Orthopedics

Feature Article 

Operative Versus Nonoperative Treatment in Complex Proximal Humeral Fractures

Zhi Mao, MD; Lihai Zhang, MD, PhD; Licheng Zhang, MD, PhD; Xiantao Zeng, MD; Shuo Chen, MS; Daohong Liu, MD; Zhirui Zhou, MD; Peifu Tang, MD, PhD

Abstract

This updated meta-analysis investigated whether operative treatment is superior to nonoperative treatment in complex proximal humeral fractures. The authors searched the Cochrane Central Register of Controlled Trials, PubMed, and EMBASE. Randomized controlled trials that evaluated operative vs nonoperative treatment for exclusively 3- or 4-part proximal humeral fractures were considered. Six studies with a total of 287 patients who had proximal humeral fractures were included. According to the meta-analysis, no statistically significant differences were found between operative and nonoperative treatment in Constant-Murley shoulder scores (Constant scores); Disabilities of the Arm, Shoulder, and Hand scores; total complication events; mortality; infection; nonunion; avascular necrosis; osteoarthritis; redisplacement of fractures; or dislocation or resorption of tuberosity. For health-related quality of life, EuroQol-5D (EQ-5D) favored operative treatment, but 15D scores showed no significant difference. Compared with nonoperative treatment, open reduction and internal fixation required significantly more additional surgeries (risk ratio, 6.50; 95% confidence interval, 1.54–27.50; P=.01), and more penetrations into joint space occurred (risk ratio, 9.56; 95% confidence interval, 2.27–40.13; P=.002). The limited evidence suggests that no convincing findings support the use of either open reduction and internal fixation or hemiarthroplasty for the treatment of complex proximal humeral fractures. The findings of the current study should be interpreted cautiously because of the modest sample size and the short follow-up period.

The authors are from the Department of Orthopaedics (ZM, Lihai Zhang, Licheng Zhang, DL, PT) and the Department of Medical Information (SC), Chinese PLA General Hospital, Beijing; the Evidence-Based Medicine Center (XZ), Taihe Hospital, Hubei University of Medicine, Shiyan; and the Department of Radiation Oncology (ZZ), Cancer Hospital of Guangxi Medical University, Cancer Institute of Guangxi Zhuang Autonomous Region, Nanning, China.

Drs Mao and Lihai Zhang contributed equally to this work and should be considered as equal first authors.

The authors have no relevant financial relationships to disclose.

The authors thank Wanjie Gu, MD, Department of Anaesthesiology, First Affiliated Hospital, Guangxi Medical University; Jun Xia, Cochrane Schizophrenia Group, University of Nottingham, Nottingham, United Kingdom; and Chundng Zhang, MD, the Fourth Affiliated Hospital of China Medical University, for valuable discussions related to this study.

Correspondence should be addressed to: Peifu Tang, MD, PhD, Department of Orthopaedics, Chinese PLA General Hospital, No 28 Fuxing Rd, Beijing 100853, China ( pftang301@sina.com).

Received: May 02, 2013
Accepted: November 08, 2013

Abstract

This updated meta-analysis investigated whether operative treatment is superior to nonoperative treatment in complex proximal humeral fractures. The authors searched the Cochrane Central Register of Controlled Trials, PubMed, and EMBASE. Randomized controlled trials that evaluated operative vs nonoperative treatment for exclusively 3- or 4-part proximal humeral fractures were considered. Six studies with a total of 287 patients who had proximal humeral fractures were included. According to the meta-analysis, no statistically significant differences were found between operative and nonoperative treatment in Constant-Murley shoulder scores (Constant scores); Disabilities of the Arm, Shoulder, and Hand scores; total complication events; mortality; infection; nonunion; avascular necrosis; osteoarthritis; redisplacement of fractures; or dislocation or resorption of tuberosity. For health-related quality of life, EuroQol-5D (EQ-5D) favored operative treatment, but 15D scores showed no significant difference. Compared with nonoperative treatment, open reduction and internal fixation required significantly more additional surgeries (risk ratio, 6.50; 95% confidence interval, 1.54–27.50; P=.01), and more penetrations into joint space occurred (risk ratio, 9.56; 95% confidence interval, 2.27–40.13; P=.002). The limited evidence suggests that no convincing findings support the use of either open reduction and internal fixation or hemiarthroplasty for the treatment of complex proximal humeral fractures. The findings of the current study should be interpreted cautiously because of the modest sample size and the short follow-up period.

The authors are from the Department of Orthopaedics (ZM, Lihai Zhang, Licheng Zhang, DL, PT) and the Department of Medical Information (SC), Chinese PLA General Hospital, Beijing; the Evidence-Based Medicine Center (XZ), Taihe Hospital, Hubei University of Medicine, Shiyan; and the Department of Radiation Oncology (ZZ), Cancer Hospital of Guangxi Medical University, Cancer Institute of Guangxi Zhuang Autonomous Region, Nanning, China.

Drs Mao and Lihai Zhang contributed equally to this work and should be considered as equal first authors.

The authors have no relevant financial relationships to disclose.

