Quality of life after hip fracture: a 12-month prospective study

Introduction

Hip fracture is an important health problem due to its incidence, as well as its important physical (Marks, 2010), social and economic implications (Carpintero et al., 2014), representing an important challenge for health care in the future, since the growing population of hip fractures has increasingly complex medical, social and rehabilitation needs (Baker et al., 2014). There is an important association of hip fractures with age, as most hip fractures occur in people older than 65 years. Since life expectancy continues to increase, it is likely that the incidence of fragility fractures in Spain will also increase by almost 30% in 2030 (International Osteoporosis Foundation, 2019).

Hip fractures affect one in three women and one in five men aged 50 and over. In 2017, there were 73,381 discharges due to hip fracture, of which 71% were women (INE, 2019). Hip fracture is associated with an excess of mortality, morbidity and disability (Abrahamsen et al., 2009). Mortality after a hip fracture ranges between 8.4% and 36% (Abrahamsen et al., 2009; Roche et al., 2005; Ekegren et al., 2016; Caeiro et al., 2017) and is 5–8 times higher in the 3 months following the fracture. This increased risk persists even after 10 years (Haentjens et al., 2010).

Approximately one in three women will suffer a hip fracture during their lifetime, especially from falls or as a consequence of osteoporosis (Chami et al., 2006). From an economic perspective of health, the burden of disease covers both the cost related to the disease and its consequences on morbidity (i.e., quality of life and survival) (Borgström et al., 2013). It is estimated that the average direct costs of caring for a hip fracture in Spain are at 8,400 euros, with global figures that reach between 300 and 860 million euros (Health Information Institute, 2010). In addition, hip fractures usually require a surgical intervention followed by postoperative rehabilitation to recover the prefracture function level. Hip fractures are the cause of 8% of visits to emergency services and 42% of hospital admissions related to falls (Ruths et al., 2017). Between 50% and 71% of people who experience a hip fracture will recover the levels prior to the fracture by 1 year later (Vergara et al., 2014; Dyer et al., 2016), with the consequent decline in their independence, so that up to one in four patients will be institutionalized.

Rehabilitation programs with physical exercises have been shown to have a positive effect on functional abilities (Lima et al., 2016); however, it is common that, despite following an adequate rehabilitation treatment, recovery is incomplete, and reduced mobility, imbalance, lack of confidence or fear of falling after hip fracture remain (Shumway-Cook et al., 2005; Portegijs et al., 2012). In fact, 25–75% of people who can walk independently before the fracture will become dependent after 12 months (Tang et al., 2017).

Hip fracture has been associated with a profound deterioration in the health-related quality of life (HRQOL) of the patient (González-Zabaleta et al., 2015; Peeters et al., 2016). There are innumerable studies and reports aimed at knowing the impact of the fracture through objective clinical and socioeconomic indicators, obviating the self-perception of the patient about their own functional development. However, this traditional view is changing, and an increasing number of studies and reports incorporate subjective indicators based on the self-perception of the patient as a complement to traditional indicators to evaluate clinical interventions. To our knowledge, there are still limited studies focused on the analysis of HRQOL after a hip fracture in the Spanish population, especially with long-term follow-up (Støen et al., 2014; Gjertsen et al., 2016; Alexiou et al., 2018). Given the importance of HRQOL in successful aging (Li et al., 2014), it is necessary to adequately understand the impact of a hip fracture on HRQOL. Therefore, the objective of this study was to determine HRQOL after 12 months of hip fracture and to identify potential factors associated with HRQOL.

The self-assessment of the quality of life is highly influenced by personal experience and cultural constraints.

Methods

Results

A total of 270 patients were admitted for hip fracture during the study period. Of these patients, three (1.1%) refused to participate in the study, 43 (15.92%) met one of the exclusion criteria, and two (0.8%) died before they could be recruited for inclusion in the study. A total of 224 patients were included in the study. The mortality during the first postdischarge month was five patients (2.2%).

