- Email: [email protected]
Factors associated to institutionalization and mortality over three years, in elderly people with a hip fracture—An observational study
Maturitas 89 (2016) 9–15
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Factors associated to institutionalization and mortality over three years, in elderly people with a hip fracture—An observational study Francisco Uriz-Otano a , Jorge Pla-Vidal b , Gregorio Tiberio-López c , Vincenzo Malafarina d,∗ a
Department of Geriatrics, Hospital San Juan de Dios, Pamplona, Spain Department of Psychiatry and Medical Psychology, Clinic of the University of Navarra, Pamplona, Spain c Department of Internal Medicine, Hospital Virgen Del Camino, Pamplona, Spain d Department of Geriatrics, Clinica Los Manzanos, Grupo Viamed, Calle Hermanos Maristas, 26140 Lardero, Spain b
a r t i c l e
i n f o
Article history: Received 3 February 2016 Received in revised form 6 April 2016 Accepted 7 April 2016 Keywords: Hip fracture Mortality Nursing home Institutionalization Functional status
a b s t r a c t Objective: To identify the factors associated to institutionalization and mortality in elderly patients with hip fractures (HF). Design: Thirty-six months observational study. Setting: A post-acute rehabilitation ward. Participants: subjects living in the community or in nursing-home, above the age of 65, with HF. Measurements: The following were registered: comorbidity, intra-hospital complications, Barthel index, walking ability and Mini Mental State Examination, as well as blood samples upon admission and discharge. Destination upon discharge was recorded as well as mortality during hospital stay and over the three-year follow up. Results: a total of 430 subjects were included in the study. Twenty-three patients (5.3%) died during their stay in hospital and 152 (35.3%) during follow up after discharge. Forty-five patients (10.5%) were institutionalized upon discharge. In adjusted analysis, the factors that predict intra-hospital mortality are higher comorbidity (OR, 1.46; 95%CI, 1.06–2.01), and the number of complications (OR, 1.71; 95%CI, 1.16–2.64). Factors that predict mortality in the long term are age (HR 1.04; 95%CI, 1.01–1.06), comorbidity (HR 1.19, 95% CI, 1.09–1.30), the number of complications (HR 1.17, 95%CI, 1.05–1.31). The factors that predicted new institutionalization were age (OR 1.04, 95%CI, 0.98–1.09), living alone (OR 3.95, 95%CI, 1.38–11.3), and length of hospital stay (OR 1.02 95%CI, 1.00–1.03). Conclusions: Mortality 3 years after a hip fracture and institutionalization are associated to age, the loss of autonomy in walking, a worse cognitive status and living alone before the fracture. Identification of and, when possible, intervention upon these factors can improve care of elderly people with hip fractures. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Hip fractures are the second cause of admission to hospital in people above the age of 65 and are frequently associated to the onset of disability, institutionalization and high mortality [1]. It has been calculated that, in Spain, the incidence of HF between the years 2000 and 2010 was of 62 and 100 cases per 100,000 inhabitants for men and women, respectively, between the ages of 65 and 69 and
∗ Corresponding author. E-mail address: [email protected] (V. Malafarina). http://dx.doi.org/10.1016/j.maturitas.2016.04.005 0378-5122/© 2016 Elsevier Ireland Ltd. All rights reserved.
increased to 1330 y 2334, respectively, for people above the age of 85 [2]. Between fifteen and thirty percent of subjects living at home at the time of the fracture require institutionalization [3]. Advanced old age, higher comorbidity, dementia and decline of the ability to walk are the main risk factors for institutionalization following a hip fracture [3]. Hospital mortality ranges from 4 to 8%, increasing up to 30% a year, and remaining high over the long term, with a mortality of almost 50% after 3 years and of over 70% after 7 years [4]. Our hypothesis is that age, comorbidity, cognitive status and functional status prior to the fracture and upon discharge are the
10
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Fig. 2. Cumulative mortality during the study. The cumulative mortality, expressed with Kaplan Meier curve.
with the good clinical practice requirements of the European Union and the current revision of the Helsinki declaration. 2.2. Baseline characteristics and outcome variables
Fig. 1. Flow chart of the study population.
factors that have the most bearing on institutionalization and on mortality. Therefore, the main aim of this piece of research is to gain knowledge of the factors that influence institutionalization and mortality in elderly people with hip fractures admitted to an orthogeriatrics rehabilitation unit. 2. Methods 2.1. Study sample This was a single location observational prospective cohort study, in which all patients above 65 years of age with a low impact hip fracture, admitted to the rehabilitation unit of the San Juan de Dios hospital in Pamplona, between March 2008 and February 2013 were included. Patients with hip fractures were operated on at the trauma department of their local hospital in Pamplona. Following the operation, high clinical complexity patients (defined on the basis of higher comorbidity and the presence of complications) and patients partially responding to acute rehabilitation treatment were referred to the our post-acute rehabilitation unit at San Juan de Dios Hospital. Our unit has 20 beds post-acute rehabilitation. A geriatrician, a rehabilitating physician, nurses, physical therapists, licensed practical nurses and a social worker constitute the multidisciplinary team. Physical rehabilitation is carried out in the hospital gym. Patients perform strengthening exercises of the lower limbs, balance exercises and gait training, in sessions of 45 min a day, five days a week (Monday–Friday inclusive). This study (19/2013) was approved by the Navarra Region Clinical Research Ethics Committee and was designed in conformity
Variables for the study were compiled from patients’ digitalized clinical history, from direct interviews with patients or – when patients were unable to answer questions – their relatives. We recorded demographic data (age, sex), usual place of residence and pharmacological history. Blood samples were taken upon admission and before discharge (always following 8 h fast) in order to study haemoglobin and protein concentration as well as kidney function. The presence of comorbidities was assessed with the Charlson index, considering comorbidity to be high for scores ≥3. We registered the date and type of fracture, the date of admission to acute trauma surgery unit and waiting time until surgery, the type of surgery and complications, as well as the length of stay in hospital. We recorded patients’ destination upon discharge. When patients who previously lived at home were admitted to a care home we considered this a new institutionalization. We compared the characteristics of institutionalized patients to those of patients who were discharged to their usual environment. We used the Mini Mental State Examination (MMSE) for the cognitive assessment, which sets scores between 0 and 30 points. A higher score indicates a better cognitive status. Functional ability in Activities of Daily Living (ADL) was assessed using the Barthel Index (BI). The BI uses scores ranging from zero (totally dependent) to one hundred (completely independent). We considered patients had recovered ADL when, upon discharge from hospital, they had achieved the score they had before the fracture. Walking ability at discharge was assessed with the Functional Ambulation Category (FAC) which classifies patients into 6 categories ranging from 0 – patients cannot walk – to 5-independent ambulation. We considered patients had recovered the ability to walk if they scored FAC ≥ 3 upon discharge. 2.3. Mortality We compared the characteristics of patients who died in hospital to those who were discharged, and the characteristics of patients who died during follow-up and those who were alive at the end of the study period. Patients were included in the study until the date of their death, until 3 years had elapsed since discharge from our
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
11
Table 1 Baseline characteristics of whole population. Variable
Total (n 430)
Age (years) M ± SD Female Lives at own home Charlson Index ≥ 3 n (%) Number of Drugs: A/D m (IQR) Fracture type: Subcapital/Intertrochanteric Surgery Type: PR/IF Surgery Delay m (IQR) (days) LoS Orthopaedics/rehabilitation m (IQR) (days) MMSE m (IQR) Barthel index m (IQR): P/A/D FAC m (IQR): P/A/D
84.2 ± 7.4 333 (77.4) 412 (95.8) 191 (44.4) 9 (7–10)/8 (6–9) 153 (35.6)/277(64.4) 163 (37.9)/263 (61.2) 3.0 (2.0–5.0) 10 (8.0–13.0)/35 (23–51) 21 (17–28) 85 (60–100)/10 (5–20)/50 (10–75) 4 (3–5)/1 (0–1)/3 (1–3)
Discharge destination Home Nursing home New institutionalization Acute Hospital (due to complication) Death
323 (75.1) 13 (3.0) 45 (10.5) 26 (6.0) 23 (5.3)
Data are expressed as n (%), unless otherwise specified. M: mean, SD: standard deviation. m: median, IQR: interquartile range, P: previous to admission, A: admission, D: discharge. BI: barthel index, COPD: chronic obstructive pulmonary disease. FAC: functional ambulation category, IF: internal fixation, LoS: length of stay, MMSE: Mini Mental State Examination, PR: prosthetic replacement.
