Department of Orthopedics and Rehabilitation, UF&Shands Orthopaedics and Sports Medicine Institute, P.O. Box 112727, University of Florida, Gainesville, FL 32611. E-mail: email@example.com
Objective: This study examined whether obesity affected inpatient rehabilitation outcomes after total hip arthroplasty (THA).
Research Methods and Procedures: This was a retrospective, comparative study conducted using a computerized medical database derived from THA patients at a university-affiliated rehabilitation hospital (data from 2002 to 2005). Patients were divided into four brackets based on BMI: non-obese (<25 kg/m2), overweight (25 to 29.9 kg/m2), moderate obesity (30 to 39.9 kg/m2), and severe obesity (≥40 kg/m2). All patients completed an interdisciplinary inpatient rehabilitation program after THA. Functional independence measure (FIM) scores, length of stay (LOS), FIM efficiency scores (FIM/LOS), hospital charges, and discharge disposition location were collected.
Results: FIM scores improved from admission to discharge similarly in all groups (25 to 29.5 points). However, FIM efficiency, LOS, and total charges were curvilinearly related with BMI (all p < 0.05). Total hospital charges were highest in the severely obese group compared with the overweight group (p < 0.05). Non-homebound discharge disposition rates were lower in non-obese (13.1%) and severely obese groups (10.5%).
Discussion: Elevated BMI does not prevent FIM gains in THA patients during inpatient rehabilitation. However, BMI is related with FIM efficiency, LOS, and hospital charges in a curvilinear fashion. Severely obese patients can achieve physical improvements but at a lower efficiency and greater cost.
Obesity is one of the risk factors for development of hip osteoarthritis and future total hip replacement (1, 2). The cumulative effects of carrying excessive body weight may cause biomechanical strain to the hip joint (3) that contributes to joint degeneration (4) and exacerbation of osteoarthritis. The relative risk of total hip arthroplasty (THA)1 caused by osteoarthritis increases from 1.3 to 5.2 when BMI values increase from <22 to 35 kg/m2 (2).
Obesity is a risk factor for joint arthroplasty surgery and postoperative complications (5) that could delay rehabilitation progress and recovery. Often, obese, particularly severely obese, persons are considered unsuitable candidates for THA because of the increased infection risk and limited ambulatory ability (6). Severely obese persons have greater surgical blood transfusion requirements, lengthier surgery times (7), surgical complications, and require more staff assistance with postoperative transfers from supine to sit, sit to stand, and bed to chair positions (8). Obese arthroplasty patients are at greater risk for a non-homebound disposition after the surgery than non-obese, with an odds risk ratio of 1.45 (9).
One increasingly used hospital discharge location is the inpatient rehabilitation facility (IRF). The numbers of THA procedures and obese patients are increasing annually (10), and more obese persons with hip replacements will likely be entering IRFs. The evidence regarding the effect of obesity on early functional outcomes after THA are mixed, with some showing effect (5, 8), whereas others showed a small negative effect on functional abilities (11). There are no data evaluating the effect of obesity on THA outcomes and costs in the IRF. It is unknown whether obese persons have the capacity for similar functional improvements, rates of improvement, or costs for functional improvement during inpatient rehabilitation. Therefore, the purpose of this study was to retrospectively examine the effect of increasing BMI on functional and financial outcomes in patients with THA undergoing inpatient rehabilitation.
Research Methods and Procedures
The design was a retrospective study using data from the computerized medical records system of a university-affiliated rehabilitation hospital, with patient admission dates between January 1, 2002 and March 31, 2005. This study was approved by the Human Investigation Committee at the University of Virginia.
Consecutive obese and non-obese patients with diagnoses of either primary THA or revision THA were identified as participants for this analysis by the investigators (N = 339). Subjects were stratified into four groups based on the degree of obesity using BMI: non-obese (BMI < 25 kg/m2), overweight (BMI between 25 and 30 kg/m2), obese (BMI = 30 to 39.9 kg/m2), and severely obese (BMI ≥ 40 kg/m2) (12).
