Routine Blood Test Predicts Stroke Outcomes in Patients Over 80

For the very oldest stroke patients, the question families ask most urgently is also the hardest for doctors to answer: will my mother or father walk out of this hospital able to live independently, or will the next six months bring dependency or death? A new study from China suggests that part of the answer may already sit in the routine blood tests drawn within hours of admission. Researchers at Jiujiang First People’s Hospital report that a composite laboratory score, built entirely from standard admission measurements, was strongly associated with how patients aged 80 or older fared 180 days after an acute ischemic stroke. The work, published in Frontiers in Neurology, offers a technically careful addition to a growing field of prognostic biomarker research in the oldest-old stroke population.

The score is called the Prognostic Inflammation-Nutrition Reserve Index, or PINRI, and its architecture reflects a deliberate attempt to capture two systemic domains that acute cerebral ischemia destabilizes. The first is inflammation: acute ischemia triggers peripheral immune activation, leukocyte redistribution, and endothelial injury, and poststroke immune dysregulation is closely tied to pneumonia and other infections that can derail recovery. The second is nutritional reserve: albumin and lymphocyte-based measures reflect the catabolic and immune vulnerability that dysphagia and frailty impose on older patients. PINRI combines an internally standardized logarithmic C-reactive protein-to-albumin ratio, the logarithmic systemic inflammatory response index, which integrates neutrophil, monocyte, and lymphocyte counts, and the inverse of the prognostic nutritional index, calculated from albumin plus five times the lymphocyte count. Each component was log-transformed where distributions were right-skewed, converted to z-scores, and summed with equal weights to avoid overfitting in a single-center derivation cohort.

The study screened 1,436 consecutive adult admissions with suspected acute ischemic stroke from January 2021 through December 2024, of whom 320 were aged 80 or older. After excluding duplicates, stroke mimics, hemorrhagic diagnoses, patients with missing baseline laboratory components, and those without ascertainable 180-day outcomes, 278 analyzable patients remained. Time zero was the first registered hospital-arrival timestamp, and baseline blood values had to be drawn within 24 hours, before any albumin infusion or blood products that could distort interpretation. The primary endpoint was an ordinal three-level functional survival state at 180 days: modified Rankin Scale 0 to 2 indicating functional independence, 3 to 5 indicating dependency among survivors, and death, assigned modified Rankin Scale 6. The authors argue that preserving this ordinal structure retains information that dichotomized endpoints discard, particularly in a population where complete independence is uncommon.

The outcome distribution itself underscores how heterogeneous this age group is. At 180 days, 111 patients, or 39.9 percent, were functionally independent; 114, or 41.0 percent, were alive but dependent; and 53, or 19.1 percent, had died. In total, 167 patients, or 60.1 percent, experienced dependency or death. Baseline characteristics shifted sharply across PINRI tertiles: admission stroke severity measured by the National Institutes of Health Stroke Scale rose from a median of 4 in the lowest tertile to 8 in the highest, baseline infection or fever jumped from 5.4 percent to 23.7 percent, dysphagia screen positivity climbed from 34.4 percent to 62.4 percent, and albumin, estimated glomerular filtration rate, and the prognostic nutritional index all declined. The unadjusted outcome distribution moved toward dependency and death across increasing tertiles with a global p-value below 0.001.

The headline association survived rigorous adjustment. Each one-standard-deviation increase in admission PINRI was associated with roughly twofold higher adjusted odds of a worse functional survival state, with an adjusted common odds ratio of 1.98 (95 percent confidence interval, 1.48 to 2.65; p < 0.001) after accounting for age, sex, prestroke disability, admission severity, atrial fibrillation, renal function, admission glucose, and reperfusion therapy. The contrast between extremes was starker still: patients in the highest tertile had 3.43 times the adjusted odds of the lowest tertile (95 percent confidence interval, 1.74 to 6.74). Secondary binary models were directionally consistent for dependency or death, severe dependency or death, and all-cause mortality, and a sensitivity analysis using the full seven-level modified Rankin Scale produced a similar estimate.

