Solely sutures is a reliable fixation for valgus-impacted proximal humeral fractures

Abstract

Purpose

Valgus impacted proximal humeral fractures with substantial displacement may severely compromise shoulder function and open reduction can therefore be considered. Internal fixation hardware may eventually however constitute problems. In a small subset of these fractures, in which there is no medial comminution and intact blood supply it is possible to use a least possible fixation method with solely sutures through the tendons of the rotator cuff and the shaft as described by Bigliani 1990. The aim of the present study was to investigate the outcome of this fixation method in a retrospective series of patients.

Methods

Twenty-seven patients, 23 women and 4 men, with a mean age of 66 years (53–83) and minimum follow-up time of 2 years (2–11 years) accepted follow-up with standard shoulder radiographs and 22 were examined with Oxford Shoulder Score (OSS), Constant-Murley (CM) shoulder score and Numeric Rating Scale for pain (NRS). Reoperations or adverse events were recorded.

Results

The median OSS was 45 (range 5–48) and CM 63 (range 21–98). The relative CM score was 81% of the uninjured side and median NRS 0 (range 0–5). No reoperations were undertaken. One patient had a complete avascular necrosis (AVN), 3 patients had a partial AVN of the humeral head and two a partial resorption of the greater tuberosity.

Conclusion

The use of sutures only for fixation of displaced valgus-impacted proximal humeral fractures provided a relatively reliable outcome with good functional scores and no reoperations. There appears to be no need for the use of hardware for fixation of this rare subset of fractures.

1. Introduction

In a small subset of comminuted proximal humeral fractures (PHF’s) the head is displaced in severe valgus. They may be classified as both Neer 3-part and 4-part (or AO/OTA B1.1 and C2.1) fractures. Non-operative treatment in this type of fracture configuration appears reasonable alternative in elderly, less active patients.^1,2 On the other hand, a prospective study on non-operatively treated fractures by Foruria et al. suggested that this very predictably results in a poor range of motion and occasionally severe pain.³ Hence, open reduction and osteosynthesis may be considered. Previous descriptions in the literature include fixation using pins and metal cerclage, plates and screws with or without bone substitute and primary arthroplasty.^{4, 5, 6, 7} Based on previous reports it appears that, regardless of the technique of fixation, the functional outcome of an osteosynthesis may be relatively good if the blood supply to the humeral head is preserved.^5,8,9 Hardware problems from usage of locking plates are however a quite common reason for reoperation of PHF’s, in some reports up to 30%, especially if avascular necrosis ensues.^{7,10, 11, 12} At our institution, we therefore have used only non-absorbable sutures for fixation of valgus impacted, PHFs, based on a description by Bigliani 1990.¹³

The aim of this study was to retrospectively review the clinical and radiological outcomes after using open reduction and no other devices than non-absorbable sutures between the rotator cuff insertions and the humeral shaft for fixation of valgus impacted PHF’s.

2. Materials and methods

2.1. Study subjects

Patients were retrospectively identified in our local database through ICD-codes (S42.20 and NBJ49 – pins were/are never used in adults at our unit) from 2005 to 2017. To be suitable for the method patients would typically have a three- or four-part proximal humeral fracture according to the Neer classification with valgus impaction of the humeral head resulting in a head-shaft angle of at least 160° degrees with an impaction of the head onto the shaft (Fig. 1a).^14,15 Factors indicative of a preserved blood perfusion into the humeral head was sought for, such as a medial calcar fragment of at least 7 mm in length in continuation with the humeral head, a non-displaced lesser tuberosity or no or minimal ad latus displacement of the humeral head.⁹ A split humeral head fracture was not considered to be suitable for this fixation method nor a PHF with a severely comminuted greater tuberosity. These factors were assessed with plain radiographs and when considered necessary with a CT scan.

Fig. 1.

Displaced valgus-impacted fracture (1a) and healed fracture at 5-year follow-up (1b).

To be considered for operative treatment the patients had to be active and independent with relatively high functional demands. The choice between operative and non-operative treatment was ultimately decided by a consultant consensually with the patient.

Thirty-three patients eligible patients were offered a re-examination (Fig. 2). Twenty-seven patients with a follow-up time of more than 2 years (range 2–10 years) after surgery; twenty-three women and four men, accepted (Table 1). Mean age was 66 years (range 53–83 years) at the time of injury. Five of the 27 patients only came for radiographs leaving 22 patients for clinical follow-up (Fig. 2). At follow-up no patient could recall that they had had any shoulder problems before the injury occurred. There were two smokers. Fifteen patients were retired when sustaining the fracture, 2 were clerical workers and 5 had been heavy laborers.

Fig. 2.

Flowchart of the patients eligible for the study and those included.

Table 1.

Summary of essential findings on pre-operative radiographs.

