Abstract
Purpose:
There is no consensus in the literature on nonoperative treatment of displaced and multipart fractures of proximal humerus as those are normally treated operatively. Our aim was to compare the functional results of nonoperative management and open reduction internal fixation with the proximal humerus internal locking system of 2-, 3-, and 4-part proximal humerus fractures, among themselves and with a healthy control group.
Methods:
Between 2014 and 2018, 92 proximal humerus fractures constituting a nonoperative group (n = 47) and an operative group (n = 45) together with healthy control subjects (n = 45) were analyzed in a tertiary care referral center. The American Shoulder and Elbow Surgeons (ASES) shoulder score and visual analog scale (VAS) pain score were used for subjective functional analysis. Range of motion and muscle strength were analyzed objectively for all patients and healthy control subjects.
Results:
In 2-part fractures, VAS scores and hand grip strength were determined as lower in the nonoperative group (p = 0.033 and p = 0.034, respectively). In 3- and 4-part fractures, there was no difference between the two groups in terms of ASES and VAS scores. Patients who underwent surgery had more muscle strength than those in the nonoperative group, but only arm extensor and forearm flexor muscle strengths were statistically significant for 3-part fractures. In cases of 4-part fractures, objective functional results were similar between the two groups.
Conclusions:
With insufficient functional results and high complication rates in surgery, nonoperative management is still the preferred choice for proximal humerus fractures, especially in case of multipart fractures.
Introduction
Following distal radius and hip fractures, proximal humerus fractures are the third most common fractures with an incidence of 105 of 100,000. 1,2 They are commonly seen not only in elderly osteoporotic patients but also in young patients. 3 –7 With expanded life spans, it is expected that proximal humerus fracture occurrence will also increase; thus, determining the optimum treatment becomes much more important. 8,9 Non-displaced or minimally displaced fractures could be treated well nonoperatively, but obtaining angular stability in displaced fractures is only possible by surgery. Management of these fractures is complicated, and surgeons base their decisions subjectively on a list of factors such as general health status, preference of the surgeon, morphology of fracture, age, and hand dominance. 10,11 However, subjective and objective functional results may not be at the level desired and complication rates may reach 50%. 12 –14 This is why nonoperative management is the only choice in these patients. However, there is very limited information in the literature about the results of nonoperative management of displaced fractures. The purpose of this study was to compare the functional results of nonoperative management in 2-, 3-, and 4-part proximal humerus fractures with those of open reduction internal fixation with the proximal humerus internal locking system (PHILOS; Mescomed, Double Medical®, Turkey) and healthy controls.
Materials and methods
Between 2014 and 2018, 135 patients with 2-, 3-, or 4-part fractures, with 81 patients treated nonoperatively and 59 patients treated surgically, were analyzed retrospectively. In the nonoperative group, 12 patients were lost to follow-up, 11 patients refused to participate in the study, and 11 patients died; thus, there were a total of 47 patients. In the operative group, 4 patients were lost to follow-up, 7 patients refused to take part in the study, and 3 patients died; there were thus a total of 45 patients. In both groups, the average follow-up time was 25 months. To compare joint range of motion (ROM) and muscle strength, a control group with similar age, sex, and body mass index (BMI) was formed with 45 healthy individuals who had applied to the orthopedics outpatient clinic of our hospital for any reason other than upper extremity problems, who had normal physical examination of the upper extremities, and who agreed to participate voluntarily in the study. Exclusion criteria for the study were pseudo-arthrosis, pathological fractures, refractures, neuromuscular diseases, open fractures, and coexisting fractures of the ipsilateral extremity. Approval was given by the institutional review board, and informed consent was obtained from each subject.
Nonoperative treatment
The shoulder was immobilized in a sling for 3 weeks with passive ROM exercises starting after 2 weeks including pendulum exercises, followed by progressive exercises against resistance. Active ROM exercises started at the fourth week. After starting active ROM exercises, patients were referred to physical therapy and rehabilitation. Physical therapy ended after 6–8 weeks. Closed reduction was performed if needed at the initial evaluation by the treating surgeon. Patients came to the outpatient clinic once a week within the first month. They were then seen once a month until the third month and 3 months later after the third month visit, then 6 months later, and then yearly follow-up was performed.