The sociodemographic and clinical characteristics of the patients are shown in Table 1. The mean age of the study participants was 84.6 years (SD ± 6.1 years); the majority of patients were women (76.3%) and polymedicated (69.6%). A total of 64.3% of the patients suffered a trochanteric fracture, compared to 35.7% who suffered a neck fracture. Fracture reduction by intramedullary rods was the most common type of surgical procedure (66.2%). Most surgeries were performed under spinal anesthesia (79.5%). The time between hospital admission and surgery was 3.0 days (SD ± 2.8 days), and the hospital stay was 5.3 days (SD ± 1.2 days). Charlson’s Comorbidity Index at the time of surgery was 5.3 (SD ± 1.2).

The evolution of clinical factors and the functional ability of patients during the follow-up period are shown in Table 2. The mean scores of IADL Lawton and Brody decreased significantly (p < 0.001) from 5.1 (SD 2.7) at baseline to 2.2 (SD 1.5), 2.3 (SD 1.9) and 2.3 (SD 1.9) after 1, 6 and 12 months, respectively. Similarly, the Barthel Index decreased significantly (p < 0.001) from 87.6 (SD 16.9) at baseline to 53.5 (SD 17.7), 58.9 (SD 19.5) and 59.1 (SD 19.6). The percentage of patients indicating not being able to walk independently (FAC) increased significantly (p < 0.001) from 42.8% at baseline to 99.1% after 1 month, to 95.2% after 6 months and to 91.2% after 12 months. The mean value of the Yesavage Geriatric Depression Scale also increased significantly (p < 0.001) from 3.2 (SD 3.7) at baseline to 5.6 (SD 3.3), 6.2 (SD 3.2) and 6.6 (SD 3.2), respectively. Scores from the EQ-5D VAS decreased significantly (p < 0.001) from 72.8 (SD 15.8) at baseline to 48.3 (SD 17.2) after 1 month, to 48.2 (SD 15.4) after 6 months and to 46.1 (SD 15.2) after 12 months (Fig. 1A). The EQ-5D index score showed a similar significant reduction (p < 0.001) from 0.6 at baseline to 0.1, 0.3 and 0.3, respectively (Fig. 1B). The values for PCS-12 significantly decreased (p < 0.001) from 38.6; SD 7.8 at baseline to 31.0 (SD 5.2) after 1 month, to 33.1 (SD 5.1) after 6 months and to 33.5 (SD 5.3) after 12 months (Fig. 2). Similarly, MCS-12 decreased from 46.5 (SD 9.4) at baseline to 44.8 (SD 9.7; p = 0.022) after 6 months and to 44.3 (SD 9.4; p = 0.005) after 12 months. Changes in the dimensions of the EQ-5D questionnaire are shown in Table 3. All dimensions showed significant changes from baseline after 1, 6 and 12 months, especially by increasing the perception of “some problems”.

In the univariate analysis (Tables 4 and 5), we found nine parameters that correlated significantly with the EQ-5D index at 12 months. These parameters included patient age, CCI, Barthel Index, Yesavage Depression Scale, Lawton and Brody Scale, residential status, polymedicated status, previous hip fracture and FAC. The strongest correlations were found with the Barthel Index (Spearman’s Rho coefficient = 0.484), Lawton and Brody (Spearman’s Rho coefficient = 0.502) and Yesavage Depression Scale (Spearman’s Rho coefficient = −0.295).

The prefracture Barthel Index, prefracture Lawton and Brody Scale and prefracture FAC correlated significantly with the difference between the prefracture EQ-5D index and the EQ-5D index at 1 year after surgery, and the correlation coefficients were lower.

The strongest correlation between the EQ-5D VAS score at 1 month was with Lawton and Brody (r = 0.451), the prefracture Barthel Index (r = 0.392), and the Yesavage Depression Scale (r = −0.045) (Table 4).

