Complications of Locked Plating for Proximal Humeral Fractures in Active Patients in a Level One Trauma Center

Abstract

Background

The incidence of complications associated with locked plating of proximal humerus fractures (PHF) has been reported in up to 40% of surgeries. This study aimed to analyze the incidence and risk factors of complications and re-intervention associated with locked plating of PHF in a an young active working population.

Methods

This retrospective cohort study included patients indicated for locked plating of displaced PHF at a dedicated workers’ accident trauma center. The variables analyzed were patient comorbidities, fracture characteristics, and immediate radiologic surgical outcomes. Postoperative complications and risk factors were determined.

Results

A total of 127 patients with a median follow-up of 31 months and a median age of 52 years were included. The complication and reintervention rates were 13.4% and 12.6%, respectively. The main cause of reintervention was post-traumatic stiffness. The incidence of other complications was 4% screw protrusion, 1.6% avascular necrosis, 1.6% varus collapse. Complications were not associated with age, comorbidity, fracture classification, bone mineral density, Hertel's risk factors, presence of calcar comminution, reduction quality, and calcar screw position.

Discussion

The incidence of complications and reintervention was low. The main cause of reintervention was persistent stiffness, and no risk factors for complications were found in this study.

Introduction

Proximal humeral fractures (PHF) are more common in women, with an increase in incidence after 50 years of age.^1,2 They comprise approximately 7–8% of all adult fractures, and the incidence reportedly increases with age. ³ The treatment of PHF is among the most debated of all fracture treatments. ⁴ Most PHFs are stable, minimally displaced osteoporotic fractures in elderly patients and are due to a low-energy fall. ² Most patients with these injuries can regain shoulder function without surgery. Surgical treatment is reserved for patients with displaced fractures who require maximum shoulder function. Open reduction and internal fixation (ORIF) is generally indicated for young or older patients with high functional demands and good bone quality.^5,6

The technique of choice has been highly dependent on the surgeon's assessment of the fracture type, degree of comminution and displacement, bone quality, and comfort with the technique. Despite using multiple types of implants, ORIF with a plate has been the most widely used. When locked screw technology was introduced in the plates for surgical treatment of PHF, it was initially considered the standard for this type of fracture.^7,8 However, after ORIF procedures with locked plates, reoperation rates and global complications of up to 25% and 49%, respectively, have been reported.^9,10 This is particularly because of the loss of reduction with varus misalignment and subsequent screw penetration.^9–11

Whether there is any improvement in our complication rate after locked plate fixation for proximal humeral fractures over the last few years has yet to be determined. This study aimed to evaluate the complication and reintervention rates after locked plate fixation of PHFs in a single trauma center among active working-age patients. We hypothesize that the complication and reintervention rates have decreased due to improved knowledge regarding this fracture entity, such as identifying risk factors for failure, the use of primary arthroplasty, and the overall learning curve.

Materials and methods

A retrospective cohort study was done in a workers’ accident trauma center, which provides treatment to injured active workers under the national workers’ accident insurance law. The study included all patients who were treated between January 2015 and December 2018 with ORIF for a displaced PHF using a PHILOS (Synthes, Oberdorf, Switzerland) locking plate. The patients’ file records and imaging data were reviewed by a blinded shoulder surgeon. All patients underwent preoperative shoulder radiography and computed tomography. Only patients with a minimum clinical and radiological follow-up of one year were included. Patients were prospectively observed at the time of operation and followed up longitudinally. Cases of isolated and displaced greater tuberosity fractures were excluded from the study.

Perioperative data, including patient characteristics and fracture-specific features, were recorded during the inpatient stay. Variables analyzed were age, comorbidities (tabaquism, alcoholism, diabetes, body mass index [BMI] > 30 kg/m²), fracture characteristics (Tingart index, deltoid tuberosity index [DTI], head-shaft angulation [HSA], calcar comminution, medial head extension, medial hinge disruption), and immediate radiological surgical outcomes in a true anteroposterior shoulder radiograph (HSA, calcar screw position, reduction outcome). Fractures were classified according to the Neer and Mayo/FJD classifications. ¹² The Mayo/FJD classification considers seven fracture patterns: isolated fractures of the greater or lesser tuberosity (GT, LT), fractures of the surgical neck (SN) with or without metaphyseal extension, fractures at the anatomic neck with varus and posteromedial (VPM) or valgus (VL) displacement of the head, and fractures with head dislocation (HD), splitting (HS), or depression/impaction (HI). GT and LT fractures were initially excluded.

