SEARCH

SEARCH BY CITATION

Keywords:

  • Shoulder fractures;
  • hemiarthroplasty;
  • internal fixation;
  • meta-analysis

Abstract

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References

Background The optimal management of complex proximal humeral fractures is debatable. The present study uses meta-analytical techniques to compare conservative and operative management of proximal humeral fractures with respect to morbidity and functional outcomes.

Methods Studies published between 1970 and 2007 comparing conservative and operative treatment of proximal humeral fractures were included. The end points evaluated were morbidity and functional outcomes. A random effects model was used and sensitivity analysis was performed to account for bias in patient selection.

Results Ten studies matched the selection criteria, reporting on 486 patients. Two hundred and nineteen (45%) were managed conservatively, 174 (36%) underwent operative fixation and 93 (19%) underwent hemiarthroplasty. Mean follow-up ranged from 6 months to 156 months. No significant difference was demonstrated in post-treatment Constant scores (weighted mean difference = 1.62, −7.12, 10.36), rate of avascular necrosis [odds ratio (OR) = 0.92, confidence interval (CI) = 0.37–2.30] ongoing pain or non-union (OR = 1.28, CI = 0.11–15.46) between the groups. These findings remained consistent when considering studies matched for three- and four-part fractures and fracture dislocations.

Conclusion On the basis of available comparative data, there is no demonstrable difference in outcomes between fractures managed surgically or conservatively. There is a need for large, randomised trials to guide management of these fractures.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References

Fractures of the proximal humerus are common and can be disabling, affecting predominantly an elderly, frail population. In the UK, they are estimated to comprise 5.7% of all fractures [1]; the proximal humerus is the third commonest site of fracture after the hip and the distal radius in elderly patients [2]. The incidence of proximal humeral fractures in increasing and is projected to increase yet further within the context of an ageing population [3]. Although minimally displaced fractures appear to be amenable to conservative treatment [4], the displaced type of fracture has a poor prognosis and the optimal treatment is still disputed.

Surgical stabilization of displaced proximal humeral fractures is challenging, particularly in poor quality, osteoporotic bone. Several forms of osteosynthesis have been employed, including intramedullary devices [5], plate and screw fixation including locking plates [6], external fixation [7] and techniques such as sutures [8] and tension banding [9]. Primary hemiarthroplasty is widely used in severely comminuted fractures and appears to provide a reasonable outcome in terms of pain at the cost of a decreased range of movement [10]. Series of conservative treatment suggest that reasonable functional results can be achieved with immobilization alone, without the morbidity associated with surgical management, even in more complex fracture types [11].

Decision-making with respect to the treatment of complex proximal humeral fractures is hampered by a lack of available evidence. Although series exist, there are very few comparative studies comparing conservative management with the various surgical options. The present study aims to provide a comprehensive and systematic review of the available comparative literature, together with meta-analysis of suitable studies.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References

Study selection

A Medline, Embase, Ovid and Cochrane database search was performed for studies between 1970 and 2007, which compared conservative with operative management for proximal humeral fractures. Mesh search headings used were: ‘proximal humeral fracture and operative’, ‘proximal humeral fracture and conservative’, ‘meta-analysis', ‘fracture humerus', ‘conservative versus operative’, ‘fracture dislocation humerus', ‘meta-analysis' and ‘comparative study’. The above terms, and their combinations were also searched as text-words, as were ‘proximal humerus fracture’, ‘avascular necrosis humerus' and ‘humeral head’. The ‘related articles' function was used to broaden the search, and all abstracts, studies and citations scanned were reviewed. References of the articles acquired were also searched by hand. Results were not restricted to the English language.

Data extraction

Three reviewers (T.G.N., A.D.L. and P.S.) independently extracted the following data from each study: first author, year of publication, study population characteristics, study design, inclusion and exclusion criteria, matching criteria and number of subjects operated on with each technique. There was 100% agreement between the three reviewers.

