Introduction

Studies on proximal humeral fractures usually focus on the management of displaced fractures. There has been much less interest on minimally displaced fractures, since they are generally accepted to have a good prognosis [1, 6, 11]. We studied minimally displaced proximal humeral fractures and provide a detailed analysis of the functional outcome.

Patients and methods

We evaluated retrospectively 64 adult patients with proximal humeral fractures treated conservatively between 1995 and 1999. As described by Neer [11], fractures that had less than 45° angulation and less than 1 cm displacement between the major fragments were considered to be minimally displaced. Patients with previous shoulder problems and patients with fracture dislocations were excluded. Out of 32 patients, five were not able to co-operate with isokinetic muscle-strength testing. The study therefore finally included 27 patients with a mean age of 51 (25–75) years who had sustained a simple fall or a traffic accident.

There were 14 women with a mean age of 55 years and 13 men with a mean age of 41 years. Sixteen patients were younger than 60 years of age. In all patients, the right extremity was dominant, and in 12, the fracture involved the right shoulder. Fifteen fractures showed only a single fracture line, of which seven involved the surgical neck and eight the greater tubercle. Twelve fractures showed multiple fracture lines.

The fractured arm was immobilised and oral non-steroidal anti-inflammatory drugs were prescribed. As soon as the pain permitted, passive exercises of the shoulder together with active exercises of the elbow and wrist were started under supervision of a physiotherapist. After active range of motion (ROM) had been regained, isometric strengthening and then resistive exercises were added. A forceful stretching programme was initiated 12 weeks after the fracture.

At the final follow-up, all patients were evaluated according to the Constant and Murley scoring system [3]. Instead of spring-balance testing of strength, we modified the system by introducing the concept of isokinetic testing. Measurement of muscle strength was done using the Cybex-Norm dynamometer ( Cybex, Inc., Ronkonkoma, NY, USA). Peak torque values of abduction and adduction were obtained for both fractured and non-fractured shoulders. Peak torque values of non-fractured shoulders were accepted as 25 points, and the fractured shoulders were scored proportionately. Measurements of deltoid and supraspinatus muscles are especially important to evaluate shoulder strength.

Prior to isokinetic testing, calibration was performed via standard weights. The position of the patient and dynamometer was settled according to standard measurement technique, i.e. dynamometer was tilted 30°, rotated 10° and elevated 13°. After the seat belt had been fastened, length of the adaptor was settled at 90° abduction of the shoulder. This brought the axis of dynamometer and the axis of glenohumeral joint to the same projection for abduction. ROM was limited mechanically at 0° and 90° (Fig. 1). The unaffected shoulder was tested initially. After co-operation had been established with four trials, real testing was performed using 60°/s angular velocity.

Fig. 1
figure 1

Positioning of subject for shoulder abduction/adduction

Full size image

In order to evaluate the rotator cuff tendons, impingement tests were performed (subacromial tenderness, impingement sign I, II and painful arc), which were followed by ultrasonography (US) with a 7.5–10.0°MHz “linear array” transducer (General Electric Logic 700 MR) of both shoulders. US was carried out in the sitting position with the patient facing the examiner. Muscular and tendinous structures were evaluated at both transverse and longitudinal planes. While the presence of focal hypoechoic zones inside tendons and a decrease in the width of cuff characterised partial tears, visualisation of tear ends or a segmental loss of tendon structure were the criteria to diagnose complete tears [8]. The results were classified as excellent, good, moderate and poor according to pain, daily activities including ROM and isokinetic muscle strength (Table 1). SPSS version 11 was used for statistical analysis. Mann-Whitney U test was utilised to compare the means of groups.

Table 1 Classification of results. ADL activities of daily living, ROM range of movement
Full size table

Results

The mean follow-up period was 25 (12–34) months. All fractures were united without displacement. While 17 patients had no pain, six had mild pain that did not interfere with routine activities. Pain was moderate in three patients, and only one had severe pain that necessitates regular use of oral analgesics. Twenty patients were able to perform all activities of daily living, and seven had mild trouble in overhead activities and weight carrying. At the latest follow-up, mean ROM of all fractured shoulders reached 90% of the non-fractured side.

