Prognostic Value of HALP Score, PNI, and SII in Predicting 1-Year Mortality in Geriatric Femoral Fractures: A 5-Year Emergency Department Cohort Study

Abstract

Background

With the global rise in geriatric populations, femoral fractures in elderly individuals have become a major health burden. This study aimed to evaluate the prognostic value of 3 readily available laboratory-based indices – HALP (hemoglobin, albumin, lymphocyte, platelet), Prognostic Nutritional Index (PNI), and Systemic Immune-Inflammation Index (SII) – in predicting 1-year mortality among elderly patients with femoral fractures.

Material/Methods

This retrospective cohort study included 309 patients aged ≥65 years who presented to the emergency department of a tertiary university hospital between 2018 and 2023 with low-energy femoral fractures and who underwent surgery. Demographic, clinical, and laboratory data were collected. HALP, PNI, and SII were calculated preoperatively. In-hospital, 30-day, and 1-year mortality were recorded. ROC analysis was used to assess the predictive accuracy of each index.

Results

One-year mortality was 32.4%, while in-hospital and 30-day mortality were 5.5% and 11%, respectively. Lower albumin and PNI scores were significantly associated with higher mortality at all time points (P<0.001). PNI demonstrated moderate predictive accuracy (AUC=0.659), while HALP had limited but statistically significant predictive value (AUC=0.577, P=0.030). SII did not show statistically significant prognostic value (AUC=0.549, P=0.166). Multivariate analysis showed PNI and HALP are independent predictors of long-term mortality.

Conclusions

PNI and HALP are practical, cost-effective tools with prognostic utility in elderly patients with femoral fractures. Early identification of high-risk individuals using these biomarkers may facilitate targeted interventions and improved outcomes. Future multicenter prospective studies are warranted for external validation.

Introduction

With the global increase in the elderly population, the incidence of femoral (hip) fractures among geriatric individuals has risen significantly, posing substantial challenges to healthcare systems due to prolonged hospitalization, increased postoperative complications, and high mortality rates. Epidemiological data estimate that the annual number of hip fractures worldwide will rise from approximately 1.6 million in 2000 to 6.3 million by 2050 [1]. Current data suggest that 30-day mortality rates range between 1% and 7%, while 1-year mortality rates are approximately 23% to 30% [2–4]. Beyond the direct impact on patients, hip fractures in the elderly contribute to increased healthcare expenditures, greater demand for rehabilitation services, and loss of functional independence, which collectively impose a heavy socioeconomic burden. These high rates underscore the critical need for early risk stratification and the development of clinical interventions aimed at improving long-term survival in elderly patients with hip fractures.

Despite the availability of various prognostic scoring systems (eg, Nottingham Hip Fracture Score, POSSUM), a common criticism is that these tools inadequately incorporate biological markers, potentially limiting their accuracy in certain patient populations [5]. Recent studies have increasingly focused on laboratory-based biomarkers that reflect nutritional and inflammatory status, as these factors are strongly associated with post-fracture outcomes. For example, nutritional depletion and systemic inflammation are known to impair fracture healing, delay mobilization, and increase vulnerability to infections, thereby directly affecting both short- and long-term prognosis.

Patient-specific factors such as inflammation, immune suppression, and nutritional status exert a significant influence on post-fracture mortality. Accordingly, laboratory-based prognostic indices such as the HALP score (hemoglobin, albumin, lymphocyte, and platelet), PNI (Prognostic Nutritional Index), and SII (Systemic Immune-Inflammation Index) have been investigated for their potential utility in predicting 1-year mortality.

The PNI reflects nutritional and immunologic status. Recent studies have demonstrated that elderly patients with low PNI scores following hip fractures have significantly increased 1-year mortality and postoperative complications [6,7]. For instance, in a prospective study involving 124 patients over 70 years of age, the PNI was found to be an independent risk factor for 1-year postoperative mortality (ROC-AUC: 0.764, cutoff ≤38.4, sensitivity 83.9%) [8]. In a larger cohort of 3351 patients, those with higher PNI scores exhibited a 34% to 39% reduction in 2-year mortality risk [9]. These findings suggest that optimizing nutritional status could be a modifiable strategy to reduce mortality risk. Nevertheless, conflicting evidence exists regarding the optimal PNI cutoff values across populations, indicating that further validation is required.