Inclusion criteria

To be included in the analysis, studies had to fulfil the following criteria: (i) comparison of conservative to operative approaches in patients treated for proximal humeral fractures; (ii) inclusion of at least one of the outcome measures; (iii) documentation of the specific operative technique and protocol for conservative management; and (iv) when two studies were reported by the same institution and/or authors, either the one of higher quality, or the most recent publication was included in the analysis, unless the study outcomes were mutually exclusive or measured at different time intervals.

Exclusion criteria

Studies were excluded if they failed to fulfil the inclusion criteria or there was considerable overlap between cohorts previously evaluated.

Outcomes of interest and definitions

The main analysis compared conservative versus operative management of all types of proximal humeral fractures including fracture dislocations. Fractures were classified according to Neer [12] and the following outcomes were used to compare all groups and subgroups of patients:

  • 1
    Adverse events including, the overall rates of avascular necrosis (AVN), constant/severe pain, non-union and pseudoarthrosis.
  • 2
    Functional outcomes: unsatisfactory/poor Neer scores and Constant scores.

The operative group includes patients treated with the techniques listed earlier. It was not possible to sub-classify patient outcomes according to operative procedure as a result of insufficient data. In addition, this group does not include patients who underwent arthroplasty because these were analyzed separately. Fracture dislocations were all anterior dislocations, except one, with associated comminuted fractures.

Subgroup analysis

Subgroup analyses were performed to isolate and examine the outcome of patients with more severe fractures. Three- and four-part fractures and fracture dislocations were examined separately comparing conservative versus operative treatment. Within the three- and four-part nondislocated fracture subgroup, we also compared conservative management versus arthroplasty and open reduction with internal fixation versus arthroplasty.

Statistical analysis

Meta-analysis was performed in line with recommendations from the Cochrane Collaboration and the Quality of Reporting of Meta-analyses (QUORUM) guidelines [13,14]. The effect measures estimated were odds ratio (OR) for dichotomous data and weighted mean difference (WMD) for continuous data [15], both reported with 95% confidence intervals (CI). This ratio represented the odds of an adverse event occurring in conservative group compared to the operative group. An OR was considered statistically significant at p < 0.05 or if the 95% CI did not include the value 1. Studies that contained a zero in one cell for the number of events of interest in one of any two groups compared resulted in problems with the computation of ratio measurement. The Haldane correction was therefore applied, whereby a value of 0.5 was added in both groups for that particular study [16].

For categorical variables, the ORs were combined with the Mantel–Haenszel chi-square method using a ‘random effects' meta-analytical technique [17]. In a random effects model, it is assumed that there is variation between studies and that the calculated odds ratio thus has a more conservative value [18,19]. In the Forest plot, the squares indicate point estimates of treatment effect (OR), with the size of the square representing the weight attributed to each study and 95% CIs indicated by horizontal bars. The diamond represents the summary OR from the pooled studies with 95% CIs. For studies that presented continuous data as mean and range values (e.g. Constant scores), the SD was calculated using statistical algorithms and checked using ‘bootstrap’ re-sampling techniques. Thus, all continuous data were standardized for the analysis. The quality of the non-randomised studies was assessed by using the Newcastle–Ottawa Scale with some modifications to match the needs of this study [20]. The quality of the studies was evaluated by examining three items: patient selection, comparability of the two study groups and assessment of outcome. Studies achieving five or more stars were considered as being of higher quality.

Two strategies were employed to quantitatively assess heterogeneity:

  • 1
    Graphical exploration with funnel plots to evaluate publication bias [15,21].
  • 2
    Sensitivity analysis for the following subgroups: (i) prospective studies including randomised controlled trials (RCTs) and (ii) RCTs alone. Analysis was conducted using Review Manager, version 4.2 (The Cochrane Collaboration, Oxford, UK).

RESULTS

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References

Eligible studies

Our search revealed 1422 studies published between 1970 and 2007, of which 22 matched the selection criteria and were retrieved in full. Twelve of these matched the inclusion criteria and 10 were finally analyzed [9,22–30]. Two of the included studies were RCTs [22,23], three were prospective non-randomised studies [24–26], one included both prospective and retrospective groups[26] and the rest were retrospective studies [9,28–30]. Two studies were excluded because they did not provide sufficient raw data for analysis [31,32].

