Abstract
Purpose
Hip fractures in the elderly are associated with high 1-year mortality rates. Identifying reliable prognostic markers is essential for improving outcomes. The albumin haemoglobin index (AHI), previously studied in gastrointestinal surgery, has not been evaluated in hip fracture patients. This study aimed to assess AHI’s prognostic significance in predicting 1-year mortality, with additional hazard model analyses at 3 and 6 months, and compared its performance to other preoperative blood parameters.
Method
This retrospective study included 775 patients aged ≥ 65 years who underwent hip fracture surgery from January 2020 to December 2021. Patients underwent a Fracture Liaison Service. Preoperative blood parameters, including haemoglobin (Hb), albumin (Alb), creatinine (Cr), alkaline phosphatase (ALP), alanine aminotransferase (ALT), red cell distribution width (RDW), and RDW/albumin ratio (RAR), were analysed. AHI was calculated as the product of albumin and haemoglobin. Receiver operating characteristic (ROC) curves assessed the prognostic performance of blood predictors for 1-year mortality, while Cox proportional hazards models evaluated the association between preoperative blood parameters and mortality risk at 3, 6, and 12 months postoperatively.
Results
The 1-year mortality rate was 16.6% (129/775). ROC analysis showed preoperative AHI as a strong predictor of 1-year mortality (AUC: 0.75, 95% CI: 0.71–0.80, p < 0.001), comparable to RAR (AUC: 0.77, 95% CI: 0.73–0.82) and superior to RDW (AUC: 0.72, 95% CI: 0.67–0.77). Cox models confirmed AHI’s predictive value for 1-year mortality (adjusted HR: 3.67, 95% CI: 2.44–5.51, p < 0.001), similar to RAR (adjusted HR: 3.94, 95% CI: 2.62–5.90) and higher than RDW (adjusted HR: 2.12, 95% CI: 1.43–3.14). AHI < 3331, RAR > 0.468, and RDW > 14.15% were significant predictors, with consistent trends at 3 and 6 months in hazard models.
Conclusion
AHI is a potential novel predictor of 1-year mortality in elderly hip fracture patients, with prognostic performance comparable to RAR and superior to RDW. Its integration into preoperative assessments may enhance risk stratification and guide interventions to reduce mortality, offering a simple and accessible biomarker for clinical use.
Summary
Preoperative albumin haemoglobin index (AHI) demonstrated strong prognostic accuracy for 1-year mortality in elderly hip fracture patients (AUC: 0.75, 95% CI: 0.71–0.80, p < 0.001; adjusted HR: 3.67, 95% CI: 2.44–5.51, p < 0.001), with an optimal cutoff value of < 3331. AHI is a potential novel biomarker for predicting 1-year mortality in this population.
Keywords: Albumin haemoglobin index, Hip fracture, One-year mortality, Prediction, RAR, RDW
Introduction
Fragility hip fractures in the elderly are associated with high mortality and morbidity. Despite aggressive intervention with timely operations, orthogeriatric co-care and fracture liaison services, mortality rates remain high worldwide, reaching up to 36% in 1 year [1–3]. With the ageing population, the number of hip fractures is also expected to increase, especially in Asia, which has projections from 1,124,060 in 2018 to 2,564,888 in 2050 [4, 5]. This has also contributed to the very high risk of a re-fracture shortly after an initial one, also known as an imminent risk of fracture [6–8]. Therefore, fracture liaison services worldwide and global initiatives have been tackling these problems to decrease overall mortality [9]. Based on this, the socioeconomic burden of hip fractures poses a huge threat to society. One of the major reasons for mortality is due to various comorbidities and frailty that exist among these patients [10, 11]. Therefore, novel approaches are still needed to improve survival and quality of life, as well as identify predictors of mortality.
Routine blood tests during hip fracture admissions provide opportunities to identify prognostic markers [12]. Previous studies have shown that markers, including low serum albumin and haemoglobin levels, are established predictors of postoperative mortality in hip fracture patients [13]. Recent studies have also identified red blood cell distribution width (RDW) and ratio of RDW/serum albumin (RAR) as validated biomarkers in predicting all-cause mortality in the general population as well as post-hip fracture [14, 15]. However, consensus on other markers or indices remains limited. The albumin haemoglobin index (AHI) has recently been identified as a novel marker for predicting mortality and hospital stay in gastrointestinal surgery patients [16]. The rationale of this index is due to hypoalbuminemia being a surrogate marker for malnutrition and a predictor of post-operative complications, as well as anaemia being frequently observed in malnutrition and chronic disease [16]. Given the identification of this new valuable marker for survival, it may have the potential to be applied to hip fracture patients, which can help clinicians predict those who have worse outcomes, such that early precautions can be provided. These can potentially aid in guiding clinical decision-making.
