- Email: [email protected]
A systematic review of growth hormone for hip fractures
Growth Hormone & IGF Research 22 (2012) 97–101
Contents lists available at SciVerse ScienceDirect
Growth Hormone & IGF Research journal homepage: www.elsevier.com/locate/ghir
Review
A systematic review of growth hormone for hip fractures Shengping Yang a, b, Liying Cao a, b, Shumei Cai a, b, Jinqiu Yuan c, Jianmin Wang a, b,⁎ a b c
The First Clinical Medicine College of Lanzhou University, Lanzhou 730000, Gansu, China The First Hospital of Lanzhou University, Lanzhou 730000, Gansu, China School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong, China
a r t i c l e
i n f o
Article history: Received 5 January 2012 received in revised form 8 February 2012 accepted 5 March 2012 Available online 1 April 2012 Keywords: Growth hormone Hip fractures Systematic review
a b s t r a c t Objective: To evaluate the clinical effectiveness of growth hormone (GH) for hip fractures. Design: Based on the principles and methods of Cochrane systematic reviews, we searched the Cochrane Library (2011, 3 issue), PubMed (1966 to October 2011), EMBASE (1974 to October 2011), OVID database (1963 to October 2011), Chinese Bio-medicine database (1978 to October 2011), China Journal Full-text Database (1979 to October 2011), and VIP database (1989 to October 2011). Randomized controlled trials (RCTs) of GH treatment for hip fractures were included. We assessed the quality of included trials according to the Cochrane Handbook for Systematic Reviews of Interventions Version. The Cochrane Collaboration's software RevMan 5.0 was used for meta-analysis. Results: 3 RCTs with a total of 162 hip fractures patients were included. Results of GH compared to placebo treatment showed that IGF-I levels significantly increased in the short term, but no significant differences in the long term. Additionally, there was no statistical difference in adverse events. Only one trial reported the measurement of bone mineral content (BMC), which showed that GH group was unchanged while the placebo group lost BMC at both 4 weeks and 8 weeks. Only one study measuring BMC showed no significant differences in the change of the MBI scores between GH group and placebo group; however, the changes from baseline in the hGH group were less than that in the placebo group among the older than 75 years group. Owing to the difference of measurement indexes in those studies, we could not perform a meta-analysis. Conclusions: With the low quality of current evidence, GH may be effective in hip fractures. More carefully designed, double-blinded and placebo-controlled randomized trials with large numbers of participants about GH in the treatment of hip fractures are required. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Growth hormone (GH) is a kind of protein hormone secreted by eosinophilic granulocyte of anterior pituitary [1], and it could affect the development of several tissues including liver, kidney, muscle and bone. Several studies both in vivo and in vitro have showed that GH plays an important role in the regulation of bone formation and restoration [2]. On the one hand, GH is involved in the induction of differentiation of pre-osteoblast and increases the activity of the osteoblast. On the other hand, GH has an indirect action on the stimulation of bone remodeling through producing insulin-like growth factor-I (IGF-I) locally [3]. Additionally, serum IGF levels are relatively stable, and can be an indicator of changes of serum GH level. Therefore, the injection of GH could increase bone mineral content (BMC) and bone mineral density (BMD) [4], and it seems to be a promising treatment for promoting fracture healing.
Hip fractures consist of fractures of the femur head, the femur neck, the trochanters or the inter- or subtrochanteric region. They are associated with a one-year mortality rate ranging from 14% to 36% among elderly patients [5]. Hip fracture injuries have been identified as one of the most serious health care problems affecting older people for many years [6]. At present, systematic review has been recognized as the best evidence in evidence-based medicine [7–9]. Owing to the lack of highquality evidence produced by evidence-based medicine, we want to know whether GH has therapeutic effects on fracture healing of the hip. And we expect this systematic review of randomized controlled trials (RCTs) to identify the effects of GH treatment on hip fractures, which could be a reference for clinicians.
2. Material and methods 2.1. Criteria for considering studies for this review
⁎ Corresponding author. Tel: + 86 13919803289; fax: + 86 0931 8619797. E-mail address: [email protected] (J. Wang). 1096-6374/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.ghir.2012.03.002