The authors thank Wanjie Gu, MD, Department of Anaesthesiology, First Affiliated Hospital, Guangxi Medical University; Jun Xia, Cochrane Schizophrenia Group, University of Nottingham, Nottingham, United Kingdom; and Chundng Zhang, MD, the Fourth Affiliated Hospital of China Medical University, for valuable discussions related to this study.

Correspondence should be addressed to: Peifu Tang, MD, PhD, Department of Orthopaedics, Chinese PLA General Hospital, No 28 Fuxing Rd, Beijing 100853, China ( pftang301@sina.com).

Received: May 02, 2013
Accepted: November 08, 2013

Proximal humeral fracture is a common skeletal injury that comprises 4% to 9% of all fractures of the human body.1–5 It is the 3rd most common fracture in patients aged more than 65 years, after hip and wrist fractures. The annual incidence of these fractures is approximately 66 per 10,000 persons, and the incidence is increasing by 15% every year.6–8 Fragile bones and a risk of falling are considered the 2 predominant risk factors.8

Most proximal humeral fractures are not displaced and can be treated nonoperatively. However, nonoperative treatment has not shown consistently satisfactory results,4,9 especially with complex fractures, such as displaced 3- and 4-part proximal humeral fractures classified by Neer,10 which account for approximately 12.6% of all proximal humeral fractures.11 In addition to nonoperative treatment, a broad spectrum of operative treatments are available, including open reduction and internal fixation (ORIF) with a variety of devices, hemiarthroplasty, and percutaneous fixation.5 Technical advances in locking plate fixation have led to an increased interest in ORIF in patients with osteoporosis.4,12,13 According to a recent study, the rates of operative treatment and ORIF increased by 25.6% and 28.5%, respectively, from 1999 to 2005 in the United States; the rates of repeat surgery also increased significantly.14

Several earlier systematic reviews summarized the results of treatment for 3- and 4-part proximal humeral fractures.15,16 However, they did not improve evidence-based decision making because of lack of randomized controlled trials.15 Recently, several randomized controlled trials have investigated whether operative treatment may provide greater benefits than nonoperative treatment.17–22 These studies had modest sample sizes and yielded conflicting results. Optimal treatment of these fractures remains controversial, and functional outcomes can be disappointing.5,21,23 Therefore, the authors updated previous systematic reviews and performed a meta-analysis of randomized controlled trials to evaluate the efficacy and safety of operative and nonoperative treatment for complex proximal humeral fractures.

Materials and Methods

Eligibility Criteria

Studies were considered acceptable for inclusion in the systematic review if they met the following criteria: (1) displaced 3- and 4-part proximal humeral fractures15; (2) operative treatment vs nonoperative treatment; (3) at least 1 of the following outcomes reported: functional scores, such as the Constant-Murley shoulder score (Constant score), Disabilities of the Arm, Shoulder, and Hand (DASH) score, American Shoulder and Elbow Surgeons (ASES) score, health-related quality of life (HRQoL) score, range of motion (ROM), or complications; and (4) randomized controlled trial study design.

Studies were excluded if they (1) were abstracts, letters, or meeting proceedings; (2) had repeated data; or (3) enrolled patients with pathologically or metabolically induced fractures or open fractures.

Search Strategy

In accordance with guidelines from the Cochrane Handbook for Systematic Reviews of Interventions,24 the authors conducted a computer search of the Co-chrane Central Register of Controlled Trials (CENTRAL, March 2013), PubMed (1950 to March 2013), and EMBASE (1980 to March 2013). The following key words were used: “shoulder fractures,” “proximal humerus fractures,” “orthopedic procedures,” “fracture fixation,” and “arthroplasty.” No language restriction was applied. In addition, the reference lists of the included studies were manually searched for potentially eligible studies.

Data Extraction and Quality Assessment

Two reviewers independently extracted relevant data from the included studies, and discrepancies were resolved by a 3rd reviewer. Data included demographic information, study design, interventions, allocation concealment process, blinding, time to last follow-up, rate of follow-up, intention to treat, functional scores, HRQoL score, ROM, and complications. Some corresponding authors were contacted by e-mail to obtain missing information.

The risk of bias of each eligible study was independently assessed by 2 authors, in accordance with the Cochrane risk of bias tool,24,25 which defines 9 aspects: (1) random sequence generation (selection bias); (2) allocation concealment (selection bias); (3) blinding of participants (performance bias); (4) blinding of treatment providers (performance bias); (5) blinding of outcome assessors (detection bias); (6) intention to treat (attrition bias); (7) selective reporting (reporting bias); (8) comparable study groups; and (9) other bias. A qualification of risk of bias, including low risk, unclear risk, or high risk, was provided. The final qualification for each study was determined by consensus among 3 authors.

