Abstract
Background
Elderly patients carry high risks of postoperative serious medical complications (PSMC) and mortality, but reliable perioperative predictive indicators remain limited.
Objective
To evaluate the predictive value of perioperative hematological markers for PSMC and 1-year mortality in elderly surgical patients.
Methods
A total of 590 patients aged 65–85 years were enrolled. Perioperative blood parameters were measured preoperatively and on postoperative days 1, 3, 5, 7. Patients were grouped by PSMC occurrence and 1-year survival status. ROC analysis was used to assess diagnostic performance.
Results
PSMC occurred in 58 patients (9.8%), and 89 patients (15.08%) died within 1 year. Perioperative red blood cells, hemoglobin, albumin, creatinine, procalcitonin, troponin, D-dimer, and D-dimer to albumin ratio (DAR) showed significant dynamic changes (all p < 0.01). The PSMC group had lower albumin and higher D-dimer and DAR than the non-PSMC group (all p < 0.01), with markedly higher 1-year mortality (86.21% vs. 7.33%, p < 0.001). Preoperative DAR showed the comparable predictive power for PSMC (AUC = 0.82), followed by D-dimer (AUC = 0.80) and albumin (AUC = 0.69). For 1-year mortality, D-dimer and DAR both had an AUC of 0.74, with DAR exhibiting higher sensitivity (85.39%). Elevated DAR was significantly associated with increased PSMC and mortality risks (both p < 0.001). Multivariate logistic regression showed high DAR was an independent risk factor for PSMC (OR = 26.82, p < 0.001) and 1-year mortality (OR = 21.75, p < 0.001).
Conclusion
Perioperative DAR is a promising prognostic biomarker for PSMC and 1-year mortality in elderly surgical patients. High DAR indicates poor prognosis and may help identify high-risk individuals for targeted intervention.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12877-026-07989-4.
Keywords: Hip fracture, D-dimer to albumin ratio, Postoperative serious medical complications, Survival rate, Predictive value
Introduction
Aging is a major global social and health challenge [1]. Hip fractures account for about 30% of all fragility fractures, with over 80% occurring in individuals aged ≥ 65 years [2]. In China, the rising elderly population and age-related diseases make hip fractures particularly concerning due to high morbidity, disability, and mortality [3]. Elderly individuals have increased fall risk from visual, neurological, and muscular impairments, and their lower bone density and energy absorption further heighten hip fracture risk after falls [4, 5]. Approximately one-third of elderly fracture patients sustain hip fractures, and 86% of these occur in those aged ≥ 65 [6]. After hip fracture, quality of life declines sharply: 35% of survivors cannot walk independently, 25% require long-term care, 6-month mortality is 10–20%, and 1-year mortality 20–30%, earning it the term “the last fracture of life” and imposing heavy burdens on patients and healthcare systems [7–9].
Currently, surgical treatment is widely regarded as the primary intervention for improving the prognosis of elderly patients with hip fractures. It can effectively alleviate pain, restore some functionality, and improve survival rates [10]. However, compared to their non-fractured peers, elderly hip fracture patients face a significantly higher postoperative mortality risk, particularly in the first two years after surgery. This risk is 4–8 times greater than that of the general elderly population and may persist for several years post-surgery [11–13]. In addition to impaired motor function and muscle strength, these patients often suffer from multiple comorbidities, including hypertension, coronary heart disease, diabetes, and cerebral infarction, which collectively contribute to slow recovery and frequent complications post-surgery [14]. Recent research has focused on identifying modifiable risk factors to optimize perioperative management and improve patient outcomes. Among various biomarkers, D-dimer and albumin (ALB) have garnered considerable attention due to their critical roles in coagulation function, inflammatory response, and nutritional status [15].
D-dimer is a degradation product of cross-linked fibrin, and elevated levels suggest a hypercoagulable state and ongoing fibrinolysis, making it a valuable marker for assessing thrombotic diseases and predicting prognosis [16]. In elderly hip fracture patients, DD levels are often markedly elevated due to endothelial damage, blood stasis, and immobilization [17]. Studies have shown that a DD level ≥ 4.12 µg/mL serves as an independent risk factor for acute ischemic stroke within 90 days post-surgery and is a significant predictor of one-year postoperative survival [18]. In contrast, albumin, synthesized by hepatocytes, plays a crucial role in maintaining plasma oncotic pressure and material transport. Its levels reflect the body’s nutritional status and inflammatory burden [19, 20]. Elderly patients often experience hypoalbuminemia due to aging, declining liver and kidney function, malnutrition, and surgical stress [21]. Numerous studies have confirmed that preoperative hypoalbuminemia is a strong predictor of postoperative mortality in elderly hip fracture patients, with albumin levels below 2.9 g/dL at admission significantly correlating with higher 30-day postoperative mortality [22].
Although both D-dimer and albumin independently have significant predictive value, single markers exhibit limitations when assessing clinical outcomes influenced by multiple factors. Recently, composite indices such as the DAR have demonstrated superior predictive performance compared to individual markers [23]. As a combined indicator, DAR reflects three core pathophysiological processes: coagulation activation (represented by D-dimer), inflammation, and nutritional status (represented by albumin). It has been validated as a reliable prognostic marker in multiple diseases, including gastric cancer and aneurysmal subarachnoid hemorrhage, where it outperforms D-dimer or albumin alone [23]. Our unpublished preliminary clinical data found that elderly lower extremity fracture patients who experienced cardiovascular events post-surgery had significantly higher preoperative median DAR compared to those who did not experience such events.
Accordingly, we hypothesized that perioperative DAR would be a stronger predictor of PSMC and 1-year mortality in elderly hip fracture patients than single biomarkers. To test this hypothesis, we conducted a large-sample retrospective analysis to evaluate the dynamic changes in DAR across perioperative time points, compare its predictive performance with conventional biomarkers, and assess its potential utility in early risk stratification.
Materials and methods
Study population
A total of 847 consecutive elderly patients aged 65 to 85 years who underwent hip fracture surgery at the Affiliated Jiangning Hospital of Nanjing Medical University between January 2015 and December 2022 were initially screened. After applying the inclusion and exclusion criteria, 590 patients were finally enrolled. The reasons for exclusion were as follows: multiple fractures or severe trauma (n = 42), concomitant malignancy, hematologic disease, or severe liver disease (n = 38), pathological fracture or high-energy injury (n = 29), cardiovascular or cerebrovascular events within 3 months prior to surgery (n = 31), incomplete clinical data or loss to follow-up (n = 117). A detailed flowchart of patient selection is presented in Fig. 1.
Fig. 1.
The flowchart of patient selection
Data collection
Relevant preoperative, intraoperative, and postoperative data of consecutive elderly patients undergoing hip fracture surgery from 2015 to 2022 were systematically recorded. Data collected included age, sex, body mass index (BMI), operative duration, ASA classification, surgical approach, preoperative comorbidities (e.g., hypertension, diabetes, coronary heart disease, dementia, hemiplegia), important auxiliary examinations and biochemical test results, and PSMC. The preoperative auxiliary examinations included routine 12-lead electrocardiography, Cardiac ultrasound, chest radiograph/lung computed tomography (CT), cranial CT, and lower extremity venous ultrasound, which were used to evaluate preoperative cardiopulmonary, cerebrovascular and venous thrombosis status of patients. Hematologic parameters during hospitalization were also documented. The serial perioperative hematological and biochemical parameters collected in this study included red blood cell (RBC), hemoglobin (HB), total protein, albumin (ALB), albumin-globulin (A/G) ratio, creatinine, urea, procalcitonin, troponin, D-dimer, and calculated D-dimer to albumin ratio (DAR = D-dimer (mg/L) / Albumin (g/dL)). All blood samples were tested by the clinical laboratory of our hospital using standard clinical detection kits and instruments. The primary outcome was PSMC during hospitalization, which was strictly defined as confirmed severe clinical events with consistent clinical severity, including hemodynamically unstable arrhythmia, cerebral infarction, postoperative delirium, angina pectoris, pulmonary embolism [24, 25]. The other main outcome was 1-year all-cause mortality confirmed by telephone follow-up and outpatient medical records. The D-dimer/albumin ratio (DAR) was calculated as D-dimer (mg/L)/Alb (g/dL). The occurrence and timing of each event were retrospectively identified by systematically reviewing inpatient medical records, including progress notes, nursing records, and medication orders. The exact postoperative day (POD) of each event was recorded to clarify the timing of onset. All patients underwent general anesthesia. All patients received uniform perioperative thromboprophylaxis with subcutaneous low-molecular-weight heparin 4000 IU once daily.