2.1.1. Types of studies We included all relevant RCTs and controlled clinical trials (CCTs).
98
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
2.1.2. Types of participants We included people with hip fractures diagnosed by any criteria without discrimination of age, sex and race. 2.1.3. Types of interventions The treatment group's intervention was GH therapy with any dose, including human GH (hGH) or recombinant hGH (r-hGH). The control group's interventions were placebo, no treatment, or other active therapies. 2.1.4. Types of outcome measures Primary outcomes a) Hormonal response: mainly manifested as changes from baseline in IGF and Insulin-like Growth Factor Binding Proteins (IGFBP) or the IGF or IGFBP levels during trial. b) Adverse events. Secondary outcomes c) Measurement of BMC, BMD. d) Quality of life. 2.2. Search methods for identification of studies We searched the Cochrane Library (2011, 3 issue), PubMed (1966 to October 2011), EMBASE (1974 to October 2011), OVID database (1963 to October 2011), Chinese Bio-medicine database (1978 to October 2011), China Journal Full-text Database (1979 to October 2011), and VIP database (1989 to October 2011). We used the following search strategy to search PubMed, and used the similar search strategy in other databases. Search terms used in Chinese databases were in Fig. 1. The search strategy for identifying randomized trials was performed according to the designing search strategies described in the Cochrane Handbook for Systematic Reviews of Interventions [10]. References of included studies were extra searched. There were no language restrictions. Search strategy in PubMed #1 randomized controlled trial [pt] #2 controlled clinical trial [pt] #3 “Clinical Trials as Topic” [mh] #4 randomized [ti/ab] #5 placebo [ti/ab] #6 randomly [ti/ab] #7 trial [ti] #8 OR/#1–#7 #9 “Growth Hormone” [mh] #10 growth hormone #11 somatotropin #12 GH #13 “Human Growth Hormone” [mh] #14 human growth hormone #15 hGH #16 recombinant human growth hormone #17 recombined human growth hormone #18 rhGH #19 r-hGH #20 OR/#9-#19 #21 “Hip Fractures” [mesh] #22 hip fractures #23 hip fracture #24 OR/#21–#23 #25 #8 and #20 and #24
would be assessed further, then we inspected the full articles to decide whether they met the review criteria. When disagreement occurred, we resolved this by discussion. Contacting with the authors was initially by e-mail or telephone if any information was not available. Study selection was not blinded as blinding is not necessary in systematic reviews of randomized clinical trials [11]. 2.3.2. Assessment of risk of bias The risk of bias of included RCTs was assessed as described in the Cochrane Handbook for Systematic Reviews of Interventions [10] by two reviewers independently. It was judged by using the following criteria: sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting, and other sources of bias. With a judgment of ‘Yes’ indicating low risk of bias, ‘No’ indicating high risk of bias, and ‘Unclear’ indicating unclear or unknown risk of bias. 2.3.3. Data extraction and analysis Two reviewers independently extracted data from selected trials. When disputes arose, we attempted to resolve these by discussion. When this was not possible we did not enter the data, but added the outcome of the trial to the list of those awaiting assessment. Statistical analysis was performed by using the Review Manager software 5.0. For dichotomous data, relative risk (RR) and associated 95% confidence interval (CI) were calculated, while mean difference (MD) and 95%CI were calculated for continuous data. Heterogeneity between studies was assessed by the χ 2 test with p b 0.10 used to indicate statistical significance. I 2 also was calculated to measure the quantity of heterogeneity, with I 2 > 50% indicating significant heterogeneity [12]. The meta-analysis was conducted using the fixedeffect model if there was no statistically significant heterogeneity (p ≥ 0.10, I 2 ≤ 50%); otherwise, the possible reasons were explored or the random-effect model was used for the significant heterogeneity (p b 0.10, I 2 > 50%). When data could not be extracted for metaanalysis, the data from these trials was assessed as descriptive data and still considered in the results of the review. 3. Results 3.1. Description of studies We identified 842 references through the search. After reading titles and abstracts, we excluded 833 articles because they were duplicates, non-clinical studies or inconsistent with study objectives for our review. A total of 9 references were retrieved for further assessment, and then 6 references [13–18] were excluded because they did not meet our inclusion criteria. Finally, 3 studies [19–21] were included. The flowchart of literature screening is presented in Fig. 2 as described in the PRISMA statement [22]. A total of 162 hip fracture patients were included and the follow-up period ranged from 4 days to 2 months. Groups were well matched at the baseline from the information in the majority of trials. Further details are given in the characteristics of included studies (Table 1). 3.2. Risk of bias in included studies The risks of bias in the included trials have been summarized in Table 2 and a “Risk of bias graph” judgment about each risk of bias
2.3. Data collection and analysis 2.3.1. Selection of studies Two review authors independently scanned the titles, abstract sections and keywords of every record to determine whether the studies
Fig. 1. Search terms used in Chinese databases.
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
842 of records identified through database searching
0 of additional records identified through other sources
816 of records after duplicates removed
816 of records screened
807 of records excluded 9 of full-text articles assessed for eligibility
6 of full-text articles excluded: non-GH studies n=4 outcomes of relevance not reported n=1 reviews n=1
3 of studies included in qualitative synthesis
3 of studies included in qualitative synthesis (meta-analysis)
Fig. 2. The flowchart of literature screening.
item presented as percentages across all included trials in Fig. 3. All trials were described as randomized whereas they didn't describe the method of generating the allocation sequence. The method of concealment was not described, either. Double-blinds were reported in all trials although no description was provided. And all trials reported the missing data or patients lost to follow-up. We sent e-mails to the authors for the further information, but no responses were received.
3.3. Effects of interventions 3.3.1. Hormonal response All of included trials reported the hormonal response. Yeo's study [19] reported the mean IGF-I levels and mean IGFBP-3 levels during the study (days 4, 5, 6, 8, 10, 14 and 18) in two figures, without concrete numerical value, and no response when we sent an e-mail to the author to inquire. In this study, the author got a result that there was a significant increase in both r-hGH groups but no differences between the two r-hGH groups both for IGF-I and IGFBP-3.
