Controversies remain in the optimal management of complex proximal humerus fractures. The literature is replete with studies reporting successful outcomes of nonsurgical treatment of minimally or mildly displaced proximal humerus fractures, but traditionally surgical treatment in the form of either fracture fixation or arthroplasty has been advocated for more complex fracture types.1–4 Despite the remarkable advances made in the surgical techniques and instruments, indications for surgery in complex proximal humerus fractures remain in dispute. Recently, investigators have reported good outcomes of nonsurgical treatment of complex proximal humerus fractures such as 3-part and 4-part fractures.1,5–10
It is not uncommon that orthopedic surgeons encounter patients who sustained a severe proximal humerus fracture that would normally necessitate surgical treatment but cannot be considered for such treatment due to various circumstances such as severe medical comorbidities, advanced age, socioeconomic situations, and personal preference of patients. While there have been studies that looked at the outcomes of nonsurgical treatment of complex proximal humerus fractures, these studies were not randomized clinical trials and only included the patients who had been treated nonsurgically, excluding those treated surgically. Patient selection was based on fracture classification systems (eg, 3-part or 4-part fractures by Neer classification11) rather than the fracture severity based on the actual necessity of surgical treatment. Although fracture classification systems do provide general treatment guidelines, ultimately a decision on treatment is made by surgeons’ collective analysis on the extent of fracture displacement, angulation, and comminution. Not all 3-part or 4-part fractures are severe and require surgical treatment, and thus, study results from pooled data can be biased toward more favorable outcomes than they actually are if only severe fractures requiring surgical treatment are included.
The purpose of the present study was to investigate the clinical outcomes of nonsurgically treated severe proximal humerus fractures that would have been treated surgically had it not been for circumstances precluding surgical treatment. To this end, we performed a retrospective cohort study where patients with a proximal humerus fracture meeting specific inclusion criteria were identified and underwent surveys for shoulder function, pain, strength, range of motion, demographic information, and radiographic features.
MATERIALS AND METHODS
After obtaining approval from our institutional review board, the departmental billing records were searched for all patients treated for a proximal humerus fracture either surgically or nonsurgically at our institution from January 2005 to October 2016. Patient medical records were then reviewed to include only nonsurgically treated patients. The initial injury shoulder x-rays of each patient were downloaded and put in a presentation program (Microsoft PowerPoint) after de-identification. Three orthopedic attending surgeons (2 shoulder and elbow-fellowship and 1 trauma-fellowship trained) independently reviewed the x-ray images of each patient and decided whether surgical treatment (eg, fracture fixation, hemiarthroplasty, reverse arthroplasty) would be necessary purely on the basis of the x-ray findings. The surgeons were blinded to the patients’ demographic and medical information. Patients who were identified as requiring surgical treatment from at least 2 of the 3 surgeons were included in the study. The surgeons decided their treatment choice using their subjective criteria including the fracture patterns on the basis of the AO/OTA classification, displacement/angulation of the fracture, and their years of anecdotal experience. The selected patients were then contacted for study participation. They were either mailed a questionnaire to be returned or met in the clinic for an interview and physical examination. The survey included questions regarding American Shoulder and Elbow Surgeons (ASES) score,12 Quick Disabilities of the Arm, Shoulder and Hand (QuickDASH) score,13 range of motion, abduction strength, forward elevation strength, and visual analog scale (VAS) pain score. For the patients who were surveyed through a mailed questionnaire, we used a validated patient self-assessment method reported by Yang et al.14 This method used diagram-based questions, in which patients were asked to identify the image representing their own active shoulder range of motion in various planes. Patients were also asked to perform a strength examination using premeasured zip-lock bags filled with water that corresponded to predetermined weights up to 2.72 kg. This method has been validated by the study of Yang et al.14 The demographic information, medical history, mechanism of injury, injury side, reason for nonsurgical treatment, the method and duration of nonsurgical treatment, and ASA score based on American Society of Anesthesiologists physical status classification15 were obtained from patients’ medical records. Imaging studies of each patient were reviewed to determine the fracture pattern on the basis of the AO/OTA classification16 and the Neer classification, and the presence of nonunion, avascular necrosis of the humeral head, arthritic changes, and the head-shaft angle of the humerus in the most recent x-rays.