Outcome measures
The primary composite endpoint was PSMC during hospitalization, which included confirmed severe postoperative serious medical complications: hemodynamically unstable arrhythmia, cerebral infarction, postoperative delirium, angina pectoris, pulmonary embolism. To enhance the clinical interpretability of the composite endpoint, we additionally reported the incidence of each individual component event separately in the result section, providing detailed stratified information for different pathological types of adverse events, rather than only presenting the overall composite outcome. For long-term prognosis, the main outcomes included 1-year all-cause mortality, which was confirmed by telephone follow-up and outpatient medical records. We also supplemented the assessment and analysis of 1-year cardiovascular-specific mortality to provide more comprehensive prognostic information.
Statistical analysis
Data were analyzed using SPSS 26.0(IBM Corp., NY, USA). Continuous variables were presented as mean ± standard deviation or median (interquartile range) and compared using t-test or Mann-Whitney U test. Categorical variables were presented as percentages (%) and compared using the χ² test. Repeated measures analysis of variance (ANOVA) was used to compare preoperative and postoperative levels of each parameter at 1, 3, 5, and 7 days, evaluating group, time, and interaction effects. Normality was assessed by the Shapiro-Wilk test, and sphericity by Mauchly’s test (Greenhouse-Geisser correction applied if violated). Descriptive statistics and partial η² were reported to indicate effect size; Bonferroni post-hoc tests were performed when interaction effects were significant. Receiver operating characteristic (ROC) curves were plotted, and the area under the curve (AUC), sensitivity, and specificity were calculated. The optimal cutoff value of preoperative DAR for predicting PSMC and 1-year mortality was determined by ROC curve analysis using the Youden index. Patients were dichotomized into high-DAR and low-DAR groups based on this preoperative DAR cutoff. For sensitivity analysis, patients were further divided into quartiles according to preoperative DAR levels to verify the robustness of the results. To evaluate the robustness of predictive performance, internal validation was performed using bootstrap resampling with 1,000 iterations to estimate optimism-corrected AUC values for preoperative DAR. Multivariate logistic regression analyses were performed to identify independent risk factors for PSMC and 1-year mortality, with adjustment for gender, age, BMI, operative duration, ASA classification, surgical approach, preoperative comorbidities and DAR. Results were reported as odds ratio (OR) with 95% confidence interval (CI). A post-hoc sample size assessment was performed. For the PSMC endpoint (58 events, 9.8% incidence), the events per variable (EPV) was 7.25 (8 predictors), which is below the traditional threshold of 10 [26] but meets the less stringent criterion of EPV ≥ 5 suggested for strong effect sizes [27]. Bootstrap internal validation showed minimal optimism (corrected AUC 0.81 vs. original 0.82), indicating acceptable model stability. For the 1-year mortality endpoint (89 events, 15.1% incidence), the EPV was 11.13, satisfying the conventional requirement. Thus, the sample size is adequate for mortality analysis, while the PSMC analysis should be interpreted with caution and validated in external cohorts. Statistical significance was defined as α = 0.05, with p < 0.05 considered statistically significant.
Results
Baseline characteristics of the study population
A total of 590 elderly patients were included in this study. The cohort consisted of 251 male (42.54%) and 339 female (57.46%) patients, with a median age of 74 years and a mean body mass index of 23.18 ± 2.98 kg/m2. Hypertension was the most prevalent preoperative comorbidity, affecting 20.68% of patients, followed by cerebral infarction (4.58%), diabetes (3.56%), and coronary heart disease (1.69%). During the postoperative period, 58 patients (5.25%) experienced PSMC, with hemodynamically unstable arrhythmia being the most common. The one-year mortality rate was 15.08% (89 patients). Specific details are presented in Table 1.
Table 1.
Baseline characteristics of patients
| N = 590 | Proportion | |
|---|---|---|
| Gender | ||
| Male | 251 | 42.54% |
| Female | 339 | 57.46% |
| Age | 74(69, 79) | |
| BMI (kg/m2) | 23.18 ± 2.98 | |
| Operative Duration (hours) | 2.62(2.25, 3.08) | |
| ASA Classification | ||
| 2 | 99 | 16.78% |
| 3 | 473 | 80.17% |
| 4 | 18 | 3.05% |
| Surgical Approach | ||
| Hip Arthroplasty | 574 | 97.29% |
| PFNA | 16 | 2.71% |
| Preoperative Comorbidities | ||
| Hypertension | 122 | 20.68% |
| Cerebral Infarction | 27 | 4.58% |
| Diabetes | 21 | 3.56% |
| Coronary Heart Disease | 10 | 1.69% |
| PSMC* | 58 | 9.83% |
| Hemodynamically unstable arrhythmia | 33 | 5.59% |
| Cerebral Infarction | 8 | 1.36% |
| Postoperative Delirium | 12 | 2.03% |
| Angina Pectoris | 8 | 1.36% |
| Pulmonary Embolism | 8 | 1.36% |
| 1-Year Postoperative Survival | ||
| Survival | 501 | 84.92% |
| Death | 89 | 15.08% |
PSMC Postoperative serious medical complications
*Some patients have multiple postoperative complications simultaneously. PFNA Proximal Femoral Nail Anti-rotation
Perioperative blood test results
According to the blood test results presented in Table 2, all indicators showed significant changes at different postoperative time points (p < 0.01). RBC and hemoglobin began to decline on the first day post-surgery, reaching their lowest values on the third day. They slightly recovered thereafter, but remained lower than preoperative levels. Total protein and albumin exhibited a downward trend post-surgery, with the lowest values observed on the third and fifth days, and the albumin/globulin ratio showed a similar decline pattern. Creatinine and urea levels were significantly elevated on the first and third days post-surgery, but gradually decreased, returning to preoperative levels by the seventh day. Procalcitonin and troponin levels significantly increased on the first and third days post-surgery, then rapidly declined, reaching baseline levels by the seventh day. D-dimer remained elevated post-surgery, peaking on the fifth day, and then decreased by the seventh day. The DAR gradually increased post-surgery, peaking on the fifth day, and decreased by the seventh day, showing a trend toward higher levels consistent with a perioperative hypercoagulable profile initially but then alleviated.
Table 2.
Preoperative and postoperative blood test results
| Preoperative | Postoperative Day 1 | Postoperative Day 3 | Postoperative Day 5 | Postoperative Day 7 | F | p | |
|---|---|---|---|---|---|---|---|
| RBC | 4.70(4.30, 5.05) | 4.04(3.80, 4.26) | 3.51(3.28, 3.71) | 3.72(3.45, 3.96) | 3.81(3.54, 4.07) | 1073 | < 0.01 |
| HB | 140(132, 150) | 104(97, 112) | 92(85, 99) | 99.50(92, 107) | 104(98, 112) | 2348 | < 0.01 |
| Total Protein | 68.9(63.8, 75) | 70.3(67.5, 72.8) | 62.3(58.58, 65.90) | 57.10(53.48, 60.93) | 59.80(55.60, 63.53) | 656.6 | < 0.01 |
| Albumin | 47.45(43.70, 51.23) | 39.80(37.38, 42.10) | 34.70(32.38, 37.20) | 33.00(30.40, 35.40) | 33.40(31.00, 35.90) | 912.1 | < 0.01 |
| Albumin/Globulin Ratio | 2.03(1.76, 2.27) | 2.02(1.75, 2.25) | 1.60(1.41, 1.81) | 1.41(1.21, 1.60) | 1.50(1.32, 1.68) | 757.2 | < 0.01 |
| Creatinine | 69.21(57.66, 80.78) | 115.86(108.10, 122.64) | 117.04(110.02, 124.87) | 109.80(102.06, 117.08) | 74.19(61.59, 85.89) | 2033 | < 0.01 |
| Urea | 5.73(4.38, 6.97) | 8.59(7.67, 9.29) | 9.08(8.25, 9.76) | 8.05(7.28, 8.84) | 6.62(5.80, 7.30) | 1072 | < 0.01 |
| Procalcitonin | 0.27(0.12, 0.43) | 1.24(0.83, 1.63) | 1.97(1.46, 2.51) | 0.93(0.61, 1.21) | 0.25(0.12, 0.37) | 2137 | < 0.01 |
| Troponin | 0.02(0.01, 0.04) | 0.12(0.09, 0.16) | 0.19(0.13, 0.24) | 0.09(0.06, 0.12) | 0.02(0.01, 0.04) | 1870 | < 0.01 |
| D-dimer | 0.55(0.26, 0.80) | 1.17(0.91, 1.58) | 3.47(2.74, 4.17) | 4.89(3.86, 5.91) | 2.85(2.39, 3.41) | 1540 | < 0.01 |
| DAR | 0.01(0, 0.02) | 0.03(0.02, 0.04) | 0.10(0.08, 0.12) | 0.15(0.12, 0.18) | 0.09(0.07, 0.10) | 1158 | < 0.01 |
RBC Red blood cells, HB Hemoglobin, DAR D-dimer to albumin ratio
Comparison of baseline data and one-year mortality between postoperative Patients with and without PSMC
A total of 590 patients were included, of whom 58 (9.8%) experienced PSMC within one year postoperatively. The two groups were generally comparable at baseline, with no statistically significant differences observed in age, sex, body mass index, operative duration, ASA classification, surgical approach, or the distribution of preoperative comorbidities (all p > 0.05) (Table 3). Notably, however, the PSMC group demonstrated a substantially higher one-year mortality rate compared to the non-PSMC group, with the difference reaching statistical significance (p < 0.001). These findings suggest that the occurrence of PSMC is associated with a marked increase in postoperative mortality risk, despite the absence of significant differences in most baseline characteristics between the two groups.
