1.  Introduction

The relationship between hyperglycemia and acute illness is complex and is often attributed to fluctuations in circulating stress hormones [1-3]. Regardless of cause, hyperglycemia during hospitalization for acute illness is an independent risk factor for poor outcome [4]. In-hospital mortality is even higher among patients with newly diagnosed hyperglycemia than it is for those with previously diagnosed diabetes [5]. For example, hyperglycemia among patients with diabetes admitted for acute myocardial infarction predicted both in-hospital and 1-year mortality [6].

Data regarding target glucose levels during hospitalization and the preferred method of glucose control are inconclusive. While some clinical studies have demonstrated benefit from tight glucose control [7], others, such as the NICE-SUGAR study, showed no effect or even increased risk of hypoglycemia and mortality [8].

Data regarding the effect of diabetes and glycemic control on rehabilitation outcomes is limited. Reistetter et al. demonstrated that diabetic patients tended to have longer rehabilitation admissions following hip fracture compared to their non-diabetic counterparts [9]. However, other population characteristics were not taken into account such as diabetes treatment or glycemic control. Lieberman et al. concluded that the odds for successful rehabilitation among diabetic patients were four-fold poorer compared to their non-diabetic peers [10], while a later publication by the same group of authors did not reaffirm the role of diabetes in rehabilitation following hip fracture [11]. Adunsky et al. found that diabetes had no effect on rehabilitation outcomes following hip fracture [12]. This study examined the effect of glycemic control during rehabilitation for surgical repair of hip fracture on short-term and 3-month outcomes.

2.  Study Design

The study included patients aged 65 years and older who were admitted to the Geriatric Rehabilitation Unit affiliated with Meir Medical Center, a tertiary care facility. The Geriatric Rehabilitation Unit patient population includes neurologic (post-CVA) and orthopedic patients. The duration of hospitalization ranges from two weeks for post-surgical hip fracture repair to up to one month for neurologic patients. Patients included in our study were admitted to the unit following surgical repair of hip fracture (the index event) from January 2014 until June 2015. The study was approved by the hospital Ethics Committee. Informed consent was not required.

Data collected from hospital electronic medical records included demographics, diabetic treatment prior to hospitalization and during rehabilitation, comorbidities, type of surgical procedure, readmission, mortality during rehabilitation and mortality within 3 months of discharge. All patients had been diagnosed with type II diabetes mellitus prior to the hip fracture. Patients with a diagnosis of dementia at admission to rehabilitation were excluded, as were those on systemic glucocorticoid therapy.

The rehabilitation unit used a unified glucose control protocol. Diabetic patients were transferred to a basal-bolus insulin therapy regimen if they had at least 2 glucose values > 180 mg/dl. Patients on insulin therapy continued their regimen with adjustments according to glucose levels. Glucose control was achieved without intervention by the investigators.

Glycemic control was assessed using hemoglobin A1c (HbA1c) and glucose measurements. If HbA1c value within the 3 months preceding hospitalization was unavailable, it was measured at admission to rehabilitation using the D-10TM hemoglobin testing system (Bio-Rad, Hercules, CA, USA). Capillary glucose levels were measured at bedside using a glucometer and documented in the electronic medical record. Minimum, average and maximum glucose levels were abstracted from the patient’s glucose chart. Hypoglycemia was defined as glucose level < 70 mg/dl. Severe hypoglycemia was defined as glucose < 50 mg/dl.

Patient pre- and post-fracture functional status was defined as independent (able to conduct all Activities of Daily Living [ADL]), dependent (need assistance with at least 3 ADL) or frail (neither independent nor dependent). Functional outcomes were measured using the Functional Independence Measure (FIM) scale, which includes18 items composed of 13 motor tasks (FIM_motor) and 5 cognitive tasks (FIM_cognitive). Tasks are rated on a 7-point ordinal scale that ranges from18 (indicating total assistance or complete dependence) to 126 (indicating complete independence). Patients are usually evaluated at admission and discharge.

To evaluate changes in functional status, a surrogate marker, termed ΔFIM_motor, was defined and calculated as the difference between the FIM_motor score at the beginning of rehabilitation and at discharge. Patients who were admitted to rehabilitation from December 2014 through June 2015 were contacted by telephone 3 months post-discharge. Functional status at this point was defined as independent, frail or dependent according to self-reported ADL function.

