Systematic review on the effects of exercise in people with upper‐limb fractures
Review ID	Search date	Studies	Findings	Conclusion	Comment
Bruder 2017	January 2011 and July 2016 Updated review for Bruder 2011	22 RCTs (1299 participants with an upper‐limb fracture); 6 RCTs with proximal humeral fracture: Agorastides 2007; Bertoft 1984; Hodgson 2003a; Lefevre‐Colau 2007; Lundberg 1979; Revay 1992	"There was insufficient evidence from 13 trials to support or refute the effectiveness of home exercise therapy compared with therapist‐supervised exercise or therapy that included exercise following distal radius or proximal humeral fractures. There was insufficient evidence from three trials to support or refute the effectiveness of exercise therapy compared with advice/no exercise intervention following distal radius fracture. There was moderate evidence from five trials (one examining distal radius fracture, one radial head fracture, and three proximal humeral fracture) to support commencing exercise early and reducing immobilisation in improving activity during upper limb rehabilitation compared with delayed exercise and mobilisation... Less than 40% of included trials reported adequate exercise program descriptions to allow replication according to the TIDieR checklist."	"There is emerging evidence that current prescribed exercise regimens may not be effective in reducing impairments and improving activity following an upper limb fracture. Starting exercise early combined with a shorter immobilisation period is more effective than starting exercise after a longer immobilisation period."	Separate subgroup analysis for proximal humeral fractures. The study questions were recast in terms of exercise and thus different from comparisons in our review; we also separated primary non‐surgical treatment / rehabilitation and post‐surgical rehabilitation. Out‐of‐date search. Missing RCTs for early versus delayed mobilisation: Kristiansen 1989; Ring 2019; Torrens 2012
Systematic reviews of surgical versus non‐surgical treatment (main focus)
Review ID	Search date	Studies	Findings (data not included)	Conclusion	Comment
Beks 2018	September 2017	22 studies (1743 participants): 7 RCTs: Boons 2012; Fjalestad 2012; Olerud 2011a; Olerud 2011b; ProFHER 2015; Stableforth 1984; Zyto 1997, and 15 observational studies	"There was no difference in functional outcome between operative and nonoperative treatment, ... . Major reinterventions occurred more often in the operative group. Pooled effects of RCTs were similar to pooled effects of observational studies for all outcome measures."	"We recommend nonoperative treatment for the average elderly patient (aged > 65 years) with a displaced proximal humeral fracture. Pooled effects of observational studies were similar to those of RCTs, and including observational studies led to more generalizable conclusions."	Separate analysis for different study designs Out‐of‐date search Missing new RCTs: Launonen 2019a; Lopiz 2019 Also missing old RCT: Kristiansen 1988
Launonen 2015a	April 2014	9 studies (409 participants): 8 RCTs Boons 2012; Fjalestad 2010a (separate publications counted twice); Fialka 2008; Olerud 2011a; Olerud 2011b; Voigt 2011; Zyto 1997; and one controlled clinical trial (Carbone 2012)	"No statistically significant differences were found between nonoperative treatment and operative treatment with a locking plate for any disability, for quality‐of‐life score, or for pain, in patients with 3‐ or 4‐part fractures. In 4‐part fractures, 2 trials found similar shoulder function between hemiarthroplasty and nonoperative treatment. 1 trial found slightly better health‐related quality of life (higher EQ‐5D scores) at 2‐year follow‐up after hemiarthroplasty. Complications were common in the operative treatment groups (10–29%)."	"Nonoperative treatment over locking plate systems and tension banding is weakly supported. 2 trials provided weak to moderate evidence that for 4‐part fractures, shoulder function is not better with hemiarthroplasty than with nonoperative treatment. 1 of the trials provided limited evidence that health‐related quality of life may be better at 2‐year follow‐up after hemiarthroplasty. There is a high risk of complications after operative treatment."	No meta‐analysis. Contradictory regarding scope and inclusion: included trials comparing different surgical interventions (e.g. Fialka 2008). Out‐of‐date search Missing RCTs: Launonen 2019a; Lopiz 2019; ProFHER 2015; Stableforth 1984 Also missing old RCT: Kristiansen 1988