Table 3.
Baseline data and one-year postoperative survival status of patients with and without PSMC
| Non-PSMC Group (n = 532) | PSMC Group (n = 58) | Statistic Value | p Value | |
|---|---|---|---|---|
| Gender | 0.008 | 0.928 | ||
| Male | 226(42.48%) | 25(43.10%) | ||
| Female | 306(57.52%) | 33(56.90%) | ||
| Age (year) | 74(69, 79) | 77(70, 82) | -1.662 | 0.096 |
| BMI (kg/m2) | 22.89(20.81, 25.33) | 23.13 ± 2.60 | -0.140 | 0.889 |
| Operative Duration (hours) | 2.62(2.25, 3.08) | 2.59(2.12, 3.19) | -0.404 | 0.686 |
| ASA Classification | 1.016 | 0.602 | ||
| 2 | 90(16.92%) | 9(15.52%) | ||
| 3 | 427(80.26%) | 46(79.31%) | ||
| 4 | 15(2.82%) | 3(5.17%) | ||
| Surgical Approach | 0.238 | 0.626 | ||
| Hip Arthroplasty | 517(97.18%) | 57(98.28%) | ||
| PFNA | 15(2.82%) | 1(1.72%) | ||
| Preoperative Comorbidities | 3.874 | 0.275 | ||
| Hypertension | 100(18.80%) | 22(37.93%) | ||
| Cerebral Infarction | 25(4.70%) | 2(3.45%) | ||
| Diabetes | 18(3.38%) | 3(5.17%) | ||
| Coronary Heart Disease | 10(1.88%) | 0(0%) | ||
| One-Year Postoperative Survival Status | 254.012 | < 0.001 | ||
| Survived | 493(92.67%) | 8(13.79%) | ||
| Died | 39(7.33%) | 50(86.21%) |
BMI Body mass index, PSMC Postoperative serious medical complications, PFNA Proximal Femoral Nail Anti-rotation
Comparison of perioperative blood test parameters between postoperative patients with and without PSMC
Perioperative hematological parameters were compared between the non-PSMC group (n = 532) and PSMC group (n = 58) at preoperative stage and postoperative days 1, 3, 5 and 7 (Table 4). No significant intergroup differences were observed in red blood cells (RBC), hemoglobin (HB), total protein, albumin/globulin (A/G) ratio, creatinine, urea, procalcitonin and troponin at all time points (all p > 0.05).
Table 4.
Comparison of perioperative blood test parameters between postoperative patients with and without PSMC
| Non-PSMC Group (n = 532) | PSMC Group (n = 58) | Statistic Value | p Value | |
|---|---|---|---|---|
| RBC | 0.823 | 0.511 | ||
| Preoperative | 4.70(4.29, 5.05) | 4.63(4.28, 5.05) | ||
| Postoperative Day 1 | 4.05(3.80, 4.26) | 4.01(3.81, 4.26) | ||
| Postoperative Day 3 | 3.51(3.27, 3.71) | 3.50(3.31, 3.80) | ||
| Postoperative Day 5 | 3.72(3.45, 3.97) | 3.63(3.44, 3.95) | ||
| Postoperative Day 7 | 3.81(3.54, 4.08) | 3.79(3.60, 4.04) | ||
| HB | 1.092 | 0.360 | ||
| Preoperative | 140(132, 149) | 139(133, 152) | ||
| Postoperative Day 1 | 104(97, 112) | 104(97, 111) | ||
| Postoperative Day 3 | 92(85, 99) | 89(85, 95) | ||
| Postoperative Day 5 | 99(92, 107) | 99(91, 108) | ||
| Postoperative Day 7 | 104(98, 112) | 104(98, 114) | ||
| Total protein | 0.307 | 0.580 | ||
| Preoperative | 68.9(63.5, 75.0) | 69(59.1, 75) | ||
| Postoperative Day 1 | 70.4(67.5, 72.8) | 69(66.6, 72.8) | ||
| Postoperative Day 3 | 62.3(58.5, 65.7) | 63(58.9, 66.6) | ||
| Postoperative Day 5 | 57.1(53.4, 61) | 58.2(54.2, 61.4) | ||
| Postoperative Day 7 | 59.8(55.5, 63.6) | 58.9(55.3, 64.3) | ||
| Albumin | 4.666 | 0.001 | ||
| Preoperative | 47.5(43.8, 51.4) | 46.3(43.0, 50.8) | ||
| Postoperative Day 1 | 39.9(37.4, 42.2) | 38.2(35.5, 40.9) | ||
| Postoperative Day 3 | 34.8(32.5, 37.4) | 32.7(31.4, 34.1) | ||
| Postoperative Day 5 | 33.1(30.5, 35.4) | 31.0(29.1, 34.4) | ||
| Postoperative Day 7 | 33.6(31.2, 36) | 31.1(29.4, 34.3) | ||
| A/G ratio | 0.391 | 0.815 | ||
| Preoperative | 2.03(1.76, 2.27) | 1.98(1.62, 2.25) | ||
| Postoperative Day 1 | 2.01(1.75, 2.23) | 2.01(1.71, 2.30) | ||
| Postoperative Day 3 | 1.60(1.41, 1.81) | 1.60(1.35, 1.83) | ||
| Postoperative Day 5 | 1.41(1.20, 1.60) | 1.41(1.26, 1.60) | ||
| Postoperative Day 7 | 1.50(1.32, 1.68) | 1.47(1.29, 1.69) | ||
| Creatinine | 0.125 | 0.973 | ||
| Preoperative | 69.19(57.62, 80.70) | 68.89(57.26, 80.93) | ||
| Postoperative Day 1 | 115.93(108.33, 122.88) | 114.64(105.41, 120.64) | ||
| Postoperative Day 3 | 116.92(109.92, 124.89) | 117.03(110.03, 125.42) | ||
| Postoperative Day 5 | 109.55(101.95, 116.78) | 111.52(103.76, 118.91) | ||
| Postoperative Day 7 | 74.37(61.01, 86.13) | 74.78(65.05, 84.78) | ||
| Urea | 0.840 | 0.500 | ||
| Preoperative | 5.73(4.40, 6.93) | 5.67(3.76, 7.02) | ||
| Postoperative Day 1 | 8.58(7.68, 9.27) | 8.85(7.59, 9.57) | ||
| Postoperative Day 3 | 9.04(8.24, 9.75) | 9.24(8.39, 9.97) | ||
| Postoperative Day 5 | 8.06(7.34, 8.84) | 7.66(6.92, 8.94) | ||
| Postoperative Day 7 | 6.60(5.78, 7.30) | 6.86(6.24, 7.57) | ||
| Procalcitonin | 1.922 | 0.166 | ||
| Preoperative | 0.28(0.12, 0.44) | 0.11(0.03, 0.38) | ||
| Postoperative Day 1 | 1.23(0.82, 1.64) | 1.25(0.83, 1.60) | ||
| Postoperative Day 3 | 1.95(1.44, 2.51) | 2.39(1.77, 2.72) | ||
| Postoperative Day 5 | 0.94(0.62, 1.22) | 0.79(0.48, 1.11) | ||
| Postoperative Day 7 | 0.25(0.11, 0.37) | 0.28(0.13, 0.41) | ||
| Troponin | 0.072 | 0.789 | ||
| Preoperative | 0.02(0.01, 0.04) | 0.03(0.02, 0.04) | ||
| Postoperative Day 1 | 0.12(0.09, 0.16) | 0.10(0.08, 0.17) | ||
| Postoperative Day 3 | 0.19(0.13, 0.24) | 0.19(0.11, 0.24) | ||
| Postoperative Day 5 | 0.09(0.06, 0.12) | 0.09(0.06, 0.12) | ||
| Postoperative Day 7 | 0.02(0.01, 0.04) | 0.03(0.01, 0.04) | ||
| D-dimer | 3.511 | 0.008 | ||
| Preoperative | 0.47(0.21, 0.76) | 0.56(0.31, 0.81) | ||
| Postoperative Day 1 | 1.16(0.9, 1.57) | 1.36(1.04, 1.71) | ||
| Postoperative Day 3 | 3.46(2.74, 4.17) | 3.49(2.89, 4.34) | ||
| Postoperative Day 5 | 4.81(3.84, 5.85) | 5.83(4.51, 6.67) | ||
| Postoperative Day 7 | 2.83(2.37, 3.40) | 3.27(2.71, 3.66) | ||
| DAR | 3.757 | 0.005 | ||
| Preoperative | 0.01(0, 0.02) | 0.01(0.01, 0.02) | ||
| Postoperative Day 1 | 0.03(0.02, 0.04) | 0.04(0.03, 0.05) | ||
| Postoperative Day 3 | 0.10(0.08, 0.12) | 0.11(0.09, 0.13) | ||
| Postoperative Day 5 | 0.15(0.12, 0.18) | 0.18(0.15, 0.21) | ||
| Postoperative Day 7 | 0.09(0.07, 0.10) | 0.10(0.09, 0.12) |
RBC Red blood cells, HB Hemoglobin, DAR D-dimer to albumin ratio, A/G ratio Albumin-globulin ratio
Albumin levels were significantly lower in the PSMC group than in the non-PSMC group across the entire perioperative period (all p = 0.001). D-dimer levels in the PSMC group were markedly elevated compared with the non-PSMC group during the perioperative period (P = 0.008). Consistently, the D-dimer to albumin ratio (DAR) was significantly higher in the PSMC group at all perioperative time points, with a statistically significant difference (p = 0.005).