This study aims to evaluate the (i) prognostic significance of albumin haemoglobin index (AHI) for 1-year mortality post hip fracture surgery, with additional hazard analyses at 3 and 6 months to identify its predictive consistency across time points, and (ii) compare AHI with other preoperative blood parameters, including red cell distribution width (RDW) and RDW/albumin ratio (RAR) for 1-year mortality risk prediction.
Material and methods
Study population
In this retrospective cohort study, a total of 775 patients aged 65 and older suffering from a fragility hip fracture admitted to the Department of Orthopaedics and Traumatology, Prince of Wales Hospital, Hong Kong, China, under the Fracture Liaison Service (FLS) between January 2020 and December 2021 were included for analysis. All hip fracture patients admitted to our institution will be arranged for FLS. The FLS is comprised of a multidisciplinary team, including orthopaedic surgeons, geriatricians, physiotherapists, occupational therapists, dieticians, and nurses, focusing on surgery, rehabilitation, osteoporosis management and secondary fracture prevention [2]. Data from all our hip fracture operation records are recorded into the Clinical Management System database prospectively [10]. Patients treated conservatively were excluded. This study was approved by the Joint Chinese University of Hong Kong—New Territories East Cluster Clinical Research Ethics Committee (CRE Ref No.: 2021.622).
Demographics (gender, age, smoking, alcohol use, comorbidities, fracture type, surgery type, and living arrangement), and preoperative serum levels of haemoglobin (Hb), albumin (Alb), creatinine (Cr), alkaline phosphatase (ALP), alanine aminotransferase (ALT), and RDW were collected. The albumin haemoglobin index (AHI) was calculated as the product of serum albumin and haemoglobin levels [AHI = Alb (g/L) *Hb (g/L)], and RAR was calculated as the ratio of RDW to serum albumin concentration [RAR = RDW (%)/Alb (g/L)]. Haemoglobin, albumin, and RDW were measured using the Beckman Coulter UniCel DxH 800 haematology analyser and Beckman Coulter AU5800 chemistry analyser with standardised protocols. The primary outcome was the prognostic performance of AHI for 1-year mortality after hip surgery, assessed via ROC curve analysis. Secondary analyses at 3-month, 6-month, and 1-year mortality risk were conducted using hazard models. Secondary outcomes included a comparison of AHI with other preoperative blood parameters.
Statistical analysis
Data were analysed using IBM SPSS Statistics, version 27. Continuous data that were normally distributed were expressed as mean (standard deviation), while continuous data that were non-normally distributed were expressed as median (interquartile range). Histograms and the Kolmogorov–Smirnov test were used for normality analyses. Categorical data were expressed as frequency (percentage). Continuous data between survivors and non-survivors were compared using two-sample t-tests or Mann–Whitney U-tests for normally distributed data and non-normally distributed data, respectively. Categorical data were compared using Pearson’s chi-squared test or Fisher’s exact test when appropriate. Receiver operating characteristic (ROC) curve analyses were used to assess the survival discriminating power of each blood parameter for 1-year mortality after patients receiving hip surgery. The area under the curve (AUC), specificity and sensitivity were obtained from the ROC curve with optimal cut-off values determined as the value with the highest Youden index (calculated as sensitivity + specificity − 1). Cox proportional hazards models were used to evaluate the association of individual blood parameters and the risk of 3-month, 6-month, and 1-year mortality post-surgery. The proportional hazards assumption was assessed using log [− log S(t)] plots of individual blood parameters; the plots showed no strong signs of assumption violation for all parameters. According to the calculated optimal cutoff values, the pre-operative blood parameters including haemoglobin < 111.5 g/L, albumin < 31.5 g/L, creatinine > 88.5 µmol/L, ALP > 102.5 IU/L, HB*ALB < 3331, ALB/CRE < 0.382, RDW > 14.15, and RAR > 0.468 were included in the univariate analysis adjusted for no covariate (model 1). Variables with p < 0.10 from the univariate analyses (Pearson’s chi-squared test, two-sample t-tests or Mann–Whitney U-tests) were included in the multivariate Cox proportional hazards model (model 2). Presentation of results for categories with < 5 deaths was suppressed because of concerns about the accuracy of confidence intervals and p-values. Hazard ratios (HRs) and 95% confidence intervals (CIs) were computed. P-value < 0.05 (two-tailed) is considered statistically significant.
Results
Among the 775 patients, 646 survived (83.4%) and 129 died (16.6%) within one-year post-surgery. The cohort included 544 females (70.2%) and 231 males (29.8%), with a mean age of 83.45 years (survivors) and 87.23 years (non-survivors), respectively (p < 0.001). Non-survivors were generally older than survivors. Males (23.8%) had a significantly higher percentage of non-survivors than females (13.6%) (p < 0.001). Fracture types included 414 trochanteric fractures and 360 neck of femur fractures. Surgeries included 410 cephalomedullary nails, 236 hemiarthroplasties, 96 hip screws fixation, 30 total hip replacements, and 3 dynamic hip screws. There was no statistically significant difference in the type of fracture and type of surgery between survivors and non-survivors (Table 1). Significant differences were observed in several blood parameters between survivors and non-survivors. Preoperatively, non-survivors had lower haemoglobin (103.0 vs. 114.5 g/dL, p < 0.001) and albumin (29.0 vs. 33.0 g/L, p < 0.001), and higher creatinine (93.0 vs. 74.0 µmol/L, p < 0.001) and ALP (84.0 vs. 66.0 IU/L, p < 0.001) levels. ALT showed no significant difference between the two groups (Table 1).