unit or until the end of the study on 01/07/2015. Cause of Death Certificates were obtained from the local health authority. 2.4. Statistical analysis Categorical variables are shown as absolute values and relative frequency. Continuous variables are shown as mean and standard deviation (SD), whenever there is no normal distribution, as median and 25 and 75 percentiles. We calculated the incidence of mortality upon admission and during the three-year follow up after admission to the rehabilitation unit, with a confidence interval of 95%. Regarding the bivariate analysis, we used the Chi-square test for categorical variables and the Student t-test or Mann-Whitney U test for quantitative variables. Graphic representation of the survival data was done with Klapan Meier curves. We used an univariate study with a binary logistic regression for the detection of factors associated to intrahospital mortality and institutionalization upon discharge and a multiple logistic regression in order to identify posible predictive factors. In order to identify the factors associated to and prognostic of mortality in the long term, we did a univariate and multivariate study with the Cox proportional risk method. The dependent variable was vital status 3 years after discharge from the rehabilitation unit. All significant variables (at the p < 0.05 level) were included in the multivariable models with no selection. If the variable was still significant after adjusting for the other variables at p < 0.05 then it was declared predictive of the outcome. We used the Statistical Package for the Social Science (SPSS), version 22.0 (SPSS inc., Chicago, IL). 3. Results Fig. 1 shows the flow chart of the study population. The basal characteristics of our population are summarized in Table 1 (see also Table A1, in supplementary data). The study included 430 patients (77% women), 154 of whom had a history of osteoporotic fracture (40 of which were previous hip fractures). Mean age was 84.2 ± 7.4 years (82.4 ± 9.1 for men and 84.7 ± 6.8 for women), with 120 subjects (28.1%) 80 years of age or younger, 224 (52.1%) aged between 81 and 90, and 85 subjects (19.8%) above the age of 90.
3.1. Mortality The mean length of follow up in order to calculate mortality was 854.9 days (range 1–1,120 days). During the research period 175 patiets died (40.7%). When plotting the risk function survival curve (Kaplan Meier) (Fig. 2) we observed 30.3% of deaths occurred in the first six months, reaching 49% in the first year. Twenty-three patients (5.3%) died during their stay at the rehabilitation unit due to a respiratory infection (55%), heart failure (30%) and the remaining 15% due to other causes. Factors associated to intra-hospital mortality can be found in Table 2. Of the 407 patients discharged, 152 (37.3%) died during the 3 year follow up, mostly because of imbalances in their heart condition (41%) respiratory infection (34%), dementia (13.7%), or tumours (11.2%). In the Cox univariable study, the factors associated to an increase in long term mortality can be found in Table 3. The hypoalbuminemia was one of the most important factors associated with mortality. The albumin concentration is significantly lowest in both, in patients who die during the admission (Table 2), and in patients who die during follow-up (Table 3). The most important factors that predict mortality in the long term are age (Fig. 3, panel A), comorbidity (Fig. 3, panel B), recovery of walking autonomy (Fig. 3, panel C) and cognitive status (Fig. 3, panel D). The type of fracture, type of surgery, delay until surgery, and length of stay at the rehabilitation unit were not associated to long term mortality (p > 0.05). 3.2. New institutionalization Upon discharge, 362 patients (84.2%) returned to their usual environment and 45 patients (10.5%) were admitted to a care home. The characteristics of institutionalized patients compared to those who returned to their usual environment can be found in Table 4. The results of the adjusted multivariable study of the factors that predict intra-hospital mortality, mortality in the long term and new institutionalization are shown in Table 5. 4. Discussion The main aim of this study is to identify the factors that influence new institutionalization and mortality in both the short and long
12
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Table 2 Factors Associated with Short Term Mortality in Univariable Logistic Regression Analysis. Variable
Death (n 23)
Discharge (n 407)
Crude OR (95% CI)
p
Age (y) M ± SD Male Charlson index ≥ 3 LoS Orthopaedics m (IQR) (days) Complications ≥ 3 Heart failure Respiratory infection Delirium Pressure ulcers Admission Albumin mg/dL m (IQR) MMSE m (IQR) Barthel index m (IQR) Previous Admission FAC m (IQR) Previous Admission Inpatient rehabilitation
86.3 ± 6.9 11 (47.8) 21 (91.3) 16 (9.0–30.0) 19 (82.6) 10 (43.5) 15 (65.2) 17 (73.9) 12 (52.2) 2.5 (2.0–2.9) 16 (10–20)
84.1 ± 7.4 321 (78.9) 170 (41.8) 8 (10.0–13.0) 144 (35.4) 86 (21.1) 89 (21.9) 177 (43.5) 55 (13.5) 3.1 (2.8–3.4) 24 (17–28)
2.58 (1.04–6.41) 3.42 (1.45–8.42) 14.6 (3.38–63.26) 1.18 (1.09–1.24) 6.57 (2.39–18.1) 2.87 (1.22–6.77) 6.69 (2.75–16.3) 3.68 (1.42–9.53) 6.98 (2.93–16.6) 0.04 (0.01–0.14) 0.92 (0.88–0.96)
0.035 0.005 <0.001 <0.001 <0.001 0.016 <0.001 0.007 <0.001 <0.001 <0.001
70 (70–90) 0 (0–0)
85 (60–100) 10 (0–20)
0.98 (0.97–0.99) 0.75 (0.64–0.88)
0.007 <0.001
3.0 (2.0–4.0) 1.0 (0.0–1.0) 1 (4.3)
4.0 (3.0–5.0) 1.0 (0.0–1.0) 350 (86.0)
0.51 (0.38–0.71) 0.043 (0.01–0.31) 0.007 (0.01–0.31)
<0.001 <0.001 0.002
Data are expressed as n (%), unless otherwise specified, M: mean, SD: standard deviation, m: median, IQR: interquartile range. FAC: function ambulation category, LoS: length of stay.
Table 3 Factors associated with three years mortality in the univariable cox regression analysis. Variable
Death (n 152)
Alive at the end of the study (n 255)
Age (years) M ± SD Male Charlson Index ≥ 3
86.5 ± 6.6 41 (27.0) 96 (63.2)
82.6 ± 7.6 45 (17.6) 74 (29.0)
Crude HR (95% CI) 1.06 (1.04–1.09) 1.57 (1.10–2.25) 3.09 (2.22–4.30)
p <0.001 0.013 <0.001
Medical comorbidities Cardiovascular disease Diabetes Stroke Dementia
59 (38.8) 54 (35.5) 50 (32.9) 83 (54.6)
100 (39.2) 51 (20.0) 48 (18.8) 72 (28.2)
1.88 (1.36–2.60) 1.46 (1.04–20.7) 1.48 (1.04–2.10) 2.07 (1.51–2.85)
<0.001 0.029 0.019 <0.001
Drugs Benzodiazepines Neuroleptics
48 (31.6) 80 (52.6)
142 (55.7) 26 (10.2)
1.39 (1.03–1.87) 2.63 (1.85–3.75)
0.030 <0.001
Complications ≥ 3 Heart failure Respiratory infection Delirium Pressure ulcers Anaemia transfused Kidney insufficiency
74 (48.7) 45 (29.6) 24 (15.8) 93 (61.2) 37 (24.3) 129 (84.9) 24 (15.8)
70 (27.5) 41 (16.1) 46 (18.0) 84 (32.9) 18 (7.1) 111 (43.5) 16 (6.3)
2.17 (1.58–2.98) 1.95 (1.38–2.77) 1.63 (1.14–2.32) 2.54 (1.93–3.52) 3.00 (2.44–4.75) 2.22 (1.42–3.46) 2.12 (1.37–3.29)
<0.001 <0.001 0.007 <0.001 <0.001 <0.001 0.001
Haemoglobin mg/dL m (IQR) Admission Discharge
10.0 (11.0–12.0) 11.6 (10.1–12.5)
11.1 (10.4–12.0) 12.0 (11.0–12.8)
0.88 (0.78–0.99) 0.83 (0.73–0.94)
0.035 <0.001
Albumin mg/dL m (IQR) Admission Discharge
2.7 (2.5–3.1) 2.9 (2.6–3.2)
2.8 (2.6–3.2) 3.1 (2.8–3.4)
0.51 (0.34–0.77) 0.61 (0.42–0.90)
0.011 0.012
MMSE m (IQR)
20 (10–25)
24 (20–30)
0.94 (0.92–0.95)
<0.001
Barthel index m (IQR) Previous Admission Discharge
75 (35–90) 0 (0–0) 20 (5–63)
90 (70–100) 10 (5–20) 65 (20–80)
0.94 (0.92–0.95) 0.92(0.93–0.97) 0.98 (0.97–0.98)
<0.001 <0.001 <0.001
FAC m (IQR) Previous Discharge
4 (3–4) 2 (1–3)
4 (4–5) 3 (2–3)
0.72 (0.63–0.82) 0.72 (0.63–0.82)
<0.001 <0.001
Inpatient rehabilitation Walk Recoverya Functional Recoverya
120 (78.9) 58 (38.2) 54 (35.5)
230 (90.2) 178 (69.8) 142 (55.7)
0.46 (0.31–0.69) 0.35 (0.25–0.49) 0.60 (0.42–0.86)
<0.001 <0.001 0.006
Data are expressed as n (%), unless otherwise specified. FAC: function ambulation category, IQR: interquartile range, M: mean, m: median, MMSE: Mini Mental State Examination, SD: standard deviation. a Expressed as a percentage of the total number of patients who received rehabilitation. (n 350).
term in elderly people with hip fractures admitted to a geriatric rehabilitation unit. Total mortality was 40% over 3 years follow up, with a lower mortality in the first year than in previous studies [5]. This dif-
ference may stem from the fact that most of these studies were carried out in acute trauma surgery units and not in post acute rehabilitation units, as was ours.