Preoperative hematological indices predicting PSMC: ROC analysis
This study evaluated the diagnostic performance of various preoperative hematological indices for the PSMC (Fig. 2). The diagnostic performance of preoperative hematological indices for predicting PSMC was assessed via area under the curve (AUC) with 95% confidence intervals (CIs), sensitivity and specificity (Table 5). Most single indices had poor predictive value for PSMC, with AUCs near 0.5 and non-significant P values, including red blood cells (RBC, AUC = 0.50), hemoglobin (HB, AUC = 0.52), total protein (AUC = 0.51), albumin/globulin (A/G) ratio (AUC = 0.52), creatinine (AUC = 0.53), urea (AUC = 0.53), procalcitonin (AUC = 0.55) and troponin (AUC = 0.50) (all p > 0.05). Albumin was the only significant single index (AUC = 0.69, 95%CI:0.62–0.77, p < 0.001), with a sensitivity of 53.45% and specificity of 80.26%. D-dimer showed good predictive efficacy (AUC = 0.80, 95%CI:0.74–0.86, p < 0.001), with 81.03% sensitivity and 72.74% specificity. The DAR exhibited favorable and comparable diagnostic performance (AUC = 0.82, 95%CI:0.76–0.89, p < 0.001), with identical sensitivity (81.03%) and specificity (72.74%) to D-dimer. Internal validation using bootstrap resampling (1,000 iterations) demonstrated that the optimism-corrected AUC for preoperative DAR was 0.81 for predicting PSMC, closely approximating the original estimates (0.82), indicating robust predictive performance without substantial overfitting.
Fig. 2.
Receiver operating characteristic (ROC) curves for various biomarkers in predicting PSMC. A ROC curve for red blood cell (RBC) count. B ROC curve for hemoglobin (HB). C ROC curve for total protein. D ROC curve for albumin. E ROC curve for albumin-to-globulin (A/G) ratio. F ROC curve for creatinine. G ROC curve for urea. H ROC curve for procalcitonin. I ROC curve for troponin. J ROC curve for D-dimer. K ROC curve for D-dimer-to-albumin ratio (DAR). AUC, area under the curve; CI, confidence interval
Table 5.
Diagnostic performance of preoperative hematological indices in predicting PSMC
| AUC(95%CI) | p value | Sensitivity | Specificity | |
|---|---|---|---|---|
| RBC | 0.50(0.43 to 0.57) | 0.96 | 62.07 (49.20% to 73.44%) | 49.25 (45.02% to 53.49%) |
| HB | 0.52(0.44 to 0.60) | 0.66 | 39.66 (28.09% to 52.51%) | 71.99 (68.03% to 75.64%) |
| Total protein | 0.51(0.43 to 0.58) | 0.90 | 44.83 (32.75% to 57.55%) | 60.71 (56.50% to 64.77%) |
| Albumin | 0.69(0.62 to 0.77) | < 0.001 | 53.45 (40.80% to 65.67%) | 80.26 (76.67% to 83.42%) |
| A/G ratio | 0.52(0.44 to 0.60) | 0.64 | 67.24 (54.42% to 77.92%) | 45.86 (41.68% to 50.11%) |
| Creatinine | 0.53(0.45 to 0.60) | 0.53 | 81.03 (69.15% to 89.07%) | 28.76 (25.08% to 32.75%) |
| Urea | 0.53(0.45 to 0.60) | 0.50 | 55.17 (42.45% to 67.25%) | 52.63 (48.39% to 56.84%) |
| Procalcitonin | 0.55(0.46 to 0.63) | 0.25 | 43.1 (31.18% to 55.88%) | 72.74 (68.81% to 76.35%) |
| Troponin | 0.50(0.42 to 0.58) | 0.94 | 32.76 (22.08% to 45.58%) | 70.68 (66.67% to 74.39%) |
| D-dimer | 0.80(0.74 to 0.86) | < 0.001 | 81.03 (69.15% to 89.07%) | 72.74 (68.81% to 76.35%) |
| Preoperative DAR | 0.82(0.76 to 0.89) | < 0.001 | 81.03 (69.15% to 89.07%) | 72.74 (68.81% to 76.35%) |
RBC Red blood cells, HB Hemoglobin, DAR D-dimer to albumin ratio. All sensitivity and specificity values were calculated using the optimal cutoff values determined by ROC analysis with the Youden index (maximizing sum of sensitivity and specificity)
Multivariate logistic regression analysis for PSMC
Multivariate logistic regression analysis adjusted for gender, age, BMI, operative duration, ASA classification, surgical approach and preoperative comorbidities demonstrated that perioperative high DAR was an independent risk factor for PSMC (OR = 26.82, 95%CI: 19.89–33.75, p < 0.001). Gender, age, BMI, operative duration, ASA classification, surgical approach and preoperative comorbidities were not independently associated with PSMC (all p > 0.05) (Table 6).
Table 6.
Multivariate logistic regression analysis of the factors associated with PSMC
| B | Std. Error | Beta | t | Sig. | 95%CI for B | |
|---|---|---|---|---|---|---|
| Gender | -0.02 | 0.05 | -0.03 | -0.46 | 0.65 | -0.12-0.07 |
| Age | 0.01 | 0.01 | 0.14 | 2.05 | 0.05 | 0.00 -0.02 |
| BMI | 0.01 | 0.01 | 0.05 | 0.77 | 0.44 | -0.01-0.02 |
| Operative Duration | -0.02 | 0.03 | -0.06 | -0.85 | 0.40 | -0.07-0.03 |
| ASA Classification | -0.02 | 0.04 | -0.04 | -0.59 | 0.56 | -0.10 -0.05 |
| Surgical Approach | 0.05 | 0.06 | 0.06 | 0.83 | 0.41 | -0.07 -0.16 |
| Preoperative Comorbidities | 0.34 | 0.22 | 0.10 | 1.54 | 0.13 | -0.10-0.78 |
| Preoperative DAR | 26.82 | 3.51 | 0.50 | 7.64 | < 0.001 | 19.89–33.75 |
DAR D-dimer to albumin ratio, ASA American Society of Anesthesiologists, BMI Body mass index, PSMC Postoperative serious medical complications
Comparison of baseline data and PSMC between survival and death groups
Based on the data presented in Table 7, a comparative analysis was conducted between the survival and death groups to evaluate baseline characteristics and the presence of PSMC. Among the 590 patients, 89 died within one year postoperatively. No statistically significant differences were observed between the two groups regarding gender, age, body mass index, or preoperative comorbidities, including hypertension, cerebral infarction, diabetes, and coronary heart disease (all p > 0.05). However, a marked disparity was noted in the incidence of PSMC, which occurred in 56.18% of patients in the death group compared with only 1.60% in the survival group (p < 0.001). These findings indicate that the occurrence of PSMC is strongly associated with increased mortality, despite comparable baseline profiles between the two groups.
Table 7.
