Introduction

A demographic shift is currently underway worldwide, with a growing proportion of geriatric individuals in the population. This trend can be attributed to various factors, such as advancements in medicine and improved living conditions. In developed countries, the geriatric population is growing at a faster rate than the general population and is expected to continue to do so [1,2,3]. This demographic shift is also evident in the incidence of pelvic fractures, which has been steadily increasing, especially among older individuals [4,5,6]. This is particularly worrying considering the elevated risk of adverse outcomes after traumatic insults and subsequent surgical interventions in this patient population [7]. The mortality rate following low energy pelvic ring fractures can be as high as 27%, comparable to that following hip fracture in geriatric patients [5,6,7,8,9,10,11,12,13]. Despite the significant disease burden associated with pelvic fractures, they have not been as thoroughly studied in the geriatric population [5, 14].

Geriatric pelvic fracture patients present a heterogeneous patient population with varying prognoses following injury based on fracture pattern and preexisting comorbidities [5, 6, 8,9,10,11,12]. One potential avenue for guiding patient care is stratification of patients based on their intrinsic risk of adverse outcomes. Frailty, a condition characterized by a reduced physiological reserve to withstand external stressors [15,16,17,18], has previously been found to be relatively common in the geriatric trauma patients. It has also been associated with an increased risk of both morbidity and mortality [8, 19,20,21]. The aim of the current investigation was therefore to determine the association between frailty, measured using the Orthopedic Frailty Score (OFS), and adverse outcomes in geriatric pelvic fracture patients. We hypothesize that an increasing OFS is associated with increasing risk of morbidity and mortality in geriatric patients who have sustained a pelvic fracture.

Methods

The need for ethical approval for the current analysis was waived by the relevant ethical review board due to the retrospective, anonymized nature of the utilized dataset. All aspects of the current investigation adhered to the Declaration of Helsinki and STrengthening the Reporting of OBservational studies in Epidemiology guidelines [22]. Data were obtained from the American College of Surgeons’ 2013–2019 Trauma Quality Improvement Program (TQIP) database. This dataset allowed for the retrieval of variables pertaining to patient demographics, clinical characteristics, injury severity measured using the Abbreviated Injury Severity Score (AIS), interventions, and outcomes. All geriatric (65 years or older) patients who suffered an isolated pelvic fracture subsequent to a blunt trauma were eligible for inclusion. An isolated pelvic fracture was defined as any fracture in the ilium, ischium, pubis, sacrum, coccyx, or acetabulum with a lower extremity AIS ≥ 2, any abdomen AIS, and an AIS ≤ 1 in all other regions. Patients with any abdomen AIS were included in order to avoid excluding patients with intraabdominal vascular injuries resulting from the pelvic fracture. Patients were excluded if they had an AIS of 6 in the lower extremity or abdomen, as these injuries are generally not considered survivable.

Calculating the Orthopedic Frailty Score

The degree of frailty for each patient was estimated using the Orthopedic Frailty Score. This score makes use of five dichotomous variables: age ≥ 85 years old, institutionalization, non-independent functional status, congestive heart failure, and a history of malignancy (excluding non-invasive neoplasms of the skin). Institutionalization was defined as the admission origin being from an institution such as a nursing home, long-term care facility, or other group living arrangement. Patients received one point for each variable present, resulting in a maximum total score of 5 [23].

Statistical analysis

Patients were separated into groups based on their OFS: non-frail (OFS = 0), pre-frail (OFS = 1), and frail (OFS ≥ 2) [24]. Continuous variables, being non-normally distributed, were summarized as medians and interquartile ranges (IQRs), while categorical variables were presented as counts and percentages. Statistical analysis was performed using the Kruskal–Wallis test for continuous variables and either the Chi-squared test or Fisher’s exact test for categorical variables, as appropriate. The primary outcome of interest was in-hospital mortality, and secondary outcomes included in-hospital complications (myocardial infarction, cardiac arrest with CPR, stroke, deep vein thrombosis, pulmonary embolism, acute respiratory distress syndrome, urinary tract infection, pneumonia, surgical site infection, sepsis, decubitus ulcer, unplanned intubation, unplanned admission to the OR, unplanned admission to the ICU), failure-to-rescue (FTR), and ICU admission. FTR was defined as in-hospital mortality following a complication.

To assess the association between frailty and adverse outcomes, Poisson regression models were employed in order to minimize potential confounding. Explanatory variables included frailty (measured using the OFS), age, sex, race, highest AIS in each region, type of fracture, surgical and angiographic procedures, hypertension, previous myocardial infarction, history of peripheral vascular disease, cerebrovascular disease, dementia, chronic obstructive pulmonary disease, smoking status, chronic renal failure, diabetes mellitus, cirrhosis, coagulopathy, drug use disorder, alcohol use disorder, major psychiatric illness, and advanced directives limiting care. The response variable was either in-hospital mortality, complications, FTR, or ICU admission. Results were presented as adjusted incidence rate ratios (IRRs) and corresponding 95% confidence intervals (CIs) calculated using robust standard errors [25]. These analyses were also repeated in several subgroup analyses. The first subgroup consisted of all patients who suffered a pelvic fracture after a ground-level fall (GLF). The second subgroup consisted of patients who were managed conservatively, defined as not undergoing surgical fracture fixation, vascular surgery, any angiographic procedure, or laparotomy/pelvic packing for hemorrhage control. The final subgroup consisted of patients with an abdomen AIS ≤ 1.

