Geriatric hip fracture is a common complication of osteoporosis and a major problem worldwide. The total number of cases is expected to rise to approximately 2.6 million by 2025, with an increase to 7.3-21.3 million by 2050 (1, 2). The reported 1-year mortality rate is 22% (3). Therefore, the apparent consensus among researchers and surgeons is that as life expectancy improves and the overall age of the population increases, the associated burden on health services will rise in conjunction with the projected increase in the elderly population (4–7).

Many factors can predict mortality in patients with geriatric hip fractures including age, serum albumin, sodium, hemoglobin, arrhythmia, pneumonia, and heart failure (8, 9). A systematic review of preoperative predictors of mortality identified strong evidence for 12 predictors, including advanced age, male sex, poor preoperative ambulation status, higher American Society of Anesthesiologists (ASA) status, dementia, diabetes, cancer, cardiac disease, and multiple comorbidities (10).

Although there has been extensive focus and research on the predictors of post-hip fracture mortality, research regarding the use of risk prediction models is relatively limited. Previous studies have shown the Charlson comorbidity index (CCI) to predict postoperative complications in patients with surgically treated hip fractures (11) as well as mortality at 1-year follow-up (12). As patient age was subsequently determined to be correlated with prognosis (8, 9), the age-adjusted Charlson comorbidity index (aCCI), a modified version of the CCI, was introduced into clinical practice in 1994. The aCCI incorporates age as a correction variable of the final score by adding 1 point for every decade over 40 years of age (13), therefore it is especially suitable for geriatric patient populations. Even though the Elixhauser comorbidity index (ECI) has been used to predict mortality (14), it was reported that CCI provided a better prediction for in-hospital mortality than the ECI (15). However, the role of comorbidities in hip fractures has not yet been thoroughly evaluated. A better understanding of comorbidities can promote recognition of their prognostic implications for hip fractures. Moreover, whether the aCCI shows predictive performance in geriatric patients with hip fractures requires further verification. Thus, the purpose of this study was to use aCCI scores to predict long-term mortality and, consequently, help reduce post-hip fracture mortality. We hypothesized that a higher aCCI score would be associated with higher postoperative mortality and that aCCI score at admission could predict prognosis.

The results of this study demonstrated a linear association between aCCI score and mortality after geriatric hip fracture, with a 31% increase in mortality (HR = 1.31, 95% CI: 1.21–1.41; P < 0.0001) and a one-unit increase in aCCI. Patients in Q2 showed greater mortality (HR = 1.60, 95% CI: 1.23–2.09; P = 0.0005) than those in Q1, and patients in Q3 showed greater mortality (HR = 2.18, 95% CI: 1.66–2.87; P < 0.001) than those in Q1. Clinically, we suggest using the aCCI score at admission to predict prognosis in elderly patients with hip fracture. By improving the management of potentially controllable disorders, we can improve patient survival rates.

The aCCI is a significant predictor of survival after cancer (19), infective endocarditis (20), transcatheter mitral valve repair (21), and COVID-19 pneumonia (22). Presently, there are numerous studies on the association between CCI and mortality in patients with hip fracture, all having reported that high CCI scores were associated with mortality or considered high CCI score to be a risk factor (8, 12, 23–25). In one study, Pan et al. (8) reported that the risk variables for mortality after hip fracture surgery in geriatrics were age, albumin level, sodium level, hemoglobin level, and CCI score (HR = 1.38) in 45 patients (8, 12). Additionally, Hjelholt et al. (26) developed a user-friendly prediction tool for 1-year mortality in patients with hip fractures. The final model included nursing home residency, CCI score, cumulative ambulation score, body mass index, and age; had acceptable discrimination and calibration; and predicted one-year mortality risk ranging from 5 to 91% depending on the combination of predictors in the individual patient.