Comparison of baseline data and postoperative serious medical complications between survival and death groups
| Survival Group(n = 501) | Death Group(n = 89) | Statistic Value | p Value | |
|---|---|---|---|---|
| Gender | 0.247 | 0.619 | ||
| Male | 211(42.12%) | 40(44.94%) | ||
| Female | 290(57.88%) | 49(55.06%) | ||
| Age | 74(69, 79) | 76(70, 82) | -1.543 | 0.123 |
| BMI(kg/m2) | 22.89(20.81, 25.31) | 22.68(21.32, 24.82) | -0.084 | 0.933 |
| Operative Duration (hours) | 2.62(2.25, 3.08) | 2.62(2.25, 3.08) | -0.039 | 0.969 |
| ASA Classification | 0.111 | 0.946 | ||
| 2 | 85(16.97%) | 14(15.73%) | ||
| 3 | 401(80.04%) | 72(80.90%) | ||
| 4 | 15(2.99%) | 3(3.37%) | ||
| Surgical Approach | 0.172 | 0.678 | ||
| Hip Arthroplasty | 488(97.41%) | 86(96.63%) | ||
| PFNA | 13(2.59%) | 3(3.37%) | ||
| Preoperative Comorbidities | ||||
| Hypertension | 105(20.96%) | 17(19.10%) | 0.159 | 0.690 |
| Cerebral Infarction | 24(4.79%) | 3(3.37%) | 0.349 | 0.555 |
| Diabetes | 16(3.19%) | 5(5.62%) | 1.294 | 0.255 |
| Coronary Heart Disease | 9(1.80%) | 1(1.12%) | 0.205 | 0.651 |
| Combined with PSMC | 254.012 | < 0.001 | ||
| No | 493(98.40%) | 39(43.82%) | ||
| Yes | 8(1.60%) | 50(56.18%) |
PFNA Proximal Femoral Nail Anti-rotation
Postoperative blood index dynamic changes and their relationship with patient prognosis
Based on the dynamic changes in postoperative blood indices presented in Table 8, no statistically significant differences were observed between the survival and death groups across the majority of hematological and biochemical parameters, including red blood cell count, hemoglobin, total protein, albumin, albumin-to-globulin ratio, creatinine, urea, procalcitonin, troponin, and D-dimer throughout the perioperative period (all p > 0.05). Notably, however, the D-dimer-to-albumin ratio (DAR) demonstrated a significant difference between the two groups (p = 0.007), suggesting its potential as a discriminating marker. Overall, while most laboratory parameters followed similar dynamic trajectories in both groups, the DAR emerged as a distinguishing factor associated with postoperative mortality.
Table 8.
Analysis of dynamic changes in postoperative blood indices between the survival and death groups
| Survival Group(n = 501) | Death Group(n = 89) | Statistic Value | p Value | |
|---|---|---|---|---|
| RBC | 1.328 | 0.250 | ||
| Preoperative | 4.72(4.30, 5.06) | 4.63(4.34, 5.04) | ||
| Postoperative Day 1 | 4.05(3.80, 4.26) | 3.99(3.80, 4.27) | ||
| Postoperative Day 3 | 3.52(3.28, 3.73) | 3.47(3.29, 3.71) | ||
| Postoperative Day 5 | 3.73(3.44, 3.96) | 3.68(3.53, 3.98) | ||
| Postoperative Day 7 | 3.80(3.54, 4.08) | 3.85(3.55, 4.06) | ||
| HB | 0.958 | 0.328 | ||
| Preoperative | 140(132, 150) | 139(132, 154) | ||
| Postoperative Day 1 | 105(97, 112.5) | 104(97, 111) | ||
| Postoperative Day 3 | 92(85, 99) | 92(85.5, 97) | ||
| Postoperative Day 5 | 99(92, 107) | 100(93, 108) | ||
| Postoperative Day 7 | 104(98, 113) | 103(96, 111) | ||
| Total protein | 0.184 | 0.668 | ||
| Preoperative | 68.3(63.5, 75) | 69.00(62.35, 75.50) | ||
| Postoperative Day 1 | 70.4(67.5, 72.9) | 69.90(67.35, 71.90) | ||
| Postoperative Day 3 | 62.1(58.5, 65.6) | 63.10(59.85, 66.85) | ||
| Postoperative Day 5 | 57.0(53.3, 60.8) | 58.20(55.40, 61.90) | ||
| Postoperative Day 7 | 59.8(55.6, 63.55) | 59.00(55.55, 63.65) | ||
| Albumin | 0.012 | 0.911 | ||
| Preoperative | 47.50(43.95, 51.40) | 46.20(42.70, 49.70) | ||
| Postoperative Day 1 | 39.80(37.40, 42.25) | 39.60(36.55, 41.15) | ||
| Postoperative Day 3 | 34.8(32.5, 37.35) | 33.70(32.00, 36.35) | ||
| Postoperative Day 5 | 33.10(30.40, 35.40) | 32.50(30.20, 35.45) | ||
| Postoperative Day 7 | 33.80(30.40, 35.40) | 32.50(30.20, 34.50) | ||
| A/G ratio | 0.009 | 0.925 | ||
| Preoperative | 2.03(1.75, 2.28) | 2.02(1.80, 2.21) | ||
| Postoperative Day 1 | 2.01(1.75, 2.24) | 2.07(1.77, 2.26) | ||
| Postoperative Day 3 | 1.60(1.40, 1.81) | 1.60(1.42, 1.83) | ||
| Postoperative Day 5 | 1.41(1.21, 1.60) | 1.424(1.20, 1.61) | ||
| Postoperative Day 7 | 1.50(1.33, 1.68) | 1.47(1.26, 1.67) | ||
| Creatinine | 0.007 | 0.933 | ||
| Preoperative | 69.19(57.56, 80.74) | 69.31(58.84, 80.88) | ||
| Postoperative Day 1 | 115.51(107.89, 122.50) | 116.39(108.98, 123.38) | ||
| Postoperative Day 3 | 117.06(110.04, 124.96) | 116.66(109.55, 124.61) | ||
| Postoperative Day 5 | 109.78(102.402.04, 117.24) | 109.82(101.94, 116.33) | ||
| Postoperative Day 7 | 74.06(61.27, 85.79) | 74.34(62.29, 87.22) | ||
| Urea | 0.027 | 0.869 | ||
| Preoperative | 5.72(4.37, 7.00) | 5.92(4.57, 6.88) | ||
| Postoperative Day 1 | 8.57(7.71, 9.30) | 8.64(7.59, 9.26) | ||
| Postoperative Day 3 | 9.03(8.25, 9.75) | 9.23(8.34, 9.77) | ||
| Postoperative Day 5 | 8.05(7.24, 8.84) | 8.04(7.37, 8.88) | ||
| Postoperative Day 7 | 6.60(5.81, 7.30) | 6.64(5.78, 7.31) | ||
| Procalcitonin | 0.604 | 0.438 | ||
| Preoperative | 0.28(0.12, 0.44) | 0.21(0.08, 0.41) | ||
| Postoperative Day 1 | 1.23(0.82, 1.63) | 1.24(0.90, 1.64) | ||
| Postoperative Day 3 | 1.97(1.46, 2.50) | 2.01(1.39, 2.64) | ||
| Postoperative Day 5 | 0.92(0.61, 1.20) | 0.97(0.72, 1.28) | ||
| Postoperative Day 7 | 0.25(0.12, 0.37) | 0.24(0.11, 0.38) | ||
| Troponin | 0.008 | 0.930 | ||
| Preoperative | 0.02(0.01, 0.04) | 0.02(0.01, 0.04) | ||
| Postoperative Day 1 | 0.12(0.09, 0.16) | 0.12(0.08, 0.17) | ||
| Postoperative Day 3 | 0.18(0.13, 0.24) | 0.19(0.13, 0.24) | ||
| Postoperative Day 5 | 0.09(0.06, 0.12) | 0.09(0.06, 0.12) | ||
| Postoperative Day 7 | 0.02(0.01, 0.04) | 0.02(0.01, 0.04) | ||
| D-dimer | 1.544 | 0.215 | ||
| Preoperative | 0.40(0.19, 0.72) | 0.77(0.57, 0.89) | ||
| Postoperative Day 1 | 1.17(0.90, 1.57) | 1.23(0.92, 1.62) | ||
| Postoperative Day 3 | 3.40(2.70, 4.15) | 3.56(2.96, 4.33) | ||
| Postoperative Day 5 | 4.88(3.84, 5.90) | 4.94(4.00, 6.09) | ||
| Postoperative Day 7 | 2.81(2.33, 3.38) | 3.27(2.72, 3.67) | ||
| DAR | 7.372 | 0.007 | ||
| Preoperative | 0.01(0.00, 0.02) | 0.02(0.01, 0.02) | ||
| Postoperative Day 1 | 0.03(0.02, 0.04) | 0.03(0.02, 0.04) | ||
| Postoperative Day 3 | 0.10(0.08, 0.12) | 0.11(0.09, 0.13) | ||
| Postoperative Day 5 | 0.15(0.12, 0.18) | 0.15(0.12, 0.19) | ||
| Postoperative Day 7 | 0.08(0.07, 0.10) | 0.09(0.08, 0.11) |
Predictive performance of blood indices for mortality risk within 1 year postoperatively
The diagnostic efficacy of preoperative hematological indices for predicting 1-year mortality was further evaluated, with the area under the curve (AUC) of receiver operating characteristic (ROC) curve as the main evaluation index (Fig. 3). Most preoperative indices showed poor predictive performance for 1-year mortality, with their AUC values close to 0.5 and no statistically significant difference (all p > 0.05). Among them, albumin presented a relatively higher AUC (0.56) yet still failed to reach statistical significance (p = 0.08), and all other indices including RBC, HB, total protein, A/G ratio, creatinine, urea, procalcitonin and troponin had AUC values ranging from 0.50 to 0.54, indicating no practical predictive value. In contrast, preoperative D-dimer and preoperative DAR exhibited favorable and statistically significant diagnostic performance for 1-year mortality (both p < 0.0001), with identical AUC values of 0.74 (95%CI: 0.69 to 0.79). Specifically, preoperative DAR had a higher sensitivity (85.39%) than D-dimer (79.78%), while D-dimer showed a slightly higher specificity (61.88%) compared with preoperative DAR (59.28%) (Table 9), suggesting both indices as effective preoperative predictors for 1-year mortality, with preoperative DAR having an advantage in identifying positive cases. Internal validation using bootstrap resampling (1,000 iterations) demonstrated that the optimism-corrected AUC for predicting 1-year mortality, closely approximating the original estimates (0.74), indicating robust predictive performance without substantial overfitting.