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
13
Fig. 3. Factors associated with long term mortality. Kaplan Maier curve. Panel A. Association between long-term mortality with age. Patients were split into: age ≤ a 80 (solid line), between 81 and 90 (dotted line) and >90 (dashed line). Panel B. Long-term mortality and comorbidity (Charlson Index). Low comorbidity was expressed by Charlson Index <3, and higher comorbidity for Charlson Index ≥3. Panel C. Association between long-term mortality with discharge FAC (Functional Ambulation Category). Patients were split into: independent walkers (FAC ≥ 3) or patients who required aid in order to walk (FAC <3). Panel D. Association between long-term mortality and cognitive status (MMSE, Mini Mental State Examination). Patients were classified as; cognitively whole (MMSE ≥ 24, solid line), with mild cognitive impairment (MMSE between 23 and 16, dotted line) or with severe cognitive impairment (MMSE < 16, dashed line).
Tabla 4 Significant factors for new institutionalitation in the univariable logist regression analysis. Variable
New Institutionalitation (n = 45)
Discharge to usual environment (n = 362)
Age (years) M ± SD Living alone before the fracture LoS rehabilitation m (IQR) (days) n of complications, m (IQR) Urinary retention Urinary infection MMSE m (IQR) Discharge Barthel Index m (IQR)
86.4 ± 6.8 22 (48.9) 43 (32–54) 2 (1.5–4) 13 (28.9) 20 (44.4) 20 (11.5–26.5) 30 (5–65)
84.1 ± 7.4 80 (22.1) 35 (23–52) 1.5 (1.0–3.0) 58 (16.0) 83 (22.9) 24 (17–28) 85 (60–100)
Crude OR (95% CI) 1.06 (1.01–1.11) 2.93 (1.45–8.42) 1.18 (1.09–1.24) 1.21 (1.01–1.47) 1.10 (1.07–1.24) 1.04 (1.01–1.23) 0.96 (0.93–0.99) 1.99 (1.01–3.96)
p 0.032 0.003 0.010 0.045 0.032 0.037 0.002 0.049
Data are expressed as n (%), unless otherwise specified. FAC: function ambulation category, IQR: interquartile range, M: mean, m: median, MMSE: Mini Mental State Examination, SD: standard deviation.
The population included in the study presented advanced old age. Several studies showed that age is a factor influencing mortality [6,7] and new institutionalization [4]. In our study, age did
not predict intra-hospital mortality, contrary to the observations of Eschbach el al. [8], or institutionalization, with a low association to long term mortality. As has been previously demonstrated [9],
14
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Table 5 Multivariable models of predictors factors for mortality and new institutionalization. Adjusted multivariable regression (95% Confidence Interval) Variable
Died in hospitala,c
Age Female sex Live alone FAC Discharge MMSE Charlson Index n◦ of complication Discharge Albumin mg/dL Length of Stay (d)
– – – – – 1.46 (1.06–2.01) 1.75 (1.16–2.64) 0.15 (0.34–0.68) –
p
0.020 0.022
Died 36 monthsb,d
p
New institutionalizationa,e
1,04 (1.01–1.06) 0.59 (0.40–0.86) – 0.74 (0.63–0.87) 0.98 (0.96–1.00) 1.19 (1.09–1.30) 1.17 (1.05–1.31) 0.60 (0.41–0.86) –
0.003 0.007
–
0.027 0.038 <0.001 0.003 0.005
3.95 (1.38–11.3) – 0.93 (0.89–0.97) – – – 1.02 (1.01–1.03)
p
0.003 0.001
0.028
FAC: Function Ambulation Category; MMSE: Mini-Mental State Examination. a Binary Logistic Regression, Odd Ratio (OR). b Cox regression, Hazard Ratio (HR) Adjusted Factors: c Age, Sex, Charlson index, Tumor, use of antidepressant, Traumatology Unit Length of Stay, Number and Types of Complications (Cardiac failure, Respiratory infection, Delirium and Pressure Ulcers), Albumin Concentration, MMSE, Previous Barthel Index, Previous FAC. d Age, Sex, Previous Residence, Charlson index, Comorbidities (Cardiac failure, Dementia, Stroke, Tumor, Diabetes), use of Benzodiazepines and Neuroleptics, Number and Types of Complications (Cardiac failure, Respiratory infection, Delirium, Renal Failure, anemia and Pressure Ulcers), Albumin Concentration, MMSE, Previous and Discharge Barthel Index, Discharge FAC, Inpatient Rehabilitation. e Age, Caregiver, Orthopaedics Unit Length of Stay, MMSE, Albumin Concentration, Number and Types of Complications (Urinary infection and retention), Discharge Barthel Index.
our results may be due to the fact that we took other correction factors other than age which may influence mortality into account (comorbidity, complications, cognitive impairment and functional status). Comorbidity measured with the Charlson index [9,10], and the number and type of complications [11–14] were also important factors associated to an increase in mortality in our study, whether in hospital or during follow up after discharge. Hypoalbuminemia represents a good marker for mortality [15,16] and was, in our study, one of the strongest predictive factors for mortality over 3 years. Further research will be needed in order to assess whether intensive treatment aimed at reinstating normal albumin levels can reduce mortality. The association we found between haemoglobin concentration and mortality lost statistical significance in the multivariable analysis. We believe this weak association depends on other factors such as comorbidity, complications or age, as shown in previous research [17]. Almost half of the included patients had received a blood transfusion during their stay at the acute trauma unit. Like in previous studies, we found no differences in mortality between patients who had received blood transfusions and those who had not [18]. There is controversy about the effects of delays in surgery [19,20], as well as of the type of fracture [21] on mortality. Surgical delay was greater than 72 h for 38% of the patients included in this study but, contrary to other studies, [6,12] we found no link to mortality. Hip fractures are one of the most important causes of loss of function and most studies view functional status prior to fracture as a predictor of mortality [22,23] and institutionalization [24]. We observed that ability to walk after the fracture is a long term mortality predictor. We recently researched the factors associated to functional recovery in elderly patients with hip fractures admitted to a geriatric rehabilitation unit, and observed how functional status and ability to walk before the fracture predict the recovery of walking ability upon discharge [25]. Eighty two percent of patients included carried out rehabilitation; fifty-five percent recovered their prior functional status and 64% recovered the ability to walk. We observed lower mortality among these patients, as other studies have also shown [26–28]. Results from previous research, together with our recent findings lead us to believe that, in addition to the prevention of falls, it would be necessary to admit
these patients to specialized hip fracture rehabilitation units in order to reduce disability and mortality [29]. The prevalence of cognitive impairment is high among elderly people with hip fractures, and is associated to a high mortality [9]. In our study, the correlation between the MMSE and mortality is weak; nevertheless, it is one of the main factors predicting institutionalization. The precarious balance maintained by people with cognitive impairment in a family setting is often shattered by an acute traumatic event such as a hip fracture. The other great problem for elderly people is living alone, a factor that increases the risk of institutionalization following a hip fracture almost fourfold, a result which coincides with what other studies found [30]. Finally, in previous studies, institutionalization appears to be associated to higher mortality [28], whereas we found no such link in our multivariable analysis. Our study has a number of limitations. Firstly, it was carried out in a post-acute rehabilitation unit, and data on mortality during stay in the acute trauma unit and the total number of subjects who suffered a hip fracture in our community were not available. Secondly, we did not assess the evolution of functional state after discharge and its possible link to mortality. Finally, the study was carried out within a specific geriatric post-acute rehabilitation unit, with a population whose characteristics may limit the possibility of extrapolating our results to other settings or environments. Despite these limitations, we believe this study represents a real snapshot of a geriatric rehabilitation unit where we come across elderly, complex, heterogeneous patients with higher comorbidity and a high risk of complications in our day to day work. The association of all these factors complicates both care of these patients and their inclusion in research studies.
5. Conclusions Bearing in mind the complexity of elderly people with hip fractures; age, comorbidity and complications should not be considered isolated variables in the prediction of mortality or institutionalization. Other factors such as functional and cognitive status should be taken into account in the identification of patients at risk. To conclude, integrated care in specific rehabilitation units that takes into account the cumulative effect of these factors could reduce adverse health results.