Navarro 2018	December 2016	18 RCTs and 21 non‐randomised studies on humerus fractures. RCTs for proximal fractures: Agorastides 2007; Boons 2012 Buecking 2014; Chen 2016 (in our studies awaiting assessment: SAA); Fialka 2008; Fjalestad 2010a (2012 and 2014 papers treated as if separate); Gracitelli 2016; Liu 2011 (in our SAA); Lopiz 2014; Olerud 2011a; Olerud 2011b; ProFHER 2015); Sebastiá‐Forcada 2014; Voigt 2011; Zhang 2011; Zyto 1997	"For hemiarthroplasty (HA) and non‐operative treatment, there was no clinically important difference for moderately displaced PHF at one‐year follow‐up regarding patient rated outcomes, (...). The intervention cost for HA was at least USD 5500 higher than non‐surgical treatment. The trend in Sweden is that surgical treatment of PHF is increasing. When functional outcome of percutaneous fixation/plate fixation/prosthesis surgery and non‐surgical treatment was compared for PHF there were no clinically relevant differences, (...). There was not enough data for interpretation of quality of life or complications. Evidence was scarce regarding comparisons of different surgical options for humerus fracture treatment."	"There is moderate/low certainty of evidence that surgical treatment of moderately displaced PHF in elderly patients has not been proven to be superior to less costly non‐surgical treatment options. Further research of humerus fractures is likely to have an important impact."	"The objective of this Health Technology Assessment was to evaluate effectiveness, complications and cost‐effectiveness of surgical or non‐surgical treatment for proximal, diaphyseal or distal fractures of the humerus in elderly patients." Just the surgical versus non‐surgical treatment comparison was considered to have enough evidence. Out‐of‐date search Missing new RCTs: Launonen 2019a and Lopiz 2019 Also missing old RCTs: Kristiansen 1988 and Stableforth 1984
Sabharwal 2016b	1 May 2015	7 RCTs (528 participants): Boons 2012; Fjalestad 2010a; Olerud 2011a; Olerud 2011b; ProFHER 2015; Stableforth 1984; Zyto 1997	"The overall meta‐analysis found that there was no difference in clinical outcomes. However, subgroup and sensitivity analyses found improved patient outcomes for more complex fractures managed surgically. Four‐part fractures that underwent surgery had improved long‐term health utility scores (...). They were also less likely to result in osteoarthritis, osteonecrosis and non/malunion (...). Another significant subgroup finding was that secondary surgery was more common for patients that underwent internal fixation compared with conservative management within the studies with predominantly three‐part fractures (...)."	"This meta‐analysis has demonstrated that differences in the type of fracture and surgical treatment result in outcomes that are distinct from those generated from analysis of all types of fracture and surgical treatments grouped together. This has important implications for clinical decision making and should highlight the need for future trials to adopt more specific inclusion criteria".	Over‐interpretation of subgroup analysis. Letter from Handoll et al in 2016 pointing out this and other major flaws of this review and authors' response are available in review supplementary materials (Handoll 2016b). Out‐of‐date search. Missing new RCTs: Launonen 2019a; Lopiz 2019 Also missing old RCT: Kristiansen 1988
Xie 2016	July 2015	7 RCTs (518 participants): Boons 2012; Fjalestad 2010a; Olerud 2011a; Olerud 2011b; ProFHER 2015; Stableforth 1984; Zyto 1997	"No statistical differences were found between operative and non‐operative treatment in CS [Constant] scores at 12 mo (months) [...] and 24 mo [...]. There are also no statistical differences between operative and non‐operative treatment in DASH scores at 12 mo [...] and 24 mo [...]. Statistical differences were found between operative and non‐operative treatment in total complication rates [...]. Statistical differences in EQ‐5D at 24 mo [...] were found between operative and non‐operative treatment but no statistical differences were found in ED‐5D at 12 mo [...], 15D at 12 mo [...] and 15D at 24 mo [...]"	"Operative treatments did not significantly improve the functional outcome and healthy‐related quality of life in elderly patients. Instead, Operative treatment for [complex proximal humeral fractures]CPHFs led to higher incidence of postoperative complications."	Out‐of‐date search. Missing new RCTs: Launonen 2019a; Lopiz 2019 Also missing old RCT: Kristiansen 1988