The characteristics of the included studies are summarized in Table 1. Analysis was performed on 486 patients with proximal humeral fractures, 219 (45%) treated conservatively, 174 (36%) with internal fixation and 93 (19%) with replacement arthroplasty. Patients with three- and four-part nondislocated fractures were analyzed separately, 169/335 (50.4%) were treated conservatively and 166/336 (49.4%) were treated operatively.

Table 1.  Characteristics of included studies
Study (year)Study typeCons casesIF casesAP cases% FemaleMean overall age (years)Mean follow-up (months)Inclusion criteriaMatchingStudy quality (star rating)
  1. Inclusion criteria: 1, Displaced three- and four-part fractures; 2, nonpathological fractures; 3, at least 30% contact between the head and the humeral shaft; 4, no other fractures in the upper limbs; 5, no co-morbidity affecting the end result; 6, ability for patient to co-operate; 7, avulsion of at least one tuberosity; 8, fracture dislocations (anterior); 9, two-part fractures; 10, nondisplaced four-part fractures.

  2. Matching: 1, Age; 2, sex; 3, fracture pattern; 4, related injuries; 5, classification of fractures by the same surgical and radiological teams; 6, mechanism of injury; 7, follow-up; 8, right/left side or dominant limb; 9, same operative fixation/same surgeon.

  3. RCT, randomised controlled trial; PNR, prospective non-randomised; R, retrospective; Cons, conservative group; IF, internal fixation group; AP, arthroplasty.

Fjalestad (2005)PNR/R5515717512–1411,5,6,74
Ilchmann (1998)R161876.565.266.31,81,2,510
Kristiansen (1989)RCT16157169121,91,2,39
Kollig (2003)PNR91358.55279.219
Lim (1983)PNR9936662.2331,7,87
Neer (1970)R31434355.357.61,856
Schai (1995)R24402968.85949.21,84
Stableforth (1984)PNR161678.167.9611,2,5,8,910
Zyto (1997)RCT151486.274501,2,3,4,5,61,3,4,5,6,7,89
Zyto (1995)R28737136–43142

Three studies reported on fracture dislocations with 19/71 (26.8%) treated conservatively and 52/71 (73.2%) treated with internal fixation [9,25,30]. The conservative group patients underwent either open or closed reduction of their dislocation under anaesthetic; in the operative group, all reductions were performed at the time of internal fixation. Patients who underwent arthroplasty within this group were examined separately. Mean follow-up varied greatly, from 6 months to 156 months between studies.

Analysis of all types of proximal humeral fractures excluding dislocations

Adverse events. The rate of avascular necrosis and ongoing pain were observed to favour the conservative group, whereas the rate of non-union was lower in the operative fixation group. None of these results reached statistical significance. They are presented in Table 2.

Table 2.  Comparison of conservative versus internal fixation of nondislocated proximal humeral fractures
Outcome of interestNumber of studiesNumber of patientsOutcomes by group % (Cons, IF)OR/WMD*95% CIp-valueHeterogeneity χ2Heterogeneity p-value
  1. AVN, avascular necrosis; Cons, conservative group; IF, internal fixation group; OR, odds ratio; CI, confidence interval; WMD, weighted mean difference.

  2. *Constant score is continuous and therefore Weighted Mean Difference was used rather than Odds Ratio.

Adverse events
AVN935411.2, 13.70.800.31, 2.050.6414.060.08
Non-union417514.3, 10.02.040.74, 5.670.172.040.56
Pseudoarthrosis21015.63, 6.70.720.11, 4.620.730.420.52
Constant/severe pain3M28.6, 35.00.700.25, 1.950.491.580.45
Functional outcomes
Constant score31441.62*−7.12, 10.360.160.360.72
Unsatisfactory/poor Neer scores412551.3, 48.91.250.30, 5.270.767.630.05

Functional outcomes. In the reviewed literature shoulder function was assessed in a variety of ways, with the Neer and Constant scores being the most consistently used. Although there was a small trend favouring the operative group, our analysis did not reveal any statistical difference when function was assessed by either scoring system; for example, an unsatisfactory/poor Neer score was achieved by 24/49 (49%) patients treated operatively compared to 39/76 (51.3%) treated conservatively. There was insufficient data on the range of movement, although this would have been a desirable analysis. These results are presented fully in Table 2.