Peak torque deficit in abduction in patients older than 60 years of age was found to be significantly higher than in patients younger than 60 years (p<0.05). In the older age group, the mean Constant score was significantly lower than in the younger group (p<0.05). There was no significant difference between peak torque deficits in patients having the fracture on the dominant side and non-dominant sides. Isokinetic test results revealed that the number of fractured shoulders that had reached 75–100%, 50–74%, 25–49% and <24% of abduction peak torque strength of contralateral shoulders were four, nine, ten and four respectively. Only in one patient did the fractured shoulder reach the same peak torque in abduction as the opposite one. For adduction, on the other hand, peak torque values of the fractured shoulders were close to the values of the non-fractured shoulders.

Mean Constant score was 81 (54–100). Statistical analysis revealed that, dominance and fracture geometry had no significant effect on scores of pain, activities of daily living, ROM, isokinetic test results of abduction or overall Constant score. In five patients with complete rotator cuff tears as diagnosed with US, clinical impingement signs were also positive. In patients with partial tears, impingement signs were only positive in one out of four. Data about patients with tears are given in Tables 2 and 3.

Table 2 Results in patients with complete rotator cuff tears. ROM range of motion
Full size table
Table 3 Results in patients with partial rotator cuff tears. ROM range of motion
Full size table

Overall, the results were graded as excellent in 11 patients, good in nine, moderate in five and poor in two. While 16/17 patients without pain had good or excellent results, the ratio for those with pain was only 4/10. The ratio of good and excellent results for those with and without cuff tears were 3/9 and 17/18 respectively.

Discussion

Minimally displaced fractures of the proximal humerus are generally known to have low morbidity and high patient satisfaction after conservative treatment. Studies have generally yielded excellent results. Kristiansen and Christensen [7], Mills and Horne [9] and Young and Wallace [13] reported 94%, 95% and 97% good and excellent results respectively. However, in a detailed analysis by Koval et al [6], good and excellent functional healing rate was only 77%. Re-trauma and pre-existing cuff diseases were the major causes of deterioration. According to our evaluation system, which incorporated Cybex test results and Constant and Murley scoring, the ratio of good and excellent results was rather low (74%), which is in accordance with the study of Koval et al. [6]. When only the muscle strength of fractured shoulders was taken into account, the results were even more unsatisfactory. Only 13 patients reached a muscle strength of 50% or higher. Muscular weakness was significantly more pronounced in patients older than 60 years of age.

In the different series, there are various methods of evaluation. Usually, data on muscle strength has been neglected. Our study is the first to include Cybex isokinetic measurements in the study of proximal humeral fractures with minimal displacement. We experienced that some shoulders that were thought to have normal muscular strength with manual testing indeed had a significant muscular weakness when tested with Cybex. Dominance did not affect Cybex measurements in different age and gender groups and life styles in healthy people [2, 4, 10]. Our findings support the data from Koval et al. [6] that, dominance did not affect the overall results after minimally displaced proximal humeral fractures. Our results are also in agreement with the study of Koval et al, that fracture geometry was not a significant factor [6]. Our study also revealed that, although the shoulders were painless, functional and strong enough for daily living, they had serious loss of muscular strength, which was more pronounced in the older age group. This was in contrast to Koval et al., who reported that age was not a major determinant for final outcome [6].

Co-existence of proximal humeral fractures and rotator cuff tears were identified. Rotator cuff tears may accompany displaced fractures [11]. In a study by Kim et al. [5], partial cuff tears were documented with arthroscopy in 23 patients suffering from chronic shoulder pain 6 months after minimally displaced tuberculum majus fractures. Rotator cuff tears were the major cause of painful shoulders in our study. Since asymptomatic cuff tears are rather frequent in advanced age groups [12], it is not possible to establish the timing of the tears. The patients might have had a silent cuff tear before the trauma, or the tear might have occurred together with the fracture. Whatever the mechanism of the tear, it seems to be a detrimental factor. Treating physicians should consider the possibility of a cuff tear when the patient fails to proceed in rehabilitation or the standard therapy programme ends in a painful shoulder.