In the multiple regression analysis (adjusted R² = 0.318), the Yesavage Depression Scale baseline (B = −0.023 (−0.34 to −0.013), p < 0.001) and the patient age (B = −0.011 (−0.18 to −0.04), p < 0.002) were significantly associated with a lower EQ-5D index at 1 year after surgery, while the Lawton and Brody baseline (B = 0.41 (0.25–0.58), p < 0.003) was associated with higher EQ-5D index levels (Table 6). For the difference between the EQ-5D Index just before surgery and 1 year after surgery, the Barthel Index (B = −0.006 (−0.009 to −0.003), p < 0.001) was identified as an independent variable (adjusted R² = 0.077) (Table 7).

In the multiple regression analysis of factors influencing the EQ-5D VAS score at 1 year (adjusted R² = 0.257), a higher presurgery Lawton and Brody score was associated with a higher EQ-5D VAS (B = 1.683 (0.914–2.451), p < 0.001), while the Yesavage Depression Scale (B = −1.184 (−1.688 to −0.680), p = 0.001) and the patient age (B = −0.340 (−0.670 to −0.10), p = 0.002) were associated with lower EQ-5D VAS levels (Table 8). Regarding the difference between the EQ-5D Index just before surgery and at 1 year after surgery, the patient age (B = −0.518 (−0.902 to −0.134), p < 0.008) was identified as an independent variable (adjusted R² = 0.029) (Table 9).

Discussion

Hip fractures are the most frequent cause of admission to trauma units for older people, which often results in reduced mobility and a loss of independence. One of the goals of our study was to employ several methods to statistically demonstrate the impairment in the HRQOL: the EQ-5D questionnaire (VAS and index score) and the SF-12 health survey (PCS-12 and MCS-12). Most studies designed to evaluate the HRQOL have employed the EQ-5D index score (social tariff). However, this score has limitations, such as a bimodal or trimodal distribution or a ceiling effect, and it does not provide information on the change in HRQOL (Ostendorf et al., 2004). Another goal of our study was to separately evaluate the distribution of the EQ-5D dimensions.

The use of the HRQOL, as a tool for understanding the specific effect of a condition (in our case, hip fracture) on the life of the patient, has increased significantly in recent decades. However, to our knowledge, there is still limited information on the quality of life after hip fracture in the Spanish population in the medium and long term. An international study (Borgström et al., 2013) with data from 11 countries evaluated the quality of life at 4 months after hip fracture using the EQ-5D questionnaire, and the Spanish section consisted of 46 patients. The multicenter study on (Caeiro et al., 2017) the quality of life 12 months post hip fracture (Martínez et al., 2014), evaluated the quality of life related to health in patients with subcapital fracture of femur subjected to different hemostatic treatments, where the EQ-5D scale was used with five levels of severity, while the Ubeda study (Úbeda et al., 2015) evaluated the quality of life in patients with hip arthroplasty secondary to osteoarthritis, where the fracture was an exclusion criterion.

The population of our study mainly consists of women older than 80 years who have a low index of institutionalization and that are consistent with the other studies. (Martínez et al., 2014; Úbeda et al., 2015; Ramírez-Pérez et al., 2014; Milte et al., 2018; Parsons et al., 2018; Moerman et al., 2016; Tidermark et al., 2003; Sugeno et al., 2008). The length of hospital stay was 5.3 days (±1.2), which is considerably less than that shown in other studies (Úbeda et al., 2015) and (Caeiro et al., 2017) at 11.8 days (±7.9). Notably, there was a 21% increase in the rate of institutionalization of elderly individuals who suffered a hip fracture at 12 months.

Mortality from all causes a year after the fracture is lower than that evidenced in the study by (Caeiro et al., 2017) (8.5 vs 15.8), although the comorbidity index was much higher in our study (5.3 ± 1.2 vs 1.9 ± 1.3). In the study of (Ruths et al., 2017), using the ASA risk as an indirect indicator of comorbidity, the results were very similar to those of our study, (ASA III—60.7% vs. 53%) (ASA IV 7.6% vs 6.7%), although mortality was much higher (25.3%).