3.  Statistical Analysis

Data were presented as numbers and percentages for nominal variables, and mean and standard deviation for continuous parameters. Chi-Square or Fishers' Exact test were conducted for non-metric parameters (each when appropriate) and t-test or Mann-Whitney non-parametric test for quantitative variables, according to variable distribution. P < 0.05 was considered statistically significant. All statistical analyses were performed with SPSS-23 software.

4.  Results

4.1.  Patient Characteristics and Outcomes

A total of 64 patients with diabetes were recruited to the study. Two patients were lost to follow-up and 33 were available for long-term functional analysis. The demographics of the study population are presented in (Table 1). Functional status prior to hip fracture was the main determinant of rehabilitation outcome. Higher pre-fracture functional status correlated with lower mortality during hospitalization (p = 0.032). There was a trend toward an inverse correlation between pre-fracture functional status and mortality at 3 months (p = 0.058).

Higher FIM scores at discharge, as well as greater ΔFIM_motor were also significantly correlated with lower 3-month mortality (p < 0.001, p < 0.001 and p < 0.001 respectively). FIM_motor at the beginning of rehabilitation was inversely correlated with mortality at 3 months (p = 0.009).

Higher pre-fracture functional status correlated with greater ΔFIM_motor (p = 0.009). Patients with subcapital fracture had higher FIM_motor scores at discharge and ΔFIM_motor compared with patients with pertrochanteric fracture (65.24±11.16 vs. 53.08±19.07, p < 0.006; 25.04±11.12 vs. 13.47±17.36, p < 0.004). Jewish patients had greater ΔFIM_motor compared with Arab patients (19.98±13.67 vs. 2.71±25.85, p < 0.007).

Long-term functional status reported at telephone interview correlated with pre-fracture functional status (p = 0.029). FIM scores at the end of rehabilitation and ΔFIM_motor also correlated with functional status 3 months after rehabilitation.

Total mortality rate in our study was 17.2% (5 patients died during rehabilitation and 6 died during the 3-month follow-up). No association was found between epidemiologic characteristics of the patients and the outcomes measured, including mortality during rehabilitation, mortality at 3 months, rate of readmission to the general hospital and duration of rehabilitation.

4.2.  Glycemic Control and Outcomes

Anti-diabetic drugs most frequently used during rehabilitation were metformin (57.81%), short-acting insulin analogues (25.00%) and long-acting insulin analogues (39.10%). Four patients (6.25%) did not require anti-diabetic agents before hospitalization. One of these patients had hyperglycemia during rehabilitation and was started on metformin. Sulfonylureas and metformin were discontinued in two patients (3.13%) due to adequate glycemic control.

HbA1c levels were available for 58 patients (90.63%). Patients were stratified into 2 groups: HbA1c ≤ 6.5% and HbA1c > 6.5%. HbA1c data are presented in Table 1. There was no correlation between the HbA1c and the rate of hypoglycemia (p = 1.000), the rate of readmission to the general hospital (p = 0.915), mortality during rehabilitation (p = 0.378), 3-month mortality (p = 1.00) or duration of rehabilitation (p = 0.174).

Data regarding correlation of glucose levels with rehabilitation outcomes and hypoglycemic events are summarized in (Table 2). There was no correlation between the minimum, average or maximum glucose levels and mortality during rehabilitation, 3-month mortality or duration of rehabilitation. Minimum glucose levels were not correlated with rate of readmission to the general hospital. Patients with lower maximum glucose levels were less likely to be readmitted to the general hospital (p = 0.041). There was a trend towards correlation between lower average glucose level and lower rate of readmission to the general hospital (p = 0.054).

The incidence of hypoglycemia during hospitalization was 14.06%. Only 2 patients (3.12%) experienced severe hypoglycemia, with no clinical consequences. There was no correlation between hypoglycemia and mortality during rehabilitation (p = 0.141), 3-month mortality (p = 0.177) or patient functional status at the end of the follow-up period (p = 0.883).

With regard to functional outcomes, there was no correlation between HbA1c, minimum, average or maximum glucose values and FIM scores at discharge (Tables 1 and 2). For patients who participated in the follow-up telephone interview, there was no correlation between glycemic control variables and status 3 months post-discharge (p = 0.288).

5. Discussion

Diabetic patients are at increased risk for hip fractures [13]. Some clinical factors were associated with this increase including HbA1c above 9% [14], insulin treatment [15] and glucose variability [16]. Moreover, mortality after hip fracture is up to 30% higher in diabetic patients compared to non-diabetic patients [17-19].