Statistical Analysis

Statistical analysis was performed with Review Manager 5.1 software (Cochrane Collaboration, Oxford, United Kingdom) for outcome measures. P<.05 was considered statistically significant. Heterogeneity was evaluated by visual inspection of the forest plot along with consideration of the test for heterogeneity and the I2 statistic.26 An I2 value of greater than 50% was considered substantial heterogeneity. A fixed effects model was used in the meta-analysis unless there was significant heterogeneity among studies. The random effects model of DerSimonian and Laird27 was used, regardless of heterogeneity. Continuous variables were presented as mean difference (MD) and dichotomous variables were presented as risk ratio (RR), both with a 95% confidence interval (CI). When the RR of outcomes was statistically significant, the authors calculated and reported numbers needed to treat or numbers needed to harm (NNH). To define sources of heterogeneity, subgroup analyses based on ORIF vs hemiarthroplasty were defined during the analysis design phase. Publication bias was tested by funnel plots when possible.

Results

Study Selection and Characteristics

The initial search identified 422 references. After duplicate references were removed and the titles, abstracts, and contents of the full text were examined, 6 randomized controlled trials were included in the meta-analysis. Figure 1 shows the process of literature selection. Table 1 shows the characteristics and demographic data of each included study. A total of 287 patients with complex humeral fractures were included in this study. The 6 randomized controlled trials were performed in the Netherlands,20 Norway,18,28 Sweden,17,19,21 and the United Kingdom22 during the period between 1984 and 2012. The total number of patients in each study ranged from 32 to 60. Mean age ranged from 65.5 to 79.9 years. The percentage of female patients in the study populations ranged from 75% to 96%. The studies followed patients for periods of 6 to 60 months, and the rate of patient follow-up ranged from 72.5% to 98%. Two studies17,18 included displaced 3- or 4-part fractures. Three studies20–22 included only displaced proximal humeral 4-part fractures, and the remaining study19 included only displaced 3-part fractures. The studies investigated the following implants: Global 1 FX shoulder fracture endoprosthesis (DePuy, Leeds, United Kingdom)20; angular stable locking plates (nonspecific locking compression plates of the AO basic type; Synthes, Bettlach/Solothurn, Switzerland)18; locking plates (Philos plates)19; Global Fx prostheses (DePuy, Sollentuna, Sweden)21; Neer prostheses, uncemented22; and cerclage wiring or surgical tension band.17

Flow chart summarizing the selection process of randomized controlled trials (RCTs). Abbreviation: CENTRAL, Cochrane Central Register of Controlled Trials.

Figure 1:

Flow chart summarizing the selection process of randomized controlled trials (RCTs). Abbreviation: CENTRAL, Cochrane Central Register of Controlled Trials.

Characteristics of Included Studies

Table 1:

Characteristics of Included Studies

Study Quality

Figure 2 shows the methodologic quality of the randomized controlled trials as independently assessed by 2 authors. All of the studies were single-center studies. Four of the studies18–21 were judged as having used sufficient allocation concealment. One study17 used closed envelopes without reporting adequate safeguards. Only 1 study19 was reported to have blinded the outcome assessors. For incomplete outcome data, 1 included study did not have adequate follow-up.22 Four studies18–21 reported a proper intention-to-treat analysis and clearly stated interventions. The comparability of baseline characteristics was generally acceptable.

Risk of bias summary. Abbreviations: +, low risk; ?, unclear risk; −, high risk; HA, hemiarthroplasty; ORIF, open reduction and internal fixation.

Figure 2:

Risk of bias summary. Abbreviations: +, low risk; ?, unclear risk; −, high risk; HA, hemiarthroplasty; ORIF, open reduction and internal fixation.

Outcomes

There were no statistically significant differences in Constant scores, DASH scores, or ASES scores between operative and nonoperative treatment at either 12 or 24 months. For HRQoL, operative treatment yielded significantly better Euro-Qol-5D (EQ-5D) scores at both 12 and 24 months of follow-up compared with non-operative treatment. However, no statistically significant difference was detected for 15D (Table 2).

Functional Status and Health-related Quality of Life

Table 2:

Functional Status and Health-related Quality of Life

Meta-analysis was considered inappropriate for ROM because of the obvious clinical heterogeneity and considerable variations in measurement of results. ROM did not differ significantly between operative and nonoperative treatment in the randomized controlled trials that examined this outcome, except that in 1 study with a small sample size22 from 1984 operative treatment with hemiarthroplasty was favored. Table 3 shows the results for ROM.

Range of Motion

Table 3:

Range of Motion

The 6 randomized controlled trials17–22 included in this meta-analysis included a total of 287 fractures, and they all provided information on total complication events after at least 1 year of follow-up. Meta-analysis of total complication events showed no significant difference between operative and nonoperative treatment (RR, 1.81; 95% CI, 0.87–3.76; P=.11). Because of the significant heterogeneity (P<.00001; I2=85%), subgroup analysis of ORIF vs hemiarthroplasty was performed. Subgroup analysis did not show specific sources of heterogeneity when divided by type of intervention. Therefore, the random effects model was used (Figure 3).