Table 1.
Comparison of characteristics between survivors and non-survivors within 1-year post-hip surgery
| Parameters | Total (n = 775) |
Survivors (n = 646) |
Non-survivors (n = 129) |
P-value |
|---|---|---|---|---|
| Age (years)a | 84.08 (7.880) | 83.45 (7.801) | 87.23 (7.535) | < 0.001 |
| Gender, n (%) | < 0.001 | |||
| Female | 544 (70.2) | 470 (72.7) | 74 (57.4) | |
| Male | 231 (29.8) | 176 (27.3) | 55 (42.6) | |
| Smoking, n (%) | 0.056 | |||
| Non-smokers | 617 (79.6) | 524 (81.1) | 93 (72.1) | |
| Current smokers | 36 (4.6) | 29 (4.5) | 7 (5.4) | |
| Ex-smokers | 122 (15.7) | 93 (14.4) | 29 (22.5) | |
| Alcohol drinking, n (%) | 0.862 | |||
| Non-drinkers | 726 (93.7) | 606 (93.8) | 120 (93.0) | |
| Current drinker | 17 (2.2) | 14 (2.2) | 3 (2.3) | |
| Ex-drinker | 32 (4.1) | 26 4.0) | 6 (4.7) | |
| Preoperative living arrangement, n (%) | < 0.001 | |||
| Live alone | 106 (13.7) | 101 (15.6) | 5 (3.9) | |
| Live with someone | 582 (75.1) | 484 (74.9) | 98 (76.0) | |
| Institutional care | 87 (11.2) | 61 (9.4) | 26 (20.2) | |
| Comorbidities, n (%) | ||||
| Hypertension | 539 (69.5) | 445 (68.9) | 94 (72.9) | 0.369 |
| Diabetes mellitus | 264 (34.1) | 217 (33.6) | 47 (36.4) | 0.534 |
| Malignancy | 92 (11.9) | 71 (11.0) | 21 (16.3) | 0.090 |
| Liver disease | 46 (5.9) | 42 (6.5) | 4 (3.1) | 0.135 |
| Chronic renal disease | 90 (11.6) | 75 (11.6) | 15 (11.6) | 0.995 |
| Neurovascular disease | 134 (17.3) | 116 (18.0) | 18 (14.0) | 0.272 |
| Congestive heart failure | 43 (5.5) | 32 (5.0) | 11 (8.5) | 0.106 |
| Myocardial infarction | 17 (2.2) | 15 (2.3) | 2 (1.6) | 0.585 |
| Dementia | 139 (17.9) | 114 (17.6) | 25 (19.4) | 0.640 |
| Type of fracture, n (%) | 0.325 | |||
| TOF | 414 (53.4) | 340 (82.1) | 74 (17.9) | |
| NOF | 361 (46.6) | 306 (84.8) | 55 (15.2) | |
| Type of surgery, n (%) | 0.264 | |||
| Cephalomedullary nail | 410 (52.9) | 337 (52.2) | 73 (56.6) | |
| Hemiarthroplasty | 236 (30.5) | 194 (30.1) | 42 (32.6) | |
| Hip screw fixation | 96 (12.4) | 85 (13.2) | 11 (8.5) | |
| Total hip replacement | 30 (3.9) | 28 (4.3) | 2 (0.3) | |
| Dynamic hip screw | 3 (0.4) | 2 (0.3) | 1 (0.1) | |
| Pre-operationb | ||||
| Haemoglobin, g/L | 114.0 (101–126) | 114.5 (103–126) | 103.0 (94–113) | < 0.001 |
| Albumin, g/L | 33.0 (29.0–36.0) | 33.0 (30.0–36.0) | 29.0 (26.0–32.0) | < 0.001 |
| Creatinine, µmol/L | 77.0 (60.0–101.0) | 74.0 (59.0–99.0) | 93.0 (70.0–136.0) | < 0.001 |
| ALT, IU/L | 17.0 (14.0–23.0) | 17.0 (13.0–22.0) | 16.0 (12.0–21.5) | 0.015 |
| ALP, IU/L | 78.0 (62.0–97.0) | 66.0 (52.0–82.0) | 84.0 (67.0–110.5) | < 0.001 |
| RDW, % | 13.9 (13.2–15.0) | 13.8 (13.2–14.7) | 14.9 (14.0–16.7) | < 0.001 |
| Hospital length of stay (days) | 10.0 (8.0–13.0) | 9.0 (7.0–13.0) | 12.0 (9.0–18.0) | < 0.001 |
TOF, trochanteric fractures; NOF, neck of femur fractures; ALT, alanine aminotransferase; ALP, alkaline phosphatase, RDW, red cell distribution width
aData expressed in mean (standard deviation).bData expressed in median (25–75%)