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Further research shall be needed in order to assess, among other points, the economic aspects of functional loss and institutionalization of elderly people who have suffered a hip fracture, as well as its impact on quality of life. Author contributions Francisco Uriz-Otano and Vincenzo Malafarina: study concept and design, analysis and interpretation of data, and preparation of manuscript. Francisco Uriz-Otano acquisition of participants and data. Gregorio Tiberio Lopez and Jorge Pla Vidal collaborated actively in the scientific basis of the paper, and in drafting and editing the manuscript. Conflict of interest None declared. Funding None. Ethics This study (19/2013) was approved by the Navarra Region Clinical Research Ethics Committee. Provenance and peer review This article was peer reviewed. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.maturitas.2016. 04.005. References [1] M.L. Alvarez-Nebreda, M.T. Vidan, J.A. Serra, Hip fracture management and outcomes in Spain, Eur. Geriatr. Med. 2 (2010) 108–111. [2] E. Cirera, K. Perez, E. Santamarina-Rubio, A.M. Novoa, M. Olabarria, Improvements in hip fracture incidence counterbalanced by the rise of other fracture types: data from Spain 2000–2010, Injury 12 (2014) 2076–2083. [3] E.K. Osnes, C.M. Lofthus, H.E. Meyer, J.A. Falch, L. Nordsletten, I. Cappelen, I.S. Kristiansen, Consequences of hip fracture on activities of daily life and residential needs, Osteoporos. Int. 7 (2004) 567–574. [4] T. Hagino, S. Ochiai, E. Sato, Y. Watanabe, S. Senga, H. Haro, Prognostic prediction in patients with hip fracture: risk factors predicting difficulties with discharge to own home, J. Orthop. Traumatol. 2 (2011) 77–80. [5] J.J. Roche, R.T. Wenn, O. Sahota, C.G. Moran, Effect of comorbidities and postoperative complications on mortality after hip fracture in elderly people: prospective observational cohort study, BMJ 7529 (2005) 1374–1376A. [6] B.D. Chatterton, T.S. Moores, S. Ahmad, A. Cattell, P.J. Roberts, Cause of death and factors associated with early in-hospital mortality after hip fracture, Bone Joint J. 2 (2015) 246–251. [7] A. Soderqvist, W. Ekstrom, S. Ponzer, H. Pettersson, T. Cederholm, N. Dalen, M. Hedstrom, J. Tidermark, Stockholm hip fracture group prediction of mortality in elderly patients with hip fractures: a two-year prospective study of 1, 944 patients, Gerontology 5 (2009) 496–504. [8] D.A. Eschbach, L. Oberkircher, C. Bliemel, J. Mohr, S. Ruchholtz, B. Buecking, Increased age is not associated with higher incidence of complications, longer stay in acute care hospital and in hospital mortality in geriatric hip fracture patients, Maturitas 2 (2013) 185–189.
15
[9] D. Norring-Agerskov, A.S. Laulund, J.B. Lauritzen, B.R. Duus, S. van der Mark, M. Mosfeldt, H.L. Jorgensen, Metaanalysis of risk factors for mortality in patients with hip fracture, Dan. Med. J. 8 (2013) A4675. [10] J. Karres, N.A. Heesakkers, J.M. Ultee, Vrouenraets BC. Predicting 30-day mortality following hip fracture surgery: evaluation of six risk prediction models, Injury 2 (2015) 371–377. [11] A. Ozturk, Y. Ozkan, S. Akgoz, N. Yalcyn, R.M. Ozdemir, S. Aykut, The risk factors for mortality in elderly patients with hip fractures: postoperative one-year results, Singapore Med. J. 2 (2010) 137–143. [12] K.A. Lefaivre, S.A. Macadam, D.J. Davidson, R. Gandhi, H. Chan, H.M. Broekhuyse, Length of stay, mortality, morbidity and delay to surgery in hip fractures, J. Bone Joint Surg. Br. 7 (2009) 922–927. [13] M. Harstedt, C. Rogmark, R. Sutton, O. Melander, A. Fedorowski, Impact of comorbidity on 6-month hospital readmission and mortality after hip fracture surgery, Injury 4 (2015) 713–718. [14] M. Mariconda, G.G. Costa, S. Cerbasi, P. Recano, E. Aitanti, M. Gambacorta, M. Misasi, The determinants of mortality and morbidity during the year following fracture of the hip: a prospective study, Bone Joint J. 3 (2015) 383–390. [15] A.S. Laulund, J.B. Lauritzen, B.R. Duus, M. Mosfeldt, H.L. Jorgensen, Routine blood tests as predictors of mortality in hip fracture patients, Injury 7 (2012) 1014–1020. [16] S. Cabrerizo, D. Cuadras, F. Gomez-Busto, I. Artaza-Artabe, F. Marin-Ciancas, V. Malafarina, Serum albumin and health in older people: review and meta analysis, Maturitas 1 (2015) 17–27. [17] L.J. Potter, B. Doleman, I.K. Moppett, A systematic review of pre-operative anaemia and blood transfusion in patients with fractured hips, Anaesthesia 4 (2015) 483–500. [18] J.L. Carson, M.L. Terrin, H. Noveck, D.W. Sanders, B.R. Chaitman, G.G. Rhoads, G. Nemo, K. Dragert, L. Beaupre, K. Hildebrand, W. Macaulay, C. Lewis, D.R. Cook, G. Dobbin, K.J. Zakriya, F.S. Apple, R.A. Horney, J. Magaziner, FOCUS investigators. Liberal or restrictive transfusion in high-risk patients after hip surgery, N. Engl. J. Med. 26 (2011) 2453–2462. [19] E. Castronuovo, P. Pezzotti, A. Franzo, D. Di Lallo, G. Guasticchi, Early and late mortality in elderly patients after hip fracture: a cohort study using administrative health databases in the Lazio region, Italy, BMC Geriatr. (2011) 37 (2318-11-37). [20] N. Simunovic, P.J. Devereaux, S. Sprague, G.H. Guyatt, E. Schemitsch, J. Debeer, M. Bhandari, Effect of early surgery after hip fracture on mortality and complications: systematic review and meta-analysis, CMAJ 15 (2010) 1609–1616. [21] W.P. Lin, C.J. Wen, C.C. Jiang, S.M. Hou, C.Y. Chen, J. Lin, Risk factors for hip fracture sites and mortality in older adults, J. Trauma 1 (2011) 191–197. [22] G. Bellelli, P. Mazzola, M. Corsi, A. Zambon, G. Corrao, G. Castoldi, G. Zatti, G. Annoni, The combined effect of ADL impairment and delay in time from fracture to surgery on 12-month mortality: an observational study in orthogeriatric patients, J. Am. Med. Dir. Assoc. 7 (2012) 664 (e9, 664. e14). [23] T.C. Marufu, A. Mannings, I.K. Moppett, Risk scoring models for predicting peri-operative morbidity and mortality in people with fragility hip fractures: qualitative systematic review, Injury 12 (2015) 2325–2334. [24] F.J. Tarazona-Santabalbina, A. Belenguer-Varea, E. Rovira Daudi, E. Salcedo Mahiques, D. Cuesta Peredo, J.R. Domenech-Pascual, H. Gac Espinola, J.A. Avellana Zaragoza, Severity of cognitive impairment as a prognostic factor for mortality and functional recovery of geriatric patients with hip fracture, Geriatr. Gerontol. Int. 3 (2015) 289–295. [25] F. Uriz-Otano, J.I. Uriz-Otano, V. Malafarina, Factors associated with short-term functional recovery in elderly people with a hip fracture. Influence of cognitive impairment, J. Am. Med. Dir. Assoc. 3 (2015) 215–220. [26] D.P. Seitz, G.M. Anderson, P.C. Austin, A. Gruneir, S.S. Gill, C.M. Bell, P.A. Rochon, Effects of impairment in activities of daily living on predicting mortality following hip fracture surgery in studies using administrative healthcare databases, BMC Geriatr. 9 (2014) (2318-14-9). [27] T. Torpilliesi, G. Bellelli, S. Morghen, S. Gentile, E. Ricci, R. Turco, M. Trabucchi, Outcomes of nonagenarian patients after rehabilitation following hip fracture surgery, J. Am. Med. Dir. Assoc. 1 (81) (2012) e1–e5. [28] A.W. Ireland, P.J. Kelly, R.G. Cumming, Risk factor profiles for early and delayed mortality after hip fracture: analyses of linked Australian Department of Veterans’ affairs databases, Injury 6 (2015) 1028–1035. [29] P. Saez Lopez, N. Sanchez Hernandez, S. Paniagua Tejo, J.A. Valverde Garcia, M. Montero Diaz, N. Alonso Garcia, A. Freites Esteve, Clinical pathway for hip fracture patients, Rev. Esp. Geriatr. Gerontol. 4 (2015) 161–167. [30] A.J. Vochteloo, S.T. van Vliet-Koppert, A.B. Maier, W.E. Tuinebreijer, M.L. Roling, M.R. de Vries, R.M. Bloem, R.G. Nelissen, P. Pilot, Risk factors for failure to return to the pre-fracture place of residence after hip fracture: a prospective longitudinal study of 444 patients, Arch. Orthop. Trauma Surg. 6 (2012) 823–830.