Correlations between shoulder function scores and other variables were calculated by obtaining Spearman rank correlation coefficient (ρ). The head-shaft angle of the humerus was categorized into 3 groups: varus: <130 degrees, normal: 130 to 150 degrees, and valgus: >150 degrees for statistical comparisons. The shoulder function scores, range of motion, and strength were compared between the 3 head-shaft angle groups using 1-way analysis of variance. They were also compared between the patients with a head-split fracture and patients with no head-split fracture using unpaired t test. The data were presented as mean±SD. The statistical significance was set at P=0.05.
The search found 556 patients who had been treated nonsurgically for a proximal humerus fracture over the study period. Among them, 96 patients received agreement for surgical treatment by ≥2 attending surgeons. Of the 96 patients, 33 patients have deceased, 7 declined to participate in the study, 31 could not be reached, 1 patient was admitted to a hospice care facility, 1 patient had a stroke on the side of the proximal humerus fracture, and 1 patient had dementia. This left 22 patients (4 male individuals and 18 female individuals) in total who were able to participate in the study. Of the 22 patients, 5 patients came in for a study visit, and 17 patients answered the study survey at home.
The mean age of the 22 patients at the time of injury was 69±12.3 years (range, 46 to 91 y). The mean time from injury to the final surveys was 59.1±35.8 months (range, 22 to 133 mo). The mean time from injury to the final radiographs was 8.8±5.8 months (range, 6 to 16 mo). In regards to the work status, 15 were retired, 3 patients had a sedentary job, 2 had unknown job status, 1 had a self-employed manual job, and 1 was unemployed at the time of the survey. The mean number of medical comorbidities were 2.6±1.4 (range, 0 to 5) (Table 1). The dominant-side shoulder was involved in 11 patients. The mean ASA physical status score was 2.5±0.9 (range, 1 to 4): 4 patients with a grade 4, 4 with a grade 3, 13 with a grade 2, and 1 with a grade 1. The mechanism of injury was a fall in all patients except one who was injured in a motor vehicle accident. The reason for nonsurgical treatment was (1) consensual decision to not pursue surgical treatment after discussing surgical risks with the surgeon in 10 patients, (2) surgeon’s recommendation against surgical treatment due to severe medical comorbidities in 8 patients, (3) advanced age in 2 patients, (4) a concern for noncompliance in 1 patient, and (5) a concern for infection because of a skin lesion at the projected surgical site in 1 patient. The severity of medical comorbidities was determined subjectively on the basis of the combined opinions of the orthopedic surgeons, primary care physicians, and patients themselves. There was no patient with a pathologic fracture, previous neurologic disorders, or workers’ compensation status.
Nonsurgical Treatment and Radiographic Findings
All patients underwent shoulder immobilization in an arm sling except 1 patient who had a long arm coaptation splint. The immobilization period was 4.5+1.2 weeks (range, 2 to 6 wk). All patients underwent a course of physical therapy following the initial immobilization. The average duration of therapy was 6.4 weeks (range, 4 to 12 wk). The fracture type classification is summarized in Table 2. There were six 2-part fractures, seven 3-part fractures, five 4-part fractures, and 4 head-split fractures. All fractures had severe displacement and/or angulation of the fractured bone fragment(s), which were determined by at least 2 surgeons to warrant surgical treatment. All patients went on to radiographic union of their fracture in the latest x-rays. Except for 1 patient with a head-split fracture who developed mild arthritic changes a year after the initial injury, none of the patients developed significant glenohumeral arthritis or avascular necrosis of the humeral head within the study follow-up period. The mean head-shaft angle of the humerus measured on the latest Grashey view x-ray was 129±34 degrees (range, 84 to 202 degrees). The head-shaft angle could not be measured in the 4 patients with a head-split fracture. Ten patients healed in an abnormal varus position (<130 degrees), 3 healed in a normal head-shaft angle (between 130 and 150 degrees), and 5 patients healed in an abnormal valgus position (>150 degrees).