ROC analysis identified preoperative AHI as a strong predictor of 1-year mortality (AUC; 0.75, 95% CI: 0.71–0.80, p < 0.001), comparable to RAR (AUC: 0.77, 95% CI: 0.73–0.82, p < 0.001) and superior to RDW (AUC: 0.72, 95% CI: 0.67–0.77, p < 0.001). The calculated cutoff values revealed that patients with preoperative haemoglobin < 111.5 g/L, albumin < 31.5 g/L, creatinine > 88.5 µmol/L, ALP > 102.5 IU/L, HB*ALB < 3331, ALB/CRE < 0.382, RDW > 14.15, and RAR > 0.468 had a higher risk of 1-year mortality. With the calculated cutoff values, the sensitivities of the AHI, haemoglobin and albumin were 71.3, 72.9, and 72.1%, respectively. The specificities of the AHI, haemoglobin and albumin were 71.2, 60.8, and 65.6%, respectively. Additional details of AUCs and prognostic performances of other parameters are provided in Table 2. The ROC curve of blood parameters for predicting 1-year mortality in patients with hip fracture surgery is shown in Fig. 1.
Table 2.
The prognostic performance of preoperative blood parameters in predicting 1-year mortality of patients undergoing hip fracture surgery
| AUC (95% CI) | P-value | Optimum cutoff value | Sensitivity (%) | Specificity (%) | |
|---|---|---|---|---|---|
| Haemoglobin (g/L) | 0.69 (0.64–0.74) | < 0.001 | < 111.5 | 72.9 | 60.8 |
| Albumin (g/L) | 0.73 (0.69–0.78) | < 0.001 | < 31.5 | 72.1 | 65.6 |
| Creatinine (umol/L) | 0.65 (0.60–0.71) | < 0.001 | > 88.5 | 57.4 | 67.9 |
| ALP (IU/L) | 0.59 (0.54–0.65) | 0.001 | > 102.5 | 34.1 | 82.8 |
| HB*ALB | 0.75 (0.71–0.80) | < 0.001 | < 3331 | 71.3 | 71.2 |
| ALB/CR | 0.71 (0.66–0.76) | < 0.001 | < 0.382 | 69.0 | 65.8 |
| RDW (%) | 0.72 (0.67–0.77) | < 0.001 | > 14.15 | 70.5 | 62.8 |
| RAR | 0.77 (0.73–0.82) | < 0.001 | > 0.468 | 72.9 | 72.0 |
HB, haemoglobin; ALB, albumin; CR, creatinine; ALP, alkaline phosphatase; RDW, red cell distribution width; RAR, red blood cell distribution width to albumin ratio; AUC, area under the curve; CI, confidence interval
Fig. 1.
The comparison of ROC curve of albumin, albumin/creatinine ratio, ALP, creatinine, haemoglobin, AHI, RAR, RDW before hip surgery
Cox proportional hazards models confirmed AHI < 3331 as a robust predictor of 1-year mortality (HR: 5.17, 95% CI: 3.53–7.58, p < 0.0001 in the crude model (model 1); HR: 3.67, 95% CI: 2.44–5.51, p < 0.001 in the adjusted model (model 2)), similar to RAR > 0.468 (HR: 5.60, 95% CI: 3.81–8.23, p < 0.001 in model 1; HR: 3.94, 95% CI: 2.62–5.90, p < 0.001 in model 2). RDW > 14.15% was also an independent predictor for mortality (HR: 3.50, 95% CI: 2.40–5.11, p < 0.001 in model 1; HR: 2.12, 95% CI: 1.43–3.14, p < 0.001 in model 2). In multivariable analyses, no significant association was found between creatinine, ALP, and mortality (3-month, 6-month, and 1-year mortality) after adjusting for age, gender, smoking status, malignancy and other parameters. AHI, RAR, RDW and other blood parameters showed consistent predictive trends at 3 and 6 months in the hazard models (Table 3).
Table 3.