Nevertheless, data pertaining to the effect of glycemic control on rehabilitation outcomes following hip fracture is limited. It has been shown that diabetic patients tend to remain in rehabilitation longer compared to non-diabetic patients. Studies addressing the influence of glycemic control on rehabilitation outcomes have presented conflicting results [9,12].

In the current study, several different parameters of glycemic control did not correlate with rehabilitation outcomes after hip fracture. We established a link between higher maximal glucose levels and increased hospital readmission rate. Although poor glycemic control is considered a risk factor for perioperative complications such as cardiovascular events and infections and readmissions [20,21], it is difficult to ascertain whether hyperglycemia underlies the higher rate of complications or reflects the severity of the patient’s illness secondary to circulating stress hormones.

The total mortality rate in our study was 17.2% (5 patients died during rehabilitation and 6 died during the 3-month follow-up). An association between glucose control and mortality was not found, possibly due to the homogeneity of the patients; 68.42% had HbA1c levels ≤ 7%. These relatively low HbA1c levels are surprising considering the high prevalence of uncontrolled diabetes mellitus in the general population and the lower fracture risk among elderly patients with low HbA1c levels (6.5-6.9%) [14,21]. However, strict glycemic control may also come at a risk of hypoglycemia, which can lead to recurrent falls and fractures [22]. The relatively low HbA1c levels in our study may be partially explained by post-operative anemia and administration of blood transfusions, which could have diluted patients’ hemoglobin. This hypothesis may be refuted by the good correlation of the HbA1c levels with the glucose levels (Table 1).

In accordance with previous studies [22-25], pre-fracture functional status in our study was a strong predictor of patient outcome following hip fracture. Independent patients with diabetes tended to survive longer, to achieve a greater change in FIM scores and to return to their pre-fracture status compared with frail or dependent patients with diabetes. This may be linked to better mental status at the time of fracture [26], a factor that represents a commitment to acquire new abilities, such as using assistive devices for walking, and to regain previous activity levels. The correlation between higher FIM_motor and FIM_cognitive scores at discharge and survival and long-term functional status found in our study supports this concept.

The strengths of this study include the analysis of the glycemic control and its effect on rehabilitation using HbA1c and minimum, average and maximum glucose levels. In addition, glycemic control was also evaluated in relation to patients’ functional status pre- and post-fracture. To the best of our knowledge, previous studies did not analyze glycemic control to this depth. Another strength of the study is the 3-month follow-up after rehabilitation.

The main limitation of this study is the small number of participants. In addition, patients with poor glucose control were under-represented in our cohort. Moreover, evaluation of 3-month functional status by telephone interview which was limited to 3 categories (independent, frail or dependent) rather than using formal FIM scoring may have led to imprecision. Self-reporting may also be a cause of bias.

6.  Conclusion

In this study, we demonstrated that pre-fracture status rather than glucose control parameters was the main determinant of rehabilitation outcome among patients with diabetes following hip fracture. However, higher maximal glucose levels were correlated with an increased rate of readmissions to the general hospital. This may reflect the severity of the medical condition necessitating the readmission. Future studies with larger populations from multiple centers, with longer follow-up periods and more sophisticated evaluation of post-rehabilitation functional status is necessary to establish these results.

7.  Declaration of Interest: The authors report no conflict of interests.

Table 1: Demographic data according to hemoglobin A1c values.

Table 2: The relationship between the average and maximum glucose levels and rehabilitation outcomes.