	Patients age at follow-up	Medial calcar >7 mm8	Ad latus displacement between humeral head and shaft (mm).∗	Number of fragments according to Neer∗∗
1	57	No	Yes < 5	4
2	67	No	No	3
3	57	Yes	No	3
4	83	Yes	No	3
5	86	Yes	Yes >5	3
6	70	No	Yes <5	4
7	68	Yes	No	3
8	77	Yes	Yes <5	4
9	86	No	No	3
10	70	Yes	Yes <5	3
11	70	Yes	No	3
12	77	Yes	No	3
13	59	Yes	No	3
14	76	No	Yes <5	4
15	75	Yes	No	3
16	61	Yes	Yes <5	3
17	78	Yes	Yes <5	3
18	84	No	No	3
19	73	Yes	No	3
20	71	Yes	Yes <5	3
21	80	Yes	Yes <5	3
22	57	Yes	Yes <5	3
23	86	Yes	Yes <5	3
24	76	Yes	Yes <5	3
25	73	No	No	3
26	77	Yes	No	3
27	86	No	No	4

∗ Greatest displacement ∗∗ Neer fracture classification 1970.

2.2. Surgical procedure and rehabilitation

All patients were operated through a deltopectoral approach. Reduction was achieved using a blunt elevator to disimpact and elevate the valgus displaced humeral head, after which sutures were used to reduce and fix the displaced greater and, if fractured, lesser tuberosity fragments (Fig. 3). Reduction of the humeral head was very gently performed in order not to further jeopardize a presumably preserved blood perfusion. For the fixation we used 1–2 non-absorbable sutures (Ethibond #6) between the subscapularis and the infraspinatus tendons of the rotator cuff and supplemented by at least 2 sutures between two 2.5 mm holes in the shaft and the respective rotator interval and superior-posterior cuff tendons in a figure-of-eight configuration. The needle was put deep into and through the tendons close to the insertions at the minor and greater tuberosities, respectively. The arm was immobilized in a sling for 2–4 weeks but the patients were allowed to do early passive pendulum exercises. Thereafter they all followed a standard rehabilitation program guided by a physiotherapist including active supported motion exercises in all planes during 6 weeks.

Fig. 3.

A. Suture placement and reduction with a blunt elevator. B. Final construct.

2.3. Outcome measures

With regards to functional outcome shoulder specific, validated instruments; the Oxford Shoulder Score (OSS) questionnaire and Constant-Murley score (CM) were used.^{16, 17, 18} The patients were re-examined by 2 of the authors. CM strength was measured with a Nottingham Mecmesin myometer (Mecmesin, Slinfold, UK). A Numeric Rating scale (NRS) in which the patients graded pain during activity, at rest and at night on a ordinal scale between 0 and 10 was used and the patients also filled in the Short Portable Mental Status Questionnaire (SPMSQ) according to Pfeiffer, to assess cognitive function.^19,20 All plain standard shoulder radiographs in three projections were reviewed and independently assessed by three of the authors. In case of disagreement, the radiographs were re-assessed and a consensus decision was made. There were only minor disagreements before consensus was reached. Data was collected regarding fracture union, re-displacement, avascular necrosis (AVN) and resorption of the greater tuberosity.

2.4. Statistics

Wilcoxon’s signed rank test was used for Constant-Murley score comparison.

3. Results

3.1. Clinical outcomes

All patients were found to have intact rotator cuff tendons during surgery. The median OSS was 45 (range 5–48). The median CM score on the injured side (63, range 21–98) was significantly lower than on the uninjured (81, range 46–100), p = 0.001. The median NRS (pain) was 0 (range 0–6) during activity, 0 (range 0–5) at rest and 0 (0–7) during sleep. Only one patient used pain medication (paracetamol) regularly for shoulder pain (Table 2). All seven patients who had an occupation before the injury could return and perform the same work tasks after rehabilitation. The Short Portable Mental Status Questionnaire (SPMSQ) showed overall high scores: median 10 points of 10 (range 8–10)²⁰.

Table 2.

Patients with clinical data at follow-up.

Patient number	Age at follow-up	OSSa (n = 21)	CMb score injured side (n = 20)	CMb score non-injured side	CMb score ratio injured/non-injured (%)	Pain activity NRS (n = 21)	Pain at rest NRS	Pain at sleep NRS	AVNc T = total P = partial
1	56	45	72	86	84	0	0	1
2	66	48	60	46	100d	0	0	1
3	54	48	82	82	100	0	0	0
4	78	48	32	67	48	0	0	0	P
5	75	48	62	77	81	0	0	0
6	67	37	70	87	80	0	0	0
7	63	48	82	83	99	0	0	0
8	73	28	21	70	30	4	1	1	T
9	76	48	73	59	100d	0	0	0
10	64	–	64	84	76	–	–	–
11	60	48	58	81	72	0	0	0
12	71	47	75	85	88	0	0	0
13	53	42	56	81	69	1	1	0	P
14	65	34	58	91	64	1	0	1	P
15	66	48	80	82	97	0	0	0
16	56	45	98	100	98	0	0	0
17	75	5	57	74	77	3	1	5
18	81	36	55	68	81	2	1	1
19	70	43	76	76	100d	1	0	1
20	68	48	–	–	–	0	0	0
21	76	48	–	–	–	0	0	0
22	54	45	55	83	66	6	5	7
Median values	67	45	63	81	81	0	0	0

^aOxford shoulder score.