Cox proportional hazard regression of preoperative blood parameters in predicting 3-month, 6-month, and 1-year mortality among patients undergoing hip fracture surgery
| Parameters | Hazard ratio (95% confidence interval) | |||||
|---|---|---|---|---|---|---|
| 3 months | P-value | 6 months | P-value | 12 months | P-value | |
| Model 1a | ||||||
| Age | 1.07 (1.03–1.12) | 0.001 | 1.05 (1.02–1.09) | < 0.001 | 1.07 (1.04–1.09) | < 0.001 |
| Gender | 1.51 (0.85–2.69) | 0.158 | 2.08 (1.37–3.16) | < 0.001 | 1.86 (1.31–2.64) | < 0.001 |
| Haemoglobin < 111.5 g/L | 3.53 (1.87–6.66) | < 0.001 | 3.58 (2.24–5.78) | < 0.001 | 3.69 (2.50–5.44) | < 0.001 |
| Albumin < 31.5 g/L | 4.69 (2.44–8.99) | < 0.001 | 3.44 (2.20–5.39) | < 0.001 | 4.27 (2.90–6.27) | < 0.001 |
| Creatinine > 88.5 µmol/L | 2.38 (1.35–4.19) | 0.003 | 2.04 (1.35–3.09) | < 0.001 | 2.54 (1.79–3.60) | < 0.001 |
| ALP > 102.5 IU/L | – | – | 2.25 (0.31–16.13) | 0.421 | 1.59 (0.22–11.40) | 0.643 |
| HB*ALB < 3331 | 5.19 (2.75–9.79) | < 0.001 | 4.51 (2.90–7.08) | < 0.001 | 5.17 (3.53–7.58) | < 0.001 |
| ALB/CR < 0.382 | 3.49 (1.90–6.41) | < 0.001 | 3.15 (2.03–4.88) | < 0.001 | 3.73 (2.57–5.42) | < 0.001 |
| RDW > 14.15% | 2.34 (1.31–4.18) | 0.004 | 2.75 (1.78–4.28) | < 0.001 | 3.50 (2.40–5.11) | < 0.001 |
| RAR > 0.468 | 4.81 (2.59–8.94) | < 0.001 | 4.92 (3.12–7.77) | < 0.001 | 5.60 (3.81–8.23) | < 0.001 |
| Model 2b | ||||||
| Haemoglobin < 111.5 g/L | 2.06 (1.05–4.03) | 0.035 | 2.31 (1.41–3.80) | < 0.001 | 2.11 (1.40–3.19) | < 0.001 |
| Albumin < 31.5 g/L | 3.32 (1.69–6.52) | < 0.001 | 2.39 (1.50–3.81) | < 0.001 | 2.86 (1.92–4.26) | < 0.001 |
| Creatinine > 88.5 µmol/L | 1.37 (0.74–2.52) | 0.319 | 1.11 (0.71–1.74) | 0.849 | 1.34 (0.92–1.96) | 0.129 |
| ALP > 102.5 IU/L | – | – | 2.47 (0.33–18.31) | 0.376 | 1.98 (0.27–14.51) | 0.501 |
| HB*ALB < 3331 | 4.00 (2.04–7.84) | < 0.001 | 3.56 (2.20–5.77) | < 0.001 | 3.67 (2.44–5.51) | < 0.001 |
| ALB/CR < 0.382 | 2.07 (1.07–4.02) | 0.031 | 1.70 (1.05–2.77) | 0.031 | 2.01 (1.34–3.03) | < 0.001 |
| RDW > 14.15% | 1.35 (0.74–2.48) | 0.334 | 1.70 (1.07–2.69) | 0.24 | 2.12 (1.43–3.14) | < 0.001 |
| RAR > 0.468 | 3.37 (1.75–6.50) | < 0.001 | 3.47 (2.14–5.62) | < 0.001 | 3.94 (2.62–5.90) | < 0.001 |
ALP, alkaline phosphatase; HB, haemoglobin; ALB, albumin; CR, creatinine; RDW, red cell distribution width; RAR, red blood cell distribution width to albumin ratio. Values not shown, indicated as “–”, for categories with number of deaths less than 5
aModel 1, univariate analysis
bModel 2, adjusted for age gender, smoking status, malignancy, haemoglobin < 111.5 g/L, albumin < 31.5 g/L, creatinine > 88.5 µmol/L, ALP > 102.5 IU/L, HB*ALB < 3331, ALB/CR < 0.382, RDW > 14.15%, RAR > 0.468
Discussion
Hip fractures are the most serious form of fragility fractures, and despite advances in modern medicine, mortality rates remain high. The costs of a hip fracture are also substantial, with health and social care costs at US$ 43,669 at 12 months [17]. To improve clinical outcomes and quality of life for hip fracture patients, many initiatives have been established, including orthogeriatric co-care, multidisciplinary care teams and fracture liaison services [2]. During in-patient stay, pre-operative and post-operative, blood taking for monitoring is often performed. This study aimed to provide pragmatic data and analysis to aid in the prediction of mortality for hip fracture patients at 1-year post-operation.