Surgical treatment
All surgeries were performed in the beach chair position, with a lateral deltoid split approach and using AO reduction techniques by two surgeons experienced in upper extremity trauma. In all operations, PHILOS (Mescomed, Double Medical) plates were used. After surgery, immobilization was performed with a sling, and passive ROM exercises were started the day after surgery with pendulum exercises. Patients without any wound problems were discharged 3–5 days postoperatively. The mean hospital stay was 3.3 ± 1.2 days (range 2–6 days). Sutures were removed between the 14th and 21st postoperative days. Active mobilization started at 2–4 weeks postoperatively, relying on the stability of the osteosynthesis and quality of bone, and then patients were referred to outpatient physiotherapy clinics. Physical therapy was terminated 6–8 weeks after surgery. In the first month after the operation, patients were seen weekly until the postoperative first month. They were then seen once a month up to the third month and were followed up 3 months later, 6 months later, and yearly.
Shoulder and elbow active ROM of all patients and the control group were measured using a universal long-arm goniometer. Arm flexor, extensor, abductor, and adductor muscles; forearm flexor and extensor muscles; and hand grip strength were measured by dynamometer. All the parameters were measured three times, and average values were determined. All the measurements were performed by the same surgeon. Dominant sides were determined for all three groups, and comparisons were done accordingly. For subjective functional analysis, the American Shoulder and Elbow Surgeons (ASES) shoulder score and visual analog scale (VAS) pain score were used. 15 Screw counts in the humeral head and complications were also noted. In routine outpatient follow-up visits, patient complaints and physical examination and radiological findings were recorded in the computer system used in our hospital for documentation purposes. These records were examined retrospectively, and complications were determined.
Statistical analyses
Statistical analysis was performed using SPSS software. The variables were investigated using visual (histogram, probability plots) and analytical methods (Kolmogorov–Smirnov test) to determine whether they were normally distributed. Since all the variables were not normally distributed, the Mann–Whitney U test was used to compare ASES and VAS scores between the nonoperative and the operative groups. Kruskal–Wallis tests were also conducted to compare ROM, muscle power, and hand grip strength between the nonoperative, operative, and control groups. The Mann–Whitney U test was performed to test the significance of pairwise differences using Bonferroni correction to adjust for multiple comparisons. An overall 5% type I error level was used to infer statistical significance.
Results
Patient characteristics according to subgroups are given in Table 1.
Number of patients in nonoperative group, operative group, and healthy control group according to gender, mean age, and subgroups of fractures.
2-Part fractures
For 2-part fractures, ASES scores were similar after nonoperative treatment to those after PHILOS plates with 82 versus 93 points (p = 0.62). VAS scores were seen to be lower in the nonoperative group (p = 0.033) (Table 2). A difference between the three groups was determined in terms of shoulder internal rotation. It was higher in the operative group than the nonoperative group (p = 0.011). There was no statistical difference in terms of arm and forearm muscle strength, but hand grip strength was identified as lower in the nonoperative group (p = 0.032). Thus, in cases of 2-part fractures, patients treated nonoperatively had less pain but also had less hand grip strength and limited shoulder internal rotation (Tables 3 and 4).
Comparison of ASES and VAS scores for both groups according to fracture type.
ASES: American Shoulder and Elbow Score; VAS: visual analog scale.
Comparison of shoulder ROM for 2-, 3-, 4-part fractures.
ROM: range of motion; FE: forward elevation; IR: internal rotation; ER: external rotation.
Comparison of muscle strength (in pounds) for 2-,3- and 4-part fractures.
3-Part fractures
For 3-part fractures, there were no differences between the treatment groups in terms of ASES and VAS scores (p = 0.098 and p = 0.304, respectively) (Table 2). The only significant difference regarding ROM between operative and nonoperative treatment was better shoulder abduction after nonoperative treatment, with 114.4° versus 88.6° in surgical treatment (p = 0.033). When comparing muscle strength, there was a tendency toward better results in the operative group, but only arm extensor and forearm flexor muscle strengths were statistically significant (p > 0.05) (Tables 3 and 4).