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Factors associated to institutionalization and mortality over three years, in elderly people with a hip fracture—An observational study Francisco Uriz-Otano a , Jorge Pla-Vidal b , Gregorio Tiberio-López c , Vincenzo Malafarina d,∗ a
Department of Geriatrics, Hospital San Juan de Dios, Pamplona, Spain Department of Psychiatry and Medical Psychology, Clinic of the University of Navarra, Pamplona, Spain c Department of Internal Medicine, Hospital Virgen Del Camino, Pamplona, Spain d Department of Geriatrics, Clinica Los Manzanos, Grupo Viamed, Calle Hermanos Maristas, 26140 Lardero, Spain b
a r t i c l e
i n f o
Article history: Received 3 February 2016 Received in revised form 6 April 2016 Accepted 7 April 2016 Keywords: Hip fracture Mortality Nursing home Institutionalization Functional status
a b s t r a c t Objective: To identify the factors associated to institutionalization and mortality in elderly patients with hip fractures (HF). Design: Thirty-six months observational study. Setting: A post-acute rehabilitation ward. Participants: subjects living in the community or in nursing-home, above the age of 65, with HF. Measurements: The following were registered: comorbidity, intra-hospital complications, Barthel index, walking ability and Mini Mental State Examination, as well as blood samples upon admission and discharge. Destination upon discharge was recorded as well as mortality during hospital stay and over the three-year follow up. Results: a total of 430 subjects were included in the study. Twenty-three patients (5.3%) died during their stay in hospital and 152 (35.3%) during follow up after discharge. Forty-five patients (10.5%) were institutionalized upon discharge. In adjusted analysis, the factors that predict intra-hospital mortality are higher comorbidity (OR, 1.46; 95%CI, 1.06–2.01), and the number of complications (OR, 1.71; 95%CI, 1.16–2.64). Factors that predict mortality in the long term are age (HR 1.04; 95%CI, 1.01–1.06), comorbidity (HR 1.19, 95% CI, 1.09–1.30), the number of complications (HR 1.17, 95%CI, 1.05–1.31). The factors that predicted new institutionalization were age (OR 1.04, 95%CI, 0.98–1.09), living alone (OR 3.95, 95%CI, 1.38–11.3), and length of hospital stay (OR 1.02 95%CI, 1.00–1.03). Conclusions: Mortality 3 years after a hip fracture and institutionalization are associated to age, the loss of autonomy in walking, a worse cognitive status and living alone before the fracture. Identification of and, when possible, intervention upon these factors can improve care of elderly people with hip fractures. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Hip fractures are the second cause of admission to hospital in people above the age of 65 and are frequently associated to the onset of disability, institutionalization and high mortality [1]. It has been calculated that, in Spain, the incidence of HF between the years 2000 and 2010 was of 62 and 100 cases per 100,000 inhabitants for men and women, respectively, between the ages of 65 and 69 and
∗ Corresponding author. E-mail address: [email protected] (V. Malafarina). http://dx.doi.org/10.1016/j.maturitas.2016.04.005 0378-5122/© 2016 Elsevier Ireland Ltd. All rights reserved.
increased to 1330 y 2334, respectively, for people above the age of 85 [2]. Between fifteen and thirty percent of subjects living at home at the time of the fracture require institutionalization [3]. Advanced old age, higher comorbidity, dementia and decline of the ability to walk are the main risk factors for institutionalization following a hip fracture [3]. Hospital mortality ranges from 4 to 8%, increasing up to 30% a year, and remaining high over the long term, with a mortality of almost 50% after 3 years and of over 70% after 7 years [4]. Our hypothesis is that age, comorbidity, cognitive status and functional status prior to the fracture and upon discharge are the
10
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Fig. 2. Cumulative mortality during the study. The cumulative mortality, expressed with Kaplan Meier curve.
with the good clinical practice requirements of the European Union and the current revision of the Helsinki declaration. 2.2. Baseline characteristics and outcome variables
Fig. 1. Flow chart of the study population.
factors that have the most bearing on institutionalization and on mortality. Therefore, the main aim of this piece of research is to gain knowledge of the factors that influence institutionalization and mortality in elderly people with hip fractures admitted to an orthogeriatrics rehabilitation unit. 2. Methods 2.1. Study sample This was a single location observational prospective cohort study, in which all patients above 65 years of age with a low impact hip fracture, admitted to the rehabilitation unit of the San Juan de Dios hospital in Pamplona, between March 2008 and February 2013 were included. Patients with hip fractures were operated on at the trauma department of their local hospital in Pamplona. Following the operation, high clinical complexity patients (defined on the basis of higher comorbidity and the presence of complications) and patients partially responding to acute rehabilitation treatment were referred to the our post-acute rehabilitation unit at San Juan de Dios Hospital. Our unit has 20 beds post-acute rehabilitation. A geriatrician, a rehabilitating physician, nurses, physical therapists, licensed practical nurses and a social worker constitute the multidisciplinary team. Physical rehabilitation is carried out in the hospital gym. Patients perform strengthening exercises of the lower limbs, balance exercises and gait training, in sessions of 45 min a day, five days a week (Monday–Friday inclusive). This study (19/2013) was approved by the Navarra Region Clinical Research Ethics Committee and was designed in conformity
Variables for the study were compiled from patients’ digitalized clinical history, from direct interviews with patients or – when patients were unable to answer questions – their relatives. We recorded demographic data (age, sex), usual place of residence and pharmacological history. Blood samples were taken upon admission and before discharge (always following 8 h fast) in order to study haemoglobin and protein concentration as well as kidney function. The presence of comorbidities was assessed with the Charlson index, considering comorbidity to be high for scores ≥3. We registered the date and type of fracture, the date of admission to acute trauma surgery unit and waiting time until surgery, the type of surgery and complications, as well as the length of stay in hospital. We recorded patients’ destination upon discharge. When patients who previously lived at home were admitted to a care home we considered this a new institutionalization. We compared the characteristics of institutionalized patients to those of patients who were discharged to their usual environment. We used the Mini Mental State Examination (MMSE) for the cognitive assessment, which sets scores between 0 and 30 points. A higher score indicates a better cognitive status. Functional ability in Activities of Daily Living (ADL) was assessed using the Barthel Index (BI). The BI uses scores ranging from zero (totally dependent) to one hundred (completely independent). We considered patients had recovered ADL when, upon discharge from hospital, they had achieved the score they had before the fracture. Walking ability at discharge was assessed with the Functional Ambulation Category (FAC) which classifies patients into 6 categories ranging from 0 – patients cannot walk – to 5-independent ambulation. We considered patients had recovered the ability to walk if they scored FAC ≥ 3 upon discharge. 2.3. Mortality We compared the characteristics of patients who died in hospital to those who were discharged, and the characteristics of patients who died during follow-up and those who were alive at the end of the study period. Patients were included in the study until the date of their death, until 3 years had elapsed since discharge from our
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
11
Table 1 Baseline characteristics of whole population. Variable
Total (n 430)
Age (years) M ± SD Female Lives at own home Charlson Index ≥ 3 n (%) Number of Drugs: A/D m (IQR) Fracture type: Subcapital/Intertrochanteric Surgery Type: PR/IF Surgery Delay m (IQR) (days) LoS Orthopaedics/rehabilitation m (IQR) (days) MMSE m (IQR) Barthel index m (IQR): P/A/D FAC m (IQR): P/A/D
84.2 ± 7.4 333 (77.4) 412 (95.8) 191 (44.4) 9 (7–10)/8 (6–9) 153 (35.6)/277(64.4) 163 (37.9)/263 (61.2) 3.0 (2.0–5.0) 10 (8.0–13.0)/35 (23–51) 21 (17–28) 85 (60–100)/10 (5–20)/50 (10–75) 4 (3–5)/1 (0–1)/3 (1–3)
Discharge destination Home Nursing home New institutionalization Acute Hospital (due to complication) Death
323 (75.1) 13 (3.0) 45 (10.5) 26 (6.0) 23 (5.3)
Data are expressed as n (%), unless otherwise specified. M: mean, SD: standard deviation. m: median, IQR: interquartile range, P: previous to admission, A: admission, D: discharge. BI: barthel index, COPD: chronic obstructive pulmonary disease. FAC: functional ambulation category, IF: internal fixation, LoS: length of stay, MMSE: Mini Mental State Examination, PR: prosthetic replacement.
unit or until the end of the study on 01/07/2015. Cause of Death Certificates were obtained from the local health authority. 2.4. Statistical analysis Categorical variables are shown as absolute values and relative frequency. Continuous variables are shown as mean and standard deviation (SD), whenever there is no normal distribution, as median and 25 and 75 percentiles. We calculated the incidence of mortality upon admission and during the three-year follow up after admission to the rehabilitation unit, with a confidence interval of 95%. Regarding the bivariate analysis, we used the Chi-square test for categorical variables and the Student t-test or Mann-Whitney U test for quantitative variables. Graphic representation of the survival data was done with Klapan Meier curves. We used an univariate study with a binary logistic regression for the detection of factors associated to intrahospital mortality and institutionalization upon discharge and a multiple logistic regression in order to identify posible predictive factors. In order to identify the factors associated to and prognostic of mortality in the long term, we did a univariate and multivariate study with the Cox proportional risk method. The dependent variable was vital status 3 years after discharge from the rehabilitation unit. All significant variables (at the p < 0.05 level) were included in the multivariable models with no selection. If the variable was still significant after adjusting for the other variables at p < 0.05 then it was declared predictive of the outcome. We used the Statistical Package for the Social Science (SPSS), version 22.0 (SPSS inc., Chicago, IL). 3. Results Fig. 1 shows the flow chart of the study population. The basal characteristics of our population are summarized in Table 1 (see also Table A1, in supplementary data). The study included 430 patients (77% women), 154 of whom had a history of osteoporotic fracture (40 of which were previous hip fractures). Mean age was 84.2 ± 7.4 years (82.4 ± 9.1 for men and 84.7 ± 6.8 for women), with 120 subjects (28.1%) 80 years of age or younger, 224 (52.1%) aged between 81 and 90, and 85 subjects (19.8%) above the age of 90.