Fig. 3.
Receiver operating characteristic (ROC) curves of various blood biomarkers for predicting death within one year after surgery. A ROC curve of RBC. B ROC curve of HB. C ROC curve of Total protein. D ROC curve of Albumin. E ROC curve of A/G ratio. F ROC curve of Creatinine. G ROC curve of Urea. H ROC curve of Procalcitonin. I ROC curve of Troponin. J ROC curve of D-dimer. K ROC curve of DAR (D-dimer to albumin ratio). AUC, area under the curve; CI, confidence interval
Table 9.
Diagnostic performance of preoperative hematological indices in predicting mortality within 1 year
| AUC(95%CI) | p value | Sensitivity | Specificity | |
|---|---|---|---|---|
| RBC | 0.53(0.47 to 0.59) | 0.34 | 61.80(51.41% to 71.21%) | 51.30(46.93% to 55.65%) |
| HB | 0.50(0.44 to 0.57) | 0.84 | 21.35(14.11% to 30.95%) | 87.23(84.02% to 89.87%) |
| Total protein | 0.51(0.44 to 0.58) | 0.74 | 34.83(25.75% to 45.17%) | 74.46(68.38% to 76.18%) |
| Albumin | 0.56(0.49 to 0.62) | 0.08 | 76.40(66.61% to 84.02%) | 35.33(31.27% to 39.61%) |
| A/G ratio | 0.53(0.47 to 0.59) | 0.33 | 65.17(54.83% to 74.25%) | 45.11(40.81% to 49.49%) |
| Creatinine | 0.51(0.44 to 0.57) | 0.86 | 69.66(59.46% to 78.24%) | 38.32(34.17% to 42.65%) |
| Urea | 0.50(0.44 to 0.57) | 0.91 | 42.70(32.93% to 53.06%) | 63.87(59.58% to 67.96%) |
| Procalcitonin | 0.54(0.48 to 0.61) | 0.19 | 46.07(36.09% to 56.37%) | 66.07(61.81% to 70.08%) |
| Troponin | 0.51(0.44 to 0.58) | 0.72 | 14.61(8.74% to 23.40%) | 89.82(86.86% to 92.17%) |
| D-dimer | 0.74(0.69 to 0.79) | < 0.0001 | 79.78(70.28% to 86.81%) | 61.88(57.55% to 66.02%) |
| Preoperative DAR | 0.74(0.69 to 0.79) | < 0.0001 | 85.39(76.60% to 91.26%) | 59.28(54.92% to 63.50%) |
RBC Red blood cells, HB Hemoglobin, DAR D-dimer to albumin ratio. All sensitivity and specificity values were calculated using the optimal cutoff values determined by ROC analysis with the Youden index (maximizing sum of sensitivity and specificity)
Multivariate logistic regression analysis of 1-year mortality
After adjustment for gender, age, BMI, operative duration, ASA classification, surgical approach and preoperative comorbidities, multivariate logistic regression revealed that elevated preoperative DAR was an independent predictor of 1-year mortality (OR = 21.75, 95%CI: 14.41–29.09, p < 0.001). No significant association was observed between gender, age, BMI, operative duration, ASA classification, surgical approach, preoperative comorbidities and 1-year mortality (all p > 0.05) (Table 10).
Table 10.
Multivariate logistic regression analysis of the factors associated with one-year mortality
| B | Std. Error | Beta | t | Sig. | 95%CI for B | |
|---|---|---|---|---|---|---|
| Gender | -0.07 | 0.05 | -0.10 | -1.40 | 0.16 | -0.17-0.03 |
| Age | 0.01 | 0.01 | 0.04 | 0.54 | 0.59 | -0.01-0.01 |
| BMI | 0.01 | 0.01 | 0.01 | 0.07 | 0.94 | -0.01-0.02 |
| Operative Duration | 0.02 | 0.03 | 0.06 | 0.82 | 0.42 | -0.03-0.08 |
| ASA Classification | 0.00 | 0.04 | -0.01 | -0.08 | 0.94 | -0.08-0.08 |
| Surgical Approach | 0.01 | 0.06 | 0.01 | 0.15 | 0.88 | -0.11-0.13 |
| Preoperative Comorbidities | -0.17 | 0.24 | -0.05 | -0.74 | 0.46 | -0.64-0.29 |
| Preoperative DAR | 21.75 | 3.72 | 0.41 | 5.85 | < 0.001 | 14.41–29.09 |
Association between preoperative DAR and PSMC
A significant association was observed between the preoperative DAR and the occurrence of PSMC. When stratified by the optimal cutoff value derived from ROC analysis, the proportion of patients with a high preoperative DAR was significantly greater in the PSMC group compared with the non-PSMC group (χ2 = 45.933, p < 0.001). Similarly, analysis by preoperative DAR quartiles revealed a progressively higher prevalence of PSMC across increasing quartiles, with the highest proportion observed in the fourth quartile (χ2 = 66.820, p < 0.001) (Table 11). Baseline characteristics did not differ significantly between high-DAR and low-DAR groups (Supplementary Table 1). These findings indicate that elevated preoperative DAR is strongly associated with the occurrence of PSMC.
Table 11.
Association between preoperative DAR and PSMC
| Non-PSMC Group (n = 532) | PSMC Group (n = 58) | χ2 | p value | |
|---|---|---|---|---|
| Preoperative DAR | 45.933 | < 0.001 | ||
| Low(n = 532) | 304(57.14%) | 228(42.86%) | ||
| High(n = 58) | 6(10.34%) | 52(89.66%) | ||
| Preoperative DAR(Quartile) | 66.820 | < 0.001 | ||
| Q1(n = 147) | 143(97.28%) | 4(2.72%) | ||
| Q2(n = 148) | 146(98.65%) | 2(1.35%) | ||
| Q3(n = 148) | 135(91.22%) | 13(8.78%) | ||
| Q4(n = 147) | 108(73.47%) | 39(26.53%) |
The cutoff value for preoperative DAR (0.01447) was derived from ROC analysis for predicting PSMC using the Youden index
Association between preoperative DAR and one-year mortality
A strong association was identified between the preoperative DAR and one-year postoperative mortality. Using the optimal cutoff value determined by ROC analysis, the proportion of patients with a high preoperative DAR was significantly greater in the death group than in the survival group (χ2 = 60.490, p < 0.001). This association was further supported by quartile analysis, which demonstrated a progressive increase in mortality rates across ascending preoperative DAR quartiles, with the highest proportion of deaths observed in the fourth quartile (χ2 = 57.921, p < 0.001) (Table 12). Baseline characteristics did not differ significantly between high-DAR and low-DAR groups (Supplementary Table 2). Collectively, these results indicate that elevated preoperative DAR is significantly associated with increased one-year mortality risk.
Table 12.
Association between preoperative DAR and one-year mortality
| Survival Group(n = 501) | Death Group(n = 89) | χ2 | p value | |
|---|---|---|---|---|
| Preoperative DAR | 60.490 | < 0.001 | ||
| Low | 297(95.81%) | 13(4.19%) | ||
| High | 204(72.86%) | 76(27.14%) | ||
| Preoperative DAR(Quartile) | 57.921 | < 0.001 | ||
| Q1(n = 147) | 141(95.92%) | 6(4.08%) | ||
| Q2(n = 148) | 141(95.92%) | 7(4.76%) | ||
| Q3(n = 148) | 117(79.59%) | 31(21.09%) | ||
| Q4(n = 147) | 102(69.39%) | 45(30.61%) |
The cutoff value for preoperative DAR (0.01066) was derived from ROC analysis for predicting one-year mortality using the Youden index
Discussion
This study systematically explored the value of the perioperative DAR in predicting PSMC and one-year mortality in elderly hip fracture patients through a large sample retrospective analysis. The results showed that the incidence of PSMC and one-year mortality in the high DAR group were significantly higher than in the low DAR group. ROC curve analysis further demonstrated that DAR exhibited good predictive power for PSMC (AUC = 0.82), and for one-year mortality, D-dimer and DAR both had an AUC of 0.74, with DAR exhibiting higher sensitivity (85.39%), showing higher AUC values than other routine markers.
