Abstract

Objective. To test the hypothesis: Time to ambulation (walking) after hip fracture surgery impacts the frequency of postoperative complications and length of hospital stay.

Methods. A retrospective observational study of a cohort of all patients admitted to a university teaching hospital with a principal International Classification of Diseases-9 diagnosis of a hip fracture during 3 calendar years.

Results. A total of 131 participants were identified (68% were aged 65 years or older). Overall, the mean time to writing an order to ambulate a patient after a hip fracture surgery was 2 ± 1.5 days. Time to ambulation after hip fracture surgery was significantly less in patients cared for on orthopedic surgery service compared to general surgery service (1.8 ± 1 vs 2.5 ± 2, p <.05) or general internal medicine service (2.5 ± 1.5, p <.05). It did not relate, however, to patient's age, sex, or race, or to patient's functional status prior to admission, fracture site (femoral neck, intertrochanteric, or subtrochanteric), whether a femoral neck fracture is displaced or not, type of anesthesia (spinal/epidural vs general), type of surgery (open reduction and internal fixation vs hemiarthroplasty), degree of preoperative risk, number of medical conditions, or to obtaining physical therapy and/or medical consultation. Time to ambulation after surgery was an independent predictor for the development of pneumonia (1.5 OR [odds ratio]/day, p <.001), new onset delirium (1.7 OR/day, p <.001), and to prolonged length of hospital stay (B [slope coefficient] = 1.36, p <.0001) but not to the development of pressure ulcers, deep venous thrombosis, or urinary tract infection.

Conclusions. Delayed ambulation after hip fracture surgery is related to the development of new onset delirium and pneumonia postoperatively as well as to increased length of hospital stay. Early ambulation after hip fracture surgery should be encouraged.

ALTHOUGH it is well accepted that early mobilization is vital to avoiding postoperative medical complications, there is hesitancy among some physicians to allow early ambulation after hip fracture surgery because of concerns of mechanical failure (1–3). A recent study by Koval and colleagues (4), however, showed that immediate unrestricted weight-bearing after hip fracture surgery was safe and did not result in increased mechanical complications. In this study, we assessed the relationship between time to ambulation after hip fracture surgery and hospitalization outcomes in a retrospective cohort of patients who underwent surgery for an acute hip fracture.

Methods

The medical records of all consecutive admissions with a principal International Classification of Diseases (ICD)-9 diagnosis of a hip fracture during the period from January 1, 1997 through December 31, 1999 to a university teaching hospital were screened. A total of 145 participants were identified. After excluding missing medical records (n = 3), data were abstracted from the remaining medical records (n = 142) by trained research assistants. Abstractors obtained demographic data including age, sex, race, place of residence prior to admission, and length of hospital stay. The admitting service and whether patients were seen by a medical or geriatric consult during hospitalization were recorded. The time an order was written to ambulate (walk) patients after surgery was also recorded. The following were also obtained: the development of pressure ulcers, development of postoperative delirium (presence of confusion or acute change in mental status from the medical record), and development of pneumonia (X-ray criteria) or urinary tract infection. A modified Activities of Daily Living scale was used to grade prefracture functional status as either independent or dependent in one or more essential activities (the information was obtained from the caregiver in case of cognitively impaired patients) (5). Fracture site (nondisplaced femoral neck, displaced femoral neck, intertrochanteric, or subtrochanteric), type of anesthesia (spinal/epidural vs. general), type of surgery (open reduction internal fixation or hemiarthroplasty, and whether physical therapy consult was obtained or not were also recorded. The classification system of the American Society of Anesthesiologists (6), Detsky Modified Multifactorial Index (7), and the Goldman Cardiac Risk Index (8) were used to assess preoperative risk. The study was approved by the Institutional Review Board of Saint Louis University School of Medicine.

Descriptive statistics were used to describe the population demographics. The Mann-Whitney test was used to compare the medians between 2 groups. The Kruskal-Wallis test was used to compare the medians between 3 or more groups. Post hoc comparisons after a significant Kruskal-Wallis test were performed using the Mann-Whitney test. The chi-square test was used to examine significance of contingency tables. Relative risks for complications were calculated compared with the incidence among the group who was ambulated on day 1 after hip surgery. The Spearman rank correlation was used to measure degree of association between two continuous variables. Logistic regression analysis was used to adjust for level of comorbidity when dependent variable was continuous, and significance was determined using the likelihood ratio test. Multiple regression analysis was used to adjust for level of comorbidity when the dependent variable was dichotomous. The Goldman's Index was used as the marker for comorbidity unless specified otherwise. Significance was defined as p < 0.05 for two-tailed tests. All analysis was performed using Statistica software (StatSoft) (9).