99
Hedström's study [20] reported the mean IGF-I levels and mean IGFBP-1 levels during treatment (0 week, 1 week, 4 weeks and after 2 months). In this study, the author draws the conclusion that there was a significant increase in IGF-I levels (p = 0.005) and a significant decrease in IGFBP-1 levels (p = 0.04) on week 1. However, there were no differences in IGF-I levels (p = 0.7) and IGFBP-1 levels (p = 0.1) levels between the two groups after 2 months. Van der Lely's study [21] reported the changes from baseline in serum IGF-I and IGFBP-1 concentrations at different phases (weeks 2, 4, 6 and 8). In this study, the authors got results that there was a significant difference in IGF-I levels at week 2 (pb 0.001) and week 4 (pb 0.001), while no significant differences at week 8 (p= 0.397). Similar and parallel changes were observed for IGFBP-3 values (data not shown). But IGFBP-1 decreased during the study in both hGH and placebo groups without significant differences between the two groups (p> 0.01). Due to the difference of measurement indexes in those studies, we could not make a meta-analysis for the hormonal response. 3.3.2. Adverse events All of the included trials reported the adverse events. The adverse events which authors reported include acute renal failure, pulmonary emboli, myocardial infarction, edema, hypertension, urinary tract infection and constipation. Owing to the difference of statistical methods and results in those studies, we could not make a meta-analysis for the adverse events. The results of adverse events are summarized in Table 3. Two trials [19,21] reported that there were no significant differences between the GH group and placebo group in the incidence of adverse events, which revealed the same results as another study [20] did. All of the included trials mentioned death with the total of 8. In Yeo's study [19], one from the r-hGH group (0.05 mg/kg/d) died from acute renal failure on day 16, meanwhile one of the r-hGH group (0.025 mg/kg/d) died from mass pulmonary embolism 40 days after hip surgery. In Hedström's study [20], one from the r-hGH group died of myocardial infarction shortly after surgery, which was not considered to be related to GH. In Van der Lely's study [21], three patients in the hGH group (0.02 mg/kg/d) died after the end of the active treatment phase and two died in the placebo group, however, the author did not describe the cause of death. 3.3.3. Measurement of BMC, BMD Measurement of BMC was reported in only one study [20]. In this trial, total BMC was unchanged in the GH treated group while the placebo group lost. There is a significant difference both at 4 weeks (p = 0.04) and 8 weeks (p = 0.01). Yet, none of those trials reported measurement of BMD. 3.3.4. Quality of life Quality of life was reported by only one trial [21]. The Modified Barthel Index (MBI) of activities of daily living (ADL) was chosen to assess the functional status of the patients. There were no significant
Table 1 Characteristics of included studies. Study
Year
Country
Groups
Age, mean (SD), yrs
No. of patients (male/female)
Weight, mean (SD), kg
BMI, mean (SD), kg/m2
Outcomes
Yeo et al. [19]
2003
UK
2004
Sweden
Van der Lely et al. [21]
2000
Netherlands
87.9 84.6 85.7 83 (7) 85 (3) 79.2 (8.5) 77.8 (9.6)
0/11 0/10 0/10 11 9 13/42 8/48
54.7 (10.40) 54.5 (7.64) 50.6 (4.45) 61 (12) 55 (6) – –
22.3 (3.78) 24.2 (2.63) 23.2 (2.38) 22.8 (4.5) 20.4 (1.8) – –
a, b
Hedström et al. [20]
r-hGH 0.05 mg/kg/d r-hGH 0.025 mg/kg/d Placebo r-hGH 0.033 mg/kg/d Placebo hGH 0.02 mg/kg/d Placebo
a: hormonal response; b: adverse events; c: measurement of BMC, BMD; d: quality of life.
a, b, c a, b, d
100
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
Table 2 Risk of bias in included studies. Study
Sequence generation
Allocation concealment
Blinding
Incomplete outcome data
Selective outcome reporting
Other sources of bias
Yeo et al. [19] Hedström et al. [20] Van der Lely et al. [21]
Unclear Unclear Unclear
Unclear Unclear Unclear
Yes Yes Yes
Yes Yes Yes
Unclear Unclear Unclear
Unclear Unclear No
4.3. Quality of the evidence
Fig. 3. Risk of bias graph.
differences in the changes in the MBI scores between two treatment groups at any time point. However, it was found that patients older than 75 years showed a variable recovery, in which the changes from baseline in the hGH group were lesser than in the placebo group at any time point in the older groups, when the difference was the greatest after 6 weeks of therapy (− 18.6 ± 18 vs. − 28.1 ± 26, p b 0.075), while those under 75 years old had little change in the MBI scores. Therefore, it may show the potential to detect a treatment effect in favor of hGH in patients older than 75 years of age. 4. Discussions 4.1. Summary of main results This systematic review summarizes the results of 3 RCTs about the effectiveness and safety of GH treating for hip fractures. Studies incorporating hGH or r-hGH into a treatment program for hip fractures were included for this review. This descriptive systematic review [23] found GH might be effective in hip fractures. However, with the small number of participants and studies and the poor quality of this study, we could not reach reliable conclusions. Our study has only focused on effects of GH for hip fractures in short-term treatment from our included literatures. Nevertheless, it does not mean excluding the possibility of long-term effects. Further work in this area is in process. 4.2. Overall completeness and applicability of evidence The patients in the included studies are elderly and most of them are women. Our research objects are similar to the view which says that the average age of patients with hip fractures is over 80, and nearly 80% are women [24]. All of the included studies reported hormonal response. However, two [19,20] reported the mean IGF-I levels during treatment, another [21] reported the changes from baseline in serum IGF-I at different phases. Although both reflect the hormonal response, we could not perform a meta-analysis. In addition, one trial [20] reported measurement of BMC and one trial [21] reported quality of life. Different indexes make our results incomplete. However, it also indicates to us that there is a wide area of potential research as many of the articles we found indicate, with only a few RCTs. We hope high quality of RCTs in this field will be conducted in the future.
A total of 3 RCTs were identified. The quality of the reports in general was poor. None of them detailed the method of randomization and allocation concealment. Therefore, selective bias was inevitable. Blinding was reported in all trials, which showed a low risk of performance bias or detection bias. Although a comprehensive literature search was conducted, some published and unpublished trials might be missed, which would lead to nonpublication bias. The chief investigators of all studies were contacted to provide additional methodological and statistical information; however, no response was obtained. A source of bias may also have arisen from the use of the poorly defined criteria for the outcome “hormonal response”. The small number of studies within comparison limits us to explore the impact on treatment effects; in the meantime, the lack of high quality trials prevents us from further investigating the heterogeneity of the studies. 4.4. Agreements and disagreements with other studies or reviews To our knowledge, this is the first systematic review to evaluate the effect of GH on hip fractures for the present. There was only one human study done in evaluating the effect of GH on fracture healing in 2007 [25]. In that study, the authors evaluated the effect of r-hGH on 368 patients with tibial fractures by radiographic evaluation and clinically assessed healing. And the author concluded that no significant enhancement of fracture healing was observed with GH in the overall groups of open and closed tibial fractures, whereas GH accelerated healing significantly in closed tibial fractures. In the future, well-designed RCTs with large sample size are warranted to confirm or refute the current evidence. 5. Conclusions GH may be effective in hip fractures in view of the low quality evidence currently. More carefully designed, double-blinded and placebocontrolled randomized trials with large numbers of participants which use GH in the treatment of hip fractures are required. Conflict of interest The authors have no conflicts of interest that are directly relevant to the content of this study.