Pathophysiological basis and comprehensive predictive advantage of DAR
D-dimer, a marker of fibrin degradation, reflects a hypercoagulable state and secondary fibrinolysis, commonly seen in trauma, surgical stress, and inflammatory responses [28]. Albumin is not only an indicator of nutritional status but also a negative acute-phase protein; its decreased levels often signal exacerbated systemic inflammatory responses and impaired hepatic synthetic function [29]. This study showed that postoperative D-dimer levels significantly increased, peaking on the 5th postoperative day, while albumin levels exhibited a downward trend post-surgery, with the lowest values observed on the 3rd and 5th days, showing changes consistent with concurrent inflammation, hypercoagulability, and reduced nutritional status. DAR organically combines these two dimensions, providing a composite marker that reflects multi-system pathological states, thus demonstrating comparable sensitivity and specificity in predicting complex clinical outcomes.
In recent years, composite biomarkers have gained increasing attention due to their integrated advantages in prognostic evaluation. For instance, the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) have been widely used to assess inflammation and prognostic risks [30]. However, DAR is particularly suited for elderly surgical patients as it encompasses three key mechanisms: coagulation activation, inflammation response, and nutritional status. Wang et al. [31] demonstrated that DAR has strong predictive power for preoperative deep vein thrombosis (DVT) in elderly hip fracture patients compared to other biomarkers (AUC = 0.677). While AUC values were similar between DAR and D-dimer for both endpoints, DAR showed higher sensitivity for 1-year mortality, favoring its use for risk screening. Unlike D-dimer alone, DAR integrates coagulation, inflammation, and nutritional status, providing a more holistic assessment in frail elderly patients. We did not compare DAR with established clinical prediction models for hip fracture prognosis, which represents a limitation to be addressed in future studies.
Correlation between preoperative DAR and PSMC
This study systematically analyzed perioperative data from 590 elderly hip fracture patients, revealing a significant correlation between the preoperative DAR and PSMC. The results showed that the incidence of PSMC in the high DAR group was significantly higher than in the low DAR group (χ2 = 45.933, p < 0.001), and ROC curve analysis indicated that preoperative DAR demonstrated good predictive ability for PSMC [31]. This finding not only validated the clinical value of preoperative DAR in identifying high-risk patients but also provided strong evidence for its use as a predictive biomarker.
From a pathophysiological perspective, the integrative nature of preoperative DAR is closely related to the occurrence of PSMC. D-dimer, a product of fibrin degradation, reflects a hypercoagulable state and fibrinolysis, commonly observed in postoperative stress, inflammation, and endothelial injury [28]. Elderly hip fracture patients are at higher risk for blood hypercoagulability due to surgical trauma, immobility, and underlying comorbidities (e.g., hypertension, diabetes), which increase the risk of thrombosis and subsequent cardiovascular and cerebrovascular events [17]. On the other hand, albumin, as a negative acute-phase protein, reflects not only malnutrition but also intensified systemic inflammation and impaired hepatic function [29]. Hypoalbuminemia is closely associated with endothelial dysfunction, enhanced oxidative stress, and decreased immune function, all of which are significant promoters of PSMC [32, 33].
This study’s multi-timepoint monitoring found that DAR was significantly higher in the PSMC group at all perioperative time points (preoperative and postoperative days 1, 3, 5, 7) (p = 0.005). The preoperative DAR gradually increased post-surgery, peaking on the fifth day (0.15 in non-PSMC group, 0.18 in PSMC group), PSMC occurred at this high-risk period, suggesting that dynamic changes in preoperative DAR can guide clinical intervention timing. Moreover, although D-dimer and albumin each had predictive value (D-dimer AUC = 0.80, albumin AUC = 0.69), DAR showed comparable diagnostic performance (AUC = 0.82), with identical sensitivity (81.03%) and specificity (72.74%) to D-dimer. This suggests that DAR, by integrating both coagulation and inflammation-nutrition pathways, provides a more comprehensive risk assessment, particularly for elderly patients with multi-morbidity and overlapping risks [23]. The independent prognostic value of DAR was verified by multivariate analyses, which showed DAR was significantly associated with PSMC and 1-year mortality irrespective of baseline characteristics.
Predictive value of preoperative DAR for long-term survival
This study, through large-sample retrospective analysis, systematically evaluated the predictive value of preoperative DAR for one-year mortality risk in elderly hip fracture patients. The results indicated that the one-year mortality rate in the high DAR group was significantly higher than in the low DAR group (χ2 = 60.490, p < 0.001). ROC curve analysis showed that for 1-year mortality, D-dimer and DAR both had an AUC of 0.74 (all p < 0.0001), with DAR exhibiting higher sensitivity (85.39%) and specificity of 59.28%. This finding highlights preoperative DAR’s potential utility in long-term prognostic assessment for elderly hip fracture patients and provides a simple, efficient tool for clinical risk stratification. From a pathophysiological perspective, preoperative DAR’s ability to predict mortality stems from its simultaneous reflection of three key pathological processes: hypercoagulability, inflammation, and malnutrition, all of which contribute to organ function decline and increased mortality risk in elderly patients [34, 35]. Elevated D-dimer suggests sustained hypercoagulability and microthrombosis, potentially causing microcirculatory disturbances in vital organs such as the heart, brain, and kidneys, leading to multi-organ failure [31, 36]. The decrease in albumin reflects worsening nutritional status and increased inflammatory burden, leading to immune dysfunction, impaired tissue repair, and increased infection risk [37, 38]. Dynamic monitoring data from this study showed that preoperative DAR was significantly higher in the death group than the survival group (p = 0.007), and perioperative dynamic DAR levels also showed significant differences between the two groups, indicating that early postoperative DAR elevation is closely associated with persistent inflammation-hypercoagulability-malnutrition, affecting long-term survival.
Clinical applicability and limitations of DAR
DAR is a readily available and cost-effective biomarker calculated from routine coagulation and biochemical tests, without additional assays, making it convenient for clinical use and suitable for promotion in various medical settings, especially resource-limited hospitals. Dynamic perioperative monitoring of DAR may help identify high-risk patients early and guide personalized interventions, such as intensified anticoagulation, nutritional support, and anti-inflammatory management. However, this study has several limitations that should be acknowledged. First, this is a retrospective, single-center study, which may lead to selection bias and residual confounding. Although we have adjusted for key perioperative variables including gender, age, BMI, ASA classification, surgical approach, operative duration, and preoperative comorbidities in multivariate logistic regression models, some unmeasured confounders may still exist. Second, the relatively large odds ratios for DAR should be interpreted cautiously. They likely reflect the integrated burden of hypercoagulability, severe systemic inflammation, and malnutrition captured by DAR, rather than a direct causal effect. This composite biomarker identifies a subset of patients with extremely high risk of adverse outcomes, consistent with the high-risk nature of elderly hip fracture populations. Dose-response relationships across DAR quartiles and bootstrap validation supported the stability of these prognostic associations, reducing concern that they arise solely from statistical bias. Third, the predictive performance of DAR was similar to that of D-dimer alone, with comparable AUC values for both PSMC and 1-year mortality. Notably, DAR showed higher sensitivity for 1-year mortality (85.39% vs. 79.78%), which is beneficial for screening high-risk individuals. However, we did not compare the predictive value of DAR with established clinical prediction models for elderly hip fracture patients, which limits the assessment of its incremental clinical benefit. Future studies should compare DAR with existing scoring systems to clarify its added value. Fourth, the optimal cutoff values of DAR were derived from this cohort using the Youden index and may vary across populations and laboratories due to differences in detection reagents, instruments, and patient characteristics. Standardization of DAR detection and cutoff values is needed before widespread clinical application. Fifth, although internal validation using bootstrap resampling confirmed the robustness of DAR in predicting PSMC and 1-year mortality, external validation in independent, prospective, multicenter cohorts is strongly warranted before clinical implementation of the DAR cutoff values. The current results should be considered hypothesis-generating rather than definitive evidence. Sixth, anesthesia type and anticoagulant agents were uniform across all patients (general anesthesia and low-molecular-weight heparin 4000 IU daily), thus eliminating confounding from these perioperative variables. Finally, this study only evaluated the predictive value of DAR, but whether interventions based on DAR stratification can improve clinical outcomes remains unclear. Future prospective interventional studies are needed to explore whether DAR-guided risk management can reduce the incidence of PSMC and improve long-term survival in elderly patients with hip fracture.
Conclusion and outlook
In conclusion, perioperative DAR is a hypothesis-generating prognostic biomarker for PSMC and 1-year mortality in elderly hip fracture patients. High DAR indicates poor prognosis and may help identify high-risk individuals for targeted intervention. Future multicenter, prospective studies are needed to validate DAR cutoff values, compare its performance with clinical prediction models, and explore whether DAR-guided interventions improve patient outcomes.