The SII index quantifies systemic inflammation and has demonstrated prognostic value in conditions such as malignancies, cardiovascular disease, and stroke [10]. Preliminary studies in geriatric patients with hip fractures have identified high SII values as strong independent predictors of 1-year mortality. A single-center retrospective analysis found that a postoperative day 5 SII ≥1751.9 was associated with an adjusted hazard ratio of 2.16 for mortality [11]. Furthermore, systematic reviews and meta-analyses have supported the association between elevated SII and increased long-term mortality risk [12]. However, some studies have reported weaker associations, raising questions about the consistency and timing of SII measurements in this patient population.

The HALP score is a relatively new metric that summarize the systemic inflammation and hematologic status. Although few studies have evaluated its prognostic significance specifically in geriatric fracture populations, the individual components of the HALP score – ranging from albumin and hemoglobin to lymphocyte count – are well-established predictors of mortality [13]. Recent evidence from oncologic and cardiovascular cohorts has shown HALP to be a promising composite biomarker, yet its value in the context of geriatric fractures remains underexplored and warrants systematic investigation.

However, studies simultaneously evaluating and comparing the prognostic value of HALP, PNI, and SII within the same geriatric hip fracture cohort, particularly in patients presenting to the emergency department, are scarce in the literature. Moreover, the relative predictive accuracy of these indices remains uncertain due to heterogeneous study designs, varied cutoff thresholds, and limited direct comparisons. Therefore, the aim of this study was to assess and compare the predictive capacity of these 3 indices for 1-year mortality in a 5-year emergency department cohort.

The significance of this research lies in the evaluation of easily accessible, low-cost biomarkers HALP, PNI, and SII in a unified clinical context. By providing a direct comparison within the same population, our study seeks to clarify their relative prognostic strengths and contribute to the refinement of early risk stratification strategies. If 1 or more of these indices demonstrate reliable sensitivity and specificity for predicting 1-year mortality, they may offer practical utility in clinical decision-making, risk stratification, early intervention planning, and efficient allocation of healthcare resources.

Material and Methods

Study Design and Population

This retrospective cohort study included patients aged 65 years and older who presented to the emergency department of a tertiary university hospital with femoral fractures between January 2018 and December 2023. Patients were identified by systematically screening the hospital’s electronic medical records (EMR) using ICD-10 codes for femoral fractures. A total of 309 patients were identified through screening of the hospital’s electronic medical records system. Inclusion criteria were: diagnosis of femoral fracture due to low-energy trauma, undergoing operative treatment, and availability of complete preoperative laboratory data. We excluded patients with pathological fractures, polytrauma, active malignancy prior to the fracture, and those under the age of 65. Data extraction was performed independently by 2 trained researchers to ensure accuracy, with discrepancies resolved by consensus.

Data Collection

Demographic data (age, sex), laboratory parameters (hemoglobin, albumin, lymphocyte, and platelet counts), fracture types (intertrochanteric or femoral neck), time to surgery, pre-fracture mobility status, and length of hospital and ICU stay were collected from the EMR. All laboratory parameters were recorded at the time of initial admission to the emergency department, before any surgical intervention. Pre-fracture mobility was assessed retrospectively from patient files and caregiver reports and categorized as ambulatory without aid, ambulatory with aid, or non-ambulatory. Mortality outcomes (in-hospital, 30-day, and 1-year) were obtained from the hospital information system and the national death notification registry. ICU admission and length of hospital stay were also recorded.

Pre-fracture mobility status was determined retrospectively from clinical records and caregiver reports documented at the time of admission. Patients were categorized as ambulatory without aid, ambulatory with walking aid, or non-ambulatory. Due to the retrospective design, standardized frailty scores or detailed comorbidity indices (eg, Charlson Comorbidity Index) could not be consistently retrieved; this limitation has been acknowledged.

Score Calculations

To assess nutritional and inflammatory status, the following indices were calculated:

• HALP score: Hemoglobin (g/L) × albumin (g/L) × lymphocyte count (/L) ÷ platelet count (/L) [14].