Subgroup analysis

Three- and four-part nondislocated fractures. When these more severe fractures were analyzed separately, similar trends to the overall analysis were observed (Table 3). The rates of AVN tended to favour the conservative group; this was not significantly significant (Fig. 1). By contrast to the main analysis, outcome scores tended to favour the operated group (Fig. 2).

Table 3.  Subgroup analysis of outcomes for nondislocated and dislocated three- and four-part fractures
Outcome of InterestNo. of studiesNo. of patientsOutcomes by group % (Cons, IF)OR/WMD*95% CIp-valueHeterogeneity χ2Heterogeneity p-value
  1. AVN, avascular necrosis; Cons, conservative group; IF, internal fixation group; OR, odds ratio; CI, confidence interval; WMD, weighted mean difference.

  2. *Constant score is continuous and therefore Weighted Mean Difference was used rather than Odds Ratio.

Analysis of prospective studies
All fractures excluding dislocations
AVN4879.5, 13.30.820.22, 3.070.301.890.60
Constant/severe pain25025.0, 26.90.920.16, 5.240.931.480.22
Constant score2517.04*−2.66, 16.740.150.250.61
Unsatisfactory/poor Neer scores23364.7, 31.33.910.90, 16.870.070.000.97
Nondislocated three- and four-part fractures
AVN4879.3, 13.60.780.21, 2.910.471.640.65
Constant score2517.04*−2.66, 16.740.150.250.61
Analysis of RCTs
All fractures excluding dislocations
AVN2537.7, 7.41.130.14, 9.420.911.100.29
Non-union26012.9, 6.91.500.10, 21.790.771.830.18
Nondislocated three- and four-part fractures
AVN2537.4, 7.71.030.14, 7.690.980.920.34
image

Figure 1. Forrest plots evaluating rates of avascular necrosis after conservative versus operative management of nondislocated three- and four-part fractures, including sensitivity analyses examining prospective studies and randomised controlled trials.

Download figure to PowerPoint

image

Figure 2. Forrest plots evaluating Constant scores for conservative versus operative management of non-dislocated 3 and 4 part fractures including sensitivity analysis of prospective studies

Download figure to PowerPoint

Three- and four-part fracture dislocations. Three studies reported data on fracture dislocations and the rate of AVN was the only outcome with sufficient data for analysis [9,25,30]. There was a lower rate of AVN after operative fixation (17.3% versus 26.3%), although this was not statistically significant.

Conservative and internal fixation versus arthroplasty for comminuted fractures. There was limited data to perform a comprehensive analysis comparing these two groups. Functionally, however, the arthroplasty group had fewer poor Neer scores (18.8%) than the internal fixation group (54.8%) and the conservative group (49.3%).

Sensitivity analysis

The sensitivity analysis (Table 4) included four prospective studies and two RCTs. It was performed on the main analysis and the three- and four-part nondislocated fracture group. Overall, this revealed similar results to the main analysis with reduced heterogeneity throughout.

Table 4.  Sensitivity analysis comparing conservative to operative management of proximal humeral fracture
Outcome of interestNumber of studiesNumber of patientsOutcomes by group % (Cons, Op)OR/WMD*95% CIp-valueHeterogeneity χ2Heterogeneity p-value
  1. AVN, avascular necrosis; Cons, conservative group; IF, internal fixation group; OR, odds ratio; CI, confidence interval; WMD, weighted mean difference.

  2. *Constant score is continuous and therefore Weighted Mean Difference was used rather than Odds Ratio.

