly symptomatic cardiopulmonary phenotype, the researchers used a pre-specified Firth penalized logistic regression model, a technique chosen to handle the limited number of outcome events. After adjusting for age, body mass index, residential altitude, baseline hemoglobin, smoking status, and alcohol use, patients in the lowest MLR quartile had an odds ratio of 33.764 for belonging to Cluster 2 compared with those in the highest quartile, with a 95 percent confidence interval of 9.615 to 180.993. A dose-response gradient was evident across the quartiles, and neither absolute altitude nor hemoglobin level emerged as a significant independent predictor in the model.

The discriminatory capacity of the ratio was assessed with receiver operating characteristic analysis. The initial area under the curve was 0.829, and after internal validation with 1,000 bootstrap resamples the optimism-corrected value settled at 0.810, which the authors describe as favorable diagnostic discrimination for identifying this specific phenotype. Model calibration was supported by a non-significant Hosmer-Lemeshow test. The practical appeal of the MLR lies in its accessibility: it can be derived inexpensively from a routine complete blood count, which matters in resource-limited primary care settings at high altitude where advanced diagnostics such as right heart catheterization or cytokine profiling are largely unavailable.

The authors propose a mechanistic interpretation for the paradoxically low MLR in the symptomatic cluster. They hypothesize that the low circulating monocyte fraction does not reflect true immune depletion but rather an apparent shift caused by margination and transendothelial migration of monocytes into the pulmonary microvasculature. If this sequestration occurs, the subsequent differentiation of these cells into macrophages within the lung could theoretically promote microvascular remodeling, ventilation-perfusion mismatch, and right ventricular strain—mechanisms documented in acute lung injury. However, the authors are explicit that this target-organ sequestration hypothesis remains untested in HAPC and requires objective validation through imaging, hemodynamic assessment, or tissue studies.

The limitations of the study are substantial and acknowledged in detail. The cross-sectional, retrospective design precludes any causal inference between the MLR, leukocyte sequestration, and cardiopulmonary symptoms, and because corrections for multiple testing were not applied, the between-group biomarker comparisons should be read as exploratory. The definition of the severe cardiopulmonary phenotype rested entirely on subjective symptom reporting extracted from electronic medical records; the modified five-item symptom score used in the study lacks formal validation against the standard Qinghai CMS score, and the symptoms cannot be equated with structural end-organ damage. The relatively small sample size per cluster means the four immunophenotypes represent a conceptual framework rather than an established classification, and the eleven-year study window introduces potential temporal biases that the z-score transformations could only partially mitigate. County-level average altitude also serves only as a rough proxy for individual exposure, and the authors caution that their findings should not be interpreted as evidence that individual altitude exposure is unrelated to clinical heterogeneity.

Within these constraints, the study offers a reframing of how high altitude polycythemia might be understood. Rather than a single condition driven by a straightforward dose-response to hypoxemia, HAPC emerges as a heterogeneous syndrome in which distinct immune trajectories—hyperinflammatory, compensated, allergic-like, and cardiopulmonary-symptomatic—unfold independently of the degree of erythrocytosis. If future prospective, multi-center cohorts with objective cardiopulmonary testing confirm these patterns, a routinely available blood ratio could become a practical screening tool for flagging which patients with high hemoglobin at altitude deserve closer cardiopulmonary evaluation, while the four-phenotype framework may guide more targeted research into the mechanisms that separate symptom severity from blood count in chronic mountain sickness.

Subject of Research: Immune phenotyping and the monocyte-to-lymphocyte ratio as a predictor of cardiopulmonary symptoms in high altitude polycythemia

Article Title: Independent predictive value of the monocyte-to-lymphocyte ratio for severe cardiopulmonary symptoms in high altitude polycythemia

Article References: Dai, L., Wu, W., Chen, T., Sun, W., Zhang, Q., Su, Q., Ni, Y., & Zhang, X. (2026). Independent predictive value of the monocyte-to-lymphocyte ratio for severe cardiopulmonary symptoms in high altitude polycythemia. Frontiers in Physiology, 17, Article 1986276. https://doi.org/10.3389/fphys.2026.1986276

Image Credits: AI Generated

DOI: 10.3389/fphys.2026.1986276

Keywords: high altitude polycythemia, chronic mountain sickness, monocyte-to-lymphocyte ratio, immunophenotyping, K-means clustering, cardiopulmonary symptoms, hypoxia, biomarker, cross-sectional study, Firth penalized regression, leukocyte subsets, risk stratification