Patient descriptive variables included age, sex, and ethnicity. Ethnicity was recorded from patient charts. Use of tobacco and alcohol and living status (alone or with supportive relatives) were recorded from patient charts and management plans. The number and type of comorbidities were compiled from medical records using the International Code of Diseases for each patient, and these diagnoses were confirmed within each medical chart by the investigators. Physiological measures performed at the time of admission to rehabilitation included a metabolic chemistry panel, complete blood count, vital signs, height, and weight. These values were recorded by the investigators. These physiological measures confirmed any specific diagnoses from the electronic database diagnoses. The type of THA procedure (primary or revision) was obtained from medical records and confirmed from each medical chart by the surgeon's discharge summary. The THA was performed by one of three orthopedic surgeons specializing in joint arthroplasty at this institution. In the case of missing data points, the participant's record was excluded from analysis. A total of 178 complete remaining medical records were analyzed.
Study Outcome Variables
The criterion measures selected from the computer database were length of stay (LOS), functional independence measure (FIM) scores at admission and discharge, and total rehabilitation hospital charges.
Functional improvement during inpatient rehabilitation is commonly measured using the 18-item FIM (13). The FIM estimates performance during tasks that can be broadly categorized as activities of daily living, mobility, and cognitive domains (14). The FIM is an important representative measure of the overall success of the interdisciplinary rehabilitation program. The primary goal identified by almost all THA patients on their program plan within their medical records was to “get home and do things themselves.” The FIM was a validated measure to capture overall functional improvements. The total possible score on the FIM is 126, and THA patients have previously shown >20-point improvements (13, 15, 16). Cognition was included with the motor scores as the two major total scores included in the overall FIM score. Within the FIM score were other specific scores for activities of daily living. Walking ability was a FIM component that was defined as the ability to walk <15.2 (category 1), 15.2 to 45.4 (category 2), or >45.4 m (category 3). Other relevant categories included weight-bearing transfers (bed to chair, tub/shower, toilet), level of independence for walking, and stairs. A score of 0 represented no activity or change, whereas 7 represented a score of significant improvement and achieving complete independence with the activity. For the purpose of this study, selected motor scores within the mobility domain (walking, transfers) that would be affected by weight or load bearing were presented in this study.
As an estimate of the rate of functional gain made during the rehabilitation intervention, an FIM efficiency score was calculated. FIM efficiency was defined as the change in FIM from admission to discharge divided by the LOS (FIM efficiency = FIM points gained/total days). LOS and total hospital charges were obtained from medical records. Daily charges were calculated by dividing the total charges by the LOS.
Discharge disposition locations were determined from the case management portion of the patient charts. Patients were discharged to home, a skilled nursing facility, acute transfer back to the hospital, or assisted living.
Each patient completed a comprehensive interdisciplinary inpatient rehabilitation under the supervision of a physiatrist. Each patient received ∼3 hours of supervised therapy daily from both physical and occupational therapists. Standardized protocol driven therapy sessions were conducted twice daily (morning and afternoon). During therapy sessions, patients engaged in activities to increase flexibility and range of motion, improve independence with performing activities of daily living, and improve gait, balance, and proprioception. Occupational therapists instructed activities of daily living for 30 to 45 minutes in the morning session, which involved dressing, bathing, adaptive strategies, and use of assistive equipment. Also, upper extremity activities were performed in a group session (e.g., ball bouncing, stretching). Physical therapy sessions conducted during both morning and afternoon, involving 60 minutes of participation in group sessions, where one physical therapist worked with two to three patients at one time. Typical therapy protocols involved seated leg raises, isometric contractions for the ankle (gastrocnemius, quadriceps, gluteus maximus), heel slides (hip flexors, biceps femoris), terminal knee extensions, hip abduction (as long as it is not contraindicated), gait walking or walking with cane, and stair climbing. Three times a week, advanced activities of daily living were performed with the occupational therapist for 30 to 60 minutes per session. Further therapeutic interventions from speech pathology or pain psychology, although rare, were provided on a case by case basis if necessary.