Nondislocated three- and four-part fractures
Adverse events
Rate of AVN933511.2, 14.50.670.23, 1.990.4717.500.03
Functional outcomes
Unsatisfactory/poor Neer scores39847.5, 56.80.560.09, 3.520.544.670.10
Constant Score31441.62*−7.12, 10.360.160.360.72
Three- and four-part fracture dislocations
Rate of AVN37126.3, 17.31.270.32, 5.100.741.400.50

Prospective studies including RCTs. Analysing for all fractures the difference in rates of AVN and constant/severe pain remained nonsignificant, with the initial trend to favour the conservative group reduced and a significant reduction in statistical heterogeneity, OR = 0.82 (heterogeneity χ2 = 1.89), 0.92 (heterogeneity χ2 = 1.48), respectively (Table 4). In terms of function, there remained no significant difference in Neer and Constant scores between the two groups; however, the trend favoured the operative group. Heterogeneity was almost eliminated (heterogeneity χ2 = 0.00), 0.92 (heterogeneity χ2 = 0.25). Similar results were achieved for patients with three- and four-part nondislocated fractures, with no significant differences in outcomes for the conservative versus operative groups.

Randomised control trials. This produced similar results to the prospective study sensitivity analysis with reduced heterogeneity. The rate of non-union remained nonsignificant between the conservative and operative group for all fractures with reduced heterogeneity (p = 0.77, heterogeneity χ2 = 1.83).

Publication bias

Funnel plots of the studies used in this meta-analysis reporting on rate of avascular necrosis for three- and four-part fractures are shown in Fig. 3. These are scatter plots of the treatment effects estimated from individual studies plotted on the horizontal axis (OR) against the standard error of the estimate shown on the vertical axis SE(log OR). Most studies lie inside the 95% CIs, with a biased distribution around the vertical toward the left, indicating a possible effect of publication bias. When only higher quality prospective studies and RCTs were considered, all studies lay within the 95% CIs and were distributed more evenly about the vertical, showing no evidence of publication bias.

image

Figure 3. Funnel plots of AVN rates for conservative versus operative management of non-dislocated 3 and 4 part fractures before and after sensitivity analysis of prospective studies.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References

Evidence-based management of complex proximal humeral fractures is challenging as a result of limitations inherent in the available literature. In the present study, we have performed a wide-ranging review and analysis of the available comparative data in an attempt to provide an overview of the evidence for management of these fractutres.

Analysis of the totality of the data available demonstrates a high percentage of poor outcomes after reconstruction of three- and four-part proximal humeral fractures, with similar rates of AVN, pain and function achieved compared to conservative management. Restricting the analysis to better quality, prospective and RCTs showed no statistical difference in functional results, AVN or pain between the conservative and operative groups. Our subgroup analysis of three- and four-part nondislocated fractures showed similar rates of AVN and functional outcome. On the basis of the available evidence, we cannot demonstrate a significant difference in outcome between any subgroup of fractures treated conservatively or surgically.

It has been shown that a period of immobilization, without manipulation, followed by intensive graded physiotherapy, provides some excellent results in minimally displaced fractures and even multi-fragmented three- and four-part fractures [22,33]. However, joint restoration, with the additional aim of preventing articular surface collapse and avascular necrosis, in three- and four-part fractures may require open or closed reduction and fixation methods to be achieved. If the fracture position is maintained, good results can be expected [22]. It has been found that closed reduction with percutaneous stabilization may lead to good or excellent results in 70% of cases, although whether internal fixation provides the degree of stability required to allow primary bone healing and the opportunity for early mobilization has been questioned [27]. This in turn may affect functional outcome and overall postoperative satisfaction.

Hemiathroplasty has been argued to be the treatment of choice for four-part fractures or fracture dislocations [30], and although some studies have reported successful studies implementing this device [26], not all studies reflect Neer's successful results [34], despite prosthesis development by the advent of in-depth analysis of humeral osteometry, the glenohumeral joint relationship [35,36] and the application to prosthetic design [37]. Assessing the functional outcome of comminuted fractures, our findings reflected Neer's conclusions these injuries are best served with hemiarthroplasty compared to internal fixation, although there was no difference in outcome between the conservative and hemiarthroplasty groups.