4-Part fractures
For 4-part fractures there were no significant differences between the treatment groups regarding ASES (p = 0.468) and VAS (p = 0.247) scores, ROM (p > 0.05), or muscle strengths (p > 0.05). There was a tendency toward more strength after PHILOS plates, but this was not statistically significant (Tables 3 and 4).
Radiological results
Displacement rates were 19% after surgery and 42.5% after nonoperative treatment. Valgus displacements were the most common displacements in both groups. There was no difference between displaced and non-displaced fracture healing in the PHILOS plate group (p > 0.05). In the operative group, there were valgus displacements in 6 (13.3%) patients, varus displacements in 2 (4.7%) patients, and greater tuberosity displacement in 1 (2.4%) patient. Average ASES and VAS scores of the patients who had radiological displacement were 76.04 (28.3–94.9) and 2.7 (1–8), respectively. In the nonoperative group, there were valgus displacements in 14 (29.7%) patients, varus displacements in 2 (4.4%) patients, and greater tuberosity displacements in 4 (8.8%) patients. This group’s average ASES score was 86.03 (71.6–100), and the average VAS score was 2.15 (1–4). No correlation was detected between radiological and functional results of both groups (p > 0.05) (Table 5).
Radiological results of both groups.
ASES: American Shoulder and Elbow Score; VAS: visual analog scale.
Complications
One patient in the nonoperative group developed pulmonary embolism on the seventh postoperative day. The patient was treated with an anticoagulant agent (two times 0.4 ml/day enoxaparin sodium). In the nonoperative group, non-union developed in one patient in the postoperative ninth month. Although the patient was offered a nonunion operation, he did not accept the operation and was followed up with routine clinical outpatient visits. Union was observed in his fracture site radiologically in the 15th month. In the surgical group, two patients had wound infections, which were treated by oral antibiotics, and four patients had glenohumeral screw penetrations. Three of these patients had a second operation, and the screws were changed for shorter ones. The fourth patient with glenohumeral screw penetration also had a rotator cuff tear; the penetrating screw was removed and the tear was repaired in a second operation. The highest complication rate was determined in cases of 3-part fractures (33%).
Discussion
To the best of our knowledge, this is the first study comparing subgroups of proximal humerus fractures treated by PHILOS plates or nonoperative methods in addition to healthy subjects. Functional results were evaluated with objective criteria, and healthy subjects were used as the control group. The aim of this study was to compare the functional and radiological results of nonoperative management and open reduction internal fixation with PHILOS plates of 2-, 3-, and 4-part proximal humerus fractures in a retrospective fashion. We especially focused on nonoperative treatment of 3- and 4-part fractures, which is lacking in the literature. Our most important finding was that the functional results of nonoperative treatment of type 4 fractures were similar to those with surgical treatment. The subjective results such as ASES and VAS scores and objective results such as ROM and muscle power were found to be similar in both operative and nonoperative groups. While the shoulder abduction in 3-part fractures was found to be less in the operative group, there was a tendency toward better muscle power, and only arm flexor and forearm extensor muscle strength were statistically significant.
The treatment of 2-part fractures is decided according to age, displacement, and angulation. Lange et al. in a study using Targon nails found no difference between the results of treating 2-part fractures surgically or nonoperatively, 16 similarly Fjalestad and Hole in their randomized controlled trial, found no better results with surgical treatment than conservative treatment for patients with displaced proximal humeral fracture at 2-year follow-up, 17 while Tamimi et al., who used percutaneous K-wiring in their study, determined that subjective functional results in the surgical group were much better. 10 In our study, although the ASES scores were higher in the surgical group, this was statistically insignificant. On the other hand, VAS scores were lower in the operative group and that was statistically significant. One of the possible reasons for this difference may be the use of different implants in surgical stabilization. Lange et al. found no difference in the objective functional results between two groups. 16 We found that joint ROM, especially shoulder internal rotation, and hand grip strength were higher in the operative group. This could be due to gentle reduction maneuvers for the nature of 2-part fractures, which are easy to reduce compared to 3- and 4-part fractures, and also the lateral deltoid split approach, which does not harm the rotator cuff muscles, especially the subscapularis muscle. To compare the results, a healthy control group was established, similar in age and BMI to the operative and nonoperative groups of patients. This also provided a statistically stronger study design.