Table 3 Clinical adverse events. Events
GH group (n = 87)a
Placebo group (n = 75)
Studies
Acute renal failure Pulmonary embolism Edema
1 3 18
0 0 17
Hypertension Urinary tract infection Constipation Total
1 14 9 46
0 8 10 35
Yeo et al. Yeo et al. Hedström et al., Van der Lely et al. Hedström et al. Van der Lely et al. Van der Lely et al.
a
Any dose of GH.
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
References [1] Y. Kato, Y. Murakami, M. Sohmiya, M. Nishiki, Regulation of human growth hormone secretion and its disorders, Internal Medicine-Tokyo-Japanese Society of Internal Medicine 41 (1) (2002) 7–13. [2] C. Ohlsson, B.A. Bengtsson, O.G.P. Isaksson, T.T. Andreassen, M.C. Slootweg, Growth hormone and bone, Endocr Rev. 19 (1) (1998) 55–79. [3] K. Chihara, T. Sugimoto, The action of GH/IGF-I/IGFBP in osteoblasts and osteoclasts, Horm Res. 48 (1997) 45–49. [4] H.J. Bail, S. Kolbeck, G. Krummrey, et al., Systemic application of growth hormone for enhancement of secondary and intramembranous fracture healing, Horm Res. 58 (2002) 39–42. [5] J.D. Zuckerman, A.D. Rosenberg, Hip fracture, New Engl J Med. 335 (26) (1996) 1995–1996. [6] R. Marks, Hip fracture epidemiological trends, outcomes, and risk factors, 1970–2009, Int J Gen Med. 3 (2010) 1–17. [7] Guide to research methods: the evidence pyramid available from http://library. downstate.edu/EBM2/2100.htm. [8] R. Harbour, J. Miller, A new system for grading recommendations in evidence based guidelines, BMJ. 323 (7308) (2001) 334. [9] Y.P. li, T.J. Tao, D. Sun, et al., A pilot study of the classification of Chinese Medical Specialty, Chinese Journal of Evidence-Based Medicine. 4 (3) (1994) 173–180. [10] J. P. T. Higgins, S. Green, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2 [Updated September 2009] The Cochrane Collaboration (2009) Available from www.cochranehand-book.org [11] J.A. Berlin, Does blinding of readers affect the results of meta-analyses? Lancet. 350 (9072) (1997) 185–186. [12] J.P.T. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat Med. 21 (11) (2002) 1539–1558. [13] T. Chevalley, P. Hoffmeyer, J.P. Bonjour, R. Rizzoli, Early serum IGF-I response to oral protein supplements in elderly women with a recent hip fracture, Clinical Nutrition. 29 (1) (2010) 78–83. [14] K. Wang, L. Zhu, H.D. Lu, D.Z. Cai, Effect of reconstructed growth hormone on hip arthroplasty of the aged patients, Orthopedic Journal of China. 13 (4) (2005) 273–275.
101
[15] M.A. Bach, K. Rockwood, C. Zetterberg, et al., The effects of MK-0677, an oral growth hormone secretagogue, in patients with hip fracture, J Am Geriatr Soc. 52 (4) (2004) 516–523. [16] M. Hedström, O. Ljungqvist, T. Cederholm, Metabolism and catabolism in hip fracture patients: nutritional and anabolic intervention — a review, Acta orthopaedica. 77 (5) (2006) 741–747. [17] A. Adunsky, J. Chandler, N. Heyden, et al., MK-0677 (ibutamoren mesylate) for the treatment of patients recovering from hip fracture: a multicenter, randomized, placebo-controlled phase IIb study, Archives of Gerontology and Geriatrics (2010). [18] M.A. Schurch, R. Rizzoli, D. Slosman, et al., Protein supplements increase serum insulin-like growth factor-I levels and attenuate proximal femur bone loss in patients with recent hip fracture, Annals of internal medicine. 128 (10) (1998) 801. [19] A.L. Yeo, D. Levy, F.C. Martin, et al., Frailty and the biochemical effects of recombinant human growth hormone in women after surgery for hip fracture, Growth Horm Igf Res. 13 (6) (2003) 361–370. [20] M. Hedström, M. Saaf, E. Brosjo, et al., Positive effects of short-term growth hormone treatment on bean body mass and BMC after a hip fracture — a double-blind placebo-controlled pilot study in 20 patients, Acta Orthop Scand. 75 (4) (2004) 394–401. [21] A.J. Van der Lely, S.W. Lamberts, K.W. Jauch, et al., Use of human GH in elderly patients with accidental hip fracture, Eur J Endocrinol. 143 (5) (2000) 585–592. [22] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, PLoS Med. 6 (7) (2009) e1000097. [23] T.X. Wu, G.J. Liu, J. Li, Some risks of affecting the quality of published systematic reviews in China, Chinese Journal of Evidence-Based Medicine. 5 (1) (2005) 51–58. [24] G.S. Keene, M.J. Parker, G.A. Pryor, Mortality and morbidity after hip fractures, BMJ. 307 (6914) (1993) 1248–1250. [25] M. Raschke, M.H. Rasmussen, S. Govender, et al., Effects of growth hormone in patients with tibial fracture: a randomised, double-blind, placebo-controlled clinical trial, Eur J Endocrinol. 156 (3) (2007) 341–351.