Several studies have attempted to collate the available data regarding proximal humeral fractures. Misra et al. performed a systematic review of 24 eligible reports on the conservative and operative treatment outcomes of three- and four-part proximal humeral fractures [38]. Their results showed that, in comparative trials, there was no statistical difference in outcome scores between conservative and operative treatment, although hemiarthroplasty conferred greater functional of range of motion compared to both conservative and operative treatment. However, in the analysis of case studies, there was no statistical difference between conservative, fixation and arthroplasty groups. A Cochrane review was published in 2008 [39], which excluded all but two studies comparing surgical against conservative management. Neither this study, nor a broader study published the same year analysing data up to 2004 [40], could draw meaningful conclusions from the data available.

Three studies have been published examining the evidence for specific interventions. Two studies reviewed the data for hemiarthroplasty [10,41], finding good outcomes in terms of pain relief and patient satisfaction, but poor functional scores. One study reviewed the data available for the use of locking plates showing reasonable functional outcomes but at the cost of a high cut-out rate and need for reoperation in 13.7% of cases [42]. All three of these reviews examine mainly noncomparative studies.

All attempts at analysis of the available data are hampered by a number of problems. First, the classification of proximal humeral fractures is difficult. The most commonly applied Neer and AO classification systems achieve poor inter-observer and intraobserver reliability scores [43–46]. Some studies have used nonstandard classifications, making the data more heterogeneous, and description of fracture types within the literature is not always comprehensive.

The same lack of homogeneity extends to the use of postoperative functional scoring methods. Although Neer and Constant scores are the most commonly used, these scoring systems are imperfect and different studies have employed a variety of different systems. In the 10 studies included in this analysis, seven different outcome scores were used.

In the present study, we have attempted to analyse all the studies in the current literature comparing conservative to surgical management of proximal humeral fractures. Our attempt to delineate an evidence-based approach to the treatment of these fractures was hampered by the paucity of good quality comparative data. Our literature search returned only two RCTs with a total of 60 patients, and our inclusion criteria have been broadened to include prospective and retrospective non-randomised studies. Although the inclusion of these studies increases the scope of research available, it also raises the spectre of selection bias. The need for further trials in this controversial area has been stated previously [39]. Encouragingly, protocols for two large RCTs have been published within the last year, raising hopes of a better evidence-base, albeit in some years time [47,48].

Further problems with the data included the use of disparate systems for fracture classification and functional scoring, as well as wide variation in study size and length of follow-up. Given the rapid development of fracture fixation devices, some of the studies examine fixation methods and prostheses that are no longer frequently used; similarly, at the time of analysis, comparative studies examining the newer designs of locking plate were unavailable. However, the present study aimed to provide a comprehensive and systematic review of the available comparative data, and this study may be taken alongside studies such as the Cochrane review, which reviews the data available with narrower inclusion criteria.

In conclusion, on the basis of the available literature, the present analysis is unable to demonstrate any statistically significant difference in outcome for complex proximal humeral fractures managed conservatively or surgically. There is a pressing need for large RCTs aiming to inform surgical practice in this area.