Forest plot for total complication rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Figure 3:

Forest plot for total complication rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Meta-analysis of the rate of additional surgery, including 247 patients from 5 randomized controlled trials,18–22 showed a significant difference in favor of non-operative treatment (RR, 4.14; 95% CI, 1.52–11.27; P=.005; I2=0%). Given the additional surgery rate of 3.2% with non-operative treatment, 9 fractures would have to be treated operatively to sustain 1 additional surgical event (NNH, 8.7; 95% CI, 5.4-22.6). Furthermore, subgroup analysis according to the type of surgery showed that only the ORIF subgroup was significantly worse compared with non-operative treatment (RR, 6.50; 95% CI, 1.54–27.50; P=.01; I2=0%). The NNH was 5 (95% CI, 3.10–12.94). The hemiarthroplasty subgroup did not differ significantly from nonoperative treatment (RR, 2.24; 95% CI, 0.52–9.60; P=.28; I2=0%; Figure 4).

Forest plot for additional surgery rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Figure 4:

Forest plot for additional surgery rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Four studies17–19,21 that included 205 fractures provided data on the incidence of penetration of the implant into joint spaces. The rate of penetration events was significantly higher in the operative group compared with the nonoperative group (RR, 9.56; 95% CI, 2.27–40.13; P=.002; heterogeneity test, P=.74, I2=0%). Given the penetration rate of 0% with nonoperative treatment, 6 fractures would have to be treated operatively to sustain 1 additional penetration event (NNH, 6; 95% CI, 4.18–10.6; Figure 5). All of the penetration events occurred in patients who underwent ORIF; 1 patient received a locking plate after randomization to the hemiarthroplasty group.21

Forest plot for penetration of implant into joint rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Figure 5:

Forest plot for penetration of implant into joint rate between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Five randomized controlled trials17–21 that reported a total of 255 fractures provided information on the rates of avascular necrosis. Meta-analysis of the rates of avascular necrosis found no significant difference between operative and nonoperative treatment (RR, 0.65; 95% CI, 0.36–1.19; P=.16; heterogeneity test, P=.12; I2=58.6%). The authors then performed subgroup analysis of ORIF vs shoulder hemiarthroplasty. Although there was a trend favoring hemiarthroplasty over nonoperative treatment, no statistically significant difference was detected (RR, 0.17; 95% CI, 0.02–1.38; P=.1; I2=0%). Furthermore, ORIF did not significantly differ from nonoperative treatment (RR, 0.74; 95% CI, 0.39–1.37; P=.37; I2=0%; Figure 6).

Forest plot for avascular necrosis between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

Figure 6:

Forest plot for avascular necrosis between operative group and nonoperative group. Abbreviations: CI, confidence interval; HA, hemiarthroplasty; M-H, Mantel-Haenszel; NO, nonoperation; ORIF, open reduction and internal fixation.

No statistically significant difference was seen in the rates of mortality, infection, nonunion, osteoarthritis, redisplacement of fracture, or dislocation or resorption of tuberosity between operative and nonoperative treatment (Table 4).

Other Complications Reported

Table 4:

Other Complications Reported

Discussion

The authors believe that this report is the first systematic review and meta-analysis of randomized controlled trials to compare operative and nonoperative treatment in complex proximal humeral fractures. Meta-analysis of randomized controlled trials is generally considered to provide the highest level of evidence for clinical interventions.29 Therefore, this report provides an update of previous systematic reviews focusing on 3- and 4-part fractures, which were limited because of lack of randomized controlled trials.15,16 Meta-analysis yielded the following findings: (1) There were no statistically significant differences in Constant scores, DASH scores, ASES scores, or the rates of total complications, mortality, infection, nonunion, avascular necrosis, osteoarthritis, redisplacement of fracture, and dislocation or resorption of tuberosity between operative and nonoperative treatment. (2) Regarding HRQoL, operative treatment yielded significantly better EQ-5D scores compared with nonoperative treatment, but no statistically significant difference was seen for 15D. (3) For ROM, only 1 study from 1984 with a small sample size favored operation with hemiarthroplasty; the other studies showed no statistically significant difference. (4) ORIF was associated with significantly more additional surgery and more penetration of implants into joint space compared with nonoperative treatment.

A previous systematic review selected only studies dealing exclusively with 3- and 4-part fractures of the proximal humerus and found that nonoperative treatment had a poorer ROM than operative treatment.15 However, the authors indicated that their data from former published literature were inadequate for evidence-based decision making because of the lack of randomized controlled trials.15 Another systematic review including 3- and 4-part fractures of the proximal humerus also compared hemiarthroplasty and nonoperative treatment.16 The authors found that the mean Constant sore in the nonoperative group was higher than in the hemiarthroplasty group,16 and they only included 1 randomized controlled trial.22 As an update, the current meta-analysis included 6 randomized controlled trials focusing on complex proximal humeral fractures.

In this meta-analysis, participants were not restricted to the elderly, but patients selected for trials were generally older than 52 years. Additionally, many surgical options were available, and they were separated into 2 subgroups: ORIF and hemiarthroplasty. Over the past decade, locking plate fixation to treat proximal humeral fractures has gained considerable popularity.14,30–33 In this meta-analysis, 2 recent randomized controlled trials18,19 used surgical treatment with locking plates.

The flaws of methodologic quality in the included randomized controlled trials were mainly in the area of blinding. However, this risk of bias was not considered critical because of the difficulty of performing blinding in surgical research. Inadequate allocation concealment leads to overestimation of treatment, whereas adequate allocation concealment can avoid selection bias.34 Fortunately, in this meta-analysis, 4 high-quality randomized controlled trials18–21 reported adequate allocation concealment, and they were published in the 2010s. This suggests an increasing awareness of the importance of high-quality randomized controlled trial design.