• Prognostic Nutritional Index (PNI): [10× serum albumin (g/dL)] + [0.005 × total lymphocyte count (/mm³)] [15].

• Systemic Immune-Inflammation Index (SII): Platelet count × neutrophil count ÷ lymphocyte count [16].

As there are no universally established threshold values for HALP, PNI, or SII, we calculated optimal cutoff points (HALP ≤20.54, PNI ≤42.35, SII ≤1790) based on ROC curve analysis within our study population.

Outcome Measures

The primary outcome was 1-year mortality. Secondary outcomes included in-hospital and 30-day mortality, ICU admission, and length of hospital stay. Mortality data were confirmed via hospital EMR and the national death registry.

Statistical Analysis

Statistical analyses were conducted using IBM SPSS Statistics for Windows, Version 25.0 (Armonk, NY: IBM Corp.). Normality of continuous variables was assessed using the Kolmogorov-Smirnov test. Normally distributed variables are presented as mean±standard deviation (SD), and categorical variables as counts and percentages (%). Group comparisons were performed using the independent samples t test or Mann-Whitney U test for continuous variables, and the chi-square test or Fisher’s exact test for categorical variables. A P value <0.05 was considered statistically significant. Multivariate logistic regression analyses were conducted to adjust for potential confounders, including age, sex, fracture type, time to surgery, and baseline laboratory values. Missing data were handled using pairwise deletion, and sensitivity analyses confirmed the robustness of the results. A P value <0.05 was considered statistically significant.

Ethical Considerations

The study was approved by the Institutional Ethics Committee (approval date: June 18, 2024; decision no: 230). As this was a retrospective study utilizing digital records, informed consent from patients or their relatives was not required according to institutional guidelines. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Results

A total of 309 patients were included in the study, with a mean age of 80.3±8.1 years and a female predominance (60.2%). Mortality outcomes were verified using hospital records and the national death registry. Comparative analyses were performed for in-hospital mortality (n=17, 5.5%), 30-day mortality (n=34, 11%), and 1-year mortality (n=100, 32.4%) (Table 1, Figure 1).

Table 1.

Comparison of clinical and demographic characteristics of patients according to mortality outcomes.