Network meta‐analyses (NMAs) including non‐surgical treatment (NST), ORIF (locking plate), hemiarthroplasty and RTSA [RSA] (limited to 3‐part or 4‐part fractures)
Review ID	Search date	Studies	Findings (data not included)	Conclusion	Comment
Du 2017	July 2017	7 RCTs (347 participants): Boons 2012; Cai 2012; Fjalestad 2010a; Olerud 2011a; Olerud 2011b; Sebastiá‐Forcada 2014; Zyto 1997	"The rank probability plot of Constant score showed that the RSA had significantly the highest Constant score and lower reoperation than other treatments. The other way around, the efficacy of ORIF was the poorest. The rank for the Constant score was: RSA, HA, nonoperation and ORIF. The rank for the reduction in total reoperation rates was: RSA, nonoperation, HA and ORIF."	"The statistical result suggested that RSA has become a beneficial choice to treat displaced 3‐ or 4‐part fracture in elderly patients, that might result in more favorable clinical outcomes and reduction of reoperation rates than other methods performed for the same indication. But the ORIF is the worst."	Over‐interpretation of evidence from just 31 participants with RSA. Out‐of‐date search Key missing RCT within search date: ProFHER 2015 (pragmatic trial and thus does not readily fit into an NMA) Also other missing RCTs, e.g.: DelPhi 2020; Jonsson 2020; Launonen 2019a; Lopiz 2019
Orman 2020	September 2016	8 RCTs (364 participants): Boons 2012; Cai 2012; Chen 2016; Fjalestad 2010a; Olerud 2011a; Olerud 2011b; Sebastiá‐Forcada 2014; Zyto 1997	"Non‐surgical treatment was associated with a lower rate of additional surgery and adverse events compared to open reduction internal fixation. Reverse total shoulder arthroplasty resulted in fewer adverse events and a better clinical outcome score than hemiarthroplasty. Non‐surgical treatment produced similar clinical scores, adverse event rates, and additional surgery rates to hemiarthroplasty and reverse total shoulder arthroplasty."	"Non‐surgical treatment results in fewer complications and additional surgeries compared to open reduction internal fixation. Preliminary data supports reverse total shoulder arthroplasty over hemiarthroplasty, but more evidence is needed to strengthen this conclusion."	Out‐of‐date search Key missing RCT within search date: ProFHER 2015 (pragmatic trial and thus does not readily fit into an NMA) Also other missing RCTs, e.g.: DelPhi 2020; Jonsson 2020; Launonen 2019a; Lopiz 2019 We placed Chen 2016 in Studies awaiting assessment as we had concerns over Methods.
Systematic reviews on locking plate versus intramedullary nail
Review ID	Search date	Studies	Findings (data not included)	Conclusion	Comment
Li 2018	December 2016	20 studies (1384 participants): 3 listed RCTs: Gracitelli 2016; Zhu 2011 and Tian 2016 (see Comment); and 17 retrospective studies	"Analyses showed that intramedullary nails were superior to locking plates in incision length, peri‐operative bleeding time, operation time and fracture healing time. However, there were no differences between treatments in Constant score or post‐operative complications."	"The intramedullary nail is superior to locking plate in reducing the total complication, intraoperative blood loss, operative time, postoperative fracture healing time and postoperative humeral head necrosis rate of PHF. Due to the limitations in this meta‐analysis, more large‐scale, multicenter, and rigorous designed RCTs should be conducted to confirm our findings."	Mixed study meta‐analysis Out‐of‐date search. Missing RCTs: Helfen 2020 and Plath 2019 Note: we did not identify Tian 2016 (60 participants) in our search; nor in our initial check of the reference list of this review. Based on the English abstract, this study seems to report mainly on short‐term perioperative outcomes and radiological outcomes. It also does not appear to contribute data to the meta‐analyses of the review. Hence, even if it is an RCT (see Comment for Shi 2019), it seems unlikely it would contribute much evidence for this comparison.