Results

Demographics

A total of 142 participants were identified. Eleven participants were excluded from the analysis (4 did not have surgery, and 7 were either not ambulated prior to discharge or their ambulation status could not be determined from their medical records). The population demographics of the remaining 131 participants are shown in Table 1. Mean time (in days) to writing the order to ambulate a patient after hip fracture surgery was 2 ± 1.5 days, (range, 1–9 days).

Factors Affecting Time to Ambulation After Surgery

Among 21 demographic and clinical factors studied (Table 2), the time of writing an order to ambulate a patient after hip fracture surgery was found significantly related to only the treating service (orthopedic, general surgery, or general medicine). Post hoc analysis by the Mann-Whitney test showed that the time to ambulation was significantly less if the patient was cared for on orthopedic surgery service (1.8 ± 1 days) compared to general surgery service (2.5 ± 1.5 days, p =.04) or general medicine service (2.5 ± 1.5 days, p =.04). This relationship remained significant even after adjusting for American Society of Anesthesiologists Classification score, Goldman's Cardiac Risk Index score, and the number of ICD-9 medical conditions (as markers of comorbidity) using multiple regression analysis.

Relation of Time of Ambulation to Hospitalization Outcomes

Main outcome measures studied were the frequency of postoperative complications and length of hospital stay. Delirium was identified if there was a note in the medical record indicating the presence of confusion or change in mental status. The relationship between time to ambulation and in-hospital mortality was not studied since none of the participants died during hospitalization.

Time to ambulation related significantly to the development of pneumonia, pressure ulcer, and new-onset delirium, and prolonged length of hospital stay but not to the development of urinary tract infection or deep venous thromboses (Table 3). After adjusting for level of comorbidity, the odds ratio (OR)/day for developing new onset delirium was 1.72, which was significant at p <.001 using the likelihood ratio test; the OR/day for developing pneumonia was 1.5, which was significant using the likelihood ratio test at p <.001; and the OR/day for developing pressure ulcers was 1.6, which was not significant using the likelihood ratio test. In addition, after adjusting for the level of comorbidity as well as for presence of delirium, time to ambulation was an independent predictor of length of hospital stay with a B [slope coefficient] of 1.36 and p <.0001. In addition, mean length of hospital stay increased by 98% among patients ambulated on postoperative day 4 or more compared with those ambulated on postoperative day 1. The relative risks of developing individual postoperative complications on specific postoperative days compared with postoperative day 1 is shown in Table 4.

Discussion

This study is the first to look into the relationship between time to ambulation after hip fracture surgery and the development of postoperative complications and length of hospital stay. Findings from this study linked delayed ambulation to increased frequency of postoperative complications and prolonged length of hospital stay. It is worth noting that the development of one or more in-hospital postoperative complications was linked to increased mortality after a hip fracture (10–13). These findings, combined with the lack of biomechanical justification for delayed ambulation after a hip fracture surgery, should encourage early postoperative ambulation.

The results from this study should be interpreted within the context of several limitations. Time to ambulation in this study was based on the day an order was written to ambulate a patient. It is possible that some patients were not ambulated on the same day the order was written in the medical chart. Although the findings from this study linked delayed ambulation to increased frequency of postoperative pneumonia and delirium, it did not clarify the cause and effect relationship (i.e., whether delayed ambulation led to the complication or whether the complication resulted in the delay to ambulate a patient). This is particularly likely in patients who developed postoperative confusion (delirium); physicians may have delayed writing an order to ambulate a patient who had confusion for fear of falls or noncompliance of weight restrictions. Another factor that may have confounded the results noted from this study is the presence of urinary incontinence. Palmer and colleagues (14) retrospectively examined a large cohort (n = 6516) of elderly women who were hospitalized with a hip fracture and reported that 21% of the women had hospital-acquired incontinence at discharge. The presence of urinary innocence might have affected the time an ambulation order was placed and the time the patient was actually ambulated by staff. Data related to the presence of incontinence was not collected in our study.

Conclusion

The findings from this study indicate that delayed ambulation after hip fracture surgery is associated with poor hospitalization outcomes and emphasizes the importance of early ambulation after hip fracture surgery. Further study is warranted to investigate the effect of timing to ambulation after hip fracture surgery on long-term outcomes in patients who sustain hip fractures.