Our study establishes AHI as a potential novel predictor of 1-year mortality in elderly hip fracture patients. ROC analysis demonstrated AHI’s strong prognostic accuracy for 1-year mortality, with a preoperative cut-off value < 3331, which predicted nearly four times increased risk of 1-year mortality in our adjusted hazard model, indicating the potential utility in clinical settings. AHI’s performance is notable given its simplicity, combining two routine preoperative parameters—albumin and haemoglobin, both widely available in clinical settings. However, AHI’s AUC overlap with that of albumin alone (AUC: 0.73, 95% CI: 0.69–0.78), suggesting that while AHI offers strong prognostic performance, its improvement over albumin alone may not differ much. AHI’s advantage lies in its ability to combine two markers, providing a more comprehensive assessment of nutritional status. Pre-operative parameters generally showed significant predictive value, highlighting their importance in early risk assessment in a fracture liaison service. Previous studies have also shown that anaemia and hypoalbuminemia are manifestations of poor nutritional status and are risk factors for poor outcomes in other diseases, including stroke [18], gastrointestinal surgery [16], and cardiovascular disease [19]. Furthermore, the use of systemic immune-inflammation markers, including peripheral platelet, neutrophil, and lymphocyte counts, has also been analysed as prognostic markers. However, it is also noted that impairment of the immune system from chronic diseases may lead to unpredictable effects and may not be modifiable [20].
The main parameters that were found to have increased risk of mortality at 1-year post-operation were anaemia, hypoalbuminemia, raised creatinine, and increased albumin/creatinine ratio. Routine monitoring of these blood parameters helps identify high-risk patients early, enabling targeted interventions including nutritional support, anaemia management, and renal function monitoring to improve patient outcomes and reduce mortality. For instance, based on our findings, patients with AHI < 3331 may benefit from pre-operative optimisation strategies. Our new finding of AHI as a good predictor of mortality can lead to further initiatives and the finding of solutions to decreasing hip fracture mortality. Orthogeriatric co-management of hip fracture patients has been advocated due to better outcomes, including decreased time to surgery, improved clinical management and reduced 1-year mortality [21]. A previous systematic review and meta-analysis also showed that orthogeriatric care models reduce length of stay, in-hospital mortality and delirium of hip fracture patients, together with reduced complications and costs [22]. Cox hazard models also showed AHI’s predictive consistency at 3 months (HR: 4.00, 95% CI: 2.04–7.84) and 6 months (HR: 3.56, 95% CI: 2.20–5.77), supporting its broader utility across time points (Table 3). Optimising patients pre-operatively and post-operatively, as well as monitoring and intervening in blood parameters and holistic care, may be effective in reducing mortality. However, there are yet to be randomised controlled trials to show which orthogeriatric care models are most effective.
We evaluated AHI’s performance against red cell distribution width (RDW) and RDW/albumin (RAR) to determine if AHI is better or equivalent to these markers for 1-year mortality prediction. ROC analysis showed RAR had the highest AUC for 1-year mortality (0.77, 95% CI: 0.73–0.82) with adjusted hazard ratio (HR: 3.94, 95% CI: 2.62–5.90), slightly outperforming AHI (AUC: 0.75, 95% CI: 0.71–0.80; adjusted HR: 3.67). RDW had a lower AUC (0.72, 95% CI: 0.67–0.77) with HR (2.12, 95% CI: 1.43–3.14), indicating that AHI is superior to RDW as a prognostic marker for 1-year mortality. The overlapping confidence intervals of AHI and RAR suggest their prognostic performance was comparable, though RAR’s higher HR indicates a stronger association with 1-year mortality. RAR’s performance reflects its ability to capture both anaemia and inflammation, as RDW is a marker of systemic inflammation and albumin reflects nutritional status [18, 19]. However, AHI’s simplicity, requiring only albumin and haemoglobin, makes it accessible for routine clinical use compared to RAR, which requires an additional parameter of RDW. Furthermore, AHI’s direct focus on malnutrition-related factors may offer a more targeted approach for interventions in elderly hip fracture patients, where nutritional deficits are a primary concern contributing to 1-year mortality.
The strengths of this study include the large cohort and a robust AUC for AHI in predicting 1-year mortality via ROC analysis, and comprehensive Cox models that confirm its predictive value at 1 year, with additional consistency at 3 and 6 months. The inclusion of RDW and RAR provides a thorough comparison, reinforcing AHI’s utility for 1-year mortality prediction. However, AHI primarily reflects preoperative poor nutritional status, and dietary interventions post-hip fracture may not rapidly reduce mortality, as some elderly may have a chronic nature of malnutrition. However, preoperative identification of high-risk patients using AHI could still potentially give earlier interventions, such as nutritional optimisation before elective surgeries or in high-risk patients, and inform orthogeriatric co-management strategies, potentially improving outcomes over time. Interestingly, previous pre-clinical studies have also shown that osteoporotic fracture healing may be compromised [23, 24], and clinical studies have also shown an association with non-union [25]. Therefore, research has also been performed on how to accelerate bone union [26, 27]. Nutrition is also an important aspect for fracture healing [28], and further studies can assess this as an outcome as well. Additional studies should test the modifiability of AHI through targeted interventions. Limitations in our study include unmeasured confounders (e.g., frailty indices) and the lack of intervention data to assess AHI’s direct impact on outcomes.