Shi 2019	July 2018	38 "retrospective" studies (2699 participants) reported in the Abstract. However, 2 RCTs included and described in a table: Gracitelli 2016; Zhu 2011	"Meta‐analysis results show that the intramedullary nails in the treatment of proximal humeral fractures are superior to locking plates in terms of intraoperative blood loss, operative time, fracture healing time, postoperative complications, and postoperative infection. But there is no significance in constant, neck angle, VAS, external rotation, antexion, intorsion pronation, abduction, NEER, osteonecrosis, additional surgery, impingement syndrome, delayed union, screw penetration, and screw back‐out."	"The intramedullary nail is superior to locking plate in reducing the total complication, intraoperative blood loss, operative time, postoperative fracture healing time and postoperative humeral head necrosis rate of PHF. Due to the limitations in this meta‐analysis, more large‐scale, multicenter, and rigorous designed RCTs should be conducted to confirm our findings."	Inconsistently reported Mixed study meta‐analysis Out‐of‐date search Missing RCTs: Helfen 2020 and Plath 2019 Note: Tian 2016 was described as retrospective here.
Sun 2018	April 2017	13 studies (958 participants): 3 RCTs: Gracitelli 2016; Smejkal 2011; Zhu 2011; and 10 comparative studies	"A significantly greater external rotation (...) and a significantly higher penetration rate (...) were observed in the locking plate group compared with the intramedullary nail group. Constant‐Murley scores, DASH scores and total complication rate were comparable between the two groups. Moreover, there were no significant differences in forward elevation, VAS scores, and other complications."	"Current evidence indicates that locking plates and intramedullary nails have similar performance in terms of the functional scores and total complication rate. No superior treatment was suggested between locking plates and intramedullary nails for displaced proximal humeral fractures."	Mixed study meta‐analysis Out‐of‐date search. Missing RCTs: Helfen 2020 and Plath 2019 We considered Smejkal 2011 was not a locking nail and made this a separate comparison.
Wang 2015	October 2014	8 studies (615 fractures): 2 RCTs: Smejkal 2011; Zhu 2011; and 6 observational studies	"Similar Constant scores were observed between the locking plate and intramedullary nail both in randomized controlled trials (RCTs) (...) and observational studies (...). Only one RCT provided American Shoulder and Elbow Surgeons Standardized scores indicating that the locking plate was better than the intramedullary nail (...). The total complication rate did not specifically favor the locking plate or intramedullary nail both in the RCTs (...) and observational studies (...)."	"In the existing literature, limited evidence suggests that the locking plate and intramedullary nail are both valuable options for the treatment of proximal humeral fractures. Because of the observed heterogeneity and variance between the subgroups, more RCT are needed to be able to definitively recommend a locking plate or intramedullary nail for specific fracture patterns."	Separate analysis for different study designs. Out‐of‐date search. Missing RCTs: Gracitelli 2016, Helfen 2020 and Plath 2019 We considered Smejkal 2011 was not a locking nail and made this a separate comparison.
Systematic reviews on RTSA versus hemiarthroplasty
Review ID	Search date	Studies	Findings	Conclusion	Comment
Austin 2019	October 2017	17 studies: 1 RCT: Sebastiá‐Forcada 2014; and 16 comparative cohort studies 15 studies included in the meta‐analysis (913 participants)	"We found that RSA is associated with improvements in forward flexion, clinical outcome scores [Constant and Dash etc pooled], and risk of reoperation, with no differences in external rotation, tuberosity healing, and deep infection rate."	"In conclusion, the results of our meta‐analysis suggest that RSA for the treatment of acute proximal humerus fractures in patients older than 65 years of age should be strongly considered as the first‐line arthroplasty option. Our results are only applicable for the short and medium term, and thus further work will be required to determine the long‐term outcomes of RSA for proximal humerus fractures."	Mixed study meta‐analysis. Out‐of‐date search. Missing RCT: Jonsson 2020