Table 1.

Population Demographics.

Factor N (%) Mean ± SD Range 
Age group    
    <65 years 42 (32) 43 ± 16 years 16–65 years 
    ≥65 years 89 (68) 80 ± 7 years 65–96 years 
Sex    
    Male 57 (44)   
    Female 74 (56)   
Race    
    Caucasians 87 (66)   
    African Americans 40 (30)   
    Others 4 (3)   
Place of residence    
    Home or assisted living 114 (87)   
    Skilled nursing facility 11 (8.4)   
    Fractured hip while in hospital 6 (4.6)   
Caring service    
    Orthopedic service 80 (61)   
    General surgery 33 (25)   
    General Medicine 13 (10)   
    ICU (medicine or surgical) 5 (4)   
Fracture site    
    Femoral neck 66 (50)   
    Non displaced* 16 (28)   
    Displaced* 41 (72)   
    Intertrochanteric 48 (37)   
    Subtrochanteric 11 (9)   
    More than one site 6 (4)   
Time to ambulation (days)  2 ± 1.5 1–9 
LOS (days)  5.9 ± 4 2–36 
Factor N (%) Mean ± SD Range 
Age group    
    <65 years 42 (32) 43 ± 16 years 16–65 years 
    ≥65 years 89 (68) 80 ± 7 years 65–96 years 
Sex    
    Male 57 (44)   
    Female 74 (56)   
Race    
    Caucasians 87 (66)   
    African Americans 40 (30)   
    Others 4 (3)   
Place of residence    
    Home or assisted living 114 (87)   
    Skilled nursing facility 11 (8.4)   
    Fractured hip while in hospital 6 (4.6)   
Caring service    
    Orthopedic service 80 (61)   
    General surgery 33 (25)   
    General Medicine 13 (10)   
    ICU (medicine or surgical) 5 (4)   
Fracture site    
    Femoral neck 66 (50)   
    Non displaced* 16 (28)   
    Displaced* 41 (72)   
    Intertrochanteric 48 (37)   
    Subtrochanteric 11 (9)   
    More than one site 6 (4)   
Time to ambulation (days)  2 ± 1.5 1–9 
LOS (days)  5.9 ± 4 2–36 

Notes: *Information on whether a femoral neck fracture was displaced or not was available in only 57 of the 66 subjects who sustained a femoral neck fracture.

LOS = Length of hospital stay; SD = standard deviation; ICU = intensive care unit.

Table 2.

Possible Factors Affecting Time to Ambulation After Surgery.