In 3- and 4-part fractures, similar to our results, both Tamimi et al. and Lange et al. found no difference in the functional results between two groups, but they determined better radiological results in the operative group and fewer complications in the nonoperative group. 10,16 In a prospective randomized study comparing tension band wiring and nonoperative treatment, Zyto found no difference in functional results but had better radiological results in the surgical group, which is similar to our study. 18 The studies of Tamimi et al. 10 and Zyto 18 did not consider 3- and 4-part fractures separately. In the study by Lange et al., however, 3- and 4-part fractures were considered separately. 16 In 3-part fractures, similar to the study of Lange et al., 16 we found that shoulder ROM was less in the operative group, but only abduction was statistically significantly different. For muscle strength, in the operative group, the arm flexor and forearm extensor muscles were significantly more powerful. There were no differences between the two groups based on daily activity and pain scores. In 4-part fractures, similar to the study of Lange et al., we found that there was no difference between the two groups based on subjective and objective functional results. 16 In contrast, Olerud et al., based on their study of 60 patients who had 3-part fractures, found that in the PHILOS group, subjective results were better but statistically insignificant (p = 0.64). 19
Our aim in surgical treatment was to obtain anatomical reduction and begin early movement, thus obtaining functional and radiological results as close to normal as possible. It is very hard to obtain anatomical reduction by nonoperative treatment, especially in 3- and 4-part fractures. However, with the advancement of implant technologies, anatomical reduction is possible by surgery. Radiologically anatomical reduction does not always result in good functional outcomes, and conversely, functional outcomes of some cases that do not have anatomical reduction can be satisfactory. In our study, we saw that ASES and VAS scores were independent of anatomical reduction. In the nonoperative group, a 3-part fracture patient who had 15° of varus displacement had the best ASES (100) and VAS (1) scores (Table 4, patient 27). Similarly, Rangan et al. found no difference between surgical treatment compared with nonsurgical treatment in terms of patient-reported clinical results over 2 years following fracture occurrence in a randomized clinical trial of patients with displaced proximal humerus fractures. 20 Although the PROFHER study was larger and better designed, similar results were obtained with our data. This may be due to the comparison with the healthy control group and the use of objective analyses, which makes our study more prone to reality in terms of daily life. In contrast, Poeze et al., based on their study of 55 patients who had minimally displaced proximal humerus fractures, found that in the nonoperative group, a transscapular (Y) radiograph seen in the first week was a predictive finding for functional results. 21 Factors other than effective anatomical fracture healing and effective physical therapy that affect the functional outcomes of proximal humerus fractures are still unknown. More complications such as incision infection and screw penetration were observed with operative treatment. Secondary surgeries were needed in 4 (8.8%) patients of the operative group. These may have side effects in terms of functional results. In addition, nonoperative treatment does not critically impair the blood supply of the humeral head, so the risk of avascular necrosis may be less. On the other hand, the small number of patients may have also caused no difference between the two groups.
The first limitation of our study is having a relatively small number of samples. This is due to the fact that not all patients could be reached and some patients did not want to be involved in the study. On the other hand, PHILOS plates may have affected the surgical results positively or negatively. Intramedullary nails and reverse shoulder prostheses can give more effective results in certain fractures. 11 The second limitation is the lack of age restriction. It could give clearer results if the geriatric patient group and the younger patient group are compared within themselves. The third limitation of the study was collecting the patients in a retrospective fashion. More reliable outcomes can be obtained by prospective follow-up studies. The last limitation of our study is that we used the Neer classification. 22 It is based on the mechanism of injury or level of the fracture line; however, it does not consider the displacement of tuberosity, and it also has poor intra- and inter-observer reliability. On the other hand, surgeons continue to use this classification system as it is helpful in guiding treatment and grouping fracture patterns for research purposes. Future studies can be improved by increasing the number of patients and by prospective randomized clinical trials.
Conclusion
Nonoperative treatment in proximal humeral fractures, particularly in multipart fractures, appears to be an alternative to the surgical option. Targeted functional results can be achieved without obtaining radiological anatomical reduction.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
