Simple Blood Test Index Predicts One-Year Death Risk After Pulmonary Embolism

A routine blood draw performed on the day a patient arrives at the hospital with a pulmonary embolism may reveal who is most likely to die within the coming year, according to new research published in Frontiers in Cardiovascular Medicine. The study, conducted at Samsun City Hospital in Türkiye, followed 471 consecutive patients diagnosed with acute pulmonary embolism between July 2021 and December 2024 and found that a composite score known as the CALLY index—calculated from C-reactive protein, albumin, and lymphocyte count—outperformed both the Prognostic Nutritional Index and the Systemic Immune-Inflammation Index in predicting one-year all-cause mortality. The work adds to a growing body of evidence suggesting that the biology of a blood clot in the lungs cannot be understood through thrombosis alone, but must also encompass inflammation, immune competence, and the body’s underlying nutritional reserve.

Acute pulmonary embolism remains one of the most lethal cardiovascular emergencies, occurring when a thrombus, usually dislodged from a deep vein in the legs, travels through the venous system and lodges in the pulmonary arteries. Despite major advances in computed tomography pulmonary angiography, anticoagulation, and thrombolytic therapy, mortality remains substantial, and in this cohort it was striking: 102 of the 471 patients—21.7 percent—died within a single year of diagnosis. Conventional risk stratification relies on clinical scoring systems such as the Pulmonary Embolism Severity Index and its simplified variant, the sPESI, which integrate age, vital signs, comorbidities, and hemodynamic status. Yet these scores were never designed to capture the inflammatory and immunological processes that increasingly appear to shape long-term outcomes after the acute event has been treated.

The CALLY index attempts to do precisely that. It is calculated as the product of serum albumin in grams per deciliter and the absolute lymphocyte count in thousands per microliter, divided by the C-reactive protein concentration in milligrams per deciliter. Each component carries distinct biological information. C-reactive protein, synthesized by hepatocytes under the influence of the proinflammatory cytokines interleukin-1 and interleukin-6, is one of the most sensitive and standardized markers of the acute-phase response, reflecting endothelial dysfunction and heightened thrombotic tendency. Lymphocytes, by contrast, are central to immune homeostasis, and low counts signal impaired immune function and systemic stress. Albumin is far more than a nutritional marker: it exerts anti-inflammatory, antioxidant, and antithrombotic effects, and low levels indicate a catabolic state, hypercoagulability, and depleted physiological reserve. By multiplying the protective components and dividing by the harmful one, the index collapses three biological domains into a single number.

The retrospective design excluded patients with severe active infections, systemic inflammatory or autoimmune disease, or advanced liver failure, and restricted the malignancy exclusion to those with active hematological disease or terminal-stage solid tumors. After screening 536 patients, the final analysis set comprised 471 individuals whose admission laboratory panels, echocardiographic measurements, and follow-up data were extracted from electronic medical records. Patients who died within the year were significantly older than survivors, with a mean age of 69.2 versus 63.8 years, and were far more likely to have malignancy, which affected 32.4 percent of non-survivors compared with 12.2 percent of survivors, and immobility, present in 33.3 percent versus 14.1 percent.

The laboratory differences between groups were equally pronounced. Non-survivors had markedly higher C-reactive protein concentrations, averaging 5.94 milligrams per deciliter versus 2.93 in survivors, along with elevated neutrophil and white blood cell counts, higher D-dimer and troponin levels, and significantly lower lymphocyte counts, hemoglobin, hematocrit, albumin, and total protein. Echocardiography revealed greater right ventricular overload in those who died, with higher systolic pulmonary artery pressure, larger right ventricular diameter, and larger left atrial diameter. Median CALLY values told the story in a single number: 0.5 in non-survivors versus 3.4 in survivors, a difference so stark it held under the strictest statistical threshold.

To quantify how well each index discriminated between those who lived and died, the researchers performed receiver operating characteristic analysis. The CALLY index achieved an area under the curve of 0.787, with a 95 percent confidence interval of 0.732 to 0.842, significantly exceeding the PNI at 0.728 and the SII at 0.643, with pairwise DeLong tests confirming the superiority of CALLY over both competitors. The optimal cutoff, derived using the Youden index, was a CALLY value of 1.6. The clinical consequences of falling on either side of that threshold were dramatic: one-year mortality was 40.8 percent among patients with CALLY values at or below 1.6, compared with just 9.9 percent among those above it. Thirty-day mortality followed the same pattern, at 14.5 percent versus 3.8 percent, while hospital length of stay did not differ, hinting that the index captures long-term biological vulnerability rather than the mechanics of the acute admission.