Contents lists available at SciVerse ScienceDirect
Growth Hormone & IGF Research journal homepage: www.elsevier.com/locate/ghir
Review
A systematic review of growth hormone for hip fractures Shengping Yang a, b, Liying Cao a, b, Shumei Cai a, b, Jinqiu Yuan c, Jianmin Wang a, b,⁎ a b c
The First Clinical Medicine College of Lanzhou University, Lanzhou 730000, Gansu, China The First Hospital of Lanzhou University, Lanzhou 730000, Gansu, China School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong, China
a r t i c l e
i n f o
Article history: Received 5 January 2012 received in revised form 8 February 2012 accepted 5 March 2012 Available online 1 April 2012 Keywords: Growth hormone Hip fractures Systematic review
a b s t r a c t Objective: To evaluate the clinical effectiveness of growth hormone (GH) for hip fractures. Design: Based on the principles and methods of Cochrane systematic reviews, we searched the Cochrane Library (2011, 3 issue), PubMed (1966 to October 2011), EMBASE (1974 to October 2011), OVID database (1963 to October 2011), Chinese Bio-medicine database (1978 to October 2011), China Journal Full-text Database (1979 to October 2011), and VIP database (1989 to October 2011). Randomized controlled trials (RCTs) of GH treatment for hip fractures were included. We assessed the quality of included trials according to the Cochrane Handbook for Systematic Reviews of Interventions Version. The Cochrane Collaboration's software RevMan 5.0 was used for meta-analysis. Results: 3 RCTs with a total of 162 hip fractures patients were included. Results of GH compared to placebo treatment showed that IGF-I levels significantly increased in the short term, but no significant differences in the long term. Additionally, there was no statistical difference in adverse events. Only one trial reported the measurement of bone mineral content (BMC), which showed that GH group was unchanged while the placebo group lost BMC at both 4 weeks and 8 weeks. Only one study measuring BMC showed no significant differences in the change of the MBI scores between GH group and placebo group; however, the changes from baseline in the hGH group were less than that in the placebo group among the older than 75 years group. Owing to the difference of measurement indexes in those studies, we could not perform a meta-analysis. Conclusions: With the low quality of current evidence, GH may be effective in hip fractures. More carefully designed, double-blinded and placebo-controlled randomized trials with large numbers of participants about GH in the treatment of hip fractures are required. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Growth hormone (GH) is a kind of protein hormone secreted by eosinophilic granulocyte of anterior pituitary [1], and it could affect the development of several tissues including liver, kidney, muscle and bone. Several studies both in vivo and in vitro have showed that GH plays an important role in the regulation of bone formation and restoration [2]. On the one hand, GH is involved in the induction of differentiation of pre-osteoblast and increases the activity of the osteoblast. On the other hand, GH has an indirect action on the stimulation of bone remodeling through producing insulin-like growth factor-I (IGF-I) locally [3]. Additionally, serum IGF levels are relatively stable, and can be an indicator of changes of serum GH level. Therefore, the injection of GH could increase bone mineral content (BMC) and bone mineral density (BMD) [4], and it seems to be a promising treatment for promoting fracture healing.
Hip fractures consist of fractures of the femur head, the femur neck, the trochanters or the inter- or subtrochanteric region. They are associated with a one-year mortality rate ranging from 14% to 36% among elderly patients [5]. Hip fracture injuries have been identified as one of the most serious health care problems affecting older people for many years [6]. At present, systematic review has been recognized as the best evidence in evidence-based medicine [7–9]. Owing to the lack of highquality evidence produced by evidence-based medicine, we want to know whether GH has therapeutic effects on fracture healing of the hip. And we expect this systematic review of randomized controlled trials (RCTs) to identify the effects of GH treatment on hip fractures, which could be a reference for clinicians.
2. Material and methods 2.1. Criteria for considering studies for this review
⁎ Corresponding author. Tel: + 86 13919803289; fax: + 86 0931 8619797. E-mail address: [email protected] (J. Wang). 1096-6374/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.ghir.2012.03.002