Statistical analyses were performed using the Statistical Package for the Social Sciences software (version 12.0). All data are expressed as mean ± SD of measurement. Frequency distribution of diagnosis etiologies, ethnicities, smoking and alcohol use, comorbidity types, and disposition locations were analyzed using χ2 tests. Because of having four groups of different sample sizes, non-parametric Kruskal-Wallis tests were used to determine whether differences existed among the groups for outcome variables. Given that our previous work has shown that THA type (primary vs. revision) influences THA outcomes (17), we examined the potential interactions of BMI group and THA status on the main outcome variables of admission and discharge FIM scores, FIM efficiency, LOS, and total charges by applying a univariate ANOVA with group factors of BMI bracket and THA status in this exploratory study. Because these analyses did not yield any significant interactions between BMI group and THA type for all outcome measures presented in this study, we present the groups by the degree of obesity only.
Hierarchical regression analysis was performed on select physiological and anthropometric variables and BMI to determine the contribution of obesity to FIM efficiency during inpatient rehabilitation. These variables included those that we and others have previously shown to affect arthroplasty outcomes and that may influence tolerance to therapy (age, sex, living alone, comorbidity number, blood transfusion status indicating anemia, and use of abductor brace) (9, 17, 18). In this exploratory study, linear and quadratic regression models were generated to determine the contribution of BMI on THA outcomes such as LOS, FIM efficiency, and total hospital charges to determine the lines of best fit in this patient base. Pearson correlations were determined between BMI and functional outcome measures (change in FIM, LOS, FIM efficiency, total charges). A priori α levels were set at 0.050 for all statistical tests; α levels were adjusted for the Pearson correlations using a Bonferroni correction for multiple comparisons (p < 0.01).
Table 1 contains the participant characteristics. The severely obese group was younger, shorter, heavier, and had greater BMI values than the remaining groups (p > 0.05). Blood glucose levels were higher in the obese group than the non-obese group (p < 0.05). Only the severely obese group used no hip abductor braces during inpatient rehabilitation compared with the remaining groups. The percentage of persons with blood transfusions was significantly higher in the non-obese than obese and severely obese groups (p < 0.05).
Table 1. Characteristics of patients with THA at admission
Table 2 presents the number and type of comorbidities of each group. Greater proportions of persons in the severely obese group had diabetes compared with the remaining groups, whereas greater proportions of obese subjects had hypertension and hypothyroidism than the remaining groups. Rheumatoid arthritis was more prevalent in the non-obese group than the overweight and obese groups (p < 0.05). Finally, frequencies of anemia were higher in the non-obese and severely obese groups (p < 0.05).
Table 2. Comorbidities of patients with THA
Non-obese (n = 46)
Overweight (n = 62)
Obese (n = 50)
Severely obese (n = 19)
Values are total numbers and percentages of the total group.
The raw FIM scores from admission to discharge and score changes (discharge FIM – admission FIM), LOS, FIM efficiency, and hospital charges are presented in Table 3. There were no statistical differences between groups for admission FIM scores (p > 0.05). Discharge FIM scores were lower in the severely obese group compared with the non-obese group (p < 0.05). Percent changes in FIM from admission to discharge, however, were not statistically different among the four groups (range, 33.5% to 36.7%; p > 0.05). Regardless of group, main effects analysis revealed that revision THA patients had lower admission and discharge scores than primary THA patients (p < 0.05).
Table 3. Functional outcomes after THA
Non-obese (n = 46)
Overweight (n = 62)
Obese (n = 50)
Severely obese (n = 19)
Values are means ± standard deviation; 95% confidence intervals are shown in parentheses. FIM scores are expressed in points.
Table 4 contains the findings for specific components of the FIM scores for each group that are relevant to mobility and load bearing vs. cognition. Among all four groups, there were no significant differences in the admission or discharge values for any of these motor score variables between groups. Changes in these FIM component scores from admission to discharge were also not different between groups (p > 0.05). The percentage of patients who achieved the highest walking distance of >45.4 m was not different between groups (p > 0.05).