Variables	Total (n=309)	In-hospital mortality			30-day mortality			1-year mortality
		Survival (n=292)	Decased (n=17)	P value	Survival (n=275)	Decased (n=34)	P value	Survival (n=209)	Decased (n=100)	P value
Age (mean±SD)	80.3 ±8.1	80.3 ±8.2	80.7 ±7.7	0.846	80.1 ±8.1	83.2 ±7.4	0.029	79.1 ±8.0	82.9 ±7.9	0.000
Sex (Female, %)	186 (60.2%)	178 (95.7%)	8 (4.3%)	0.056	172 (92.5%)	14 (7.5%)	0.016	135 (72.6%)	51 (27.4%)	0.022
Hemoglobin (mean ±SD)	11.9 ±1.9	11.92 ±1.8	12.42 ±2.3	0.262	11.9 ±1.8	11.8 ±2.1	0.798	12.0 ±1.9	11.7 ±1.8	0.149
Albumin (mean ±SD)	35.6 ±4.1	35.8 ±4.1	33.8 ±4.2	0.046	36.0 ±3.9	33.0 ±4.8	0.000	36.4 ±3.8	34.0 ±4.5	0.000
Lymphocyte (mean ±SD)	1.21 ±0.65	1.22 ±0.64	1.16 ±0.74	0.695	1.23 ±0.69	1.09 ±0.63	0.253	1.25 ±0.6	1.14 ±0.5	0.153
Platelet (mean ±SD)	228 ±65.9	229 ±66	203 ±61	0.082	229 ±63	222 ±86	0.565	229 ±63	225 ±71	0.590
HALP Score (mean ±SD)	24.3 ±1.4	24.3 ±1.4	24.4 ±1.6	0.975	24.3 ±14	21.4 ±14	0.248	25.2 ±1.4	22.2 ±1.5	0.094
HALP Score ≤20.54 (%)	155 (50.2%)	146 (94.2%)	9 (5.8%)	0.633	134 (87%)	20 (13%)	0.267	95 (62%)	59 (38%)	0.026
PNI (mean ±SD)	41.7 ±5.3	41.9 ±5.2	39.7 ±5.6	0.070	42.0 ±5.1	38.0 ±5.6	0.000	42.7 ±5.0	39.7 ±5.3	0.000
PNI ≤42.35 (%)	155 (50.2%)	146 (94.8%)	8 (5.2%)	0.363	131 (85%)	24 (15%)	0.012	91 (59%)	64 (41%)	0.001
SII (mean ±SD)	2294 ±1918	2303 ±1956	2161 ±1275	0.750	2238 ±1822	2745 ±2552	0.146	2237 ±1924	2412 ±1908	0.455
SII ≤1790 (%)	155 (50.2%)	142 (92.3%)	12 (7.7%)	0.347	142 (92%)	13 (8%)	0.140	113 (73%)	42 (27%)	0.047
Trochanteric fracture (%)	169 (54.7%)	159 (94.1%)	10 (5.9%)	0.664	153 (90.6%)	16 (9.4%)	0.343	115 (68%)	54 (32%)	0.866
Neck fracture (%)	140 (45.3%)	130 (92.9%)	10 (7.1%)	0.512	122 (87.2%)	18 (12.8%)	0.248	94 (67%)	46 (33%)	0.903
Time to surgery ≤2 day (%)	138 (44.7%)	128 (92.8%)	10 (7.2%)	0.619	124 (90%)	14 (10%)	0.665	109 (64%)	62 (36%)	0.103
Mobilization status (Mobil%)	273 (88.3%)	256 (93.8%)	17 (6.2%)	0.630	246 (90.2%)	27 (9.8%)	0.085	187 (68%)	86 (32%)	0.373
Length of hospital stay (day, mean ±SD)	5.2 ±4.4	4.82 ±3.7	11.15 ±8.1	0.000	4.9 ±4.1	7.5 ±6.0	0.002	4.39 ±2.6	6.99 ±6.5	0.000
Length of hospital stay >7 day (%)	56 (18.1%)	40 (71.4%)	16 (28.6%)	0.000	40 (71.5%)	16 (28.5%)	0.000	24 (43%)	32 (57%)	0.000
ICU stay (%)	35 (11.3%)	19 (55%)	16 (45%)	0.000	19 (55%)	16 (45%)	0.000	10 (29%)	25 (71%)	0.000
Length of ICU stay (day, mean ±SD)	1.36 ±4.9	0.81 ±3.7	9.3 ±4.9	0.000	0.95 ±4.2	4.65 ±8.1	0.000	0.37 ±1.9	3.42 ±7.8	0.000

Figure 1.

Time-based mortality rates.

In-Hospital Mortality

Of the 309 patients enrolled, 17 (5.5%) died during hospitalization and 292 (94.5%) were discharged alive. Patients who died during hospitalization were slightly older than survivors (80.7±7.7 vs 80.3±8.2 years; P=0.846), with no significant sex differences (P=0.056). Among laboratory parameters, only serum albumin was significantly lower in patients who died during hospitalization (33.8±4.2 g/L vs 35.8±4.1 g/L; P=0.046). Other parameters, including hemoglobin, lymphocyte count, platelet count, HALP score, PNI, and SII, did not show significant differences (all P>0.05). Categorical analyses based on predefined thresholds (HALP ≤20.54, PNI ≤42.35, SII ≤1790) revealed no significant association with in-hospital mortality (P=0.633, P=0.363, and P=0.347, respectively). Fracture type (trochanteric vs neck) was also not associated with in-hospital death (Table 1).

Length of hospital stay and ICU admission were significantly higher among patients who died (11.15±8.1 vs 4.82±3.7 days; P<0.001; ICU admission 45% vs 5%; P<0.001) (Table 1, Figure 2).

Figure 2.

Proportion of ICU admissions among deceased patients.

Thirty-Day Mortality

Patients who died within 30 days were significantly older (83.2±7.4 vs 80.1±8.1 years; P=0.029) and more likely to be male (P=0.016). Serum albumin levels (33.0±4.8 vs 36.0±3.9 g/L; P<0.001) and PNI scores (38.0±5.6 vs 42.0±5.1; P<0.001) were significantly lower in deceased patients. A PNI ≤42.35 was significantly associated with higher 30-day mortality (P=0.012). HALP and SII scores were not significantly different between groups (P=0.248 and P=0.146, respectively), indicating limited short-term prognostic value for these indices. Hospital stay and ICU admission were also significantly higher in this group (Table 1, Figure 2).

