ONE OF THE MORE COMMON traumatic events among

ORIGINAL ARTICLE Inpatient Rehabilitation Outcome After Hip Fracture Surgery in Elderly Patients: A Prospective Cohort Study of 946 Patients Devora Lieberman, MD, Michael Friger, PhD, David Lieberman, MD ABSTRACT. Lieberman D, Friger M, Lieberman D. Inpatient rehabilitation outcome after hip fracture surgery in elderly patients: a prospective cohort study of 946 patients. Arch Phys Med Rehabil 2006;87:167-71. Objective: To identify the factors associated with inpatient rehabilitation outcome after surgical repair of hip fracture in elderly patients. Design: A noninterventional prospective cohort study. Setting: Geriatric inpatient rehabilitation center in a tertiary university medical center in southern Israel. Participants: Patients (N 946) aged 65 years of age or older who were hospitalized for rehabilitation after surgery for hip fracture. Interventions: Not applicable. Main Outcome Measures: Functional studies by the FIM instrument and a broad spectrum of clinical, demographic, and social variables. Stepwise multiple regression was used to assess the relative contribution of the variables to the variance of the percentage change in the FIM score during the course of rehabilitation in relation to the severity of the functional impairment at its inception. Results: Eight variables were significantly and independently associated with rehabilitation outcome. Prefracture FIM scale (standardized regression coefficient in multiple regression [ ].261, P.001), serum albumin at discharge (.222, P.001), Folstein Mini-Mental State Examination (.174, P.001), visual impairment (.089, P.002), dyspnea at mild exertion (New York Heart Association class III) (.080, P.005), age (.080, P.007), poststroke motor impairment (.072, P.011), and decreased serum folic acid (.055, P.047). The total percentage of the explained variance of the primary rehabilitation outcome measure accounted for by these 8 variables (adjusted R 2 ) was 31.9%. Conclusions: The outcome of rehabilitation of elderly patients after surgical repair of hip fracture is associated with 4 clinical and nutritional correctable parameters. The other 4 variables that are associated with the process cannot be corrected but may help predict outcomes and adjust expectations. Key Words: Frail elderly; Hip fractures; Rehabilitation. 2006 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Department of Geriatrics (Devora Lieberman) and Division of Internal Medicine (Devora Lieberman, David Lieberman), Soroka University Medical Center, Klalit Health Services, Beer-Sheva, Israel; and Department of Epidemiology (Friger) and Faculty of Health Sciences (Devora Lieberman, David Lieberman), Ben-Gurion University of the Negev, Beer-Sheva, Israel. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Devora Lieberman, MD, Dept of Geriatrics, Soroka Medical Center, POB 151, Beer-Sheva 84101, Israel, e-mail: lieberma@bgumail.bgu.ac.il. 0003-9993/06/8702-9981$32.00/0 doi:10.1016/j.apmr.2005.10.002 ONE OF THE MORE COMMON traumatic events among elderly patients is hip fracture, which can cause significant mortality and morbidity. In addition to the increase in life expectancy seen over recent decades, there has been an ageadjusted increase in the incidence of hip fracture. 1 Taken together, these 2 trends have led to an exponential rise in morbidity in the elderly age group, 2 with all the associated medical, social, and economic ramifications. The treatment of choice for hip fracture is surgical repair, which is followed by rehabilitation. Rehabilitation is designed to facilitate the return of elderly patients to their premorbid status to the greatest possible extent. Among the various outcomes of rehabilitation, the elderly patient s functional condition at the end of the process is the most valuable outcome to assess in this population. This is because the functional condition represents an integrated summation of many factors that affect the elderly patient s physical and mental health. It also has the strongest effect on the patient s well-being and is critically important in assessing his/her degree of dependence on the close environment and the community. To implement this rehabilitation process in an optimal manner, it is important to identify the factors that are associated with its functional