In all patients, anteroposterior and lateral-Y view radiographs were assessed after surgery and at every follow-up for radiographic complications, fracture alignment, and healing. Whenever suspected, additional computed tomography was performed. Calcar screw positioning was considered acceptable if the calcar distance was 12 mm or less and the calcar ratio was within the bottom 25% of the humeral head. ¹³

Postoperative complications recorded during follow-up included infections, screw protrusions, avascular necrosis, varus and valgus displacement, non-union, and any cause for reintervention. Implant failure was defined as the occurrence of any of the following: loss of reduction and fracture collapse (with or without secondary screw intra-articular penetration) and screw back-out (complete dissociation of the locking threads from the plate). Loss of reduction and fracture collapse was defined as a change in the humeral head-shaft angle of more than 20° compared with the intraoperative reduction assessed fluoroscopically. ¹⁴ Avascular necrosis was evaluated based on a method described by Hattrup and Cofield. ¹⁵ Fracture reduction was assessed by postoperative reduction by an independent evaluator. Criteria for sufficient reduction were an apex-tuberosity distance (ATD) between −3 mm and +3 mm; medial hinge displacement ≤ 2 mm, HSA between 120° and 150°, and having a one-part fracture according to Neer (< 45°, < 1 cm displacement). A partial reduction was considered as having two out of three of the following: ATD between −3 mm and +3 mm; medial hinge displacement ≤2 mm; and HSA between 120° and 150°. An insufficient reduction was considered when only one or none of the aforementioned postoperative radiological criteria were met.

Statistical analysis

Continuous variables were described as means, standard deviations, and percentages of the total numbers. Statistical analyses and graphs were performed using the SPSS 20 software (IBM Corp, Armonk, NY, USA). The Chi-square test was used to test for statistically different distributions of categorical variables, and logistic regression was performed to analyze the risk factors.

Operative technique

All surgeries were performed by a team of five shoulder surgeons in a single trauma center using the same technique. Surgery was performed under general anesthesia with the patient in the beach chair position on a radiolucent table. An interscalene block was used as the standard perioperative analgesia unless the patient refused. All patients received prophylactic intravenous antibiotics prior to the procedure. ORIF was conducted using a deltoid-pectoral approach. The rotator cuff was evaluated for full-thickness tears, and predominant tuberosities were sutured to the plate using the non-resorbable FiberWire No. 2 (Arthrex, Naples, FL, USA). Bone chip allografts were used in 13.4% of the cases to fill the void after reduction in severe valgus-impacted fractures, and no other augmentation technique was added to the plate ORIF during this period. Accurate fracture reduction and correct screw position were checked during the surgical procedure using multiplane fluoroscopy.

All patients were placed in a shoulder immobilization device for the first three weeks. Hand, elbow, and pendulum exercises were allowed from day one post-surgery. Passive-assisted mobilization was started by a physiotherapist from the third week onwards. The shoulder immobilization device was used for six weeks, and active exercises with a full range of motion, together with active scapular balance exercises, were started. Strength exercises were performed from the 12th week onwards.

Results

Demographic data

The study included 127 patients (females, 61.4%) treated with a locking plate for displaced PHF. The mean age at surgery was 52 ± 15 years (range 18–82). The median follow-up period was 31 months (12–53 months). The demographic data and fracture characteristics are shown in Table 1. None of the patients had an open or pathologic fracture.

Table 1.

Demographic data and fracture description.