One-Year Mortality

Patients who died within 1 year were significantly older (82.9±7.9 vs 79.1±8.0 years; P<0.001), and more likely to be male (P=0.022). Serum albumin (34.0±4.5 vs 36.4±3.8 g/L; P<0.001) and PNI (39.7±5.3 vs 42.7±5.0; P<0.001) were significantly lower in the deceased patients. Although mean HALP scores were lower in patients who died within 1 year (22.2±1.5 vs 25.2±1.4), this difference did not reach statistical significance (P=0.094). However, the proportion of patients with HALP ≤20.54 was significantly higher in the deceased group (38% vs 62%; P=0.026). Similarly, a PNI ≤42.35 was significantly associated with 1-year mortality (41% vs 59%; P=0.001). For SII, although mean values were not significantly different (2412±1908 vs 2237±1924; P=0.455), a significantly smaller proportion of patients with SII ≤1790 was observed in the deceased group (P=0.047). ICU admission and hospital stays longer than 7 days were significantly associated with 1-year mortality (both P<0.001). ICU admission was observed in 71% of deceased patients, and their mean ICU stay was significantly longer (3.42±7.8 vs 0.37±1.9 days; P<0.001) (Table 1, Figure 2).

ROC Curve Analysis for 1-Year Mortality

Receiver operating characteristic (ROC) analysis demonstrated that the PNI score had a statistically significant and moderate predictive value for one-year mortality (AUC=0.659, P<0.001; 95% CI: 0.593–0.726). In comparison, the HALP score also reached statistical significance (AUC=0.577, P=0.030; 95% CI: 0.508–0.645) but demonstrated weaker discriminative ability. These results suggest that PNI may be a more reliable prognostic biomarker for risk stratification than HALP (Table 2, Figure 3).

Table 2.

ROC Analysis of HALP score, PNI and SII for predicting 1-year mortality.

Area under the curve
Test result variable(s)	Area	Std. error*	Asymptotic sig.#	Asymptotic 95% confidence interval
				Lower bound	Upper bound
HALP	.577	.035	.030	.508	.645
PNI	.659	.034	.000	.593	.726
SII	.549	.035	.166	.480	.618

^*Under the nonparametric assumption;

^#Null hypothesis: true area=0.5.

Figure 3.

ROC curve analysis of HALP score and PNI for predicting 1-year mortality.

In contrast, the SII score did not show significant predictive power for 1-year mortality (AUC=0.549, P=0.166; 95% CI: 0.480–0.618), indicating a limited prognostic value in this cohort (Figure 4).

Figure 4.

ROC curve analysis of SII for predicting 1-year mortality.

Discussion

In this retrospective cohort study involving 309 elderly patients with femoral fractures, in-hospital mortality was 5.5%, 30-day mortality was 11%, and 1-year mortality was 32.4%. These findings align with previous studies showing the high mortality risk in geriatric hip fracture patients and emphasizing the need for early risk stratification.

Association Between Age, Sex, and Mortality

Advanced age and male sex were consistently associated with higher short- and long-term mortality. Older patients and males had higher mortality rates, corroborating previous research [17–19]. This aligns with the known decline in physiological reserve, increase in comorbidities, and susceptibility to postoperative complications with advancing age. Similarly, a large multicenter retrospective study by Zhang et al (2023) reported a significantly higher 1-year mortality rate in patients over 80 years old [17]. Liao et al (2023) also noted that each additional year of age increased the 30-day mortality risk by 2% [18].