2.1.1. Types of studies We included all relevant RCTs and controlled clinical trials (CCTs).
98
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
2.1.2. Types of participants We included people with hip fractures diagnosed by any criteria without discrimination of age, sex and race. 2.1.3. Types of interventions The treatment group's intervention was GH therapy with any dose, including human GH (hGH) or recombinant hGH (r-hGH). The control group's interventions were placebo, no treatment, or other active therapies. 2.1.4. Types of outcome measures Primary outcomes a) Hormonal response: mainly manifested as changes from baseline in IGF and Insulin-like Growth Factor Binding Proteins (IGFBP) or the IGF or IGFBP levels during trial. b) Adverse events. Secondary outcomes c) Measurement of BMC, BMD. d) Quality of life. 2.2. Search methods for identification of studies We searched the Cochrane Library (2011, 3 issue), PubMed (1966 to October 2011), EMBASE (1974 to October 2011), OVID database (1963 to October 2011), Chinese Bio-medicine database (1978 to October 2011), China Journal Full-text Database (1979 to October 2011), and VIP database (1989 to October 2011). We used the following search strategy to search PubMed, and used the similar search strategy in other databases. Search terms used in Chinese databases were in Fig. 1. The search strategy for identifying randomized trials was performed according to the designing search strategies described in the Cochrane Handbook for Systematic Reviews of Interventions [10]. References of included studies were extra searched. There were no language restrictions. Search strategy in PubMed #1 randomized controlled trial [pt] #2 controlled clinical trial [pt] #3 “Clinical Trials as Topic” [mh] #4 randomized [ti/ab] #5 placebo [ti/ab] #6 randomly [ti/ab] #7 trial [ti] #8 OR/#1–#7 #9 “Growth Hormone” [mh] #10 growth hormone #11 somatotropin #12 GH #13 “Human Growth Hormone” [mh] #14 human growth hormone #15 hGH #16 recombinant human growth hormone #17 recombined human growth hormone #18 rhGH #19 r-hGH #20 OR/#9-#19 #21 “Hip Fractures” [mesh] #22 hip fractures #23 hip fracture #24 OR/#21–#23 #25 #8 and #20 and #24
would be assessed further, then we inspected the full articles to decide whether they met the review criteria. When disagreement occurred, we resolved this by discussion. Contacting with the authors was initially by e-mail or telephone if any information was not available. Study selection was not blinded as blinding is not necessary in systematic reviews of randomized clinical trials [11]. 2.3.2. Assessment of risk of bias The risk of bias of included RCTs was assessed as described in the Cochrane Handbook for Systematic Reviews of Interventions [10] by two reviewers independently. It was judged by using the following criteria: sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting, and other sources of bias. With a judgment of ‘Yes’ indicating low risk of bias, ‘No’ indicating high risk of bias, and ‘Unclear’ indicating unclear or unknown risk of bias. 2.3.3. Data extraction and analysis Two reviewers independently extracted data from selected trials. When disputes arose, we attempted to resolve these by discussion. When this was not possible we did not enter the data, but added the outcome of the trial to the list of those awaiting assessment. Statistical analysis was performed by using the Review Manager software 5.0. For dichotomous data, relative risk (RR) and associated 95% confidence interval (CI) were calculated, while mean difference (MD) and 95%CI were calculated for continuous data. Heterogeneity between studies was assessed by the χ 2 test with p b 0.10 used to indicate statistical significance. I 2 also was calculated to measure the quantity of heterogeneity, with I 2 > 50% indicating significant heterogeneity [12]. The meta-analysis was conducted using the fixedeffect model if there was no statistically significant heterogeneity (p ≥ 0.10, I 2 ≤ 50%); otherwise, the possible reasons were explored or the random-effect model was used for the significant heterogeneity (p b 0.10, I 2 > 50%). When data could not be extracted for metaanalysis, the data from these trials was assessed as descriptive data and still considered in the results of the review. 3. Results 3.1. Description of studies We identified 842 references through the search. After reading titles and abstracts, we excluded 833 articles because they were duplicates, non-clinical studies or inconsistent with study objectives for our review. A total of 9 references were retrieved for further assessment, and then 6 references [13–18] were excluded because they did not meet our inclusion criteria. Finally, 3 studies [19–21] were included. The flowchart of literature screening is presented in Fig. 2 as described in the PRISMA statement [22]. A total of 162 hip fracture patients were included and the follow-up period ranged from 4 days to 2 months. Groups were well matched at the baseline from the information in the majority of trials. Further details are given in the characteristics of included studies (Table 1). 3.2. Risk of bias in included studies The risks of bias in the included trials have been summarized in Table 2 and a “Risk of bias graph” judgment about each risk of bias
2.3. Data collection and analysis 2.3.1. Selection of studies Two review authors independently scanned the titles, abstract sections and keywords of every record to determine whether the studies
Fig. 1. Search terms used in Chinese databases.
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
842 of records identified through database searching
0 of additional records identified through other sources
816 of records after duplicates removed
816 of records screened
807 of records excluded 9 of full-text articles assessed for eligibility
6 of full-text articles excluded: non-GH studies n=4 outcomes of relevance not reported n=1 reviews n=1
3 of studies included in qualitative synthesis
3 of studies included in qualitative synthesis (meta-analysis)
Fig. 2. The flowchart of literature screening.
item presented as percentages across all included trials in Fig. 3. All trials were described as randomized whereas they didn't describe the method of generating the allocation sequence. The method of concealment was not described, either. Double-blinds were reported in all trials although no description was provided. And all trials reported the missing data or patients lost to follow-up. We sent e-mails to the authors for the further information, but no responses were received.
3.3. Effects of interventions 3.3.1. Hormonal response All of included trials reported the hormonal response. Yeo's study [19] reported the mean IGF-I levels and mean IGFBP-3 levels during the study (days 4, 5, 6, 8, 10, 14 and 18) in two figures, without concrete numerical value, and no response when we sent an e-mail to the author to inquire. In this study, the author got a result that there was a significant increase in both r-hGH groups but no differences between the two r-hGH groups both for IGF-I and IGFBP-3.
