Hip fracture represents one of the most common and potentially devastating injuries in older adults. Because of a greater risk of falls and of decreased bone mineral density, the risk of sustaining a hip fracture increases sharply with aging.1
The estimated mortality from hip fracture is 2% to 7% in the acute phase of hospitalization, 6% to 12% within 30 days, and as high as 33% in the year thereafter.1,2 In survivors, the reduction of functional abilities in the year after a hip fracture can reach 15% to 20%, and even individuals who were fully functional before fracture, often lose their capability to live independently.3,4 The consequences of a hip fracture persist long term; compared with persons of the same age, those with a hip fracture walk slower 2 years after the event.5
The ability to walk is considered a meaningful indicator of functional capacity and overall well-being in several clinical conditions; in older persons, gait speed strongly predicts survival, independent of clinical variables.6,7 In survivors of hip fracture, the evaluation of walking ability is an important predictor of functional recovery and of the probability of returning home.8 However, an observational approach to gait assessment may be limited because of excessive subjectivity, poor reliability, and floor or ceiling effects.9
We hypothesized that walking speed measured over a short distance would prove to be a valid predictor of functional evolution and clinical outcomes in adults recovering from surgery after hip fracture. To this aim, we conducted the present study to investigate whether 4-m gait speed, assessed in older persons prior to discharge for hip fracture surgery, could predict the functional recovery and long-term development of major clinical outcomes, such as institutionalization or death.
Participants and Study Protocol
All persons older than 65 years, consecutively admitted with a hip fracture to the Orthogeriatrics Unit of Santa Maria Annunziata Community Hospital in Florence from November 2012 to April 2013, were considered eligible in a prospective cohort design study. As per general admission criterion to the Orthogeriatrics Unit, a prefracture Barthel Index (BI)10 of less than 50 (see later, “Data Collection”) was an exclusion criterion. Barthel Index is a well-known, widely applied tool, used to describe the functional status of older persons. Its items belong to the domains of self-care (feeding, grooming, bathing, dressing, bowel and bladder care, and toilet use) and mobility (ambulation, transfers, and stair climbing). Its summary score ranges from 0 (complete dependency, bedridden state) to 100 (full independence), with 5-point increases representing the minimum change in at least 1 activity. Barthel Index has good metric properties for face-to-face, informant, or telephone ratings.11,12 Other exclusion criteria were a need for physical assistance in walking prior to admission, polytrauma, fracture due to bone metastases, critical conditions requiring admission to intensive care unit postsurgery, life expectancy of less than 6 months, severe cognitive impairment (as assessed from a Mini-Mental State Examination [MMSE]13 score on admission below 18), or disruptive behavior that might prevent functional assessment and nonsurgical treatment.
Most of baseline information was gathered from hospital records on admission to the Orthogeriatrics Unit, whereas gait was assessed as soon as the participant was able to walk without a person's help (see later). Telephone follow-up interviews were conducted 1, 2, and 12 months after discharge by the same examiner (S.G.), who remained blinded to baseline assessment; in less than 20% of the cases, when proxies had been consulted at baseline, the interview was conducted with a proxy.
The study protocol and conduction was consistent with the Declaration of Helsinki. Participants were enrolled only after signing an approved informed consent form. Thus, the rights of human subjects were protected.
Preadmission general health status was assessed by asking the participants to report BI, presence of 24-hour home assistance, and the number of drugs taken on the 15 days prior to fracture; this last information was taken as an indirect clue to the burden of comorbidity. Proxies were consulted in the presence of person's poor cognitive status (MMSE score of <18; see previously). As soon as the participant was able to stand and walk, using walking aids as needed but with no person's help, gait speed (cm/s) on a 4-m path was recorded, following the Short Physical Performance Battery Protocol.14 To this purpose, 2 marks were placed 4-m apart in the hospital hallway, taking care they were not visible to the participant. Participants were instructed to walk at their preferred speed, as if they were going to a local shop, without pausing for resting. Timing was started when the participant began to walk after the “Go!” order was given and stopped when the end marker was crossed. The investigators provided no additional cuing once the participant began to walk. Following a brief rest, a second trial was completed. Velocity was calculated by dividing the distance walked by the time (in seconds) required to complete the trial; the best result out of 2 trials was chosen. A score of 0 was assigned to participants unable to perform the test.