3.1. Mortality The mean length of follow up in order to calculate mortality was 854.9 days (range 1–1,120 days). During the research period 175 patiets died (40.7%). When plotting the risk function survival curve (Kaplan Meier) (Fig. 2) we observed 30.3% of deaths occurred in the first six months, reaching 49% in the first year. Twenty-three patients (5.3%) died during their stay at the rehabilitation unit due to a respiratory infection (55%), heart failure (30%) and the remaining 15% due to other causes. Factors associated to intra-hospital mortality can be found in Table 2. Of the 407 patients discharged, 152 (37.3%) died during the 3 year follow up, mostly because of imbalances in their heart condition (41%) respiratory infection (34%), dementia (13.7%), or tumours (11.2%). In the Cox univariable study, the factors associated to an increase in long term mortality can be found in Table 3. The hypoalbuminemia was one of the most important factors associated with mortality. The albumin concentration is significantly lowest in both, in patients who die during the admission (Table 2), and in patients who die during follow-up (Table 3). The most important factors that predict mortality in the long term are age (Fig. 3, panel A), comorbidity (Fig. 3, panel B), recovery of walking autonomy (Fig. 3, panel C) and cognitive status (Fig. 3, panel D). The type of fracture, type of surgery, delay until surgery, and length of stay at the rehabilitation unit were not associated to long term mortality (p > 0.05). 3.2. New institutionalization Upon discharge, 362 patients (84.2%) returned to their usual environment and 45 patients (10.5%) were admitted to a care home. The characteristics of institutionalized patients compared to those who returned to their usual environment can be found in Table 4. The results of the adjusted multivariable study of the factors that predict intra-hospital mortality, mortality in the long term and new institutionalization are shown in Table 5. 4. Discussion The main aim of this study is to identify the factors that influence new institutionalization and mortality in both the short and long
12
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Table 2 Factors Associated with Short Term Mortality in Univariable Logistic Regression Analysis. Variable
Death (n 23)
Discharge (n 407)
Crude OR (95% CI)
p
Age (y) M ± SD Male Charlson index ≥ 3 LoS Orthopaedics m (IQR) (days) Complications ≥ 3 Heart failure Respiratory infection Delirium Pressure ulcers Admission Albumin mg/dL m (IQR) MMSE m (IQR) Barthel index m (IQR) Previous Admission FAC m (IQR) Previous Admission Inpatient rehabilitation
86.3 ± 6.9 11 (47.8) 21 (91.3) 16 (9.0–30.0) 19 (82.6) 10 (43.5) 15 (65.2) 17 (73.9) 12 (52.2) 2.5 (2.0–2.9) 16 (10–20)
84.1 ± 7.4 321 (78.9) 170 (41.8) 8 (10.0–13.0) 144 (35.4) 86 (21.1) 89 (21.9) 177 (43.5) 55 (13.5) 3.1 (2.8–3.4) 24 (17–28)
2.58 (1.04–6.41) 3.42 (1.45–8.42) 14.6 (3.38–63.26) 1.18 (1.09–1.24) 6.57 (2.39–18.1) 2.87 (1.22–6.77) 6.69 (2.75–16.3) 3.68 (1.42–9.53) 6.98 (2.93–16.6) 0.04 (0.01–0.14) 0.92 (0.88–0.96)
0.035 0.005 <0.001 <0.001 <0.001 0.016 <0.001 0.007 <0.001 <0.001 <0.001
70 (70–90) 0 (0–0)
85 (60–100) 10 (0–20)
0.98 (0.97–0.99) 0.75 (0.64–0.88)
0.007 <0.001
3.0 (2.0–4.0) 1.0 (0.0–1.0) 1 (4.3)
4.0 (3.0–5.0) 1.0 (0.0–1.0) 350 (86.0)
0.51 (0.38–0.71) 0.043 (0.01–0.31) 0.007 (0.01–0.31)
<0.001 <0.001 0.002
Data are expressed as n (%), unless otherwise specified, M: mean, SD: standard deviation, m: median, IQR: interquartile range. FAC: function ambulation category, LoS: length of stay.
Table 3 Factors associated with three years mortality in the univariable cox regression analysis. Variable
Death (n 152)
Alive at the end of the study (n 255)
Age (years) M ± SD Male Charlson Index ≥ 3
86.5 ± 6.6 41 (27.0) 96 (63.2)
82.6 ± 7.6 45 (17.6) 74 (29.0)
Crude HR (95% CI) 1.06 (1.04–1.09) 1.57 (1.10–2.25) 3.09 (2.22–4.30)
p <0.001 0.013 <0.001
Medical comorbidities Cardiovascular disease Diabetes Stroke Dementia
59 (38.8) 54 (35.5) 50 (32.9) 83 (54.6)
100 (39.2) 51 (20.0) 48 (18.8) 72 (28.2)
1.88 (1.36–2.60) 1.46 (1.04–20.7) 1.48 (1.04–2.10) 2.07 (1.51–2.85)
<0.001 0.029 0.019 <0.001
Drugs Benzodiazepines Neuroleptics
48 (31.6) 80 (52.6)
142 (55.7) 26 (10.2)
1.39 (1.03–1.87) 2.63 (1.85–3.75)
0.030 <0.001
Complications ≥ 3 Heart failure Respiratory infection Delirium Pressure ulcers Anaemia transfused Kidney insufficiency
74 (48.7) 45 (29.6) 24 (15.8) 93 (61.2) 37 (24.3) 129 (84.9) 24 (15.8)
70 (27.5) 41 (16.1) 46 (18.0) 84 (32.9) 18 (7.1) 111 (43.5) 16 (6.3)
2.17 (1.58–2.98) 1.95 (1.38–2.77) 1.63 (1.14–2.32) 2.54 (1.93–3.52) 3.00 (2.44–4.75) 2.22 (1.42–3.46) 2.12 (1.37–3.29)
<0.001 <0.001 0.007 <0.001 <0.001 <0.001 0.001
Haemoglobin mg/dL m (IQR) Admission Discharge
10.0 (11.0–12.0) 11.6 (10.1–12.5)
11.1 (10.4–12.0) 12.0 (11.0–12.8)
0.88 (0.78–0.99) 0.83 (0.73–0.94)
0.035 <0.001
Albumin mg/dL m (IQR) Admission Discharge
2.7 (2.5–3.1) 2.9 (2.6–3.2)
2.8 (2.6–3.2) 3.1 (2.8–3.4)
0.51 (0.34–0.77) 0.61 (0.42–0.90)
0.011 0.012
MMSE m (IQR)
20 (10–25)
24 (20–30)
0.94 (0.92–0.95)
<0.001
Barthel index m (IQR) Previous Admission Discharge
75 (35–90) 0 (0–0) 20 (5–63)
90 (70–100) 10 (5–20) 65 (20–80)
0.94 (0.92–0.95) 0.92(0.93–0.97) 0.98 (0.97–0.98)
<0.001 <0.001 <0.001
FAC m (IQR) Previous Discharge
4 (3–4) 2 (1–3)
4 (4–5) 3 (2–3)
0.72 (0.63–0.82) 0.72 (0.63–0.82)
<0.001 <0.001
Inpatient rehabilitation Walk Recoverya Functional Recoverya
120 (78.9) 58 (38.2) 54 (35.5)
230 (90.2) 178 (69.8) 142 (55.7)
0.46 (0.31–0.69) 0.35 (0.25–0.49) 0.60 (0.42–0.86)
<0.001 <0.001 0.006
Data are expressed as n (%), unless otherwise specified. FAC: function ambulation category, IQR: interquartile range, M: mean, m: median, MMSE: Mini Mental State Examination, SD: standard deviation. a Expressed as a percentage of the total number of patients who received rehabilitation. (n 350).
term in elderly people with hip fractures admitted to a geriatric rehabilitation unit. Total mortality was 40% over 3 years follow up, with a lower mortality in the first year than in previous studies [5]. This dif-
ference may stem from the fact that most of these studies were carried out in acute trauma surgery units and not in post acute rehabilitation units, as was ours.