Conclusion
This study identifies AHI as a potential novel predictor of 1-year mortality in elderly patients with hip fracture surgery. The clinical implications of this study are significant. By incorporating these easily obtainable blood parameters into pre-operative assessments, clinicians may stratify risk and tailor peri-operative management strategies more accurately, potentially improving outcomes for this vulnerable patient population.
Funding
This study was supported by the General Research Fund, Hong Kong SAR Research Grants Council (RGC) (Ref: 14,116,223).
Data availability
Research data is presented within the article. Individual data is not made available.
Declarations
Ethical approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this retrospective study, formal consent is not required.
Conflicts of interest
None.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yu Hei Choi and Pui Yan Wong contributed as first authors.
References
- 1.Bhandari M, Swiontkowski M (2017) Management of acute hip fracture. N Engl J Med 377:2053–2062 [DOI] [PubMed] [Google Scholar]
- 2.Wong RMY, Law SW, Lee KB, Chow SKH, Cheung WH (2019) Secondary prevention of fragility fractures: instrumental role of a fracture liaison service to tackle the risk of imminent fracture. Hong Kong Med J 25:235–242 [DOI] [PubMed] [Google Scholar]
- 3.Wong RMY, Ko SY, Chau WW, Lee LCY, Chow SKH, Cheung WH, Law SW (2021) The first reported fracture liaison service (FLS) for vertebral fractures in China: is muscle the missing gap? Arch Osteoporos 16:168 [DOI] [PubMed] [Google Scholar]
- 4.Wong RMY, Cheung WH, Chow SKH, et al. (2022) Recommendations on the post-acute management of the osteoporotic fracture—patients with “very-high” re-fracture risk. J Orthop Translat 37:94–99. https://www.ncbi.nlm.nih.gov/pubmed/36262963 [DOI] [PMC free article] [PubMed]
- 5.Cheung CL, Ang SB, Chadha M et al (2018) An updated hip fracture projection in Asia: The Asian Federation of Osteoporosis Societies study. Osteoporos Sarcopenia 4:16–21. https://www.ncbi.nlm.nih.gov/pubmed/30775536 [DOI] [PMC free article] [PubMed]
- 6.Wong RMY, Wong PY, Liu C, Wong HW, Chung YL, Chow SKH, Law SW, Cheung WH (2022) The imminent risk of a fracture-existing worldwide data: a systematic review and meta-analysis. Osteoporos Int 33:2453-2466. 10.1007/s00198-022-06473-0 [DOI] [PubMed]
- 7.Wong RMY, Ho WT, Wai LS, Li W, Chau WW, Chow KS, Cheung WH (2019) Fragility fractures and imminent fracture risk in Hong Kong: one of the cities with longest life expectancies. Arch Osteoporos 14:104. https://www.ncbi.nlm.nih.gov/pubmed/31659457 [DOI] [PubMed]
- 8.Johansson H, Siggeirsdottir K, Harvey NC, Oden A, Gudnason V, McCloskey E, Sigurdsson G, Kanis JA (2017) Imminent risk of fracture after fracture. Osteoporos Int 28:775–780. https://www.ncbi.nlm.nih.gov/pubmed/28028554 [DOI] [PMC free article] [PubMed]
- 9.Dreinhofer KE, Mitchell PJ, Begue T et al (2018) A global call to action to improve the care of people with fragility fractures. Injury 49:1393–1397. https://www.ncbi.nlm.nih.gov/pubmed/29983172 [DOI] [PubMed]
- 10.Wong RMY, Ng RWK, Chau WW, Liu WH, Chow SKH, Tso CY, Tang N, Cheung WH (2022) Montreal cognitive assessment (MoCA) is highly correlated with 1-year mortality in hip fracture patients. Osteoporos Int https://www.ncbi.nlm.nih.gov/pubmed/35763077 [DOI] [PubMed]
- 11.Tsui KHM, Chau WW, Liu WH, Tam CY, Yee DKH, Tso CY, Zhang N, Cheung WH, Tang N, Wong RMY (2023) COVID-19 hip fracture outcomes: the role of Ct values and D-dimer levels? J Orthop Translat 43:14–20. https://www.ncbi.nlm.nih.gov/pubmed/37920546 [DOI] [PMC free article] [PubMed]
- 12.Li DY, Zhang K, Wang H, Zhuang Y, Zhang BF, Zhang DL (2024) Preoperative serum calcium level predicts postoperative mortality in older adult patients with hip fracture: a prospective cohort study of 2333 Patients. J Am Med Dir Assoc 25:655–660. https://www.ncbi.nlm.nih.gov/pubmed/37660723 [DOI] [PubMed]
- 13.Li S, Zhang J, Zheng H, Wang X, Liu Z, Sun T (2019) Prognostic role of serum albumin, total lymphocyte count, and mini nutritional assessment on outcomes after geriatric hip fracture surgery: a meta-analysis and systematic review. J Arthroplasty 34:1287–1296 [DOI] [PubMed] [Google Scholar]