On the 1- and 2-month telephone interviews, the BI was administered and the need for walking aid was assessed; these data were obtained again with the 12-month interview, when information on the amount and setting of rehabilitation received after discharge, new falls, and need for permanent institutionalization or for 24-hour home assistance was also recorded. The living status was assessed from phone interviews.
The statistical package SPSS for Windows v. 20 was used for data management and statistical analysis. Interval variables were expressed as mean (standard error of the mean), categorical variables as percent frequencies. Descriptive statistics included Student t test to compare the mean values of interval variables and Pearson r coefficient to evaluate the correlation between speed and age.
Changes in BI across 4 time points (prefracture, 1-, 2-, and 12-month follow-up) were analyzed with analysis of variance for repeated measures, where Bonferroni's correction was applied to evaluate the significance of differences between values at each individual time point and baseline values. Similarly, SPSS GLM procedure was applied to assess whether the time course of BI changes, from prefracture to 12 months as a repeated within-subject factor, differed between the 2 groups of participants identified on the basis of the median value of walking speed. In this model, the significance of the phase × group (below or above median speed) interaction term was tested, adjusting for the following covariates: age, gender, MMSE, number of drugs on admission, type of surgery (fixation vs arthroplasty), and days of rehabilitation postdischarge. Contrasts of individual time point values with baseline values were analyzed, separately in the 2 groups, applying Bonferroni's correction.
A distinct analytic approach was also used to verify whether 4-m walking speed as a continuous variable could predict 2 long-term outcomes. The first of these outcomes was represented by functional decline, as indicated by a decrease in BI score greater than 5 points, from prefracture through the 12-month follow-up.15,16 We considered as appropriate for this study a threshold more permissive than the minimum 5-point change that the BI detects, because some limitations are to be expected after hip fracture. Secondarily, a composite endpoint was created as a dichotomous variable, registering any of the followings: death, new falls, institutionalization, new need for 24-h home assistance, or long-term functional decline (ie, more than 5 points reduction in BI). Predictors of each dichotomous outcome were identified in separate models of multivariable logistic analysis, where 4-m walking speed was entered as a continuous variable; age, gender, MMSE, number of drugs on admission, type of surgery, and days of rehabilitation postdischarge were entered as covariates and backward removed when redundant. The prediction of either dichotomous outcome was also assessed as the area under the receiver operating characteristic curve, where the optimal combination between sensitivity and specificity cutoffs was obtained as the Youden's J statistics. Statistical significance was set at P values less than .05.
A total of 62 adults, mean age of 84.7 (0.9) years (range: 66-94 years), were enrolled. Of the enrolled, 48 were women (77%). A hip prosthesis was applied in 26 participants (42%), whereas the remaining 36 cases received fixation surgery. Two other participants, potentially eligible for the study, died before enrollment and baseline evaluation; therefore, they were not included in the final study sample.
Prior to admission, all persons were able to walk without a person's help. On average, prefracture BI was close to 100 and MMSE was only moderately reduced (Table 1). The number of drugs taken on admission was slightly above 5.
Predischarge, on average 6.1 (0.2) days after surgery, 47 out of 62 participants (75.8%) were able to perform the walking test, while 15 were not. In those who could walk, the range of the 4-m walk speed was 7.7 to 66.7 cm/s and the mean speed was 28.0 (2.0) cm/s. When persons who could not perform the gait test and were assigned a speed of 0 were considered, the mean speed was 21.2 (2.2) cm/s; the median value of 20.5 cm/s divided the sample in 2 groups of 31 persons each.
Mean speed was comparable between men (23.0 [5.1] cm/s) and women (20.7 [2.4] cm/s, P = .662) and was unrelated to age (r = −0.144, P = .265). Participants receiving hip prosthesis surgery had a significantly greater speed (27.1 [3.7] cm/s) than those undergoing fixation surgery (17.0 [2.4] cm/s, P = .020). No individuals reported side effects from execution of the walking test.
All baseline participants were interviewed by telephone at 1- and 2-month follow-ups, while only 57 persons (91.9%) were interviewed after 12 months, because 3 participants (5%) had died and 2 were untraceable.
Over the entire duration of the follow-up, 6 of 57 (11%) had received home- or office-based physiotherapy; the remaining 51 (89%) were admitted to a rehabilitation facility. Independent of the setting where physiotherapy was delivered, its average duration was 26.7 days (standard error of the mean = 1.6, range: 7-60).