Regarding functional status, analyses of Constant scores, DASH scores, and ASES scores showed no statistically significant differences. These results contradict a previous systematic review.16 However, the current meta-analysis differs from the previously published review because it is based on randomized controlled trials and did not detect any significant heterogeneity (I2=0).

In terms of HRQoL, the results of meta-analysis of EQ-5D19,21 showed that operative treatment led to better scores than nonoperative treatment. The latest diagnostic study35 reported that EQ-5D had good internal and external responsiveness in patients with proximal humeral fractures. Additionally, other recent studies32,36,37 of proximal fractures reported EQ-5D scores for HRQoL. However, other studies have suggested that 15D has better responsiveness than EQ-5D because 15D allows assessors to detect small changes.38 One randomized controlled trial that was included reported that the 15D data showed no differences between operative and nonoperative treatment.28 Therefore, it is impossible to draw a clear conclusion about optimal treatment to yield a better HRQoL.

Regarding ROM, a meta-analysis could not be performed because of the variety of types of data. As a qualitative review, no clinical conclusions can be reached that are in agreement with previous systematic reviews.5,15,25,30,39 Another systematic review of hemiarthroplasty reported obvious functional limitations and ROM restrictions at 3.7 years of follow-up.40 However, the follow-up of most randomized controlled trials in the current study was less than 2 years. Thus, the results are inadequate for interpreting long-term ROM.

A recent systematic review of locking plates that included proximal humeral intra-articular fractures showed marked complication and additional surgery rates31; this finding was also supported by the previous 2 systematic reviews.30,39 A higher additional surgery rate was con-firmed in patients who underwent ORIF compared with those who received non-operative treatment (23.6% vs 3.2%). This rate was also higher than in the previous systematic review (23.6% vs 14%).30

No statistically significant differences were detected in rates of total complications, mortality, avascular necrosis, infection, nonunion, osteoarthritis, redisplacement of fracture, or dislocation or resorption of tuberosity. Infection was not clearly more prevalent in the surgical group. There were 2 possible explanations. One was that the borderline result (P=.07)29 may be caused by the limited study power, and the key weakness was the small sample size (132 patients). Therefore, this result cannot be used as strong evidence. The other explanation was that the 2 randomized controlled trials that reported infection were performed before 2000,17,22 whereas the 3 randomized controlled trials18,20,21 that reported no infection were all published after 2010, which may suggest advances in surgical technique, operative condition, and perioperative management to prevent infection.

Clear heterogeneity was shown for total complications, the total complications subgroup (ORIF vs NO), osteoarthritis, and dislocation or resorption of tuberosity. This heterogeneity probably originated from conflicting data among trials. It is clear that there was clinical bias as well because locking plate fixation and tension bands were combined into the ORIF group. Subgroup analysis within the ORIF group was not conducted because of the small number of randomized controlled trials.

This analysis has certain limitations. First, only 6 randomized controlled trials, with a total of 287 fractures, could be included. Obviously, the sample size was modest. Publication bias analysis could not be conducted because of the small number of included trials. Second, there is no entirely reliable classification of proximal humeral fractures. The Neer classification has poor inter- and intra- observer reliability. For complex proximal humeral fractures, however, 3- and 4-part fractures classified by Neer were widely used in clinical trials and previous systematic reviews.15–22,28 No other classification is an acceptable alternative to represent the complex proximal humeral fracture in a systematic review. Nonetheless, unreliable classification of the fractures may bias the results. Third, the follow-up of most randomized controlled trials in the current study was less than 2 years. Therefore, further large long-term follow-up randomized controlled trials of sound methodologic quality should be conducted and more reliable classification of fractures should be used. Fortunately, such randomized controlled trials have already been designed.41–43

Conclusion

Current limited evidence suggests that no convincing findings support the use of either ORIF or hemiarthroplasty to treat complex proximal humeral fractures. The current findings should be interpreted cautiously because of the modest sample size and short-term follow-up. Future randomized controlled trials should be large, with a longer follow-up period, and they should be of sound methodologic quality and use a reliable classification of fractures.