Shoulder Function Score, Pain, Range of Motion, Strength
The mean ASES score at the latest follow-up was 82.7±24.3 (range, 30 to 100), and the mean QuickDASH score was 21.9±25.3 (range, 0 to 88) (Table 3). The mean VAS pain score at the latest follow-up was 1.1±2.2 (range, 0 to 8). The range of motion is summarized in Table 3. There were significant correlations between VAS pain scores and shoulder function scores, and between range of motion/strength and shoulder function scores (P<0.05). In contrast, there were no significant correlations between age, shoulder function scores, number of comorbidities, ASA scores, and the humeral head-shaft angle (P>0.05). When compared between the 3 different humeral neck-shaft angle groups, there was no significant difference in any of the study variables with the given study sample size (P>0.05) (Fig. 1). The type of fracture based on Neer classification did not show statistically significant differences in any of the study variables (P>0.05) with the given study sample size except for the number of comorbidities (P=0.01) (Figs. 2–4) and Table 4). When compared between the patients with a head-split fracture and those without such a fracture, there was no significant difference in any of the study variables with the given study sample size (P>0.05) (Table 5).
Although most proximal humerus fractures are minimally displaced and can be treated conservatively with good functional outcomes, treatment of displaced fractures is more challenging. For more complex fractures, such as severely displaced fractures, or 3-part and 4-part proximal humerus fractures, treatment with open reduction and internal fixation or arthroplasty is often the treatment of choice on the basis of reasons such as achieving better anatomic reconstruction and early motion in the shoulder.2–4,17 Although surgery is typically the treatment of choice for complex humerus fractures, several studies have assessed the outcomes of patients treated conservatively for these fractures and found satisfactory results.1,5–10 Our study sought to investigate the clinical outcomes of nonsurgically treated severe proximal humerus fractures that would have been treated surgically had it not been for circumstances precluding surgical treatment. Unlike other previously published studies, the present study used the clinical judgments of 2 shoulder and elbow surgeons and 1 orthopedic trauma surgeon who had been practicing for >5 years at a level-I trauma center as the criteria for proximal humerus fractures that would necessitate surgical treatment.
Our study found overall favorable clinical outcomes following nonsurgical treatment of proximal humerus fractures that would have normally necessitated surgical treatment. The mean VAS pain score was 1 (range, 0 to 8), which was far better than we had expected in these patients with a severely angulated and displaced fracture. The patient who rated her pain 8 had sustained a fall injury during her hospital admission for other medical conditions. The mean ASES score was 83 with only 3 of the 22 patients having a score lower than 60. Although it is difficult to directly compare the shoulder function results because of different shoulder function scoring systems used, our study results are in agreement with previous studies that have reported favorable outcomes following nonsurgical treatment of complex proximal humeral fractures.5,6,8,9,18,19 Yüksel et al8 reported the outcomes of nonsurgical treatment of 3-part and 4-part proximal humerus fractures in 18 patients who had refused surgery or could not undergo surgery because of medical conditions. The mean age of the patients was 68 years, and the patients were divided into 2 groups: patients younger than 65 years and those 65 years or older. The mean Constant score was 61.3. Among the patients younger than 65 years, those with a 3-part fracture showed significantly higher Constant scores than those with a 4-part fracture, but this difference was not observed in patients 65 years or older. Rasmussen et al6 reported satisfactory outcomes following nonsurgical treatment of 42 patients with a 2, 3, or 4-part fractures. Ilchmann et al5 determined that conservative treatment provided better outcomes compared with tension-band fixation for 3-part fractures, but 4-part fractures had better function and range of motion with tension-band fixation. The PROFHER study19—a recently conducted randomized clinical trial evaluating the clinical effectiveness of surgical and nonsurgical treatment for 250 adults with displaced surgical neck fractures of the proximal humerus—found no significant difference between surgical treatment compared with nonsurgical treatment in patient-reported clinical outcomes at 2-year follow-up. Although the PROFHER study reports similar finding to our study, there are important differences between the 2 studies. In PROFHER study, the degree of displacement had to be sufficient for the treating surgeon to consider surgical intervention but did not have to meet the displacement criteria of Neer (1 cm or 45 degrees angulation of displaced parts, or both) for inclusion in the trial. The authors acknowledged that this relaxing of the displacement criteria reflected the arbitrariness of these thresholds. Consequently, a significant number of minimally and mildly displaced fractures were included (eg, 18 Neer 1-part fractures). Furthermore, unlike our study, no head-split fractures were included in that study. Therefore, the PROFHER study had broader inclusion criteria than our study and might have included a number of fractures that could have been successfully treated with nonsurgical treatment.