One of the study’s most technically interesting features is its dynamic exposure analysis. PINRI was recalculated at day 3 and day 7 or discharge, and the first-week burden was computed as the trapezoidal area under the PINRI-time curve divided by observed days. Because patients who die early cannot contribute follow-up trajectories, the authors restricted this analysis to a day-7 landmark cohort of 241 survivors with repeated laboratory data, a design that limits immortal time and survival bias. Even among these survivors, each one-standard-deviation increase in first-week PINRI burden was associated with worse 180-day outcomes (adjusted common odds ratio, 1.80; 95 percent confidence interval, 1.19 to 2.72; p = 0.005), suggesting that the trajectory of systemic inflammation and nutritional depletion carries prognostic information beyond the admission snapshot.

Did the index actually add value beyond what clinicians already know? The base clinical model, containing age, sex, prestroke modified Rankin Scale, admission severity, renal function, glucose, atrial fibrillation, and reperfusion therapy, achieved an optimism-corrected C-statistic of 0.810 for dependency or death, validated with 1,000 bootstrap resamples. Adding PINRI raised the C-statistic to 0.834, a difference of 0.024 with a paired bootstrap p-value of 0.033, and improved the Brier score from 0.170 to 0.158. The gain was real but modest, and the authors are candid about it: comparator single-marker models using the C-reactive protein-to-albumin ratio, the systemic inflammatory response index, the prognostic nutritional index, or the systemic immune-inflammation index produced C-statistics of 0.821 to 0.825, and none of the pairwise differences between PINRI and these simpler indices remained statistically significant after false discovery rate control. The composite showed numerical, not established, superiority.

The authors also confront the confounding and mediation problems that shadow biomarker research in acute stroke. Baseline infection or fever and dysphagia were strongly imbalanced across PINRI tertiles, and both could partly drive the CRP, albumin, and leukocyte patterns that feed the index, as well as the aspiration risk and dependency that follow. Sensitivity analyses excluding patients with baseline infection or fever were directionally consistent, but reverse causation cannot be eliminated. PINRI may also reflect unmeasured frailty, stroke subtype, large-vessel occlusion, or preexisting inflammatory disease. Because albumin and lymphocyte counts contribute to more than one component, equal-weight summation may amplify overlapping information rather than represent three distinct biological domains. The index should be read, the authors write, as an early systemic-risk marker rather than proof that inflammation or nutrition is the sole driver of poor outcome.

What would clinical use look like? The authors propose that if externally validated, reported with fixed derivation constants, and locally calibrated, a high PINRI could be tested as a supportive-care trigger: earlier attention to swallowing, infection surveillance, nutrition assessment, rehabilitation readiness, and family counseling. Crucially, they warn against therapeutic nihilism. A high PINRI should never be used in isolation to withhold reperfusion therapy, rehabilitation, or supportive care; older patients with high systemic vulnerability may be precisely the group needing more structured supportive assessment, not less treatment. The index is currently a prognostic and risk-stratification marker, not a validated clinical decision-making threshold, and the internally standardized z-score constants are not transportable outside the derivation cohort.

The limitations are substantial and the authors enumerate them with unusual transparency. This was a single-center derivation study, so generalizability requires multicenter external validation. Imaging predictors were intentionally excluded to maintain feasibility, limiting comparison with models that include infarct volume or vascular occlusion. The seven-day onset-to-arrival eligibility window means admission biomarkers may represent different biological phases across patients, and the day-7 burden analysis may be affected by measurement-selection bias. The sample size was adequate for the primary prognostic-factor analysis but not for definitive interaction or threshold-transportability claims. Still, the study illustrates a broader shift in stroke prognostication: the move toward routine-data composites that quantify systemic reserve in the oldest-old, a population where premorbid disability, frailty, and comorbidity shape outcomes as powerfully as the neurological deficit itself. The next step is prospective testing of whether PINRI-informed supportive-care pathways can improve patient-centered outcomes without restricting access to evidence-based stroke treatment.

Subject of Research: PINRI prognostic index for 180-day functional survival after acute ischemic stroke in patients aged 80 or older

Article Title: A prognostic inflammation-nutrition reserve index for 180-day functional-survival outcomes after acute ischemic stroke in patients aged 80 years or older

Article References: Zhang, L., & Peng, J. (2026). A prognostic inflammation-nutrition reserve index for 180-day functional-survival outcomes after acute ischemic stroke in patients aged 80 years or older. Frontiers in Neurology, 17, Article 1910335. https://doi.org/10.3389/fneur.2026.1910335

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1910335

Keywords: prognostic, inflammation-nutrition, reserve, index, functional-survival, outcomes, acute, ischemic, stroke, patients, aged, years