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
13
Fig. 3. Factors associated with long term mortality. Kaplan Maier curve. Panel A. Association between long-term mortality with age. Patients were split into: age ≤ a 80 (solid line), between 81 and 90 (dotted line) and >90 (dashed line). Panel B. Long-term mortality and comorbidity (Charlson Index). Low comorbidity was expressed by Charlson Index <3, and higher comorbidity for Charlson Index ≥3. Panel C. Association between long-term mortality with discharge FAC (Functional Ambulation Category). Patients were split into: independent walkers (FAC ≥ 3) or patients who required aid in order to walk (FAC <3). Panel D. Association between long-term mortality and cognitive status (MMSE, Mini Mental State Examination). Patients were classified as; cognitively whole (MMSE ≥ 24, solid line), with mild cognitive impairment (MMSE between 23 and 16, dotted line) or with severe cognitive impairment (MMSE < 16, dashed line).
Tabla 4 Significant factors for new institutionalitation in the univariable logist regression analysis. Variable
New Institutionalitation (n = 45)
Discharge to usual environment (n = 362)
Age (years) M ± SD Living alone before the fracture LoS rehabilitation m (IQR) (days) n of complications, m (IQR) Urinary retention Urinary infection MMSE m (IQR) Discharge Barthel Index m (IQR)
86.4 ± 6.8 22 (48.9) 43 (32–54) 2 (1.5–4) 13 (28.9) 20 (44.4) 20 (11.5–26.5) 30 (5–65)
84.1 ± 7.4 80 (22.1) 35 (23–52) 1.5 (1.0–3.0) 58 (16.0) 83 (22.9) 24 (17–28) 85 (60–100)
Crude OR (95% CI) 1.06 (1.01–1.11) 2.93 (1.45–8.42) 1.18 (1.09–1.24) 1.21 (1.01–1.47) 1.10 (1.07–1.24) 1.04 (1.01–1.23) 0.96 (0.93–0.99) 1.99 (1.01–3.96)
p 0.032 0.003 0.010 0.045 0.032 0.037 0.002 0.049
Data are expressed as n (%), unless otherwise specified. FAC: function ambulation category, IQR: interquartile range, M: mean, m: median, MMSE: Mini Mental State Examination, SD: standard deviation.
The population included in the study presented advanced old age. Several studies showed that age is a factor influencing mortality [6,7] and new institutionalization [4]. In our study, age did
not predict intra-hospital mortality, contrary to the observations of Eschbach el al. [8], or institutionalization, with a low association to long term mortality. As has been previously demonstrated [9],
14
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Table 5 Multivariable models of predictors factors for mortality and new institutionalization. Adjusted multivariable regression (95% Confidence Interval) Variable
Died in hospitala,c
Age Female sex Live alone FAC Discharge MMSE Charlson Index n◦ of complication Discharge Albumin mg/dL Length of Stay (d)
– – – – – 1.46 (1.06–2.01) 1.75 (1.16–2.64) 0.15 (0.34–0.68) –
p
0.020 0.022
Died 36 monthsb,d
p
New institutionalizationa,e
1,04 (1.01–1.06) 0.59 (0.40–0.86) – 0.74 (0.63–0.87) 0.98 (0.96–1.00) 1.19 (1.09–1.30) 1.17 (1.05–1.31) 0.60 (0.41–0.86) –
0.003 0.007
–
0.027 0.038 <0.001 0.003 0.005
3.95 (1.38–11.3) – 0.93 (0.89–0.97) – – – 1.02 (1.01–1.03)
p
0.003 0.001
0.028
FAC: Function Ambulation Category; MMSE: Mini-Mental State Examination. a Binary Logistic Regression, Odd Ratio (OR). b Cox regression, Hazard Ratio (HR) Adjusted Factors: c Age, Sex, Charlson index, Tumor, use of antidepressant, Traumatology Unit Length of Stay, Number and Types of Complications (Cardiac failure, Respiratory infection, Delirium and Pressure Ulcers), Albumin Concentration, MMSE, Previous Barthel Index, Previous FAC. d Age, Sex, Previous Residence, Charlson index, Comorbidities (Cardiac failure, Dementia, Stroke, Tumor, Diabetes), use of Benzodiazepines and Neuroleptics, Number and Types of Complications (Cardiac failure, Respiratory infection, Delirium, Renal Failure, anemia and Pressure Ulcers), Albumin Concentration, MMSE, Previous and Discharge Barthel Index, Discharge FAC, Inpatient Rehabilitation. e Age, Caregiver, Orthopaedics Unit Length of Stay, MMSE, Albumin Concentration, Number and Types of Complications (Urinary infection and retention), Discharge Barthel Index.
our results may be due to the fact that we took other correction factors other than age which may influence mortality into account (comorbidity, complications, cognitive impairment and functional status). Comorbidity measured with the Charlson index [9,10], and the number and type of complications [11–14] were also important factors associated to an increase in mortality in our study, whether in hospital or during follow up after discharge. Hypoalbuminemia represents a good marker for mortality [15,16] and was, in our study, one of the strongest predictive factors for mortality over 3 years. Further research will be needed in order to assess whether intensive treatment aimed at reinstating normal albumin levels can reduce mortality. The association we found between haemoglobin concentration and mortality lost statistical significance in the multivariable analysis. We believe this weak association depends on other factors such as comorbidity, complications or age, as shown in previous research [17]. Almost half of the included patients had received a blood transfusion during their stay at the acute trauma unit. Like in previous studies, we found no differences in mortality between patients who had received blood transfusions and those who had not [18]. There is controversy about the effects of delays in surgery [19,20], as well as of the type of fracture [21] on mortality. Surgical delay was greater than 72 h for 38% of the patients included in this study but, contrary to other studies, [6,12] we found no link to mortality. Hip fractures are one of the most important causes of loss of function and most studies view functional status prior to fracture as a predictor of mortality [22,23] and institutionalization [24]. We observed that ability to walk after the fracture is a long term mortality predictor. We recently researched the factors associated to functional recovery in elderly patients with hip fractures admitted to a geriatric rehabilitation unit, and observed how functional status and ability to walk before the fracture predict the recovery of walking ability upon discharge [25]. Eighty two percent of patients included carried out rehabilitation; fifty-five percent recovered their prior functional status and 64% recovered the ability to walk. We observed lower mortality among these patients, as other studies have also shown [26–28]. Results from previous research, together with our recent findings lead us to believe that, in addition to the prevention of falls, it would be necessary to admit
these patients to specialized hip fracture rehabilitation units in order to reduce disability and mortality [29]. The prevalence of cognitive impairment is high among elderly people with hip fractures, and is associated to a high mortality [9]. In our study, the correlation between the MMSE and mortality is weak; nevertheless, it is one of the main factors predicting institutionalization. The precarious balance maintained by people with cognitive impairment in a family setting is often shattered by an acute traumatic event such as a hip fracture. The other great problem for elderly people is living alone, a factor that increases the risk of institutionalization following a hip fracture almost fourfold, a result which coincides with what other studies found [30]. Finally, in previous studies, institutionalization appears to be associated to higher mortality [28], whereas we found no such link in our multivariable analysis. Our study has a number of limitations. Firstly, it was carried out in a post-acute rehabilitation unit, and data on mortality during stay in the acute trauma unit and the total number of subjects who suffered a hip fracture in our community were not available. Secondly, we did not assess the evolution of functional state after discharge and its possible link to mortality. Finally, the study was carried out within a specific geriatric post-acute rehabilitation unit, with a population whose characteristics may limit the possibility of extrapolating our results to other settings or environments. Despite these limitations, we believe this study represents a real snapshot of a geriatric rehabilitation unit where we come across elderly, complex, heterogeneous patients with higher comorbidity and a high risk of complications in our day to day work. The association of all these factors complicates both care of these patients and their inclusion in research studies.
5. Conclusions Bearing in mind the complexity of elderly people with hip fractures; age, comorbidity and complications should not be considered isolated variables in the prediction of mortality or institutionalization. Other factors such as functional and cognitive status should be taken into account in the identification of patients at risk. To conclude, integrated care in specific rehabilitation units that takes into account the cumulative effect of these factors could reduce adverse health results.