Age (27), sex (28), fracture classification (27), time to operation (29), treatment strategy (30), and hospital stay (31) were reported as risk factors in previous studies. In addition, we considered the adjusted the factor of P < 0.1 in the univariate analysis: occupation, injury mechanism, time to admission, operation time and infusion. Thus, we comprehensively considered the variables that needed to be adjusted. Age is a very important factor that should be considered in geriatric hip fracture and adjusted during data analyses. In a cross-sectional study by Padrón-Monedero et al. (32), the results showed an association between age and mortality following hip fracture after adjusting for numerous comorbidities. The aCCI considers age, whereas the CCI does not; aCCI, which considered old age a risk factor, was more accurate. In their study, Jiang et al. (33) retrospectively assessed the association between the aCCI and 5-year mortality in a surgically treated hip fracture population of 1,057 patients. The results demonstrated that patients with aCCI ≥ 6 had an increased 5-year mortality rate with an odds ratio of 13.6 compared to those with aCCI ≤ 3. Moreover, a study by Gatot et al. (34) concluded that an aCCI ≥ 6 could predict higher 90-day readmission rates, poor quality of life, and poor potential for functional recovery 1-year post-operation in patients with hip fracture. As it used a prospective observational design, this study provided stronger association, allowing for better interpretation of mortality. In the present study, the longest follow-up was 84.19 months with an average of 38.9 months, which was longer than the 1 year used in a study by Chen et al. (35) and Garabano et al. (24), as well as the 2 years used in a study by Cher et al. (25), and 37.2 months in a study by Pan et al. (8). The sample size in our study was also larger than those used in these previous studies (8, 24, 25, 33, 35). In addition, the results of the sensitivity analysis in the present study were stable.

Comorbidities are quite common in elderly patients with hip fractures (36, 37), and the assessment of prognosis after injury is usually insufficient. The ASA is a typically utilized tool (38, 39); however, Varady et al. (40) found that the aCCI was more accurate than the ASA score for 1-year mortality after hip fracture surgery. The ECI is another popular tool for predicting the prognosis of hip fractures; however, Tang et al. (15) found that aCCI provided a better prediction of in-hospital mortality than ECI among elderly patients. Moreover, these data validate that aCCI can be reliably performed in the International Classification of Diseases-10 era. Because of the objective nature of these indices, the aCCI may be a useful preoperative measure for surgeons to assess mortality in hip fracture patients and should likely be used for institutional orthopedic research involving outcomes at 90 days and beyond (40).

To the best of our knowledge, this prospective study had the largest sample size used to date to explain the linear association between aCCI score and mortality in geriatric hip fractures. Our findings provide new insights into the association between aCCI score and mortality and contribute to clinical evidence on the use of comorbidities at admission in predicting the prognosis of elderly patients with hip fractures. The relationship between aCCI score and mortality was shown to be linear rather than non-linear, and there was no threshold or saturation point. Therefore, mortality risk increased following the addition of the aCCI. The higher the aCCI score was, the poorer the result. Clinically, we suggest that surgeons calculate aCCI scores at admission and use it to predict prognosis in the median term.

This was a prospective study with a large sample size, and although the loss-to-follow-up rate was 18%, we found that the patients who were lost to follow-up were randomized, and most of the variables in Table 1 were comparable between the present and absent groups. Furthermore, we included patients who were admitted in September 2019. One important reason was to avoid the effect of COVID-19 (41, 42) and another was to ensure follow-up over 2 years. During the analysis, to explore the real relationship between the two factors, we not only carried out linear regression using various adjusted models but also changed the continuous variable of access to a categorical variable or performed a trend test for the result. In addition, we explored the association with the curve relationship and found no threshold or saturation effect, which supplemented the stability of the linear association. We also performed sensitivity analysis using PSM to test the robustness of our results; a total of 1448 patients were matched successfully and the results remained quite stable.

However, this study has some limitations. First, as it was a prospective study, loss to follow-up (18%) was inevitable. We attempted to reach the patients three times by phone to obtain outcome data. Second, the results only apply to patients aged ≥65 years and not to younger patients. Third, the samples of this study were from China; thus, the conclusions have certain regional and ethnic restrictions, and the inference points for other races should be redefined.

Clinically, we suggest that aCCI score at admission should be used to predict prognosis for elderly patients with hip fracture.

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by the Ethics Committee of the Honghui Hospital (No. 202201009). The patients/participants provided their written informed consent to participate in this study.

XX and B-FZ conceived and designed the study. B-FZ, Y-XC, and YZ analyzed the data. D-LZ and B-FZ wrote the manuscript. All authors contributed to the article and approved the submitted version.