^bConstant-Murley score.

^cAvascular necrosis.

^d> 100 = 100%.

3.2. Radiological findings

Radiological fracture healing occurred in 27 of the 27 patients (Table 3, Fig. 1b). Three out of the 27 examined patients had an AVN of the humeral head, 2 partial and 1 total. One of the patients with a partial AVN also had a mal-union with a displaced humeral head. Two patients had partial resorption of the greater tuberosity. One of these 2 patients also had mal-union with re-displaced greater tuberosity. There were no post-operative infections and no re-operations.

Table 3.

Radiological findings at follow-up (n = 27).

Healed fractures	27
Re-operation	0
AVNa	4 (15%) (3 partial and 1 total)
Partial resorption GTb	3

^aAvascular necrosis.

^bGreater tuberosity.

4. Discussion

This study shows that despite severe displacement adequate fracture stability can be achieved by simple means with reliably good functional and radiological outcomes as well as few reoperations. The use of minimal implants seems a great advantage when treating these fractures by considerably reducing the risk of hardware complications such as screw penetration into the glenohumeral joint and discomfort from the plate which otherwise may call for re-operations and risk further complications.^{7,10, 11, 12}

Blood supply to the humeral head in these cases most probably comes from the posteromedial artery which may be preserved if medial ad latus displacement is small,⁹ as was the case our series. The incidence of AVN in valgus impacted humeral fractures varies between 8 and 26% in the literature.^5,21,22 We found signs of AVN in 15% if both total and partial necroses are included. This did however not seem to have a negative impact on clinical outcome. In their respective series also using only sutures Hockings and Haines (11 patients, mean age 55 years) described 1 total and 1 partial AVN and Panagopoulus et al. (16 patients, mean age 45 years) described 1 partial AVN.^23,24 In all three studies AVN seems to be a minor clinical problem which is an observation supported also by others.²⁵ One case in our series was afflicted with a complete AVN (about 4% of cases), a relatively moderate frequency when compared to other frequencies reported in the literature, which ranges from 1 to 35% in operatively treated fractures of multiple morphologies.²⁶

Conservatively treated, these displaced fractures result in relatively poor range of motion compared to other types of proximal humeral fractures.³ The Oxford and Constant ratio scores showed high values indicating good functional results which tallies with other studies.^6,21,23,24 The pain scores presented low values in general and this result may explain why no re-operations were performed.

Similar techniques with the intention of preserving the blood supply to the humeral head and have been described.^4,5,23,24 Sutures can be placed through the tuberosities, the humeral head and the humeral shaft rather than the cuff tendons as described by Panagopoulus et al.²⁴ or as Hockings and Haines connecting the tuberosities with figure-of-eight sutures.²³ We slightly modified this technique adding sutures from cuff tendons to the humeral shaft to further antagonise superior deforming forces.

Our surgical strategy in these fractures has been to take advantage of the strong tendinous tissue close to the tendon insertions at the tuberosities. All patients in our series had intact rotator cuff tendons. In our experience a full thickness rotator cuff tear is rare in any Neer 3- or 4-part fracture possibly reflecting cuff contraction being important for injury pattern progression together with glenoid compression forces.²⁷

We abstained from using cement, bone fillers or bone graft, but did not experience problems with re-displacement of the humeral head, which was only supported by the reduced tuberosities and the medial hinge. Thus, fixation with solely non-absorbable sutures seems to be a stable enough fixation method during healing in this type of fractures.

There are several limitations to this study, the study population is quite small, the study design is retrospective and there is a lack of a control group. The gender ratio may appear skewed (23 women/4 men), but this mirrors the overall incidence in the Scandinavian countries (75% women) also taking into account that more women than men in these studies undergo surgical treatment.^28,29 The radiographic assessment was standard x-rays, in which rotation of the humeral head is difficult to appreciate which makes measurements of head-shaft angle less reliable. Also we cannot reliably account for cases during these years possibly being converted during surgery, e.g. to plate fixation or arthroplasty. However, patients suitable for this method apparently do well. The strengths are a combination of both clinical and radiological follow-up and a larger series than previously reported of a similar technique.

5. Conclusions

Open reduction and the use of sutures for fixation of a valgus-impacted proximal humeral fracture provided a relatively reliable outcome with good shoulder function, no reoperations and infrequent pain. AVN was found at a relatively low rate and did not appear to significantly affect outcome. Our study supports that valgus-impacted proximal humeral fracture with minor ad-latus displacement may be fixed without metal hardware.