99
Hedström's study [20] reported the mean IGF-I levels and mean IGFBP-1 levels during treatment (0 week, 1 week, 4 weeks and after 2 months). In this study, the author draws the conclusion that there was a significant increase in IGF-I levels (p = 0.005) and a significant decrease in IGFBP-1 levels (p = 0.04) on week 1. However, there were no differences in IGF-I levels (p = 0.7) and IGFBP-1 levels (p = 0.1) levels between the two groups after 2 months. Van der Lely's study [21] reported the changes from baseline in serum IGF-I and IGFBP-1 concentrations at different phases (weeks 2, 4, 6 and 8). In this study, the authors got results that there was a significant difference in IGF-I levels at week 2 (pb 0.001) and week 4 (pb 0.001), while no significant differences at week 8 (p= 0.397). Similar and parallel changes were observed for IGFBP-3 values (data not shown). But IGFBP-1 decreased during the study in both hGH and placebo groups without significant differences between the two groups (p> 0.01). Due to the difference of measurement indexes in those studies, we could not make a meta-analysis for the hormonal response. 3.3.2. Adverse events All of the included trials reported the adverse events. The adverse events which authors reported include acute renal failure, pulmonary emboli, myocardial infarction, edema, hypertension, urinary tract infection and constipation. Owing to the difference of statistical methods and results in those studies, we could not make a meta-analysis for the adverse events. The results of adverse events are summarized in Table 3. Two trials [19,21] reported that there were no significant differences between the GH group and placebo group in the incidence of adverse events, which revealed the same results as another study [20] did. All of the included trials mentioned death with the total of 8. In Yeo's study [19], one from the r-hGH group (0.05 mg/kg/d) died from acute renal failure on day 16, meanwhile one of the r-hGH group (0.025 mg/kg/d) died from mass pulmonary embolism 40 days after hip surgery. In Hedström's study [20], one from the r-hGH group died of myocardial infarction shortly after surgery, which was not considered to be related to GH. In Van der Lely's study [21], three patients in the hGH group (0.02 mg/kg/d) died after the end of the active treatment phase and two died in the placebo group, however, the author did not describe the cause of death. 3.3.3. Measurement of BMC, BMD Measurement of BMC was reported in only one study [20]. In this trial, total BMC was unchanged in the GH treated group while the placebo group lost. There is a significant difference both at 4 weeks (p = 0.04) and 8 weeks (p = 0.01). Yet, none of those trials reported measurement of BMD. 3.3.4. Quality of life Quality of life was reported by only one trial [21]. The Modified Barthel Index (MBI) of activities of daily living (ADL) was chosen to assess the functional status of the patients. There were no significant
Table 1 Characteristics of included studies. Study
Year
Country
Groups
Age, mean (SD), yrs
No. of patients (male/female)
Weight, mean (SD), kg
BMI, mean (SD), kg/m2
Outcomes
Yeo et al. [19]
2003
UK
2004
Sweden
Van der Lely et al. [21]
2000
Netherlands
87.9 84.6 85.7 83 (7) 85 (3) 79.2 (8.5) 77.8 (9.6)
0/11 0/10 0/10 11 9 13/42 8/48
54.7 (10.40) 54.5 (7.64) 50.6 (4.45) 61 (12) 55 (6) – –
22.3 (3.78) 24.2 (2.63) 23.2 (2.38) 22.8 (4.5) 20.4 (1.8) – –
a, b
Hedström et al. [20]
r-hGH 0.05 mg/kg/d r-hGH 0.025 mg/kg/d Placebo r-hGH 0.033 mg/kg/d Placebo hGH 0.02 mg/kg/d Placebo
a: hormonal response; b: adverse events; c: measurement of BMC, BMD; d: quality of life.
a, b, c a, b, d
100
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
Table 2 Risk of bias in included studies. Study
Sequence generation
Allocation concealment
Blinding
Incomplete outcome data
Selective outcome reporting
Other sources of bias
Yeo et al. [19] Hedström et al. [20] Van der Lely et al. [21]
Unclear Unclear Unclear
Unclear Unclear Unclear
Yes Yes Yes
Yes Yes Yes
Unclear Unclear Unclear
Unclear Unclear No
4.3. Quality of the evidence
Fig. 3. Risk of bias graph.
differences in the changes in the MBI scores between two treatment groups at any time point. However, it was found that patients older than 75 years showed a variable recovery, in which the changes from baseline in the hGH group were lesser than in the placebo group at any time point in the older groups, when the difference was the greatest after 6 weeks of therapy (− 18.6 ± 18 vs. − 28.1 ± 26, p b 0.075), while those under 75 years old had little change in the MBI scores. Therefore, it may show the potential to detect a treatment effect in favor of hGH in patients older than 75 years of age. 4. Discussions 4.1. Summary of main results This systematic review summarizes the results of 3 RCTs about the effectiveness and safety of GH treating for hip fractures. Studies incorporating hGH or r-hGH into a treatment program for hip fractures were included for this review. This descriptive systematic review [23] found GH might be effective in hip fractures. However, with the small number of participants and studies and the poor quality of this study, we could not reach reliable conclusions. Our study has only focused on effects of GH for hip fractures in short-term treatment from our included literatures. Nevertheless, it does not mean excluding the possibility of long-term effects. Further work in this area is in process. 4.2. Overall completeness and applicability of evidence The patients in the included studies are elderly and most of them are women. Our research objects are similar to the view which says that the average age of patients with hip fractures is over 80, and nearly 80% are women [24]. All of the included studies reported hormonal response. However, two [19,20] reported the mean IGF-I levels during treatment, another [21] reported the changes from baseline in serum IGF-I at different phases. Although both reflect the hormonal response, we could not perform a meta-analysis. In addition, one trial [20] reported measurement of BMC and one trial [21] reported quality of life. Different indexes make our results incomplete. However, it also indicates to us that there is a wide area of potential research as many of the articles we found indicate, with only a few RCTs. We hope high quality of RCTs in this field will be conducted in the future.
A total of 3 RCTs were identified. The quality of the reports in general was poor. None of them detailed the method of randomization and allocation concealment. Therefore, selective bias was inevitable. Blinding was reported in all trials, which showed a low risk of performance bias or detection bias. Although a comprehensive literature search was conducted, some published and unpublished trials might be missed, which would lead to nonpublication bias. The chief investigators of all studies were contacted to provide additional methodological and statistical information; however, no response was obtained. A source of bias may also have arisen from the use of the poorly defined criteria for the outcome “hormonal response”. The small number of studies within comparison limits us to explore the impact on treatment effects; in the meantime, the lack of high quality trials prevents us from further investigating the heterogeneity of the studies. 4.4. Agreements and disagreements with other studies or reviews To our knowledge, this is the first systematic review to evaluate the effect of GH on hip fractures for the present. There was only one human study done in evaluating the effect of GH on fracture healing in 2007 [25]. In that study, the authors evaluated the effect of r-hGH on 368 patients with tibial fractures by radiographic evaluation and clinically assessed healing. And the author concluded that no significant enhancement of fracture healing was observed with GH in the overall groups of open and closed tibial fractures, whereas GH accelerated healing significantly in closed tibial fractures. In the future, well-designed RCTs with large sample size are warranted to confirm or refute the current evidence. 5. Conclusions GH may be effective in hip fractures in view of the low quality evidence currently. More carefully designed, double-blinded and placebocontrolled randomized trials with large numbers of participants which use GH in the treatment of hip fractures are required. Conflict of interest The authors have no conflicts of interest that are directly relevant to the content of this study.