References

  1. Baron JA, Karagas M, Barrett J, et al. Basic epidemiology of fractures of the upper and lower limb among Americans over 65 years of age. Epidemiology. 1996; 7(6):612–618. doi:10.1097/00001648-199611000-00008 [CrossRef]
  2. Court-Brown CM, Caesar B. Epidemiology of adult fractures: a review. Injury. 2006; 37(8):691–697. doi:10.1016/j.injury.2006.04.130 [CrossRef]
  3. Horak J, Nilsson BE. Epidemiology of fracture of the upper end of the humerus. Clin Orthop Relat Res. 1975; (112):250–253.
  4. Iyengar JJ, Devcic Z, Sproul RC, Feeley BT. Nonoperative treatment of proximal humerus fractures: a systematic review. J Orthop Trauma. 2011; 25(10):612–617. doi:10.1097/BOT.0b013e3182008df8 [CrossRef]
  5. Lanting B, MacDermid J, Drosdowech D, Faber KJ. Proximal humeral fractures: a systematic review of treatment modalities. J Shoulder Elbow Surg. 2008; 17(1):42–54. doi:10.1016/j.jse.2007.03.016 [CrossRef]
  6. Kannus P, Palvanen M, Niemi S, Parkkari J, Järvinen M, Vuori I. Increasing number and incidence of osteoporotic fractures of the proximal humerus in elderly people. BMJ. 1996; 313(7064):1051–1052. doi:10.1136/bmj.313.7064.1051 [CrossRef]
  7. Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M, Vuori I. Osteoporotic fractures of the proximal humerus in elderly Finnish persons: sharp increase in 1970–1998 and alarming projections for the new millennium. Acta Orthop Scand. 2000; 71(5):465–470. doi:10.1080/000164700317381144 [CrossRef]
  8. Roux A, Decroocq L, El Batti S, et al. Epidemiology of proximal humerus fractures managed in a trauma center. Orthop Traumatol Surg Res. 2012; 98(6);715–719. doi:10.1016/j.otsr.2012.05.013 [CrossRef]
  9. Resch H. Proximal humeral fractures: current controversies. J Shoulder Elbow Surg. 2011; 20(5):827–832. doi:10.1016/j.jse.2011.01.009 [CrossRef]
  10. Neer CS II, . Four-segment classification of proximal humeral fractures: purpose and reliable use. J Shoulder Elbow Surg. 2002; 11(4):389–400. doi:10.1067/mse.2002.124346 [CrossRef]
  11. Court-Brown CM, Garg A, McQueen MM. The epidemiology of proximal humeral fractures. Acta Orthop Scand. 2001; 72(4):365–371. doi:10.1080/000164701753542023 [CrossRef]
  12. Ricchetti ET, DeMola PM, Roman D, Abboud JA. The use of precontoured humeral locking plates in the management of displaced proximal humerus fracture. J Am Acad Orthop Surg. 2009; 17(9):582–590.
  13. Karataglis D, Stavridis SI, Petsatodis G, Papadopoulos P, Christodoulou A. New trends in fixation of proximal humeral fractures: a review. Injury. 2011; 42(4):330–338. doi:10.1016/j.injury.2010.10.016 [CrossRef]
  14. Bell JE, Leung BC, Spratt KF, et al. Trends and variation in incidence, surgical treatment, and repeat surgery of proximal humeral fractures in the elderly. J Bone Joint Surg Am. 2011; 93(2):121–131. doi:10.2106/JBJS.I.01505 [CrossRef]
  15. Misra A, Kapur R, Maffulli N. Complex proximal humeral fractures in adults: a systematic review of management. Injury. 2001; 32(5):363–372. doi:10.1016/S0020-1383(00)00242-4 [CrossRef]
  16. den Hartog D, de Haan J, Schep NW, Tuinebreijer WE. Primary shoulder arthroplasty versus conservative treatment for comminuted proximal humeral fractures: a systematic literature review. Open Orthop J. 2010; 17(4):87–92.
  17. Zyto K, Ahrengart L, Sperber A, Tornkvist H. Treatment of displaced proximal humeral fractures in elderly patients. J Bone Joint Surg Br. 1997; 79(3):412–417. doi:10.1302/0301-620X.79B3.7419 [CrossRef]
  18. Fjalestad T, Hole MØ, Hovden IA, Blücher J, Strømsøe K. Surgical treatment with an angular stable plate for complex displaced proximal humeral fractures in elderly patients: a randomized controlled trial. J Orthop Trauma. 2012; 26(2):98–106. doi:10.1097/BOT.0b013e31821c2e15 [CrossRef]
  19. Olerud P, Ahrengart L, Ponzer S, Saving J, Tidermark J. Internal fixation versus nonoperative treatment of displaced 3-part proximal humeral fractures in elderly patients: a randomized controlled trial. J Shoulder Elbow Surg. 2011; 20(5):747–755. doi:10.1016/j.jse.2010.12.018 [CrossRef]
  20. Boons HW, Goosen JH, van Grinsven S, van Susante JL, van Loon CJ. Hemiarthroplasty for humeral four-part fractures for patients 65 years and older: a randomized controlled trial. Clin Orthop Relat Res. 2012; 470(12):3483–3491. doi:10.1007/s11999-012-2531-0 [CrossRef]
  21. Olerud P, Ahrengart L, Ponzer S, Saving J, Tidermark J. Hemiarthroplasty versus non-operative treatment of displaced 4-part proximal humeral fractures in elderly patients: a randomized controlled trial. J Shoulder Elbow Surg. 2011; 20(7):1025–1033. doi:10.1016/j.jse.2011.04.016 [CrossRef]