- 14.Hao M, Jiang S, Tang J, Li X, Wang S, Li Y, Wu J, Hu Z, Zhang H (2024) Ratio of red blood cell distribution width to albumin level and risk of mortality. JAMA Netw Open 7:e2413213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Liu XF, Zheng YQ, Lin L, Lin ZY, Zhang HP, Huang XP, Wang ZF, Zhang JS (2025) Red blood cell distribution width is a short-term mortality predictor in middle-aged and older adults with hip fracture. BMC Musculoskelet Disord 26:261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Silahli M, Tekin M (2022) Albumin haemoglobin index: a novel pre-operative marker for predicting mortality and hospital stay in patients under one-year undergoing gastrointestinal surgeries. Afr J Paediatr Surg 19:89–96. https://www.ncbi.nlm.nih.gov/pubmed/35017378 [DOI] [PMC free article] [PubMed]
- 17.Williamson S, Landeiro F, McConnell T, Fulford-Smith L, Javaid MK, Judge A, Leal J (2017) Costs of fragility hip fractures globally: a systematic review and meta-regression analysis. Osteoporos Int 28:2791–2800. https://www.ncbi.nlm.nih.gov/pubmed/28748387 [DOI] [PubMed]
- 18.Tian M, Li Y, Wang X, et al. (2020) The Hemoglobin, Albumin, Lymphocyte, and Platelet (HALP) score is associated with poor outcome of acute ischemic stroke. Front Neurol 11:610318. https://www.ncbi.nlm.nih.gov/pubmed/33510706 [DOI] [PMC free article] [PubMed]
- 19.Zheng Y, Huang Y, Li H (2023) Hemoglobin albumin lymphocyte and platelet score and all-cause mortality in coronary heart disease: a retrospective cohort study of NHANES database. Front Cardiovasc Med 10:1241217. https://www.ncbi.nlm.nih.gov/pubmed/38028472 [DOI] [PMC free article] [PubMed]
- 20.Rutenberg TF, Hershkovitz A, Jabareen R, Vitenberg M, Daglan E, Iflah M, Drexler M, Shemesh S (2023) Can nutritional and inflammatory indices predict 90-day mortality in fragility hip fracture patients? SICOT J 9:30. https://www.ncbi.nlm.nih.gov/pubmed/37909883 [DOI] [PMC free article] [PubMed]
- 21.Zhang J, Yang M, Zhang X, et al. (2022) The effectiveness of a co-management care model on older hip fracture patients in China—a multicentre non-randomised controlled study. Lancet Reg Health West Pac 19:100348. https://www.ncbi.nlm.nih.gov/pubmed/35141666 [DOI] [PMC free article] [PubMed]
- 22.Van Heghe A, Mordant G, Dupont J, Dejaeger M, Laurent MR, Gielen E (2022) Effects of orthogeriatric care models on outcomes of hip fracture patients: a systematic review and meta-analysis. Calcif Tissue Int 110:162–184. https://www.ncbi.nlm.nih.gov/pubmed/34591127 [DOI] [PMC free article] [PubMed]
- 23.Chow SK, Chim YN, Wang JY, Wong RM, Choy VM, Cheung WH (2020) Inflammatory response in postmenopausal osteoporotic fracture healing. Bone Joint Res 9:368–385 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wong RM, Thormann U, Choy MH et al (2019) A metaphyseal fracture rat model for mechanistic studies of osteoporotic bone healing. Eur Cell Mater 37:420–430 [DOI] [PubMed] [Google Scholar]
- 25.Zura R, Xiong Z, Einhorn T et al (2016) Epidemiology of fracture nonunion in 18 human bones. JAMA Surg 151:e162775 [DOI] [PubMed] [Google Scholar]
- 26.Choy MV, Wong RM, Li MC, Wang BY, Liu XD, Lee W, Cheng JC, Chow SK, Cheung WH (2020) Can we enhance osteoporotic metaphyseal fracture healing through enhancing ultrastructural and functional changes of osteocytes in cortical bone with low-magnitude high-frequency vibration? Faseb j 34:4234–4252 [DOI] [PubMed] [Google Scholar]
- 27.Zhang N, Chim YN, Wang J, Wong RMY, Chow SKH, Cheung WH (2020) Impaired fracture healing in sarco-osteoporotic mice can be rescued by vibration treatment through myostatin suppression. J Orthop Res 38:277–287 [DOI] [PubMed] [Google Scholar]
- 28.Karpouzos A, Diamantis E, Farmaki P, Savvanis S, Troupis T (2017) Nutritional aspects of bone health and fracture healing. J Osteoporos 2017:4218472 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Research data is presented within the article. Individual data is not made available.