Table 3 Clinical adverse events. Events
GH group (n = 87)a
Placebo group (n = 75)
Studies
Acute renal failure Pulmonary embolism Edema
1 3 18
0 0 17
Hypertension Urinary tract infection Constipation Total
1 14 9 46
0 8 10 35
Yeo et al. Yeo et al. Hedström et al., Van der Lely et al. Hedström et al. Van der Lely et al. Van der Lely et al.
a
Any dose of GH.
S. Yang et al. / Growth Hormone & IGF Research 22 (2012) 97–101
References [1] Y. Kato, Y. Murakami, M. Sohmiya, M. Nishiki, Regulation of human growth hormone secretion and its disorders, Internal Medicine-Tokyo-Japanese Society of Internal Medicine 41 (1) (2002) 7–13. [2] C. Ohlsson, B.A. Bengtsson, O.G.P. Isaksson, T.T. Andreassen, M.C. Slootweg, Growth hormone and bone, Endocr Rev. 19 (1) (1998) 55–79. [3] K. Chihara, T. Sugimoto, The action of GH/IGF-I/IGFBP in osteoblasts and osteoclasts, Horm Res. 48 (1997) 45–49. [4] H.J. Bail, S. Kolbeck, G. Krummrey, et al., Systemic application of growth hormone for enhancement of secondary and intramembranous fracture healing, Horm Res. 58 (2002) 39–42. [5] J.D. Zuckerman, A.D. Rosenberg, Hip fracture, New Engl J Med. 335 (26) (1996) 1995–1996. [6] R. Marks, Hip fracture epidemiological trends, outcomes, and risk factors, 1970–2009, Int J Gen Med. 3 (2010) 1–17. [7] Guide to research methods: the evidence pyramid available from http://library. downstate.edu/EBM2/2100.htm. [8] R. Harbour, J. Miller, A new system for grading recommendations in evidence based guidelines, BMJ. 323 (7308) (2001) 334. [9] Y.P. li, T.J. Tao, D. Sun, et al., A pilot study of the classification of Chinese Medical Specialty, Chinese Journal of Evidence-Based Medicine. 4 (3) (1994) 173–180. [10] J. P. T. Higgins, S. Green, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2 [Updated September 2009] The Cochrane Collaboration (2009) Available from www.cochranehand-book.org [11] J.A. Berlin, Does blinding of readers affect the results of meta-analyses? Lancet. 350 (9072) (1997) 185–186. [12] J.P.T. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat Med. 21 (11) (2002) 1539–1558. [13] T. Chevalley, P. Hoffmeyer, J.P. Bonjour, R. Rizzoli, Early serum IGF-I response to oral protein supplements in elderly women with a recent hip fracture, Clinical Nutrition. 29 (1) (2010) 78–83. [14] K. Wang, L. Zhu, H.D. Lu, D.Z. Cai, Effect of reconstructed growth hormone on hip arthroplasty of the aged patients, Orthopedic Journal of China. 13 (4) (2005) 273–275.
101
[15] M.A. Bach, K. Rockwood, C. Zetterberg, et al., The effects of MK-0677, an oral growth hormone secretagogue, in patients with hip fracture, J Am Geriatr Soc. 52 (4) (2004) 516–523. [16] M. Hedström, O. Ljungqvist, T. Cederholm, Metabolism and catabolism in hip fracture patients: nutritional and anabolic intervention — a review, Acta orthopaedica. 77 (5) (2006) 741–747. [17] A. Adunsky, J. Chandler, N. Heyden, et al., MK-0677 (ibutamoren mesylate) for the treatment of patients recovering from hip fracture: a multicenter, randomized, placebo-controlled phase IIb study, Archives of Gerontology and Geriatrics (2010). [18] M.A. Schurch, R. Rizzoli, D. Slosman, et al., Protein supplements increase serum insulin-like growth factor-I levels and attenuate proximal femur bone loss in patients with recent hip fracture, Annals of internal medicine. 128 (10) (1998) 801. [19] A.L. Yeo, D. Levy, F.C. Martin, et al., Frailty and the biochemical effects of recombinant human growth hormone in women after surgery for hip fracture, Growth Horm Igf Res. 13 (6) (2003) 361–370. [20] M. Hedström, M. Saaf, E. Brosjo, et al., Positive effects of short-term growth hormone treatment on bean body mass and BMC after a hip fracture — a double-blind placebo-controlled pilot study in 20 patients, Acta Orthop Scand. 75 (4) (2004) 394–401. [21] A.J. Van der Lely, S.W. Lamberts, K.W. Jauch, et al., Use of human GH in elderly patients with accidental hip fracture, Eur J Endocrinol. 143 (5) (2000) 585–592. [22] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, PLoS Med. 6 (7) (2009) e1000097. [23] T.X. Wu, G.J. Liu, J. Li, Some risks of affecting the quality of published systematic reviews in China, Chinese Journal of Evidence-Based Medicine. 5 (1) (2005) 51–58. [24] G.S. Keene, M.J. Parker, G.A. Pryor, Mortality and morbidity after hip fractures, BMJ. 307 (6914) (1993) 1248–1250. [25] M. Raschke, M.H. Rasmussen, S. Govender, et al., Effects of growth hormone in patients with tibial fracture: a randomised, double-blind, placebo-controlled clinical trial, Eur J Endocrinol. 156 (3) (2007) 341–351.