It was found that the type of fracture had no significant influences on the clinical results in the particular population of patients in the present study. The shoulder function scores, pain scores, range of motion, and strength were not significantly different across different fracture types. Patients with a head-split fracture did not show inferior results compared with those with other fracture types. Patients with a head-split fracture tended to be older and had lower ranges of motion than those with other fracture types, but these differences did not reach statistical significance. The small sample size of the present study might have contributed to this statistical insignificance. Kollig et al20 also reported that the fracture type did not influence the outcomes in their study of 41 patients with a complex proximal humerus fracture, but no head-split fractures were included in their study. In contrast, the studies by Rasmussen et al,6 Yüksel et al,8 and Zyto9 have shown inferior outcomes in patients with a 4-part fracture compared with those with a 2-part or 3-part fracture. It may be that the low-demand patients with advanced age and severe medical comorbidities of our study might have felt their objectively inferior shoulder function more tolerable than the patients included in other studies.
It was also found that malunion of the proximal humerus in a varus or valgus position did not result in significant decreases in subjective shoulder function in this study’s particular patient population. Varus malunion was most common and seen in 10 patients (45%), and valgus malunion was seen in 5 patients (23%). Five patients had severe varus malunion with a neck-shaft angle <100 degrees. The greater tuberosity was positioned at the top of the humeral head in these patients, but their forward elevation and abduction were not significantly decreased compared with other patients. There was no avascular necrosis of the humeral head in the present study, which is at odds with many previous studies that had reported avascular necrosis in complex proximal humerus fractures. The original study of Neer2 reported a 16.1% rate of avascular necrosis in twenty 3-part fractures and eleven 4-part fractures in elderly patients who were treated nonoperatively. Zyto9 reported 2 cases of necrosis in nonsurgically treated seventeen 3-part or 4-part fractures (12%). The shorter follow-up periods of the present study than previous studies are likely the reason for this discrepancy as it takes some time for avascular necrosis to be recognized in radiographs.
The present study has important limitations to be noted. This is, in essence, a retrospective study, which might have been biased by various study errors. First, a large number of eligible patients were excluded from the study because of their death, lost to follow-up, or refusal of participation. Only 22 of 96 eligible patients (23%) were included in the study, which raises the possibility of selection bias. The exact effect of this low participation rate on the study results is unknown. Second, the study had a small sample size, and this might have contributed to the statistical insignificances observed in many of the statistical tests performed in the study. In fact, a post hoc power analysis revealed that the present study had only a power of 0.3 to detect a 15-point difference in ASES scores between the different fracture types. Larger prospective studies are needed in the future for more accurate assessment. Third, although it has been used and validated by previous studies,14 the accuracy of patients’ self-reported range of motion and strength has not been validated for our study. Fourth, the method of using independent clinical judgments of 3 attending orthopedic surgeons to select fractures that would need surgical treatment could be subjective and has not been validated. We believe that this method does reflect how we actually decide treatment options for patients in a real clinical setting and thus proves the clinical relevance of the methodology. Fifth, because of the substantial diversity in shoulder function assessment systems used in studies, the ASES and QuickDASH scores obtained in the present study cannot be directly compared with other published studies that had used different systems (eg, Constant-Murley score, Simple Shoulder Test score). Lastly, only subjective opinions of the orthopedic surgeons, other physicians involved in care of the patient, and patient themselves were used to determine whether medical comorbidities were severe enough to avoid surgical treatment. Because of the retrospective nature of the present study, objective criteria could not be applied to the decision-making process.