outcome. The list of factors that could affect this outcome includes modifiable and nonmodifiable ones. The former are important because treating them at the beginning of and throughout the rehabilitation process can lead to improved outcomes. The latter are important in that they can help determine the chances of successful rehabilitation. Identification of these factors was the aim of this prospective study. METHODS 167 Participants We recruited all patients 65 years and older hospitalized in the geriatric ward of the Soroka Medical Center for rehabilitation after surgical repair of hip fracture over a 92-month period between March 1, 1996, and October 31, 2003. The methods of the study have been presented in detail previously. 3 In brief, after surgery for hip fracture, patients were assessed for suitability for rehabilitation and, if found suitable, were transferred to the geriatric ward for a conventional rehabilitation program. Patients with advanced dementia and/or a low premorbid functional status were not considered suitable. The Helsinki Committee of the Soroka Medical Center approved the study, and all patients gave informed consent before enrollment. Measurements Within 48 hours of transfer to the geriatric ward, and after giving informed consent to participate in the study, patients or 1 of their family members provided demographic information and data relating to chronic comorbidity including smoking history. Blood samples drawn at this time were tested by conventional methods for levels of hemoglobin, serum albu-

168 REHABILITATION AFTER HIP FRACTURE SURGERY IN ELDERLY, Lieberman min, thyroid-stimulating hormone, folic acid, and vitamin B 12. Hemoglobin and serum albumin were retested at discharge. At this stage, each patient underwent a mental evaluation using the Folstein Mini-Mental State Examination (MMSE), with a scale ranging from 1 to 30 and a cutoff for normal of 23 and above. 4 In addition, each patient underwent an assessment of symptoms of depression using the Geriatric Depression Scale (GDS), with scores ranging from 0 to 30 and a cutoff for normal of 10 or less. 5 Functional assessments and evaluation of the progress of rehabilitation were done using the FIM instrument 6 that ranges from 18 to 126 and is based on a score of 1 to 7 for each of 18 different items, in accordance with the level of independence for each item. By using this scoring system, a patient with totally independent function would have a score of 126 points. The FIM score is applied to the following areas: eating; grooming; bathing; dressing (upper body); dressing (lower body); using the toilet; bladder management; bowel management; transferring (to go from 1 place to another) in a bed, chair, and/or wheelchair; transferring on and off a toilet; transferring into and out of a shower; locomotion (moving) for walking or in a wheelchair; locomotion going up and down stairs; comprehension; expression; social interaction; problem solving; and memory. A senior geriatrician determined the prefracture FIM score during the course of an interview with the patient or his/her family, close to admission. FIM scores on admission to rehabilitation and at its end were determined at a staff meeting of the geriatric ward in which the medical, rehabilitation, and nursing staffs took part. If the FIM scores were stable at 2 successive determinations at a 1-week interval, staff decided to end the rehabilitation process and discharge the patient. The type of fracture, surgical fixation, and type of anesthesia were recorded from the chart in the orthopedic surgery department. Various complications during the rehabilitation process were recorded, when they occurred, by the study staff. During rehabilitation, screening for corrected visual acuity in the better of the 2 eyes was performed by using the standard Snellen eye chart with the patients wearing their own eyeglasses; optimal acuity was set at 6/6 (in meters), and reduced vision was indicated by a higher denominator. Screening for hearing impairment was performed by using an audiometer a at 500, 1000, 2000, and 4000Hz. For purposes of analysis, mean decibel levels in the better of the 2 ears over the 4 frequencies were computed. Statistical Analysis Data were collected by using the Epi Info b and analyzed by using the SPSS c statistical