A two-sided p-value < 0.05 was considered statistically significant. Missing data were managed using multivariate imputation by chained equations, with a total of seven complete datasets being generated, assuming data were missing at random. Analyses were performed using the tidyverse, parallel, mice, and sandwich packages in the statistical software R 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria) [26].

Results

Subsequent to applying the inclusion and exclusion criteria, 66,404 patients were included for further analysis. 52% (N = 34,292) were classified as non-frail, 32% (N = 21,467) were pre-frail, and 16% (N = 10,645) were classified as frail. Median age tended to increase with a higher OFS (pre-frail and frail vs non-frail: 85 [79–87] and 86 [84–88] vs 76 [70–81] years old, p < 0.001). Additionally, there was a significant difference in the distribution of sex and race among the three groups, with a higher proportion of females (pre-frail and frail vs. non-frail: 74.6% and 77.3% vs 66.1%, p < 0.001) and White patients (pre-frail and frail vs. non-frail: 89.5% and 91.9% vs 88.3%, p < 0.001) in the pre-frail and frail groups, compared to the non-frail group. The prevalence of all comorbidities was also highest among frail patients, apart from cirrhosis and substance use disorders (Table 1). Frail patients were generally more severely injured than pre-frail patients (Lower extremity AIS ≥ 3: 16.5% vs 14.5%, p < 0.001). Pre-frail and frail patients were also more likely to have suffered a pubis fracture (67.7% and 67.6% vs 62.5%, p < 0.001) and less likely to have suffered an acetabulum fracture (26.6% and 26.8% vs 34.7%, p < 0.001), compared to non-frail patients. Surgical and angiographic interventions were also significantly less common the higher the degree of frailty (Table 2).

Table 1 Demographics of geriatric patients with pelvic fractures
Full size table
Table 2 Clinical characteristics of geriatric patients with pelvic fractures
Full size table

Before adjusting for confounding variables, both pre-frail and frail patients had higher rates of in-hospital mortality (1.8% and 2.5% compared to 1.2%, p < 0.001), complications (4.2% and 4.5% compared to 3.9%, p = 0.015), and FTR (0.8% and 0.9% compared to 0.6%, p = 0.003), compared to non-frail patients (Table 3). After controlling for confounding variables, pre-frail patients exhibited a 56% higher risk of in-hospital mortality [adjusted IRR (95% CI): 1.56 (1.32–1.85), p < 0.001], a 20% higher risk of complications [adjusted IRR (95% CI): 1.20 (1.09–1.32), p < 0.001], a 44% greater risk of FTR [adjusted IRR (95% CI): 1.44 (1.12–1.84), p = 0.004], and an 8% increase in the risk of ICU admission [adjusted IRR (95% CI): 1.08 (1.02–1.14), p = 0.008], compared to non-frail patients. Furthermore, frailty was associated with an approximate doubling in the risk of in-hospital mortality [adjusted IRR (95% CI): 1.88 (1.54–2.30), p < 0.001], a 25% higher overall risk of complications [adjusted IRR (95% CI): 1.25 (1.10–1.42), p < 0.001], a 56% higher risk of FTR [adjusted IRR (95% CI): 1.56 (1.14–2.14), p = 0.006], and a 10% increased risk of ICU admission [adjusted IRR (95% CI): 1.10 (1.02–1.18), p = 0.014], compared to non-frail patients. The results did not differ significantly in any of the subgroup analyses (Table 4).

Table 3 Crude outcomes in geriatric patients with pelvic fractures
Full size table
Table 4 Association between frailty and adverse outcomes in geriatric patients with pelvic fractures
Full size table

Discussion

Ground level falls constitute the most common mechanism of injury in the United States, and geriatric patients are most prone to sustaining injury via this mechanism [27]. Pelvic fractures, thus, constitute a relatively common injury pattern in geriatric patients [9, 28]. Consequently, a significant proportion of geriatric patients with pelvic fractures can be classified as frail, with 1 out of 6 patients being considered frail in the current investigation using a large national trauma database. The results of the Poisson regression analysis indicate that frail geriatric pelvic fracture patients suffer from an elevated risk of in-hospital mortality, complications, FTR, and ICU admission, highlighting the need for targeting care towards this more vulnerable patient population.

Pelvic fractures present a heterogeneous patient population with varying prognoses dependent on factors such as the type of fracture and preexisting comorbidities. The original OFS was developed for hip fracture patients; however, the current study suggests that this frailty score may also be applicable to the pelvic fracture population, given the similarities in age, burden of comorbidities, and level of frailty [7, 8, 13, 23]. The OFS itself is an objective and simple tool that makes use of five dichotomous variables that can be easily retrieved from past medical records or the patient without additional blood tests or intraoperative data. This reduces the risk of human error and intercoder variability.