References

  1. Top of page
  2. Abstract
  3. Conflicts of interest
  4. INTRODUCTION
  5. MATERIALS AND METHODS
  6. RESULTS
  7. DISCUSSION
  8. References
  • 1
    Court-Brown CM, Caesar B. Epidemiology of fractures: a review. Injury 2006; 37:6917.
  • 2
    Lauritzen JB, Schwarz P, Lund B, et al. Changing incidence and residual lifetime risk of common osteoporosis-related fractures. Osteoporos Int 1993; 3:12732.
  • 3
    Kannus P, Palvanen M, Niemi S, et al. Rate of proximal humerus fractures in older Finnish women between 1970 and 2007. Bone 2009; 44:6569.
  • 4
    Bahrs C, Rolauffs B, Dietz K, et al. Clinical and radiological evaluation of minimally displaced proximal humeral fractures. Arch Orthop Trauma Surg 2010; 130:6739.
  • 5
    Popescu D, Fernandez-Vallencia JA, Rios M, et al. Internal fixation of proximal humerus fractures using the T2-proximal humeral nail. Arch Orthop Trauma Surg 2009; 129:123944.
  • 6
    Papadopoulos P, Karataglis D, Stavridis SI, et al. Mid-term results of internal fixation of proximal humeral fractures with the Philos plate. Injury 2009; 40:12926.
  • 7
    Monga P, Verma R, Sharma V. Closed reduction and external fixation for displaced proximal humeral fractures. J Orthop Surg 2009; 17:1425.
  • 8
    Dimakopoulos P, Panagopoulos A, Kasimatis G. Transosseous suture fixation of proximal humeral fractures. J Bone Joint Surg Am 2007; 89:17009.
  • 9
    Ilchmann T, Ochsner P, Wingstrand H, et al. Non-operative treatment versus tension-band osteosynthesis in three- and four-part proximal humeral fractures. A retrospective study of 34 fractures from two different trauma centers. Int Orthop 1998; 22:31620.
  • 10
    Kontakis G, Koutras C, Tosounidis T, et al. Early management of proximal humeral fractures with hemiarthroplasty: a systematic review. J Bone Joint Surg Br 2008; 90:140713.
  • 11
    Hanson B, Neidenbach P, De Boer P, et al. Functional outcomes after nonoperative management of fractures of the proximal humerus. J Shoulder Elbow Surg 2009; 18:61221.
  • 12
    Neer CS II. Displaced proximal humeral fractures. I. Classification and evaluation. J Bone Joint Surg Am 1970; 52:107789.
  • 13
    Clarke M, Horton R. Bringing it all together: Lancet-Cochrane collaborate on systematic reviews. Lancet 2001; 357:1728.
  • 14
    Stroup DF, Berlin J, Morton S, et al. Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group. Meta-analysis of observational studies in epidemiology: a proposal for reporting. J Am Med Assoc 2000; 283:200812.
  • 15
    Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Br Med J 1997; 315:629634.
  • 16
    Haldane JB. The estimation and significance of the logarithm of a ratio of frequencies. Ann Hum Genet 1956; 20:30911.
  • 17
    Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22:71948.
  • 18
    DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7:17788.
  • 19
    Sterne JA, Davey Smith G. Sifting the evidence-what's wrong with significance tests? Br Med J 2001; 322:22631.
  • 20
    Athanasiou T, Al-Ruzzeh S, Kumar P, et al. Off-pump myocardial revascularization is associated with less incidence of stroke in elderly patients. Ann Thorac Surg 2004; 77:74553.
  • 21
    Egger M, Smith GD. Misleading meta-analysis. Br Med J 1995; 311: 7534.
  • 22
    Zyto K, Ahrengart L, Sperber A, et al. Treatment of displaced proximal humeral fractures in elderly patients. J Bone Joint Surg Br 1997; 79:4127.
  • 23
    Kristiansen B, Kofoed H. Transcutaneous reduction and external fixation of displaced fractures of the proximal humerus. A controlled clinical trial. J Bone and Joint Surg Br 1989; 70:8214.
  • 24
    Kollig E, Kutscha-Lissberg F, Roetman B, et al. Complex fractures of the humeral head: which long-term results can be expected. Zentralbl Chir 2003; 128:1118.
  • 25
    Lim T, Ochsner PE, Marti RK, et al. The results of treatment of comminuted fractures and fracture dislocations of the proximal humerus. Neth J Surg 1983; 35:13943.