Compared to prefracture, 1 month after surgery BI was markedly lower (96.3 [0.9] vs 76.5 [2.1]; P < .001), then it increased at 2 (84.1 [2.2]; P < .001) and 12 months (87.2 [2.8]; P = .001), yet it did not completely recover to baseline. Always compared with prefracture, BI was at least 5 points lower in 48 of 62 (77.4%), 28 of 62 (45.2%), and 17 of 57 (29.8%) participants at 1-, 2-, and 12-month follow-ups, respectively.
Of the 57 participants contacted after 12 months, 38 (66.7%) had regained full walking autonomy, 16 (28.1%) needed supervision, 1 (1.8%) required medium assistance with walking, and 2 (3.5%) were completely unable to walk; 28 participants (49.1%) did not need any walking aid, 12 (21.1%) reported at least 1 fall, and 10 (17.5%) were institutionalized or required 24-h home assistance. Twenty-seven (45.0%) of the 60 persons whose 12-month outcome was known reached the composite endpoint of death, falls, institutionalization, new need for 24-h home assistance, or long-term functional decline.
Walking Speed, Functional Variations, and Composite Outcome
Figure 1 shows changes in BI from prefracture through 12-month follow-up, separately for the 2 groups defined by a predischarge walking speed below/equal to or above the median value of 20.5 cm/s. Participants with a greater walking speed showed a significantly lower BI reduction across time, adjusting for age, gender, MMSE, number of drugs, type of surgery, and days of rehabilitation postdischarge. At 12 months, BI recovered to baseline values only in participants with greater baseline walking speed, whereas it remained significantly lower in those whose baseline walking speed was at or below the median.
In a multivariable logistic model, the walking speed (entered as a continuous variable) predicted BI reduction at 12 months; age, gender, MMSE, number of drugs, type of surgery, and days of physiotherapy postdischarge were all backward removed from the final model as redundant. Per each cm/s of speed, the risk of an unfavorable functional outcome was reduced by 5%, with an adjusted odds ratio (95% confidence interval) of 0.95 (0.91-0.997) and P = .038.
Similarly, the walking speed predicted the composite outcome, with an adjusted odds ratio of 0.93 (0.88-0.99), P = .013. Mini-Mental State Examination was the only covariate contributing to the prediction of this outcome, with an odds ratio (95% confidence interval) of 0.72 (0.59-0.88), P = .001, per each point increase; conversely, age, gender, number of drugs, type of surgery, and days of physiotherapy postdischarge were backward removed.
The area under the receiver operating characteristic curve was 0.73 for the dichotomous outcome of BI decline in 12 months and 0.72 for the composite outcome. A walking speed of 8 cm/s might be proposed as a low sensitivity-high specificity cutoff for both outcomes (functional decline: 0.47 sensitivity and 0.80 specificity; composite outcome: 0.41 sensitivity and 0.82 specificity). Conversely, if a high sensitivity-low specificity cutoff is to be preferred, different values should be considered, that is, 25 cm/s for functional decline (sensitivity: 0.82, specificity: 0.53) and 30 cm/s for the composite outcome (sensitivity: 0.85, specificity: 0.49).
This study shows that short-distance gait speed, measured precociously after surgery in older adults with a hip fracture, can be an independent predictor of functional changes through the first year postsurgery. This predictive ability was confirmed using 2 distinct analytic approaches, where gait speed was dichotomized upon the median and functional status was expressed as an interval variable (BI) or, vice versa, gait speed was maintained as an interval variable and functional status was represented by BI decline beyond an arbitrary threshold of 5 points. The association with a composite endpoint combining clinically relevant events further strengthened the prognostic implications of a slow gait speed.
In agreement with previous studies,17–21 the functional status of our participants showed a substantial decline, compared with premorbid conditions, soon after surgery, followed by a progressive, yet incomplete, recovery from 1 month through the first year. Thus, in this respect, our data are purely confirmatory of the dramatic impact that a hip fracture has on the global functioning of older persons. On the contrary, we report a novel finding that the 4-m walking speed, routinely measured as early as 6 to 7 days after surgical repair of hip fracture, has profound long-term prognostic implications. In this clinical setting, the walking ability is often investigated only from self-report, as in the BI or the Functional Independence Measure.22 To our knowledge, only 2 studies reported on measuring gait speed in acute care settings. The first was conducted on a sample of 16 participants, substantially younger than ours (77.9 vs 84.7 years), in whom gait speed was measured on average 4.7 days after hip fracture surgery; however, the aim of the cited study was only to assess the minimum detectable change in gait velocity, whereas no prognostic implication was drawn from this measure.23 The second investigation was conducted on 46 adults hospitalized with a variety of diagnoses: although initially very low, gait speed improved after a brief course of physiotherapy.24 Taken together, our findings and those from the studies mentioned suggest that measuring the gait speed is feasible and safe in hospitalized persons and may provide useful information.25 In particular, our data show that a short-distance test helps predict functional evolution, expressed as BI variations, over 1 year. It is conceivable that a premorbid functional impairment has an impact of predischarge performance.17 However, whether a predischarge poor walking speed depends mostly on pre- or postfracture conditions cannot be ascertained from our data.