Supplementary Information
Authors’ contributions
Yang Liu: Writing-Original Draft, Methodology, Resources. Chao Wang: Methodology, Software, Data Curation. Ruyi Xiang: Methodology, Software, Data Curation, Project administration. Qi Wang: Methodology, Software, Data Curation. Jiangang Zhang: Conceptualization, Supervision, Project administration, Funding acquisition. All authors have read and agreed to the published version of the manuscript.
Funding
The study was supported by Science Foundation of Kangda College of Nanjing Medical University (No. KD2023KYJJ231).
Data availability
The datasets generated during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This retrospective observational study was conducted in strict accordance with the Declaration of Helsinki. The study protocol was reviewed and approved by the Ethics Committee of The Affiliated Jiangning Hospital of Nanjing Medical University (Approval No.: 2023-03-090-K01). All participants or their legal guardians provided written informed consent prior to data collection. All clinical data included in this study were fully anonymized and de-identified before statistical analysis to protect patient privacy and confidentiality. All patient data were exclusively utilized for this academic research and were not used for any commercial purposes.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yang Liu, Chao Wang and Ruyi Xiang contributed equally to this work and should be considered co-first authors.
Contributor Information
Qi Wang, Email: 2486435394@qq.com.
Jiangang Zhang, Email: zhangjiangangNJ@163.com.
References
- 1.Ory MG, Smith ML. Editorial: Insights in Aging and Public Health: 2021. Front Public Health. 2022;10:929418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bingol I, et al. The epidemiology of geriatric fractures: A nationwide analysis of 1 million fractures. Injury. 2024;55(11):111900. [DOI] [PubMed] [Google Scholar]
- 3.Bhushan S, et al. The impact of regional versus general anesthesia on postoperative neurocognitive outcomes in elderly patients undergoing hip fracture surgery: A systematic review and meta-analysis. Int J Surg. 2022;105:106854. [DOI] [PubMed] [Google Scholar]
- 4.Zhong BX, et al. Physical performance and risk of hip fracture in community-dwelling elderly people in China: A 4-year longitudinal cohort study. Maturitas. 2021;146:26–33. [DOI] [PubMed] [Google Scholar]
- 5.Song Y, et al. Prediction models for postoperative delirium in elderly patients with machine-learning algorithms and SHapley Additive exPlanations. Transl Psychiatry. 2024;14(1):57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zhang C, et al. Incidence of and trends in hip fracture among adults in urban China: A nationwide retrospective cohort study. PLoS Med. 2020;17(8):e1003180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yang Y, et al. Inpatient cost of treating osteoporotic fractures in mainland China: a descriptive analysis. Clinicoecon Outcomes Res. 2015;7:205–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gidwani S, et al. The detection of patients with ‘fragility fractures’ in fracture clinic - an audit of practice with reference to recent British Orthopaedic Association guidelines. Ann R Coll Surg Engl. 2007;89(2):147–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Schousboe JT. Mortality After Osteoporotic Fractures: What Proportion Is Caused by Fracture and Is Preventable? J Bone Min Res. 2017;32(9):1783–8. [DOI] [PubMed] [Google Scholar]
- 10.Griffiths R, et al. Guideline for the management of hip fractures 2020: Guideline by the Association of Anaesthetists. Anaesthesia. 2021;76(2):225–37. [DOI] [PubMed] [Google Scholar]
- 11.Katsoulis M, et al. Excess mortality after hip fracture in elderly persons from Europe and the USA: the CHANCES project. J Intern Med. 2017;281(3):300–10. [DOI] [PubMed] [Google Scholar]
- 12.Leu TH, et al. Incidence and excess mortality of hip fracture in young adults: a nationwide population-based cohort study. BMC Musculoskelet Disord. 2016;17:326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Melton LJ 3, et al. Long-term mortality following fractures at different skeletal sites: a population-based cohort study. Osteoporos Int. 2013;24(5):1689–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Simunovic N, et al. Effect of early surgery after hip fracture on mortality and complications: systematic review and meta-analysis. CMAJ. 2010;182(15):1609–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen W, et al. Serum D-dimer, albumin and systemic inflammatory response markers in ovarian clear cell carcinoma and their prognostic implications. J Ovarian Res. 2020;13(1):89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wells PS, et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med. 2003;349(13):1227–35. [DOI] [PubMed] [Google Scholar]
- 17.You D, et al. Identification of hypercoagulability with thrombelastography in patients with hip fracture receiving thromboprophylaxis. Can J Surg. 2021;64(3):E324–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.He R, et al. Factors associated with 90-day acute ischemic stroke in patients >/=70 years old with total hip arthroplasty for hip fracture. BMC Geriatr. 2022;22(1):28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Qin Z, et al. Systemic Immune-Inflammation Index Is Associated With Increased Urinary Albumin Excretion: A Population-Based Study. Front Immunol. 2022;13:863640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wu B, et al. Inflammation-Targeted Drug Delivery Strategies via Albumin-Based Systems. ACS Biomater Sci Eng. 2024;10(2):743–61. [DOI] [PubMed] [Google Scholar]
- 21.Sim SD, et al. Preoperative hypoalbuminemia: Poor functional outcomes and quality of life after hip fracture surgery. Bone. 2021;143:115567. [DOI] [PubMed] [Google Scholar]
- 22.Lizaur-Utrilla A, et al. Altered seric levels of albumin, sodium and parathyroid hormone may predict early mortality following hip fracture surgery in elderly. Int Orthop. 2019;43(12):2825–9. [DOI] [PubMed] [Google Scholar]
- 23.Wu W, et al. The D-Dimer/Albumin Ratio Is a Prognostic Marker for Aneurysmal Subarachnoid Hemorrhage. Brain Sci. 2022;12(12):1700. [DOI] [PMC free article] [PubMed]
- 24.Lawrence VA, et al. Medical complications and outcomes after hip fracture repair. Arch Intern Med. 2002;162(18):2053–7. [DOI] [PubMed] [Google Scholar]
- 25.Zhou Y, et al. Modified Frailty Index Combined with a Prognostic Nutritional Index for Predicting Postoperative Complications of Hip Fracture Surgery in Elderly. J Invest Surg. 2022;35(10):1739–46. [DOI] [PubMed] [Google Scholar]
- 26.Peduzzi P, et al. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373–9. [DOI] [PubMed] [Google Scholar]
- 27.Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol. 2007;165(6):710–8. [DOI] [PubMed] [Google Scholar]
- 28.Weitz JI, Fredenburgh JC, Eikelboom JW. A Test in Context: D-Dimer. J Am Coll Cardiol. 2017;70(19):2411–20. [DOI] [PubMed] [Google Scholar]
- 29.Wu N, et al. Albumin, an interesting and functionally diverse protein, varies from ‘native’ to ‘effective’ (Review). Mol Med Rep. 2024;29(2):24. [DOI] [PMC free article] [PubMed]
- 30.Wei Y, et al. Neutrophil/lymphocyte, platelet/lymphocyte and monocyte/lymphocyte ratios in patients with affective disorders. J Affect Disord. 2022;309:221–8. [DOI] [PubMed] [Google Scholar]
- 31.Yao W, et al. D-dimer-albumin ratio (DAR) as a new biomarker for predicting preoperative deep vein thrombosis after geriatric hip fracture patients. J Orthop Surg Res. 2023;18(1):645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Soeters PB, Wolfe RR, Shenkin A. Hypoalbuminemia: Pathogenesis and Clinical Significance. JPEN J Parenter Enter Nutr. 2019;43(2):181–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Karki S, et al. Association of Hypoalbuminemia With Clinical Outcomes in Patients Admitted With Acute Heart Failure. Curr Probl Cardiol. 2023;48(11):101916. [DOI] [PubMed] [Google Scholar]
- 34.Lin GS, et al. Value of the Preoperative D-Dimer to Albumin Ratio for Survival and Recurrence Patterns in Gastric Cancer. Ann Surg Oncol. 2023;30(2):1132–44. [DOI] [PubMed] [Google Scholar]
- 35.Pan Z, et al. Prognostic value of preoperative D-dimer to albumin ratio in patients with locally advanced rectal cancer after neoadjuvant chemoradiotherapy. Discov Oncol. 2024;15(1):746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Zhang W, et al. Association of D-dimer elevation with inflammation and organ dysfunction in ICU patients with COVID-19 in Wuhan, China: a retrospective observational study. Aging. 2021;13(4):4794–810. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kaysen GA, et al. Relationships among inflammation nutrition and physiologic mechanisms establishing albumin levels in hemodialysis patients. Kidney Int. 2002;61(6):2240–9. [DOI] [PubMed] [Google Scholar]
- 38.Eckart A, et al. Relationship of Nutritional Status, Inflammation, and Serum Albumin Levels During Acute Illness: A Prospective Study. Am J Med. 2020;133(6):713–22. e7. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated during the current study are available from the corresponding author on reasonable request.