Regarding sex, females constituted the majority of the sample and had generally lower mortality rates. These results are consistent with previous research. In a meta-analysis by Wang et al (2022), male sex was identified as a significant risk factor for mortality after femoral fractures, while women tended to present at a younger age and with better functional status [19]. Factors such as sarcopenia, malnutrition, and higher frailty scores were more prominent in men [20]. The interactive effect of age and sex is also noteworthy. Particularly in elderly males, mortality risk was significantly higher, underscoring the need for more attentive postoperative care. These demographic factors, in combination with frailty or comorbidity assessments, may enhance prognostic accuracy and guide early interventions. Elderly male patients should be considered a high-risk group and monitored closely postoperatively. Although frailty has been shown to be an important predictor of mortality in geriatric patients with femoral fractures, we were unable to assess frailty scores in our retrospective dataset. This is a limitation of our study, as integrating frailty measures with age and sex could potentially improve prognostic accuracy. A national cohort study by Kim et al (2023) reported improved predictive accuracy when frailty scores were combined with age and sex in risk stratification models [21].

Serum Albumin and PNI

Low serum albumin levels and PNI scores were significantly associated with in-hospital, 30-day, and 1-year mortality. These findings are consistent with previous studies emphasizing the prognostic impact of nutritional status on mortality after hip fractures [4]. Moreover, a prospective study defined geriatric PNI as a modifiable and preventable risk factor for mortality following surgery [22]. In another study examining the relationship between PNI and mortality at 1, 3, 6, 12, and 24 months following femoral fracture surgery, low PNI was consistently associated with increased mortality at all time points. A 3-year follow-up revealed a 20% overall mortality rate, with low PNI levels correlating with higher long-term mortality [23]. PNI, reflecting both nutritional and immunologic status, may serve as a modifiable risk factor, highlighting the potential benefit of early nutritional interventions in improving outcomes.

HALP Score

The HALP score was significantly associated with 1-year mortality but not in-hospital mortality. This aligns with previous studies indicating that HALP reflects long-term immunonutritional status rather than acute post-surgical complications. Patients with HALP ≤20 had higher 1-year mortality, suggesting that HALP may help identify individuals at risk for delayed adverse outcomes. The HALP score emerged as a strong predictor of both 30-day and 1-year mortality. In the study by Vural et al (2025), a HALP score ≤17.97 was associated with a high 1-year mortality risk [24]. Similarly, in a cohort of 1707 patients, each unit increase in HALP reduced 90-day mortality by 32% and overall mortality by 39% [25].

Systemic Immune-Inflammation Index (SII)

In contrast, SII did not demonstrate significant predictive power for 1-year mortality in our cohort. This finding may reflect the influence of acute perioperative factors or patient heterogeneity that attenuate the prognostic value of baseline SII. Other studies have reported mixed results regarding SII in hip fracture populations, suggesting that its utility may depend on timing of measurement and patient characteristics. A prospective study from China reported that each 100-unit increase in SII elevated the 1-year mortality risk by 8% and follow-up mortality by 9% [26]. This trend-level association in our data suggests that the prognostic effect of SII may manifest over time. In another study evaluating all-cause mortality, a significant relationship was found between low SII and high mortality [10]. Additionally, in elderly patients with femoral fractures, low SII levels were significantly associated with increased 30-day, 1-year, and 2-year mortality [27,28].

Clinical Implementation and Implications

The positive association between age and mortality confirms the study’s focus on a high-risk population. In this context, early measurement of biomarkers such as HALP and PNI can facilitate the identification of high-risk patients, enabling more intensive postoperative care. Thresholds such as HALP ≤17–20 and PNI ≤42 may serve as predictive markers for early nutritional support, rehabilitation, and, if needed, immune-modulating therapies. Prospective randomized studies are warranted to validate these thresholds and evaluate their clinical utility.

Limitations

This study is limited by its retrospective and single-center design. Potential confounders such as socioeconomic status and baseline functional capacity could not be fully controlled. Additionally, dynamic monitoring of SII was not performed; only baseline values were analyzed. The literature suggests evaluating integrated markers such as the SII/albumin ratio, which may be considered in future research. Our study did not include an evaluation of frailty, which constitutes a limitation given its potential impact on long-term outcomes.

Conclusions

Our results support the prognostic value of immunonutritional scores such as HALP and PNI in predicting 30-day and 1-year mortality in elderly patients with femoral fractures. SII also shows potential as a long-term mortality risk indicator. Early intervention based on biomarker thresholds could guide more aggressive management strategies for high-risk patients. Prospective multicenter studies are needed to validate and generalize these results.