It is necessary to indicate that at the time of hospital admission, our patients had an adequate level of functional independence (Barthel index = 87.6), an adequate social-familiar situation (Gijon scale = 8.3), and only 16.5% were institutionalized. The indices that assess the motor abilities and functional performance experienced a decline in the first month, but in successive evaluations at 6 and 12 months, these rates experienced a slight increase or stabilization. This trend was not fulfilled in the Yesavage index, where the index at 12 months exceeds the results found 1 month after the intervention by one point. The patients in our study had a capacity to develop the BADL, in comparison with the study of (Caeiro et al., 2017), Barthel 87.6 (±16.9) vs 77.5 (±26.9), although after 12 months, there was no improvement at levels similar to those before the fracture; however, this improvement occurred in several studies (Caeiro et al., 2017; Sugeno et al., 2008).

The baseline values of the EQ-5D index show an acceptable self-assessment of quality of life, similar to that provided by other studies conducted in the Spanish population (Caeiro et al., 2017; Borgström et al., 2013; Ferrer et al., 2015), and studies conducted in different countries, such as the United Kingdom (Parsons et al., 2018), Estonia (Jürisson et al., 2016), Thailand (Amphansap & Sujarekul, 2018) and Mexico (Guirant et al., 2018), although somewhat lower than that reported by other studies in Japan (Sugeno et al., 2008) and Australia (Abimanyi-Ochom et al., 2015) and in the systematic review by Peasgood et al. (2009).

Although when evaluating each of the domains independently, the high affectation of the domains of mobility, pain or discomfort and daily activities is noteworthy, where more than 50% of the patients reported having problems in development, contrary to the results of (Gjertsen et al., 2016), although the mean age of the population in their study was considerably lower (84.6 ± 6.1 vs 77.3 ± 11.7). Our data regarding mobility and self-care activities before the fracture are in line with those obtained by (Guirant et al., 2018), where these authors evaluated the changes in HRQOL (by using EQ-5D) during 12 months in 193 patients with hip fracture, and the population showed higher basal levels of anxiety and depression. When comparing the data after 12 months, we observed that the population of our study was most greatly affected in all domains, with the domain showing greater difference with respect to the basal levels of self-care (Milte et al., 2018). A comparison of two tools to measure HRQOL (EQ-5D and ICECAP-O) in patients with hip fracture showed a significant deterioration in HRQOL at 4 months after fracture. The majority of the patients indicated problems with mobility, self-care, usual activities and pain/discomfort (information the EQ-5D dimensions). In addition, approximately one in two individuals experienced moderate or severe anxiety or depression. Moreover, most of our patients reported experiencing some problem with each of the dimensions of the EQ-5D at 12 months after hip fracture, and 32.7% and 65.8% of patients reported severe problems in self-care and usual activities, respectively. These findings agree with those described in another study (Kondo et al., 2014)

Notably, in the studies of (Parsons et al., 2018) and (Sugeno et al., 2008), the evaluation of the EQ-5D VAS showed baseline values lower than those provided in our study (67.6 and 62.6), but much higher values than those of our studies in the follow-ups ((Parsons = 67.6 basal, 33.2 at 6 months and 66.9 at 12 months) (Sugeno = 62.6 basal; one a month and 79.6 at 12 months). On the other hand, the study of Amphansap T (2018) showed higher initial values (85) but a marked decrease at 3 months (43), although these values increased significantly in months 6 to 12 (68 and 79).

Regarding the evolution of the EQ-5D index, most of the studies showed a rapid decline in short-term scores (hospital discharge or 1 month after fracture) and an improvement in the values at 12 months with respect to midterm evaluations (4 months) (Parsons et al., 2018; Sugeno et al., 2008; Amphansap & Sujarekul, 2018; Guirant et al., 2018; Abimanyi-Ochom et al., 2015; Jürisson et al., 2016; Campenfeldt et al., 2017; Støen et al., 2014; Frihagen, Nordsletten & Madsen, 2007; Keating et al., 2006), except for our study, which showed a stabilization of the values at 6–12 months of 0.3 (SD ± 0.2). This finding could be explained by the differences in the cut-off midterm follow-up, which was carried out in our study at 6 months postfracture, while in the majority of the studies this evaluation was performed at 4 months. It is possible that even the recovery margin was wider than the 6-month record.