Table 4. Changes in scores of specific FIM components in four BMI brackets from admission to discharge
Non-obese (n = 46)
Overweight (n = 62)
Obese (n = 50)
Severely obese (n = 19)
Values are means ± standard deviation; 95% confidence intervals are shown in parentheses.
FIM cognition and motor scores are expressed in points. FIM motor subscores (bed-to-chair, shower/tub and toilet transfers, stairs, and walking assistance) are expressed in points ranging from 0 to 7, where 0 = no improvement and 7 = maximal points of improvement and achieved independence in this activity.
Distance walked was determined whether the patient walked <15.2 (1), 15.2 to 45.4 (2), or >45.4 m (3); data are expressed as percentage (%) of the group in each walking category.
0.18 ± 0.79 (−0.07 to 0.43)
0.42 ± 0.07 (0.07 to 0.78)
0.21 ± 1.3 (−0.11 to 0.54)
0.06 ± 0.44 (−0.17 to 0.29)
25.9 ± 11.3 (22.2 to 29.5)
25.7 ± 7.5 (23.6 to 27.7)
25.4 ± 7.2 (23.3 to 27.5)
22.4 ± 16.7 (13.5 to 31.4)
Bed to chair
2.1 ± 0.8 (1.8 to 2.4)
2.1 ± 0.9 (1.9 to 2.4)
2.1 ± 0.8 (1.9 to 2.4)
1.9 ± 2.2 (1.0 to 3.1)
3.3 ± 2.0 (2.6 to 3.9)
3.7 ± 2.0 (3.1 to 4.1)
3.3 ± 1.9 (2.7 to 3.8)
3.3 ± 2.6 (1.8 to 4.7)
1.8 ± 1.3 (1.3 to 2.2)
1.9 ± 0.8 (1.7 to 2.1)
1.9 ± 0.8 (1.7 to 2.2)
2.1 ± 1.7 (1.1 to 2.9)
4.1 ± 1.9 (3.4 to 4.7)
3.9 ± 2.1 (3.3 to 4.4)
3.6 ± 2.2 (2.9 to 4.2)
3.3 ± 2.3 (2.0 to 4.5)
Walking without assistance
3.7 ± 1.5 (3.2 to 4.1)
3.8 ± 1.5 (3.4 to 4.2)
3.9 ± 1.4 (3.5 to 4.3)
3.6 ± 2.2 (2.3 to 4.7)
1 = 6.2, 2 = 15.5, 3 = 77.3
1 = 5.1, 2 = 10.3, 3 = 82.1
1 = 1.8, 2 = 3.6, 3 = 92.9
1 = 2.2, 2 = 19.6, 3 = 78.3
Rehabilitation hospital LOS values were significantly different in the severely obese group compared with the non-obese group (p < 0.05; Table 3). To determine the best predictive fit, both linear and quadratic regressions were performed on three of the outcome measures: LOS, FIM efficiency, and total charges. After performing the linear regressions, quadratic regression analyses showed a significant curvilinear relationship between LOS and BMI, with the lower LOS found in overweight and obese persons (R2 = 0.124, p < 0.05; Figure 1). FIM efficiency scores were also found to be significantly lower in the severely obese patients compared with non-obese patients (p < 0.05). A significant but low correlation (r = −0.185, p < 0.05) existed between BMI and FIM efficiency. However, the model fit was significantly improved with a quadratic regression that revealed a significant curvilinear relationship between BMI and FIM efficiency (r = 0.309; Figure 2).
The total charges were greater in the severely obese group compared with the overweight group (p < 0.05). A quadratic regression analysis revealed that a significant curvilinear relationship existed between BMI and total charges (R2 = 0.139, p < 0.05; Figure 3). When adjusted for LOS (daily charges), hospital charges were not significantly different between THA patient groups (p > 0.05).