Factor Time to ambulation (days) mean ± SD p 
Age   
    <65 2.1 ± 1.7 .9 
    ≥65 1.9 ± 1.3  
Sex   
    Male 2 ± 1.6 .7 
    Female 2 ± 1.4  
Race   
    Caucasian 2.2 ± 1.6 .3 
    African American 1.6 ± 1  
    Other 2 ±.8  
Functional status prior to admission   
    Independent 2 ± 1.6 .3 
    Need assistance in one or more ADL 2.1 ± 1.2  
Number of ICD-9 medical conditions r =.1 .1 
Fracture site   
    Femoral neck 2 ± 1.5 .7 
    Intertrochanteric 2 ± 1.2  
    Subtrochanteric 2 ± 1.8  
Fracture neck fracture displacement*   
    Nondisplaced 2.1 ± 1 .7 
    Displaced 3 ± 1.5  
Type of anesthesia   
    General 2.1 ± 1.3 .6 
    Spinal/epidural 2 ± 1.5  
Type of surgery   
    ORIF 2.3 ± 1.7 .7 
    Hemiarthroplasty 2 ± 1.4  
Admitting service   
    Orthopedics 1.8 ± 1 .04 
    General surgery 2.5 ± 2  
    General internal medicine 2.5 ± 1.5  
Medical or geriatric consult obtained   
    Yes 1.9 ± 1 .7 
    No 2.3 ± 1.7  
Physical therapy consult obtained   
    Yes 2 ± 1.3 .6 
    No 1.9 ± 2  
Goldman Score r =.07 .4 
Detskey Score r = –.04 .6 
ASA score r =.03 .6 
Time to surgery from admission (days) r =.03 .8 
Dementia   
    Yes 2±1 .7 
    No 2 ± 1.5  
Delirium on admission   
    Yes 2.5 ± 1.9 .8 
    No 2 ± 1.4  
Congestive heart failure   
    Yes 1.7 ± 1 .4 
    No 2 ± 1.5  
Chronic obstructive pulmonary disease   
    Yes 2 ± 1.5 .8 
    No 2 ± 1.4  
Depression   
    Yes 1.7 ± 1 .7 
    No 2 ± 1.5  
Factor Time to ambulation (days) mean ± SD p 
Age   
    <65 2.1 ± 1.7 .9 
    ≥65 1.9 ± 1.3  
Sex   
    Male 2 ± 1.6 .7 
    Female 2 ± 1.4  
Race   
    Caucasian 2.2 ± 1.6 .3 
    African American 1.6 ± 1  
    Other 2 ±.8  
Functional status prior to admission   
    Independent 2 ± 1.6 .3 
    Need assistance in one or more ADL 2.1 ± 1.2  
Number of ICD-9 medical conditions r =.1 .1 
Fracture site   
    Femoral neck 2 ± 1.5 .7 
    Intertrochanteric 2 ± 1.2  
    Subtrochanteric 2 ± 1.8  
Fracture neck fracture displacement*   
    Nondisplaced 2.1 ± 1 .7 
    Displaced 3 ± 1.5  
Type of anesthesia   
    General 2.1 ± 1.3 .6 
    Spinal/epidural 2 ± 1.5  
Type of surgery   
    ORIF 2.3 ± 1.7 .7 
    Hemiarthroplasty 2 ± 1.4  
Admitting service   
    Orthopedics 1.8 ± 1 .04 
    General surgery 2.5 ± 2  
    General internal medicine 2.5 ± 1.5  
Medical or geriatric consult obtained   
    Yes 1.9 ± 1 .7 
    No 2.3 ± 1.7  
Physical therapy consult obtained   
    Yes 2 ± 1.3 .6 
    No 1.9 ± 2  
Goldman Score r =.07 .4 
Detskey Score r = –.04 .6 
ASA score r =.03 .6 
Time to surgery from admission (days) r =.03 .8 
Dementia   
    Yes 2±1 .7 
    No 2 ± 1.5  
Delirium on admission   
    Yes 2.5 ± 1.9 .8 
    No 2 ± 1.4  
Congestive heart failure   
    Yes 1.7 ± 1 .4 
    No 2 ± 1.5  
Chronic obstructive pulmonary disease   
    Yes 2 ± 1.5 .8 
    No 2 ± 1.4  
Depression   
    Yes 1.7 ± 1 .7 
    No 2 ± 1.5  

Notes: *Information on whether a femoral neck fracture was displaced or not was available in only 56 of the 65 subjects who sustained a femoral neck fracture.

For patients admitted to a nonmedical service.

ADLs = activities of daily living; ICD-9 = International Classification of Diseases, 9th Ed.; SD = standard deviation; ORIF = open reduction and internal fixation; ASA = American Society of Anesthesiologists.

Table 3.

Relationship of Time to Ambulation and Postoperative Complications.

Condition Time to ambulation (d) mean ± SD p 
Pneumonia   
    Yes 3.3 ± 2 .03 
    No 1.8 ± 1.2  
Pressure ulcers   
    Yes 4 ± 1 .02 
    No 2 ± 1  
New-onset delirium   
    Yes 3.3 ± 2 .00004 
    No 1.8 ± 1  
Urinary tract infection   
    Yes 2.6 ± 2 .5 
    No 2 ± 1.3  
Deep Venous thrombosis   
    Yes 3 ± 0.0 
    No 2 ± 1.4  
Length of hospital stay (d) r =.33 .0001 
Condition Time to ambulation (d) mean ± SD p 
Pneumonia   
    Yes 3.3 ± 2 .03 
    No 1.8 ± 1.2  
Pressure ulcers   
    Yes 4 ± 1 .02 
    No 2 ± 1  
New-onset delirium   
    Yes 3.3 ± 2 .00004 
    No 1.8 ± 1  
Urinary tract infection   
    Yes 2.6 ± 2 .5 
    No 2 ± 1.3  
Deep Venous thrombosis   
    Yes 3 ± 0.0 
    No 2 ± 1.4  
Length of hospital stay (d) r =.33 .0001 

Note: SD = standard deviation.

Table 4.

Frequency of Complications by Days to Ambulation After Surgery.