  22. Stableforth PG. Four-part fractures of the neck of the humerus. J Bone Joint Surg Br. 1984; 66(1):104–108.
  23. Nijs S, Broos P. Outcome of shoulder hemiarthroplasty in acute proximal humeral fractures: a frustrating meta-analysis experience. Acta Orthop Belg. 2009; 75(4):445–451.
  24. Higgins JPT, Green S (eds.). Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 [Updated March 2011]. The Cochrane Collaboration. 2011. Available from www.cochrane-handbook.org.
  25. Handoll HH, Ollivere BJ, Rollins KE. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev. 2010; 12:CD000434.
  26. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003; 327(7414):557–560. doi:10.1136/bmj.327.7414.557 [CrossRef]
  27. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7(3):177–188. doi:10.1016/0197-2456(86)90046-2 [CrossRef]
  28. Fjalestad T, Hole MØ, Jørgensen JJ, Strømsøe K, Kristiansen IS. Health and cost consequences of surgical versus conservative treatment for a comminuted proximal humeral fracture in elderly patients. Injury. 2010; 41(6):599–605. doi:10.1016/j.injury.2009.10.056 [CrossRef]
  29. Hackshaw A, Kirkwood A. Interpreting and reporting clinical trials with results of borderline significance. BMJ. 2011; 343:d3340. doi:10.1136/bmj.d3340 [CrossRef]
  30. Sproul RC, Iyengar JJ, Devcic Z, Feeley BT. A systematic review of locking plate fixation of proximal humerus fractures. Injury. 2011; 42(4):408–413. doi:10.1016/j.injury.2010.11.058 [CrossRef]
  31. Brorson S, Rasmussen JV, Frich LH, Olsen BS, Hróbjartsson A. Benefits and harms of locking plate osteosynthesis in intraarticular (OTA type C) fractures of the proximal humerus: a systematic review. Injury. 2012; 43(7):999–1005. doi:10.1016/j.injury.2011.08.025 [CrossRef]
  32. Cai M, Tao K, Yang C, Li S. Internal fixation versus shoulder hemiarthroplasty for displaced 4-part proximal humeral fractures in elderly patients. Orthopedics. 2012; 35(9):e1340–e1346. doi:10.3928/01477447-20120822-19 [CrossRef]
  33. Robinson CM, Amin AK, Godley KC, Murray IR, White TO. Modern perspectives of open reduction and plate fixation of proximal humerus fractures. J Orthop Trauma. 2011; 25(10):618–629. doi:10.1097/BOT.0b013e31821c0a2f [CrossRef]
  34. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA. 1995; 273(5):408–412. doi:10.1001/jama.1995.03520290060030 [CrossRef]
  35. Olerud P, Tidermark J, Ponzer S, Ahrengart L, Bergström G. Responsiveness of the EQ-5D in patients with proximal humeral fractures. J Shoulder Elbow Surg. 2011; 20(8):1200–1206. doi:10.1016/j.jse.2011.06.010 [CrossRef]
  36. Olerud P, Ahrengart L, Söderqvist A, Saving J, Tidermark J. Quality of life and functional outcome after a 2-part proximal humeral fracture: a prospective cohort study on 50 patients treated with a locking plate. J Shoulder Elbow Surg. 2010; 19(6):814–822. doi:10.1016/j.jse.2009.11.046 [CrossRef]
  37. Wild JR, DeMers A, French R, et al. Functional outcomes for surgically treated 3- and 4-part proximal humerus fractures. Orthopedics. 2011; 34(10):e629–e633.
  38. Sintonen H. The 15D instrument of health-related quality of life: properties and applications. Ann Med. 2001; 33(5):328–336. doi:10.3109/07853890109002086 [CrossRef]
  39. Thanasas C, Kontakis G, Angoules A, Limb D, Giannoudis P. Treatment of proximal humerus fractures with locking plates: a systematic review. J Shoulder Elbow Surg. 2009; 18(6):837–844. doi:10.1016/j.jse.2009.06.004 [CrossRef]
  40. Kontakis G, Koutras C, Tosounidis T, Giannoudis P. Early management of proximal humeral fractures with hemiarthroplasty: a systematic review. J Bone Joint Surg Br. 2008; 90(11):1407–1413. doi:10.1302/0301-620X.90B11.21070 [CrossRef]
  41. Launonen AP, Lepola V, Flinkkila T, et al. Conservative treatment, plate fixation, or prosthesis for proximal humeral fracture: a prospective randomized study. BMC Musculoskelet Disord. 2012; 13:167. doi:10.1186/1471-2474-13-167 [CrossRef]
  42. 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. doi:10.1186/1471-2474-10-140 [CrossRef]
  43. Den Hartog D, Van Lieshout EM, Tuinebreijer WE, et al. Primary hemiarthroplasty versus conservative treatment for comminuted fractures of the proximal humerus in the elderly (ProCon): a multicenter randomized controlled trial. BMC Musculoskelet Disord. 2010; 11:97. doi:10.1186/1471-2474-11-97 [CrossRef]