software. Univariate statistical analysis was performed for all independent and dependent variables by using chi-square test, t tests, 1-way analysis of variance, and correlations, as appropriate. Independent variables that were found to have statistical significance were included in the multiple regression model. For multivariate analysis, forward stepwise multiple regression was used to assess the relative contribution of the various dependent variables to the variance of the absolute efficacy of rehabilitation, which was defined as follows: change in FIM during rehabilitation, divided by (126 [maximum FIM] admission FIM) 100. We decided to assess the rehabilitation outcome by absolute efficacy of rehabilitation in light of its simplicity and advantages. This measure reflects both the absolute severity of the functional impairment at the beginning of rehabilitation and the improvement in functional status achieved in its course. This important combination is unique to this index and is not expressed by any other index in use for this purpose. Each of the variables that was assessed in the course of the study was chosen based on medical logic relating to the possibility that it could be associated with the rehabilitation outcome. The 40 variables that met these criteria and, therefore, served as dependent variables for the analyses were age; sex; marital status; years of education; type of fracture; type of fixation; type of anesthesia; previous hip fracture; length of hospitalization in the orthopedic surgery department; dyspnea on mild exertion (New York Heart Association [NYHA] class III); ischemic heart disease; congestive heart failure; atrial fibrillation; chronic lung disease; hypertension; diabetes mellitus; previous stoke with motor impairment; total number of comorbid illnesses; smoking in the past only; smoking in the present; elevated serum thyroid-stimulating hormone ( 5 IU/ ml); decreased serum folic acid ( 3.0 g/l); decreased serum vitamin B 12 ( 250ng/L), hemoglobin, and serum albumin at admission and at discharge; MMSE score; GDS score; complications during rehabilitation (especially urinary tract infection, pneumonia, infection of the surgical wound, acute ischemia, deep venous thrombosis, and stroke); visual and hearing impairment; prefracture FIM score; admission FIM score; and length of rehabilitation in the geriatrics ward. Statistical significance was set as P less than.05. RESULTS Nine hundred sixty-two patients over the age of 65 were hospitalized in the geriatric ward for rehabilitation after surgery for hip fracture during the study period. A random sample over a period of 1 year in the orthopedic surgery department showed that the percentage of elderly patients that were not transferred for rehabilitation was 9.8% of all patients in this age group undergoing surgery for hip fracture. Sixteen (1.7%) of the patients who were transferred for rehabilitation died during its course. The other 946 patients who were discharged at the end of the hospitalization comprised the study group. Table 1 presents the principal characteristics of the study group, in particular sex, age, cognitive state, length of hospitalization in the orthopedic surgery department, length of hospitalization for rehabilitation in the geriatrics ward, and functional status at the 3 time points in relation to the rehabilitation process. Only 8 of the 40 tested variables were significantly and independently associated with absolute efficacy of rehabilitation. These 8 variables are listed in table 2, and their order of importance is expressed by the absolute value of the standardized regression coefficient ( ). The positive or negative value of this coefficient reflects a corresponding positive or negative Table 1: Sociodemographic Characteristics, Functional Status, and Hospitalization Variables of the Study Population (N 946) Variable Value Range Women, n (%) 717 (76) NA Mean age SD (y) 78.3 7.1 65 98 Mean MMSE score SD* 24 5.7 10 30 Mean prefracture FIM score SD 115.0 14.7 45 126 Mean days in orthopedic surgery ward SD 10.7 5.1 2 40 Mean days hospitalized for rehabilitation SD 22.1 8.9 6 47 Mean admission FIM score SD 74.0 16.2 25 108 Mean discharge FIM score SD 89.0 19.5 18 113 Abbreviations: NA, not applicable; SD, standard deviation. *Available for only 937 of the 946 patients.