Wang 2016	December 2014	8 studies (581 participants): no RCTs, 7 were retrospective	"Compared with HA, RSA was associated with a lower rate of total complications, higher American Shoulder and Elbow Surgeons (ASES) score, more healed tuberosities and improved active forward elevation. Both treatments were comparable in term of revision surgeries, mortality, subjective satisfaction and active external rotation."	"The present evidence from this meta‐analysis suggested that RSA was a more advantaged method for the treatment of complex proximal humeral fractures. Clinical decision should be preferred to RSA on the condition that patients’ medical conditions are indicated."	Mixed study meta‐analysis, mostly retrospective studies. Out‐of‐date search. Missing RCTs: Jonsson 2020 and Sebastiá‐Forcada 2014 (published before their search date)
Systematic review on deltoid‐split approach versus deltopectoral approach for plate fixation
Review ID	Search date	Studies	Findings	Conclusion	Comment
Xie 2019	December 2017	6 studies (426 participants in the analyses): 3 RCTs: Buecking 2014; Martetschlager 2012 (see Comment); Zhao 2017 (see Comment); and 3 non‐random prospective comparative studies.	"The meta‐analysis showed that the DS group had a significantly low humeral head necrosis rate and short operation time. No significant difference was found in total complication rate, functional outcome, and other Perioperative parameters between DS and DP groups."	"The prospective evidence suggested that DS approach for proximal humerus fractures had less humeral head necrosis and short operation time than DP approach. Both DS and DP approach had similar results in functional outcomes, total complication, VAS, and hospital stay."	Mixed study meta‐analysis. Out‐of‐date search. Missing RCT: HURA 2020 We excluded 2 trials included in Xie 2019: Martetschlager 2012 wasn't an RCT and Zhao 2017 had major flaws. We considered that the less or minimal invasive feature of surgery in the deltoid‐split group changed the comparison, in which Sohn 2017 also fitted.
Systematic reviews on minimally invasive plate osteosynthesis versus open plating
Review ID	Search date	Studies	Findings	Conclusion	Comment
Li 2019	April 2019	16 studies (1050 participants): 2 RCTs: Sohn 2017; Zhao 2017 (see Comment); and 14 non‐RCTs	"According to the meta‐analysis, MIPO was superior to ORIF in operation time, blood loss, postoperative pain, fracture union time, and constant score. However, MIPO was associated with more exposure to radiation and axillary nerve injury. No significant differences were found in length of hospital stays and complication except for axillary nerve injury."	"The present evidence indicates that compared to ORIF, MIPO had advantages in functional outcomes, operation time, blood loss, postoperative pain, and fracture union time for the treatment of PHFs. However, the MIPO technique had a higher rate of axillary nerve injury and longer radiation time compared to ORIF."	Mixed study meta‐analysis. Out‐of‐date search. We excluded Zhao 2017 as it had major flaws. We note that the deltoid‐split approach was used in the MIPO group and the deltopectoral approach was used in the ORIF group of Sohn 2017. We reflected this in a composite comparison that also included Buecking 2014 and HURA 2020.
Zang 2018	Not known	7 studies (unknown number of participants). I RCT found in reference list: Sohn 2017	"Meta‐analysis showed the significant differences in terms of blood loss, operative time, length of hospital stays and constant score between two groups. No significant differences were found in time to union, the union rate and complications."	"Minimally invasive plate osteosynthesis in proximal humeral fractures provided significantly shorter operative times, length of hospital stays, less blood loss and better clinical outcomes without increasing complications."	Article not obtained as behind very expensive paywall. However, reference list was available. Mixed study meta‐analysis. Out‐of‐date search.
Acute RTSA for fracture and delayed RTSA for fracture sequelae
Review ID	Search date	Studies	Findings	Conclusion	Comment
Torchia 2019	January 2018	16 studies (322 participants), 4 comparative (46 participants) whereas 12 were case series (276 participants).	"Among studies directly comparing acute versus delayed RTSA, no differences in forward flexion (P = .72), clinical outcome scores (P = .78), or all‐cause reoperation (P = .92) were found between the 2 groups. Patients undergoing delayed RTSA achieved 6° more external rotation than those undergoing acute RTSA; this difference was significant (P = .01)."	"Given the risks associated with surgery in the elderly population, consideration may be given to an initial trial of nonoperative treatment in these patients, saving RTSA for those in whom nonoperative treatment fails without compromising the ultimate outcome."	Preliminary data only but reassuring. This comparison is not in the scope of our review, which focuses on acute treatment.