In conclusion, patients in our study sustained a proximal humerus fracture that necessitated surgical treatment. They were treated nonsurgically due to circumstances precluding surgical treatment and showed favorable outcomes. Fracture pattern or articular involvement did not have a significant effect on shoulder function, pain, or strength in these patients. This finding suggests that nonsurgical management of proximal humerus fractures that would traditionally necessitate surgical treatment can be a viable treatment option in a certain group of patients when circumstances do not allow surgical treatment.
1. 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;125:310–316.
2. Neer CS. Displaced proximal humeral fractures. II. Treatment of three-part and four-part displacement. J Bone Joint Surg Am. 1970;52:1090–1103.
3. Rees J, Hicks J, Ribbans W. Assessment and management of three-and four-part proximal humeral fractures. Clin Orthop. 1998:18–29.
4. Vallier HA. Treatment of proximal humerus fractures. J Orthop Trauma. 2007;21:469–476.
5. Ilchmann T, Ochsner PE, 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;22:316–320.
6. Rasmussen S, Hvass I, Dalsgaard J, et al. Displaced proximal humeral fractures: results of conservative treatment. Injury. 1992;23:41–43.
7. Stableforth PG. Four-part fractures of the neck of the humerus. J Bone Joint Surg Br. 1984;66:104–108.
8. Yüksel HY, Yilmaz S, Akşahin E, et al. The results of nonoperative treatment for three- and four-part fractures of the proximal humerus in low-demand patients. J Orthop Trauma. 2011;25:588–595.
9. Zyto K. Non-operative treatment of comminuted fractures of the proximal humerus in elderly patients. Injury. 1998;29:349–352.
10. Zyto K, Ahrengart L, Sperber A, et al. Treatment of displaced proximal humeral fractures in elderly patients. J Bone Joint Surg Br. 1997;79:412–417.
11. Carofino BC, Leopold SS. Classifications in brief. The Neer classification
for proximal humerus fractures. Clin Orthop Relat Res. 2013;471:39–43.
12. Richards RR, An KN, Bigliani LU, et al. A standardized method for the assessment of shoulder function
. J Shoulder Elbow Surg. 1994;3:347–352.
13. Beaton DE, Wright JG, Katz JN, et al. Development of the QuickDASH: comparison of three item-reduction approaches. J Bone Joint Surg Am. 2005;87:1038–1046.
14. Yang JS, Keener JD, Yamaguchi K, et al. Reliability of patient self-assessment of shoulder range of motion and strength after shoulder arthroplasty. J Shoulder Elbow Surg. 2015;24:1–10.
15. Cohen MM, Duncan PG. Physical status score and trends in anesthetic complications. J Clin Epidemiol. 1988;41:83–90.
16. Muller M, Nazarian S, Koch P, et al. The Comprehensive Classification of Fractures of Long Bones. Berlin: Springer Verlag; 1990.
17. van den Broek CM, van den Besselaar M, Coenen JMF, et al. Displaced proximal humeral fractures: intramedullary nailing versus conservative treatment. Arch Orthop Trauma Surg. 2007;127:459–463.
18. Lill H, Bewer A, Korner J, et al. Conservative treatment of dislocated proximal humeral fractures. Zentralbl Chir. 2001;126:205–210.
19. Rangan A, Handoll H, Brealey S, et al. Surgical vs nonsurgical treatment
of adults with displaced fractures of the proximal humerus: the PROFHER randomized clinical trial. JAMA. 2015;313:1037–1047.
20. 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;128:111–118.
Keywords:Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
proximal humerus fracture; nonsurgical treatment; shoulder function; head-split fracture; neer classification