Characteristics of Included Studies

Study Intervention
Mean age, y (O vs N) No. (O vs N) % Female Follow-up, mo Rate of Follow-up Diagnosis Characteristics
Operative Treatment Nonoperative Treatment
Boons et al 201220 Global 1 FX shoulder fracture endoprosthesis (DePuy, Leeds, UK) Shoulder immobilizer 79.9 (7.7) vs 76.4 (5.6) 25 vs 25 92 vs 96 24 94 Displaced 4-part fractures
Fjalestad et al 201028 201218 Angular stable locking plates (nonspecific LCPs of the AO basic type; Synthes, Bettlach/Solothurn, Switzerland) Modified Velpeau bandages 72.2 (60–86) vs 73.1 (60–88) 25 vs 25 80 vs 94 12 98.0 Displaced 3- or 4-part fractures of OTA 11-B2 or 11-C2
Olerud et al 201119 Open reduction and locking plates (Philos plates) Slings 72.9 (56–92) vs 74.9 (58–88) 30 vs 30 80 vs 83 24 88.3 Displaced 3-part fractures
Olerud et al 201121 Global FX prosthesis (DePuy, Sollentuna, Sweden) Slings 75.8 (58–90) vs 77.5 (60–92) 27 vs 28 85 vs 86 24 89.1 Displaced 4-part fractures
Stableforth 198422 Neer prosthesis, uncemented Closed manipulation 65.6 (52–88) vs 70.1 (60–85) 16 vs 16 75 vs 81.3 6–48 93.8 Displaced 4-part fractures
Zyto et al 199717 Cerclage wiring or surgical tension band Slings 73 (7.5) vs 75 (6.7) 20 vs 20 90 vs 85 36–60 72.5 Displaced 3- or 4-part fractures

Functional Status and Health-related Quality of Life

Outcome No. of Trials No. of Patients (O vs N) MD (95% CI) P for MD I2 P for Heterogeneity
Constant scores at 12 mo 418–21 200 (99/101) 2.92 (−2.07 to 7.92) .25 0% .93
Constant scores after 24 mo 317,19,21 140 (70/70) −1.36 (−7.64 to 4.92) .67 0% .64
DASH scores at 12 mo 219,21 105 (53/52) −4.5 (−13.5 to 4.48) .33 0% .74
DASH scores at 24 mo 219,21 102 (51/51) −7.48 (−16.1 to 1.15) .09 0% .79
ASES score 118 58 (23/25) 0.20 (−0.71 to 1.11) .67 NA NA
EQ-5D at 12 mo 219,21 105 (53/52) 0.13 (0.04 to 0.22) .004 0% .45
EQ-5D at 24 mo 219,21 102 (51/51) 0.15 (0.05 to 0.24) .004 0% .65
15D at 12 mo 118 58 (23/25) 0.02 (−0.03 to 0.08) .42 NA NA

Range of Motion

Outcome for ROM Follow-up, mo Boons et al20 Fjalestad et al18 Olerud et al19 Olerud et al21 Stableforth22 Zyto et al17
CROM (O vs N)
  Mean (SD) 12 20 (8.3) vs 18 (6.9) 17.0 vs 18.2 27.0 (8.7) vs 24.1 (7.1) (P=.11) 21.2 (7.2) vs 19.2 (7.8) (P=.46) NA NA
  Mean (SD) 36 NA NA 27.3 (8.6) vs 24.7 (10.0) (P=.41) 20.2 (9.2) vs 20.1 (9.4) (P=.89) NA 26 vs 50
Forward flexion (O vs N)
  Mean (range), ° 12 98 (45–165) vs 94 (45–165) (P=.86) NA NA NA NA NA
  Mean (range), ° 24 NA NA 120 vs 111 (P=.36) 93 vs 95 (P=.85) NA NA
  >90°, No. 48 NA NA NA NA 11 vs 1 NA
  <45°, No. 48 NA NA NA NA 1 vs 7 NA
Abduction (O vs N)
  Mean (range), ° 12 77 (45–165) vs 87 (30–130) (P=.36) NA NA NA NA NA
  Mean, ° 24 NA NA 114 vs 106 (P=.28) 86 vs 87 (P=.89) NA NA
External rotation (O vs N)
  Mean (range), ° 12 17 (10–25) vs 19 (15–25) (P=.10) NA NA NA NA NA
  >25°, No. 48 NA NA NA NA 8 vs 2 NA
  <5°, No. 48 NA NA NA NA 2 vs 10 NA
Internal rotation (O vs N)
  Thumb to lumbar level 12 L3 vs L3 NA NA NA NA NA
  No. thumb to T12 spinous process 48 NA NA NA NA 15 vs 8 NA

Other Complications Reported

Outcome No. of Trials No. of Patients Operative Group Nonoperative Group RR (95% CI) P for RR I2 P for Heterogeneity
Mortality 218,22 82 3/41 1/41 2.33 (0.36–15.14) .43 0% .37
Infection 317,19,22 132 6/66 0/66 5.0 (0.9–27.89) .07 0% 1
Nonunion 517–21 253 4/125 7/128 0.66 (0.23–1.88) .44 0% .79
Osteoarthritis 217,21 95 2/47 7/48 0.39 (0.03–4.27) .44 52% .15
Redisplacement of fracture 318,20,21 155 1/77 3/78 0.51 (0.1–2.68) .42 0% .82
Dislocation or resorption of tuberosity 417,19–21 205 21/102 3/103 4.65 (0.49–44.58) .06 60% .18

10.3928/01477447-20140430-50

Sign up to receive

Journal E-contents