REHABILITATION AFTER HIP FRACTURE SURGERY IN ELDERLY, Lieberman 169 Variable Table 2: Results of the Multivariate Analysis Regression Coefficient Standard Error P Cumulative Adjusted R 2 Prefracture FIM score (recall) 0.326 0.039.261.001.176 Serum albumin at discharge (g/dl) 8.418 1.108.222.001.260 MMSE 0.565 0.102.174.001.293 Dyspnea on mild exertion, NYHA class III (1 yes, 0 no) 5.247 1.856.080.005.301 Visual impairment (1 yes, 0 no) 0.070 0.022.089.002.308 Age (y) 0.202 0.075.080.007.312 Prior stroke with motor impairment (1 yes, 0 no) 4.683 1.846.072.011.316 Decreased serum folic acid ( 3 g/l) (1 yes, 0 no) 3.999 2.009.055.047.319 NOTE. Variables independently associated with the absolute efficacy of rehabilitation in the order of their adjusted R 2 value. contribution of the specific variable to the absolute efficacy of rehabilitation. The total percentage of the explained variance of the absolute efficacy of rehabilitation accounted for by these variables (adjusted R 2 ) was 31.9%. To avoid the pitfall of overlooking or discounting other important modifiable factors that could be lost in the stepwise regression, we looked at each of the other 4 modifiable factors individually to see if any of them predicts the dependent variable after adjusting for the 4 nonmodifiable covariates (prefracture FIM score, MMSE score, age, prior stroke with motor impairment). This analysis showed that hemoglobin concentration at discharge significantly predicted the dependent variable (P.009), but elevated serum thyroid-stimulating hormone, decreased serum vitamin B 12, and hearing impairment did not. DISCUSSION Two methodologic aspects of our study need to be addressed. First, not all patients who underwent surgical repair of hip fracture were included in the study. This situation led to an inevitable selection bias in the study population. Second, we assessed rehabilitation outcome at the time of discharge from the hospital. We chose this particular point from the entire time sequence of the rehabilitation process that continues beyond discharge as well because we believe that it represents the central point in time in the process. In most cases, completion of the inpatient phase of rehabilitation enables continuation of the process in the framework of the community, where the patient returns to his/her natural environment. It also represents the end of the large expenses incurred in the hospital phase of the process. Of the 8 parameters that were found in this study to be significantly and independently associated with rehabilitation outcome, 5 were identified in previous studies. However, the point in time and the rehabilitation framework in which the rehabilitation outcome was assessed were different among the various studies and from the present one. These 5 parameters include the prefracture functional status, 7-14 the mental status, 9,10,12-16 age, 13,14,16-18 previous stroke, 19,20 and serum albumin level. 21,22 In relation to the last variable, our study introduced an important new dimension (ie, that serum albumin level at discharge rather than at admission has a significant association with the outcome of rehabilitation). This result does not enable us to determine whether this positive association represents a cause or an effect of the outcome but does enable us to propose interventions to raise the serum albumin level by appropriate dietary adjustments and supplementation. Dyspnea on mild exertion (NYHA class III), which was identified in this study as a negative, independent factor associated with rehabilitation outcome, was not reported in previous studies. This significant negative association is striking in light of the absence of such an association for cardiovascular and pulmonary diseases, which are usually the clinical cause of dyspnea, and is in contrast with the results of a previous study in which these diseases were found to have a negative prognostic effect on rehabilitation among hip fracture patients. 23 We believe that these findings reflect the fact that only the severe functional impairment caused by these diseases is associated with the outcome of rehabilitation and not the diseases per se. The negative association between visual impairment and rehabilitation outcome is not surprising. However, it should be noted that this association was not found for hearing impairment. This aspect of the association of visual and hearing impairments with the outcome of rehabilitation in hip fracture patients was described and discussed in a previous article that was dedicated to that issue. 