  • 26
    Stableforth PG. Four-part fractures of the neck of the humerus. J Bone Joint Surg Br 1984; 66:1048.
  • 27
    Fjalestad T, Strømsøe K, Blücher J, et al. Fractures in the proximal humerus: functional outcome and evaluation of 70 patients treated in hospital. Arch Orthop Trauma Surg 2005; 125:3106.
  • 28
    Zyto K, Kronberg M, Brostrom L. Shoulder function after displaced fractures of the proximal humerus. J Shoulder Elbow Surg 1995; 4:3316.
  • 29
    Schai P, Imhoff A, Preiss S. Comminuted humeral head fractures: a multicenter analysis. J Shoulder Elbow Surg 1995; 4:31930.
  • 30
    Neer CS II. Displaced proximal humeral fractures, part II. Treatment of three-part and four-part displacement. J Bone Joint Surg Am 1970; 52:1090103.
  • 31
    Yeap JS, Noor Zehan AR, Ezlan S, et al. Functional outcome of proximal humeral fractures. Med J Malaysia 2001; 56 (Suppl C):138.
  • 32
    Olsson C, Nordquist A, Petersson CJ. Long-term outcome of a proximal humerus fracture predicted after 1 year: a 13-year prospective population-based follow-up study of 47 patients. Acta Orthop 2005; 76:397402.
  • 33
    Young T, Wallace W. Conservative treatment of fractures and fracture-dislocations of the upper end of the humerus. J Bone Joint Surg Br 1985; 67:3737.
  • 34
    Sturzenegger M, Fornaro E, Jakob R. Results of surgical treatment of multifragmented fractures of the humeral head. Arch Orthop Trauma Surg 1982; 100:24959.
  • 35
    Edelson G. Variations in the retroversion of the humeral head. J Shoulder Elbow Surg 1999; 8:1425.
  • 36
    Iannotti JP, Gabriel JP, Schneck SL, Evans BG, Misra S. The normal glenohumeral relationships. An anatomical study of one hundred and forty shoulders. J Bone Joint Surg Am 1992; 74:491500.
  • 37
    Boileau P, Walch G. Three-dimensional geometry of the proximal humerus: implications for surgical technique and prosthetic design. J Bone Joint Surg Br 1997; 79:85765.
  • 38
    Misra A, Kapur R, Maffulli N. Complex proximal humeral fractures in adults—a systematic review of management. Injury 2001; 32: 36372.
  • 39
    Handoll HH, Gibson JN, Madhok R. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev 2003; 4: CD000434.
  • 40
    Lanting B, MacDermid J, Drosdowech D. Proximal humeral fractures: a systematic review of treatment modalities. J Shoulder Elbow Surg 2008; 17:4254.
  • 41
    Nijs S, Broos P. Outcome of shoulder hemiarthroplasty in acute proximal humeral fractures: a frustrating meta-analysis experience. Acta Orthop Belg 2009; 75:44551.
  • 42
    Thanasas C, Kontakis G, Angoules A. Treatment of proximal humerus fractures with locking plates: a systematic review. J Shoulder Elbow Surg 2009; 18:83744.
  • 43
    Kristiansen B, Andersen UL, Olsen CA, et al. The Neer classification of fractures of the proximal humerus: an assessment of interobserver variation. Skeletal Radiol 1988; 17:4202.
  • 44
    Bernstein J, Adler LM, Blank JE, et al. Evaluation of the Neer system of proximal humeral fractures with computer tomographic scans and plain radiographs. J Bone Joint Surg Am 1996; 78:13715.
  • 45
    Siebenrock K, Gerber C. The reproducibility of classification of fractures of the proximal end of the humerus. J Bone Joint Surg Am 1993; 75:17515.
  • 46
    Bernstein J. Fracture classifications: do they work and are they useful? J Bone Joint Surg Am 1993; 75:17434.
  • 47
    Handoll H, Brealey S, Rangan A, et al. Protocol for the ProFHER (PROximal Fracture of the Humerus: Evaluation by Randomisation) trial: a pragmatic multi-centre randomised controlled trial of surgical versus non-surgical treatment for proximal fracture of the humerus in adults. BMC Musculoskelet Disord 2009; 10:140.
  • 48
    Brorson S, Olsen BS, Frich LH. Effect of osteosynthesis, primary hemiarthroplasty, and non-surgical management for displaced four-part fractures of the proximal humerus in elderly: a multi-centre, randomised clinical trial. Trials 2009; 10:51.