F. Uriz-Otano et al. / Maturitas 89 (2016) 9–15
Further research shall be needed in order to assess, among other points, the economic aspects of functional loss and institutionalization of elderly people who have suffered a hip fracture, as well as its impact on quality of life. Author contributions Francisco Uriz-Otano and Vincenzo Malafarina: study concept and design, analysis and interpretation of data, and preparation of manuscript. Francisco Uriz-Otano acquisition of participants and data. Gregorio Tiberio Lopez and Jorge Pla Vidal collaborated actively in the scientific basis of the paper, and in drafting and editing the manuscript. Conflict of interest None declared. Funding None. Ethics This study (19/2013) was approved by the Navarra Region Clinical Research Ethics Committee. Provenance and peer review This article was peer reviewed. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.maturitas.2016. 04.005. References [1] M.L. Alvarez-Nebreda, M.T. Vidan, J.A. Serra, Hip fracture management and outcomes in Spain, Eur. Geriatr. Med. 2 (2010) 108–111. [2] E. Cirera, K. Perez, E. Santamarina-Rubio, A.M. Novoa, M. Olabarria, Improvements in hip fracture incidence counterbalanced by the rise of other fracture types: data from Spain 2000–2010, Injury 12 (2014) 2076–2083. [3] E.K. Osnes, C.M. Lofthus, H.E. Meyer, J.A. Falch, L. Nordsletten, I. Cappelen, I.S. Kristiansen, Consequences of hip fracture on activities of daily life and residential needs, Osteoporos. Int. 7 (2004) 567–574. [4] T. Hagino, S. Ochiai, E. Sato, Y. Watanabe, S. Senga, H. Haro, Prognostic prediction in patients with hip fracture: risk factors predicting difficulties with discharge to own home, J. Orthop. Traumatol. 2 (2011) 77–80. [5] J.J. Roche, R.T. Wenn, O. Sahota, C.G. Moran, Effect of comorbidities and postoperative complications on mortality after hip fracture in elderly people: prospective observational cohort study, BMJ 7529 (2005) 1374–1376A. [6] B.D. Chatterton, T.S. Moores, S. Ahmad, A. Cattell, P.J. Roberts, Cause of death and factors associated with early in-hospital mortality after hip fracture, Bone Joint J. 2 (2015) 246–251. [7] A. Soderqvist, W. Ekstrom, S. Ponzer, H. Pettersson, T. Cederholm, N. Dalen, M. Hedstrom, J. Tidermark, Stockholm hip fracture group prediction of mortality in elderly patients with hip fractures: a two-year prospective study of 1, 944 patients, Gerontology 5 (2009) 496–504. [8] D.A. Eschbach, L. Oberkircher, C. Bliemel, J. Mohr, S. Ruchholtz, B. Buecking, Increased age is not associated with higher incidence of complications, longer stay in acute care hospital and in hospital mortality in geriatric hip fracture patients, Maturitas 2 (2013) 185–189.
15
[9] D. Norring-Agerskov, A.S. Laulund, J.B. Lauritzen, B.R. Duus, S. van der Mark, M. Mosfeldt, H.L. Jorgensen, Metaanalysis of risk factors for mortality in patients with hip fracture, Dan. Med. J. 8 (2013) A4675. [10] J. Karres, N.A. Heesakkers, J.M. Ultee, Vrouenraets BC. Predicting 30-day mortality following hip fracture surgery: evaluation of six risk prediction models, Injury 2 (2015) 371–377. [11] A. Ozturk, Y. Ozkan, S. Akgoz, N. Yalcyn, R.M. Ozdemir, S. Aykut, The risk factors for mortality in elderly patients with hip fractures: postoperative one-year results, Singapore Med. J. 2 (2010) 137–143. [12] K.A. Lefaivre, S.A. Macadam, D.J. Davidson, R. Gandhi, H. Chan, H.M. Broekhuyse, Length of stay, mortality, morbidity and delay to surgery in hip fractures, J. Bone Joint Surg. Br. 7 (2009) 922–927. [13] M. Harstedt, C. Rogmark, R. Sutton, O. Melander, A. Fedorowski, Impact of comorbidity on 6-month hospital readmission and mortality after hip fracture surgery, Injury 4 (2015) 713–718. [14] M. Mariconda, G.G. Costa, S. Cerbasi, P. Recano, E. Aitanti, M. Gambacorta, M. Misasi, The determinants of mortality and morbidity during the year following fracture of the hip: a prospective study, Bone Joint J. 3 (2015) 383–390. [15] A.S. Laulund, J.B. Lauritzen, B.R. Duus, M. Mosfeldt, H.L. Jorgensen, Routine blood tests as predictors of mortality in hip fracture patients, Injury 7 (2012) 1014–1020. [16] S. Cabrerizo, D. Cuadras, F. Gomez-Busto, I. Artaza-Artabe, F. Marin-Ciancas, V. Malafarina, Serum albumin and health in older people: review and meta analysis, Maturitas 1 (2015) 17–27. [17] L.J. Potter, B. Doleman, I.K. Moppett, A systematic review of pre-operative anaemia and blood transfusion in patients with fractured hips, Anaesthesia 4 (2015) 483–500. [18] J.L. Carson, M.L. Terrin, H. Noveck, D.W. Sanders, B.R. Chaitman, G.G. Rhoads, G. Nemo, K. Dragert, L. Beaupre, K. Hildebrand, W. Macaulay, C. Lewis, D.R. Cook, G. Dobbin, K.J. Zakriya, F.S. Apple, R.A. Horney, J. Magaziner, FOCUS investigators. Liberal or restrictive transfusion in high-risk patients after hip surgery, N. Engl. J. Med. 26 (2011) 2453–2462. [19] E. Castronuovo, P. Pezzotti, A. Franzo, D. Di Lallo, G. Guasticchi, Early and late mortality in elderly patients after hip fracture: a cohort study using administrative health databases in the Lazio region, Italy, BMC Geriatr. (2011) 37 (2318-11-37). [20] N. Simunovic, P.J. Devereaux, S. Sprague, G.H. Guyatt, E. Schemitsch, J. Debeer, M. Bhandari, Effect of early surgery after hip fracture on mortality and complications: systematic review and meta-analysis, CMAJ 15 (2010) 1609–1616. [21] W.P. Lin, C.J. Wen, C.C. Jiang, S.M. Hou, C.Y. Chen, J. Lin, Risk factors for hip fracture sites and mortality in older adults, J. Trauma 1 (2011) 191–197. [22] G. Bellelli, P. Mazzola, M. Corsi, A. Zambon, G. Corrao, G. Castoldi, G. Zatti, G. Annoni, The combined effect of ADL impairment and delay in time from fracture to surgery on 12-month mortality: an observational study in orthogeriatric patients, J. Am. Med. Dir. Assoc. 7 (2012) 664 (e9, 664. e14). [23] T.C. Marufu, A. Mannings, I.K. Moppett, Risk scoring models for predicting peri-operative morbidity and mortality in people with fragility hip fractures: qualitative systematic review, Injury 12 (2015) 2325–2334. [24] F.J. Tarazona-Santabalbina, A. Belenguer-Varea, E. Rovira Daudi, E. Salcedo Mahiques, D. Cuesta Peredo, J.R. Domenech-Pascual, H. Gac Espinola, J.A. Avellana Zaragoza, Severity of cognitive impairment as a prognostic factor for mortality and functional recovery of geriatric patients with hip fracture, Geriatr. Gerontol. Int. 3 (2015) 289–295. [25] F. Uriz-Otano, J.I. Uriz-Otano, V. Malafarina, Factors associated with short-term functional recovery in elderly people with a hip fracture. Influence of cognitive impairment, J. Am. Med. Dir. Assoc. 3 (2015) 215–220. [26] D.P. Seitz, G.M. Anderson, P.C. Austin, A. Gruneir, S.S. Gill, C.M. Bell, P.A. Rochon, Effects of impairment in activities of daily living on predicting mortality following hip fracture surgery in studies using administrative healthcare databases, BMC Geriatr. 9 (2014) (2318-14-9). [27] T. Torpilliesi, G. Bellelli, S. Morghen, S. Gentile, E. Ricci, R. Turco, M. Trabucchi, Outcomes of nonagenarian patients after rehabilitation following hip fracture surgery, J. Am. Med. Dir. Assoc. 1 (81) (2012) e1–e5. [28] A.W. Ireland, P.J. Kelly, R.G. Cumming, Risk factor profiles for early and delayed mortality after hip fracture: analyses of linked Australian Department of Veterans’ affairs databases, Injury 6 (2015) 1028–1035. [29] P. Saez Lopez, N. Sanchez Hernandez, S. Paniagua Tejo, J.A. Valverde Garcia, M. Montero Diaz, N. Alonso Garcia, A. Freites Esteve, Clinical pathway for hip fracture patients, Rev. Esp. Geriatr. Gerontol. 4 (2015) 161–167. [30] A.J. Vochteloo, S.T. van Vliet-Koppert, A.B. Maier, W.E. Tuinebreijer, M.L. Roling, M.R. de Vries, R.M. Bloem, R.G. Nelissen, P. Pilot, Risk factors for failure to return to the pre-fracture place of residence after hip fracture: a prospective longitudinal study of 444 patients, Arch. Orthop. Trauma Surg. 6 (2012) 823–830.