Cox regression analysis confirmed that these were not statistical artifacts. When CALLY was entered as a continuous variable into multivariable models adjusted for age, an sPESI score of at least one, elevated cardiac troponin, and systolic pulmonary artery pressure, each one-unit increase in CALLY was associated with a hazard ratio of 0.904 for one-year mortality—a 10 percent reduction in risk per unit—highly significant even after adjustment. Separate models constructed for PNI and SII, to avoid multicollinearity from their shared components, showed that both indices were also independently associated with mortality, with higher SII values carrying a hazard ratio of 1.020 per 100-unit increase. Notably, in a sensitivity analysis restricted to 385 patients without malignancy, the CALLY association persisted virtually unchanged, with an adjusted hazard ratio of 0.899, indicating that the result was not simply driven by the cancer cases clustered among the deaths.

Perhaps most compelling is the demonstration of incremental value. The researchers built a baseline prognostic model containing the established predictors—age, sPESI of at least one, elevated troponin, and pulmonary artery pressure—and then added CALLY. The C-statistic, a measure of discrimination, rose from 0.700 to 0.739, and the likelihood-ratio test confirmed a highly significant improvement in model fit. In practical terms, a number that costs nothing to obtain, since C-reactive protein, albumin, and a differential blood count are measured routinely in nearly every hospitalized patient, added genuine prognostic information beyond what careful clinical assessment already provides. Kaplan-Meier survival curves reinforced the point, showing significantly worse one-year survival for patients below the CALLY and PNI cutoffs and above the SII threshold of 996.

The logic behind CALLY’s superiority over pure inflammation markers such as SII is that long-term survival after pulmonary embolism depends on more than the initial thrombotic insult. Inflammatory activation, right ventricular overload, endothelial injury, and persistent thrombotic processes interact with malnutrition, immune dysfunction, and reduced physiological reserve, particularly in older patients with multiple comorbidities. An index that simultaneously reads inflammation, immunity, and nutritional status captures a broader slice of that biology than any single measure. Prior studies support the concept: low CALLY values have independently predicted poor outcomes in heart failure, acute coronary syndromes, and patients undergoing transcatheter aortic valve implantation, and the index was originally developed as a prognostic tool in hepatocellular carcinoma surgery.

The authors are careful to note the limitations. The study was retrospective and single-center, temporal changes in the indices during follow-up could not be assessed because only admission values were used, and the endpoint was all-cause rather than cause-specific mortality. Residual confounding cannot be excluded in an observational design, and the 1.6 cutoff was derived and tested within the same cohort, which tends to produce optimistic performance estimates; the threshold therefore requires validation in an independent external population before it can guide clinical decisions. Even so, the message is striking in its simplicity. Three numbers that already sit in every admission chart, combined by a single division, may help clinicians identify which pulmonary embolism patients carry hidden biological fragility—and direct intensified surveillance and follow-up toward those who need it most.

Subject of Research: Prognostic value of the CALLY index for one-year mortality in acute pulmonary embolism

Article Title: Prognostic value of the C-reactive protein–albumin–lymphocyte index for predicting one-year mortality in acute pulmonary embolism

Article References: Kara, A., Erdoğan, G., Sivri, S., Yenerçağ, M., Boyacı, F., Şeker, O. O., Karagöz, A., Öztürk, E. F., & Aşcı, S. M. (2026). Prognostic value of the C-reactive protein–albumin–lymphocyte index for predicting one-year mortality in acute pulmonary embolism. Frontiers in Cardiovascular Medicine, 13, Article 1928531. https://doi.org/10.3389/fcvm.2026.1928531

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1928531

Keywords: acute pulmonary embolism, CALLY index, C-reactive protein, albumin, lymphocyte count, one-year mortality, risk stratification, inflammation, nutritional status, Prognostic Nutritional Index, Systemic Immune-Inflammation Index, biomarkers