3 The finding that decreased serum folic acid has a negative association with the success of rehabilitation after hip fracture is interesting and expresses, in our opinion, the influence of the nutritional state of the patients in the weeks before the hip fracture, in contrast to decreased serum vitamin B 12 that reflects a more prolonged nutritional deficit that was not found in our study to be an independently significant factor for the success of rehabilitation. During the course of rehabilitation, patients who were found to have decreased serum folic acid levels received supplementation. Despite this strategy, a decreased serum folic acid level at the beginning of rehabilitation had a negative association with its outcome. We assume that the short period of time in which this treatment was given in relation to the length of hospitalization for rehabilitation was not sufficient for the supplementation therapy to affect the outcome of rehabilitation. The interesting association of low folic acid levels with poor functional recovery after hip fracture may be linked, indirectly, to the association that has recently been found between risk for hip fracture and elevated homocysteine in older persons 24 and the association between low levels of folic acid and elevated levels of homocysteine. 25 The hemoglobin concentration at discharge was also found to have a significant effect on the results of rehabilitation, although this effect was identified only through secondary analyses of the results. The significance of this finding is that this variable has an effect on the results of rehabilitation but is less important than the 4 other modifiable factors that were identified in the primary study analyses. In rehabilitation of elderly patients, it is common clinical practice to correct anemia at the beginning of and throughout the rehabilitation process by transferring blood until a serum hemoglobin concentration of around 10mg/dL is achieved and by stimulating the

170 REHABILITATION AFTER HIP FRACTURE SURGERY IN ELDERLY, Lieberman blood marrow to continue correcting anemia by supplementation with iron and folic acid. It is possible that the finding on the effect of the hemoglobin concentration at discharge on the outcome of rehabilitation indicates a need to increase the hemoglobin concentration to an even higher level at the beginning of and throughout rehabilitation, but this issue should be addressed in a specifically designed study. In addition to the list of variables that had an independent, significant association with the results of rehabilitation, the 32 variables that did not have an effect are noteworthy. Intuitively, some of those variables were expected to have a significant association, but the results of our study did not support these intuitions. The model that we used to analyze the results negates variables that affect another, more significant variable. This method of analysis led to the situation in which a relatively large number of specific diseases that are common in this age group, the total number of comorbid diseases, the GDS score, hypothyroidism, and a decreased serum vitamin B 12 level that have a significant effect on the prefracture functional status were neutralized, in effect, by the recall FIM instrument, which had the most significant association with rehabilitation outcome. This result supports our contention, described previously, that the impaired functional state that stems from certain diseases, rather than the diseases per se, has the greatest association with the results of rehabilitation. Apparently, the results of our study point to the possibility of improving rehabilitation outcome by therapeutic intervention in its early phases. The aim of this intervention should be to increase the serum albumin level by appropriate dietary adjustments and supplementation, to correct those visual defects that can be corrected such as prescription of suitable eyeglasses, to improve dyspnea by optimal treatment of the diseases that cause it, and to correct decreased serum folic acid levels by supplementation. However, the results of our study can only support an association and not causality between these 4 parameters and rehabilitation outcome. This situation prevents us from concluding that intervening with these factors may alter the outcome. The issue of causality should be investigated in interventional studies specifically designed to address this question. The other 4 variables that were identified as having a significant association with rehabilitation outcome (prefracture functional state, mental status, age, prior stroke) cannot be changed by intervention during rehabilitation. However, the importance of these variables lies in their value as predictors of rehabilitation outcome. This assessment is important to help patients, families, and staff develop realistic expectations from the process and for directors of health care services for the elderly to reach reasonable decisions relating to rehabilitation strategies. Two of these 4 nonmodifiable variables (prefracture functional state, mental status) might be targets for rehabilitation in patients at high risk. Recipients of these interventions could benefit from such a preventive strategy. In relation to the variable of age, it should be noted that, although age was found to be a negative prognostic factor in this study, successful rehabilitation has been achieved in patients 85 years of age and above. 