Condition Days toAmbulation RR* 95% CI Mean ± SD p 
Pneumonia     
 4.4 .2–128   
 4.2 .06–331   
 4 or more 10.3 .8–295.6   
New-onset delirium     
 10 1.5–573   
 17 1.9–1180   
 4 or more 21 3.2–1503   
Pneumonia and or new-onset delirium     
 6.8 1.4–64   
 8.7 1.2–95   
 4 or more 14.3 2.8–137   
Length of hospital stay (d)   5 ± 2.2 .01 
   5.6 ± 2.6  
   5.6 ± 1.5  
 4 or more   9.9 ± 9  
Condition Days toAmbulation RR* 95% CI Mean ± SD p 
Pneumonia     
 4.4 .2–128   
 4.2 .06–331   
 4 or more 10.3 .8–295.6   
New-onset delirium     
 10 1.5–573   
 17 1.9–1180   
 4 or more 21 3.2–1503   
Pneumonia and or new-onset delirium     
 6.8 1.4–64   
 8.7 1.2–95   
 4 or more 14.3 2.8–137   
Length of hospital stay (d)   5 ± 2.2 .01 
   5.6 ± 2.6  
   5.6 ± 1.5  
 4 or more   9.9 ± 9  

Notes: *Relative risk (RR) was calculated based on frequency cocomplication among subjects ambulated on day 1 after surgery unless specified otherwise.

CI = confidence interval; SD = standard deviation.

Address correspondence to Hosam K. Kamel, MD, Zablocki VAMC, 5000 W. National Ave. (CC-G), Milwaukee, WI 53295. E-mail: kamel@pol.net

References

1
Kyle RF, Gustilo RB, Premer RF. Analysis of six hundred and twenty-two intertrochanteric hip fractures.
J Bone Joint Surg.
 .
1979
;
61A
:
216
-221.
2
Moller BN, Lucht U, Grymer F, Bartholdy NJ. Instability of trochanteric hip fractures following internal fixation.
Acta Orthop Scand.
 .
1984
;
55
:
517
-520.
3
Wolfgang GL, Bryant MH, O'Neill JP. Treatment of intertrochanteric fracture of the femur using sliding screw plate fixations.
Clin Orthop.
 .
1982
;
163
:
148
-158.
4
Koval KJ, Friend KD, Aharonoff GB, Zuckerman JD. Weight bearing after hip fractures: a prospective series of 596 geriatric hip fracture patients.
J Orthop Trauma.
 .
1996
;
10
:
526
-630.
5
Inouye SK, Wagner DR, Acampora D, et al. A predictive index for functional decline in hospitalized elderly medial patients.
J Gen Intern Med.
 .
1993
;
8
:
645
-652.
6
Owens WD, Flests JA, Spitznagel EL. ASA physical status classifications: a study of consistency ratings.
Anesthesiology.
 .
1978
;
49
:
239
-243.
7
Detsky AS, Abrams HB, Forbath N, Scott JG, Hilliard JR. Cardiac assessment for patients undergoing noncardiac surgery. A multifactorial clinical risk index.
Arch Intern Med.
 .
1986
;
146
:
2131
-2134.
8
Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index of cardiac risk in noncardiac surgical procedures.
N Engl J Med.
 .
1977
;
297
:
845
-850.
9
Statistica for Windows, 3rd Ed. Tulsa, OK: StatSoft, Inc.; 1977.
10
Aharonoff GB, Koval KJ, Skovron ML, Zuckerman JD. Hip Fractures in the Elderly: Predictors of one year mortality.
J Orthop Trauma.
 .
1997
;
11
:
162
-165.
11
Sexson SB, Lehner JT. Factors affecting hip fracture mortality.
J Orthop Trauma.
 .
1988
;
1
:
298
-305.
12
Myers AH, Robinson EG, Van Natta ML, Nichelson JD, Collins K, Baker SP. Hip fractures among the elderly: factors associated with in-hospital mortality.
Am J Epidemiol.
 .
1991
;
134
:
1128
-1137.
13
Magaziner J, Simonsick EM, Kashner TM, Hebel JR, Kenzora JE. Survival experience of aged hip fracture patients.
Am J Public Health.
 .
1989
;
79
:
274
-278.
14
Palmer MH, Baumgarten M, Langenberg P, Carson JL. Risk factors for hospital-acquired incontinence in elderly female hip fracture patients.
J Gerontol Med Sci.
 .
2002
;
57A
:
M672
-M677.