1. Clement S, Braithwaite SS, Magee MF, Ahmann A, Smith ER, et al. (2004) Management of diabetes and hyperglycemia in hospitals. Diabetes Care 27: 553-591.
2. Trence DL, Kelly JL, Hirsch IB (2003) The rationale and management of hyperglycemia for in-patients with cardiovascular disease: Time for change. J Clin Endocrinol Metab 88: 2430-2437.
3. Metchick LN, Petit WA Jr, Inzucchi SE (2002) Inpatient management of diabetes mellitus. Am J Med 113: 317-323.
4. Falciglia M, Freyberg RW, Almenoff PL, et al. (2009) Hyperglycemia-related mortality in critically ill patients varies with admission diagnosis. Crit Care Med 37: 3001-3009.
5. Umpierrez GE, Isaacs SD, Bazargan N, et al. (2002) Hyperglycemia: An independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 87: 978-982.
6. Malmberg K, Norhammar A, Wedel H, et al. (1999) Glucometabolic state at admission: Important risk marker of mortality in conventionally treated patients with diabetes mellitus and acute myocardial infarction: long-term results from the Diabetes and Insulin-Glucose Infusion in Acute Myocardial Infarction (DIGAMI) study. Circulation 99: 2626-2632.
7. Van den berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, et al. (2001) Intensive insulin therapy in critically ill patients. N Eng J Med 245: 1359-1367.
8. Finfer S, Chittock DR, Su SY, Blair D, Foster D, et al. (2009) Intensive versus conventional glucose control in critically ill patients. New Engl J Med 360: 1283-1297.
9. Reistetter TA, Graham JE, Deutsch A, Markello SJ, Granger CV, et al. (2011) Diabetes comorbidity and age influence rehabilitation outcomes after hip fracture. Diabetes Care 34: 1375-1377.
10. Lieberman D, Fried V, Castel H, Weitzmann S, Lowenthal MN, et al. (1996) Factors related to successful rehabilitation after hip fracture: A case control study. Diabil Rehabil 18: 224-230.
11. Lieberman D, Lieberman D (2004) Rehabilitation following hip fracture surgery: a comparative study of females and males. Disabil Rehabil 26: 85-90.
12. Mizrahi EH, Fleissig Y, Arad M, Adunsky A (2006) Functional outcome of elderly hip fracture patients: Does diabetes matter? Arch Gerontol Geriat 43: 165-173.
13. Hothersall EJ, Livingstone SJ, Looker HC, Ahmed SF, Cleland S, et al. (2014) Contemporary risk of hip fracture in type 1 and type 2 diabetes: a nation registry study from Scotland. J Bone Miner Res 29: 1054-1060.
14. Chia-Ing L, Chiu-Shong L, Wen-Yuan L, Meng NH, Chen CC, et al. (2015) Glycated hemoglobin level and risk of hip fracture in older people with type 2 diabetes: A competing risk analysis of Taiwan Diabetes Cohort Study. J Bone Miner Res 30: 1338-1346.
15. Hamilton E, Davis WA, Bruce DG, Davis TM (2017) Influence of premature mortality on the link between type 2 diabetes and hip fracture: the Fremantle Diabetes Study. J Clin Endocrinol Metab 102: 551-559.
16. Chiang JI, Li TC, Li CI, Liu CS, Meng NH, et al. (2016) Visit to visit variation of fasting plasma glucose is a predictor of hip fracture in older persons with type 2 diabetes: the Taiwan Diabetes Study. Osteoporos Int 27: 3587-3597.
17. Muraki S, Yamamoto S, Ishbashi H, Nakamura K (2006) Factors associated with mortality following hip fracture in Japan. J Bone Min Met 24: 100-104.
18. Dubey A, Aharonoff GB, Zuckerman JD, Koval KJ (2000) The effects of diabetes on outcome after hip fracture. Bul Hospit Joint Dis 59: 94-98.
19. Martinez-Laguna D, Nogues X, Abrahamsen B, Reyes C1, Carbonell-Abella C, et al. (2017) Excess of all-cause mortality after a fracture in type 2 diabetic patients: a population-based cohort study. Osteoporos Int 28: 2573-2581.
20. Marchant MH Jr, Viens NA, Cook C, Vail TP, Bolognesi MP (2009) The impact of glycemic control and diabetes mellitus on perioperative outcomes after total joint arthroplasty. J Bone Joint Surg 91: 1621-1629.
21. Conway BN, Long DM, Figaro MK, May ME (2016) Glycemic control and fracture risk in elderly patients with diabetes. Diabetes Res Clin Pract 115: 47-53.
22. Patel A, MacMahon S, Chalmers J (2008) Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med 358: 2560-2572.
23. Semel J, Gray JM, Jun Ahn H, Nasr H, Chen JJ (2010) Predictors of outcome following hip fracture rehabilitation. PM&R 2: 799-805.
24. Hu F, Jiang C, Shen J, Tang P, Wang Y (2012) Preoperative predictors for mortality following hip fracture surgery: A systematic review and meta-analysis. Injury 43: 676-685.
25. Kang JW, Kim KJ, Lee SK, et al. (2013) Predictors of mortality in patients with hip fracture for persons aging more than 65 years old. Int J Bio-Sci Bio-Tec 5: 27-34.
26. Alegre-López J, Cordero-Guevara J, Alonso-Valdivielso JL, Fernández-Melón J (2005) Factors associated with mortality and functional disability after hip fracture: An inception cohort study. Osteoporos Int 16: 729-736.