26 CONCLUSIONS The outcome of rehabilitation of elderly patients after surgical repair of hip fracture is associated with 4 correctable clinical and nutritional parameters. The other 4 associated significant variables cannot be corrected but may help in predicting outcomes and adjusting expectations of all individuals involved in the rehabilitation process. References 1. Lyons AR. Clinical outcomes and treatment of hip fractures. Am J Med 1997;103:51S-63S. 2. Cooper C, Campion G, Melton LJ 3rd. Hip fractures in the elderly: a world-wide projection. Osteoporos Int 1992;2:285-9. 3. Lieberman D, Friger M, Lieberman D. Visual and hearing impairment in elderly patients hospitalized for rehabilitation following hip fracture. J Rehabil Res Dev 2004;41:669-74. 4. Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-96. 5. Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 1983;17:37-49. 6. Keith RA, Granger CV, Hamilton BB, Sherwin FS. The functional independence measure: a new tool for rehabilitation. Adv Clin Rehabil 1987;1:6-18. 7. Koval KJ, Skovron ML, Aharonoff GB, Meadows SE, Zuckerman JD. Ambulatory ability after hip fracture. A prospective study in geriatric patients. Clin Orthop Relat Res 1995;Jan(310):150-9. 8. Myers AH, Palmer MH, Engel BT, Warrenfeltz DJ, Parker JA. Mobility in older patients with hip fractures: examining prefracture status, complications, and outcomes at discharge from the acute-care hospital. J Orthop Trauma 1996;10:99-107. 9. Marottoli RA, Berkman LF, Leo-Summers L, Cooney LM Jr. Predictors of mortality and institutionalization after hip fracture: the New Haven EPESE cohort. Established Populations for Epidemiologic Studies of the Elderly. Am J Pub Health 1994;84: 1807-12. 10. Kaehrle P, Maljanian R, Bohannon RW, Horowitz S. Factors predicting 12-month outcome of elderly patients admitted with hip fracture to an acute care hospital. Outcomes Manag Nurs Pract 2001;5:121-6. 11. Koval KJ, Skovron ML, Aharonoff GB, Zuckerman JD. Predictors of functional recovery after hip fracture in the elderly. Clin Orthop Relat Res 1998;Mar(348):22-8. 12. Cree M, Carriere KC, Soskolne CL, Suarez-Almazor M. Functional dependence after hip fracture. Am J Phys Med Rehabil 2001;80:736-43. 13. Parker MJ, Palmer CR. Prediction of rehabilitation after hip fracture. Age Ageing 1995;24:96-8. 14. Cree AK, Nade S. How to predict return to the community after fractured proximal femur in the elderly. Aust N Z J Surg 1999; 69:723-5. 15. Heruti RJ, Lusky A, Barell V, Ohry A, Adunsky A. Cognitive status at admission: does it affect the rehabilitation outcome of elderly patients with hip fracture? Arch Phys Med Rehabil 1999; 80:432-6. 16. Magaziner J, Simonsick EM, Kashner TM, Hebel JR, Kenzora JE. Predictors of functional recovery one year following hospital discharge for hip fracture: a prospective study. J Gerontol 1990; 45:M101-7. 17. Barnes B, Dunovan K. Functional outcomes after hip fracture. Phys Ther 1987;67:1675-9. 18. Broos PL, Van Haaften KI, Stappaerts KH, Gruwez JA. Hip fractures in the elderly: mortality, functional results and social readaptation. Int Surg 1989;74:191-4. 19. Lieberman D, Fried V, Castel H, Weitzmann S, Lowenthal MN, Galinsky D. Factors related to successful rehabilitation after hip fracture: a case-control study. Disabil Rehabil 1996;18:224-30. 20. Youm T, Aharonoff G, Zuckerman JD, Koval KJ. Effect of previous cerebrovascular accident on outcome after hip fracture. J Orthop Trauma 2000;14:329-34. 21. Cummings SR, Phillips SL, Wheat ME, et al. Recovery of function after hip fracture. The role of social supports. J Am Geriatr Soc 1988;36:801-6.

REHABILITATION AFTER HIP FRACTURE SURGERY IN ELDERLY, Lieberman 171 22. Van Hoang H, Silverstone FA, Leventer S, Wolf-Klein GP, Foley CJ. The effect of nutritional status on length of stay in elderly hip fracture patients. J Nutr Health Aging 1998;2:159-61. 23. Stavrou ZP, Erginousakis DA, Loizides AA, Tzevelekos SA, Papagiannakos KJ. Mortality and rehabilitation following hip fracture. A study of 202 elderly patients. Acta Orthop Scand Suppl 1997;275:89-91. 24. McLean RR, Jacques PF, Selhub J, et al. Homocysteine as a predictive factor for hip fracture in older persons. N Engl J Med 2004;350:2042-9. 25. Lowering blood homocysteine with folic acid based supplements: meta-analysis of randomised trials. Homocysteine Lowering Trialists Collaboration. BMJ 1998;316:894-8. 26. Lieberman D, Lieberman D. Rehabilitation after proximal femur fracture surgery in the oldest old. Arch Phys Med Rehabil 2002; 83:1360-3. Suppliers a. GSI-16 audiometer; Grason-Stadler Inc, 1 Westchester Dr, Milford, NH 03055. b. Epi Info, Centers for Disease Control and Prevention, Epidemiology Program Office, 1600 Clifton Rd, Mail Stop E-91, Atlanta, GA 30333. c. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.