Frailty refers to a clinical condition in which patients have a reduced physiological capacity to cope with external stressors due to the decline of several organ systems. This can lead to an increased risk of postoperative complications and mortality. Although frailty is typically associated with aging, it's worth noting that it is a distinct process and not a guaranteed consequence of getting older [15,16,17,18]. Frailty has also been proposed as a measure in orthopedic trauma to ascertain priority for advanced measures, such as preoperative optimization, multidisciplinary care, and post discharge rehabilitative therapy [29,30,31]. Previously, frailty has been found to be an independent risk factor for suffering fractures, especially pelvic fractures [20, 21]. Frail pelvic fracture patients have also been observed to exhibit an increased risk of mortality and non-home discharge as well as lower functional status at discharge [8, 19, 20].

As demonstrated by current and previous studies, frailty is relatively common among patients with pelvic fractures [8, 19]. Simultaneously, the incidence of pelvic fractures in geriatric patients is increasing, in large part due to rising life expectancy [4,5,6]. As the trend of population aging is not expected to diminish in either the short or long term [1,2,3], this cohort will increasingly strain healthcare systems around the world. To address this issue, risk stratification tools, such as the OFS, are needed to optimally allocate resources and minimize added cost. Such an approach enables early intervention by identifying patients with an elevated risk of decline.

Given the paucity of evidence for how to optimize care in this population [5, 14], there is little consensus on how to best manage pelvic fracture patients [8, 32]. Taking into account the similarity of pelvic fracture to hip fracture patients [5,6,7,8,9,10,11,12,13], it may be worth investigating interventions which have proven effective in the later cohort. Orthogeriatric care models in particular have previously been proposed as a potential option [33,34,35]. These models have been found to offer a cost-efficient method for reducing length of stay, mortality, and the risk of delirium among hip fracture patients in multiple systematic reviews [36,37,38].

Previous research conducted by Ramser et al. and Ravindrarajah et al., employing the electronic Frailty Index (eFI), has indicated an association between frailty and an elevated risk of both short- and long-term mortality among patients with pelvic fractures [8, 20]. Moreover, the findings of Perea et al. revealed that patients with pelvic fractures classified as frail according to the Clinical Frailty Scale (CFS), exhibited a lower functional status upon discharge and were more likely to be transferred to a skilled nursing facility [19]. Using the 6-Item Modified Frailty Index (MF-6), Pean et al. investigated the impact of frailty on patients with pelvic fractures, as well as other lower extremity fractures, undergoing surgical fixation. Their comprehensive analyses not only underscored the association of frailty with heightened mortality risk, major adverse events, readmission, and non-home discharge, but also with an extended length of stay [39].

Many tools have consequently been employed to measure frailty in patients who have suffered a pelvic fracture, such as the CFS, the eFI, and MF-6 [8, 19, 20, 39]; moreover, the Fried frailty phenotype (FP) has also been a common measure of frailty in orthopedic patients [40,41,42]. However, these indices have limitations that restrict their clinical utility. The, to an extent, subjective nature of the CFS, relying on a clinician’s judgment and observation of functional abilities and overall health [43], can lead to interobserver variation, especially among inexperienced users [44,45,46,47]. The need to evaluate functional ability and the limitations imposed by underlying comorbidities also pose challenges when used in emergency settings. The eFI requires a total 36 variables, while the risk score used by Ramser et al. required 35, including the comorbidities in the Elixhauser Comorbidity Index, which can make the calculation cumbersome and time consuming, leading to less usage in clinical practice [48, 49]. Additionally, the MF-6 requires blood test results to determine the presence of hypoalbuminemia [39]. The FP allows for a detailed characterization of frailty in patients [18], yet it can be difficult to assess factors such as slowness and physical activity in patients immobilized by pain or unintentional weight loss and exhaustion in a population where 15% of patients suffer from dementia. In contrast, the OFS stands out by only requiring five dichotomous variables. It can also be assessed immediately upon arrival in the emergency room, or prior to admission based on electronic medical records, without the need for further blood tests. This focus on simplicity also necessitates that some markers of frailty used in other indices, such as confusion, dementia, and the use of walking aids, have been excluded from the OFS; nevertheless, the inclusion of these variables did not result in a clinically significant improvement in the discriminatory ability of the OFS [23]. While treating physicians make individual clinical decisions, predictive tools can facilitate resource allocation, support objective decision-making, and enhance communication with patients and their families.

Nevertheless, this study has several limitations. Its retrospective nature limits the availability of certain variables such as long-term mortality, quality of life, and readmission. Additionally, it was not possible to adjust for perioperative and anesthesia-related confounders. There also remains a risk residual confounding, as well as non-differential misclassification due to reliance on ICD-10 codes for many of the variables. Despite these limitations, the study benefits from its extensive national sample size. The dataset provides a range of information that allows for adjustments to account for preadmission comorbidities and demographic and racial variations.

Conclusion

Frail patients with pelvic fractures demonstrate a disproportionately increased risk of in-hospital mortality, complications, failure-to-rescue, and ICU admission. The OFS can be used to identify this cohort. Additional investigations are warranted to determine what measures are required to mitigate adverse events in this vulnerable patient population.