Non-obese and severely obese patients were less frequently discharged to home than the remaining groups. Non-obese patients were more frequently discharged to a skilled nursing facility than the remaining groups (p < 0.05). Severely obese patients were acutely transferred back to the hospital more often than overweight and obese patients (p < 0.05). These results are presented in Table 5.
Table 5. Discharge disposition locations for persons who completed inpatient rehabilitation after THA
Non-obese (n = 46)
Overweight (n = 62)
Obese (n = 50)
Severely Obese (n = 19)
Values are expressed as a percent of the total group.
The results of the hierarchal regression analysis for FIM efficiency during inpatient rehabilitation are shown in Table 6. After controlling for possible contributors to FIM efficiency (age, sex, number of comorbidities, family support in a home environment, use of abductor brace, and blood transfusion status; total R2 = 0.090), the addition of BMI itself added 0.023 to the R2 value (R2 = 0.113; F change = 3.549, p = 0.021) and significantly contributed to the FIM efficiency model in this cohort.
Table 6. Hierarchal regression analysis for FIM efficiency after THA inpatient rehabilitation
Each step includes the listed variable and the addition of the previous factor(s) and represents a separate regression equation.
This is the first report to show that BMI influences outcomes of THA patients during inpatient rehabilitation. Discharge FIM scores and FIM efficiency were higher and LOS was shorter in overweight than severely obese persons. For many of the IRF outcomes (LOS, FIM efficiency, total charges), the responses based on BMI were not linear, but rather curvilinear, with the highest functional gains and shortest LOS made in overweight persons. The lowest FIM efficiencies occurred in persons with BMIs on the different ends of the BMI spectrum (<25 and ≥40 kg/m2). Although all persons improved physical function during rehabilitation to a similar degree, the efficiency and cost at which this was achieved differed depending on the BMI value. After controlling for several subject characteristics including age, sex, comorbidity number, use of hip abductor braces, and blood transfusions need, BMI was identified as a significant contributor to FIM efficiency in the regression models. BMI is an important consideration when predicting functional outcomes after THA.
Obesity and Inpatient Rehabilitation Outcomes
Currently, there are no comparative data in the literature regarding inpatient rehabilitation gains in obese THA patients. Data regarding the perioperative outcomes and BMI are mixed. Some studies show similar LOS and frequency of perioperative complications (blood loss, surgical time) in obese and non-obese THA patients (5, 8). Others show increased risk for surgical complications and non-homebound discharge (9). Some prospective studies indicate that obesity does not adversely affect patient outcomes over the course of 1 to 7 years (12, 19). For example, Short-Form 36 and Western Ontario and McMaster Universities Osteoarthritis Index scores on physical function were reported to be lower in THA patients with BMI values ≥40 kg/m2 compared with those <40 kg/m2 (12). Despite lesser physical activity levels after THA, (20), obese persons have higher rates of hip implant loosening and leg pain (19). The discrepancies between these studies may lie in the lack of standardization of the definitions of obesity used in each study (12); selection of specific obesity criteria may bias results to favor a specific outcome.
In this study, we examined the influence of BMI on THA outcomes. There were no differences in admission FIM, a finding that may be caused by the rigorous screening process performed by the physiatry consult team before admission to the inpatient rehabilitation hospital. Specific criteria for admission must be corroborated by the physiatrist, occupational therapist, and physical therapist before referral for inpatient rehabilitation. All patients may have been admitted in a similar functional state. Once at the rehabilitation hospital, severe obesity is associated with a lower FIM score at discharge, lower FIM efficiency, and a longer LOS than normal weight. Only one previous report has shown that obese persons (BMI ≥ 40 kg/m2) have different postoperative functional recovery levels and a greater need of assistance for transfers from supine to sit, sit to stand, and bed to chair positions in the inpatient setting than non-obese counterparts after the arthroplasty procedure (8). Recent data have shown that obese elderly adults in general have lower Short Form 36 scores for domains of physical function, change in health, vitality, and role limitations-physical than non-obese older adults. Obese older adults have reduced functional capabilities such as slower walking speed, shorter one-legged stand time, and slower speed in completing an obstacle course than non-obese counterparts (21). These general compromised capabilities in obesity could translate to slower functional recovery and improvement on specific exercises during inpatient rehabilitation. We found that changes in several motor scores for activities that involve weight bearing or transfers (e.g., walking ability, bed to chair, toilet, tub/shower) were not different among the BMI brackets, indicating that obesity itself does not impede gains in function. A possible interpretation is that the relief from the joint pain after surgery superseded any effect of excessive body weight while performing load bearing tasks in this study. This pain relief may have fostered improvement in all groups, irrespective of BMI. This difference in response from our study to others could suggest that gains in function were specific to this population at this site. These data should be corroborated by other sites with larger patient bases.