In a study on 157 hip fracture persons, of whom only 57 were reexamined at follow-up, Ingemarsson et al20 reported that the ability to walk 10 m within 15 s with no need for person's help after 1 year was predicted by the Timed Up and Go Test26 but not by 10-m or 30-m walk. Beyond the differences in the performance tests and outcome measures used, Ingemarsson's study and ours are consistent in showing that physical performance effectively predicts functional recovery long-term.
Interestingly, the MMSE score did not contribute to predicting the functional evolution over time. This might suggest that poor mental status should not represent an a priori reason for withholding aggressive treatment, complete predischarge evaluation, and postdischarge physiotherapy in older persons with hip fracture. On the contrary, it should be emphasized that MMSE was the only multivariable predictor of the composite outcome, which included death and other major clinical events: this confirms that the cognitive status is a major determinant of the overall well-being in older adults.
Attrition was negligible in this study, also compared with other reports.20 This might be attributed to our choice to conduct phone interviews (therefore preventing need for transportation) and to collect proxy information: phone interviews27 and proxy information28 can allow for reliable assessment of functional abilities in older adults. We did not enroll participants with a markedly compromised prefracture functional status, which represents a well-known limiting factor for subsequent recovery17: this might explain the limited long-term mortality we observed, compared with other studies.29 No other strict exclusion criteria were applied and, therefore, our participants were probably representative of reasonably well-functioning older persons suffering from hip fracture; this is further supported by the general clinical characteristics of our sample, shown in Table 1. Our assessment was simple and not time-consuming, easily reproducible in current clinical practice. Taken together, these characteristics should allow external validity of our findings, which represents a further strength of the study.
On the contrary, the small sample size clearly represents a major limitation of our study; caution should be exerted also in the view of the limited discrimination that our area under the receiver operating characteristic curves offered. Moreover, most variables were obtained from routine hospital records and, therefore, their reliability and validity cannot be ascertained. Thus, our findings must be viewed as preliminary: should other, larger studies confirm our results, physical performance testing could be accepted as a safe and highly informative assessment, applicable in older adults early after surgical repair of hip fracture. Besides its purely prognostic implications, this assessment approach might also prove helpful in clinical decision-making for postoperative management.
Short-distance gait speed, assessed predischarge in older persons after hip fracture surgery, predicts long-term functional changes and major clinical outcomes, independent of other factors. We, therefore, recommend the application of this simple assessment in older adults recovering from hip fracture. This assessment approach might prove helpful in clinical decision-making for postoperative management of older persons with hip fracture.
1. Negrete-Corona J, Alvarado-Soriano JC, Reyes-Santiago LA Hip fracture
as risk factor for mortality in patients over 65 years of age. Case-control study. Acta Ortop Mex. 2014;28(6):352–362.
2. Vestergaard P, Rejnmark L, Mosekilde L Increased mortality in patients with a hip fracture
-effect of pre-morbid conditions and post-fracture complications. Osteoporos Int. 2007;18(12):1583–1593.
3. Young Y, Fried LP, Kuo YH Hip fractures among elderly women: longitudinal comparison of physiological function changes and health care utilization. J Am Med Dir Assoc. 2010;11(2):100–105.
4. Rosell PAE, Parker MJ Functional outcome after hip fracture
. A 1-year prospective outcome study of 275 patients. Injury. 2003;34(7):529–532.
5. Jarnlo GB, Thorngren KG Standing balance in hip fracture
patients. 20 middle-aged patients compared with 20 healthy subjects. Acta Orthop Scand. 1991;62(5):427–434.
6. Abellan van Kan G, Rolland Y, Andrieu S, et al Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J Nutr Health Aging. 2009;13(10):881–889.