In the study developed by (Kelly-Pettersson et al., 2019) aimed at examining the influence of depression on quality of life at 1 year after a hip fracture, with follow-up evaluations at 3 and 12 months, the scores at both 3 and 12 months in the control group and the group of patients with depression were somewhat lower than the baseline scores, similar to the findings in the study of (Tidermark et al., 2002), where the HRQOL of patients over 65 years of age deteriorated significantly at 4 and 17 months after a femoral neck fracture. When using the EQ-5D questionnaire, the authors revealed a significant decrease of 0.78 in the baseline to 0.59 and 0.51 in postfracture follow-ups.

Moerman et al. (2016) reported a significant decrease in HRQOL using the SF-12 health survey in their work between the beginning and after 3 months of hip fracture. The values of MCS recovered to baseline levels after 12 months, whereas the PCS did not. In our study, both components were far from their initial values at both 6 and 12 months, and even the MSC component reduced its value with respect to 6 months. (Tseng et al., 2016) evaluated the quality of life at month, 3, 6 and 12 months after hip fracture using the abbreviated questionnaire SF36. The average PCS score was 45.53 (±5.92) at 1 month after discharge and improved to 63.67 (±10.88) at 12 months after discharge. The MCS score was 55.31 (±9.72) at 1 month after discharge and remained relatively stable during the first 3 months after discharge but decreased slightly to 51.97 (±9.53) at 12 months after discharge. All values were superior to those reported in our study.

The univariate analysis showed a correlation of the HRQOL with age, ICC, BADL, IADL, depression, previous hip fracture, baseline living status, polymedicated status and walk independently. The multivariate analysis confirmed the correlation between HRQOL and BADL, IADL, age and depression. For the EQ-5D VAS, we found similar results to the univariate analysis but did show a correlation with previous hip fracture. The multivariate analysis found correlations with BADL, depression and age.

Moerman et al. (2016) also identified factors potentially associated with the decline in HRQOL (in PCS, not in MCS): age over 80, treatment with osteosynthesis, higher prefracture level of mobility, intracapsular fracture, and ASA classification I and II. (Suwanpasu & Pongpaew, 2016) indicated that HRQOL could be improved in cases of identifying (and treating) limitations in daily living activity, frailty, and depression. In our study, a better HRQOL was associated with not having depression after 12 months, walking independently after 12 months, a higher HRQOL (EQ-5D VAS) at baseline, and younger ages.

The HRQOL at baseline has been associated with the first incident of hip fracture in postmenopausal women (Chen et al., 2018). The authors demonstrated that at baseline, the scores for the SF-36 questionnaire in women who suffered a hip fracture were significantly lower than those for the controls (same age but no hip fracture). Similarly, the HRQOL at baseline was the main factor contributing to the loss in HRQOL after 12 months, as demonstrated by Guirant et al. (2018).

The result of a recent systematic review (Peeters et al., 2016) showed that HRQOL and health status are negatively associated with female gender, comorbidity, inadequate nutritional status, low physical or psychosocial functioning prior to hip fracture, longer hospital stays, postoperative complications and pain.

Conclusions

Patients experience a significant impairment in all dimensions of the HRQOL after a hip fracture in all domains, especially self-care, usual activities and mobility. The decline in the HRQOL was more marked in the first month and increased slightly as the recovery progressed; however, these values was very low from the initial values at 12 months after the surgical intervention. Further studies involving a larger cohort of patients, control groups and longer prospective follow-up periods are required to corroborate the present results.

Supplemental Information