The majority of all patients (86.9% to 98%) returned home, with a small portion of the patients discharged to a skilled nursing facility and acutely transferred back to the hospital. The reasons for discharge to external facilities may be different between the BMI brackets, however. Non-obese persons were older and had more women than the other groups, characteristics that together have shown to increase the chance of a non-homebound discharge in THA patients (18). Specific reasons for acute transfers included atrial fibrillation, chest pain, malaise, and hyponatremia in non-obese persons and pulmonary emboli and atrial fibrillation complications in the overweight group. Severely obese persons were acutely transferred because of sepsis, unresponsiveness, and hypotension. These data agree with those of Jain et al. (9), who reported a greater likelihood of non-homebound discharge in older obese individuals than non-obese counterparts. Further research is required to determine whether there is a protective effect of BMI such that overweight and moderately obese persons retain better function with age compared with the frail, elderly non-obese, or severely obese.
The finding that the severely obese group accrued greater hospital charges during the inpatient rehabilitation is novel. Severe obesity may have increased the need for hospital resources. In this cohort, severely obese persons required a greater LOS for functional gains needed for discharge (at an average daily cost of $1100/d). Severely obese persons had greater rates of diabetes. Clinical management responsibilities for the hospital staff may have been intensified because of management of wound healing and blood glucose (e.g., accuchecks, medication dispensation, wound debridement and dressing change, wound care) (9). The additional management and staff assistance with transfer mobility for these patients and the clinical management responsibilities were the most likely contributors to the higher costs in the severely obese patients in this study.
With any retrospective study, mechanisms explaining differences among groups could not be determined but only speculated. The effects of socioeconomic status, time-course of development of hip joint pain and osteoarthritis disease course before elective THA, acute care complications, type of joint component, inflammatory laboratory values, and pain coping skills may partially explain some of these differences in rehabilitation outcomes. This study did not attempt to match all subjects and control for all possible confounders such as pre-surgical functional status and personality traits that can influence recovery processes. Also, the patients who were admitted into the inpatient rehabilitation program were rigorously screened by a consultation team to determine whether the patient could tolerate aggressive therapies postsurgery; these patients may, therefore, represent a very healthy population of THA patients and may not reflect the overall THA population who is referred to the general IRF. Finally, given that there is no standard inpatient rehabilitation therapy protocol among IRFs for the THA population, the results presented here may be limited to the average IRF. In addition, this study should be repeated with larger patient populations to tease out effects of the aforementioned variables. Prospective studies may include comparisons of BMI or body fat percentage brackets with interventions to improve FIM efficiency during inpatient rehabilitation after THA.
In conclusion, elevated BMI does not prevent FIM gains in THA patients during inpatient rehabilitation. However, BMI is related with FIM efficiency, LOS, and hospital charges in a curvilinear fashion. Severely obese patients can achieve physical improvements but at a lower efficiency and greater cost than counterparts with less body fat.
The authors thank Jennifer Martin, Lori Aylor, and Tom Cook for their assistance with data procurement. There was no funding/outside support for this study.
Nonstandard abbreviations: THA, total hip arthroplasty; IRF, inpatient rehabilitation facility; LOS, length of stay; FIM, functional independence measure.
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.