7. Studenski S, Perera S, Patel K, et al Gait speed and survival in older adults. JAMA. 2011;305(1):50–58.
8. Broos PL, Van Haaften KI, Stappaerts KH, Gruwez JA Hip fractures in the elderly. Mortality, functional results and social readaptation. Int Surg. 1989;74(3):191–194.
9. Toro B, Nester CJ, Farren PC The status of gait assessment among physiotherapists in the United Kingdom. Arch Phys Med Rehabil. 2003;84(12):1878–1884.
10. Mahoney FI, Barthel DW Functional evaluation: the Barthel Index. Md State Med J. 1965;14:61–65.
11. Duffy L, Gajree S, Langhorne P, Stott DJ, Quinn TJ Reliability (inter-rater agreement) of the Barthel Index for assessment of stroke survivors: systematic review and meta-analysis. Stroke. 2013;44(2):462–468.
12. Della Pietra GL, Savio K, Oddone E, Reggiani M, Monaco F, Leone MA Validity and reliability of the Barthel Index administered by telephone. Stroke. 2011;42(7):2077–2079.
13. 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(3):189–198.
14. Guralnik JM, Ferrucci L, Simonsick EM, Salive ME, Wallace RB Lower-extremity function in persons over the age of 70 years as a predictor of subsequent disability. N Engl J Med. 1995;332(9):556–561.
15. Palese A, Gonella S, Moreale R, et al Hospital-acquired functional decline in older patients cared for in acute medical wards and predictors: findings from a multicentre longitudinal study. Geriatr Nurs. 2016;37(3):192–199.
16. Palleschi L, De Alfieri W, Salani B, et al Functional recovery
of elderly patients hospitalized in geriatric and general medicine units. The PROgetto DImissioni in GEriatria Study. J Am Geriatr Soc. 2011;59(2):193–199.
17. Mariconda M, Costa GG, Cerbasi S, et al Factors predicting mobility and the change in activities of daily living after hip fracture
: a 1-year prospective cohort study. J Orthop Trauma. 2016;30(2):71–77.
18. Fukui N, Watanabe Y, Nakano T, Sawaguchi T, Matsushita T Predictors for ambulatory ability and the change in ADL after hip fracture
in patients with different levels of mobility before injury: a 1-year prospective cohort study. J Orthop Trauma. 2012;26(3):163–171.
19. Semel J, Gray JM, Ahn HJ, Nasr H, Chen JJ Predictors of outcome following hip fracture
rehabilitation. PMR. 2010;2(9):799–805.
20. Ingemarsson AH, Frändin K, Mellström D, Möller M Walking ability and activity level after hip fracture
in the elderly a follow-up. J Rehabil Med. 2003;35(2):76–83.
21. Koval KJ, Skovron ML, Aharonoff GB, Zuckerman JD Predictors of functional recovery
after hip fracture
in the elderly. Clin Orthop Relat Res. 1998;(348):22–28.
22. Hollman JH, Beckman BA, Brandt RA, Merriwether EN, Williams RT, Nordrum JT Minimum detectable change in gait velocity during acute rehabilitation following hip fracture
. J Geriatr Phys Ther. 2008;31(2):53–56.
23. Dodds TA, Martin DP, Stolov WC, Deyo RA A validation of the Functional Independence Measurement and its performance among rehabilitation inpatients. Arch Phys Med Rehabil. 1993;74(5):531–536.
24. Braden HJ, Hilgenberg S, Bohannon RW, Ko MS, Hasson S Gait speed is limited but improves over the course of acute care physical therapy. J Geriatr Phys Ther. 2012;35(3):140–144.
25. Bohannon RW, Ahlquist M Documentation of prestroke ambulation. Int J Rehabil Res. 2003;26(1):71–72.
26. Thompson M Performance of community dwelling elderly on the Timed Up and Go test. Phys Occup Ther Geriatr. 1995;13(3):17–30.
27. Smith PM, Illig SB, Fiedler RC, Hamilton BB, Ottenbacher KJ, Fielder RC Intermodal agreement of follow-up telephone functional assessment using the Functional Independence Measure in patients with stroke. Arch Phys Med Rehabil. 1996;77(5):431–435.
28. Magaziner J, Simonsick EM, Kashner TM, Hebel JR Patient-proxy response comparability on measures of patient health and functional status. J Clin Epidemiol. 1988;41(11):1065–1074.
29. Abrahamsen B, van Staa T, Ariely R, Olson M, Cooper C Excess mortality following hip fracture
: a systematic epidemiological review. Osteoporos Int. 2009;20(10):1633–1650.