		Number	%
Sex	Male	49	38.6%
	Female	78	61.4%
Tabaquism		18	14.4%
Alcohol Consumption		16	12.8%
Diabetes Mellitus		21	16.8%
Rheumatoid Arthritis		4	3.2%
Body Mass Index	≥ 30 kg/m2	8	6.4%
Neer	2	57	46.7%
	3	60	49.2%
	4	1	0.8%
	6	4	3.3%
Mayo	SN	22	18%
	VPM	33	27%
	VL	54	44.3%
	HD, HS HI	13	10.7%
Tingart	≤ 4 mm	1	0.8%
DTI	≤ 1.4	3	2.5%
Medial head extension	< 8 mm	44	36.1%
Medial hinge displacement	> 2 mm	108	88.5%
Anatomic neck fracture		15	12.3%
Calcar comminution	> 1 fragment	93	76.2%

SN: surgical neck; VPM: Varus posteromedial; VL: Valgus impaction; HD: Head dislocation; HS: Head splitting; HI: Head depression/impaction.

Complications and revisions

The overall complication rate of locked plating was 13.4%, and the reintervention rate was 12.6%. The incidence of each complication and the results of the surgeries are shown in Table 2. Age distribution was not associated with complications (p = 0.478) and reinterventions (p = 0.707). Postoperative HSA was not associated with complications (p = 0.839) or reinterventions (p = 0.529).

Table 2.

Incidence of complications and surgical results.

Calcar Screw		82	64.6%
Reduction	Sufficient	73	57.9%
	Partial	42	33.3%
	Insufficient	11	8.7%
Overall Complications		17	13.4%
Reintervention		16	12.6%
Infection		1	0.8%
Screw Penetration		5	4.0%
Avascular necrosis		2	1.6%
Varus collapse		2	1.6%
Non-Union		1	0.8%

Complications and reinterventions were not associated with the presence of tabaquism, alcoholism, diabetes, BMI > 30 kg/m²; Tingart index < 4 mm, deltoid tuberosity index < 1.4, Neer classification, Mayo/FJD classification, calcar comminution, medial head extension < 8 mm, medial hinge disruption, calcar screw position, and outcome of postoperative reduction (Table 3). The main cause of reintervention was post-traumatic stiffness.

Table 3.

Complication and reintervention risk factors.

		Complication	P	Reintervention	P
Risk Factors		χ2		χ2
Tabaquism		0.054	0.817	0.016	0.9
Alcohol Consumption		0.705	0.401	0.574	0.448
Diabetes Mellitus		2.741	0.098	3.333	0.068
Rheumatoid Arthritis		0.551	0.458	0.661	0.416
Obesity	BMI ≥ 30 kg/m2	0.001	0.979	0.002	0.964
Neer 2		1.036	0.309	0.629	0.428
Neer 3		0.735	0.391	0.368	0.544
Neer 4		6.228	0.013	6.68	0.01
Neer 6		0.67	0.413	0.624	0.429
Mayo SN		0.003	0.959	0.031	0.861
Mayo VPM		0.056	0.813	0.165	0.685
Mayo VL		0.076	0.782	0.341	0.559
Mayo HD, HS, HI		0.026	0.873	0.066	0.798
Tingart Index	≤ 4 mm	0.163	0.686	0.152	0.696
Deltoid tuberosity index	≤ 1.4	0.498	0.48	0.464	0.496
Medial head extension	< 8 mm	1.035	0.309	1.551	0.213
Medial hinge displacement	> 2 mm	2.56	0.11	2.387	0.122
Anatomic neck fracture		0.525	0.469	0.712	0.399
Calcar Comminution	> 1 fragment	3.488	0.062	3.12	0.077
Calcar screw position		0.283	0.595	0.553	0.457
Sufficient Reduction		0.954	0.329	0.474	0.491
Partial Reduction		0.544	0.461	0.143	0.705
Insufficient Reduction		0.227	0.634	0.327	0.567
Bone chip allograft		0.307	0.579	0.454	0.5

SN: surgical neck; VPM: varus posteromedial; VL: valgus impaction; HD: head dislocation; HS: splitting; HI: depression/impaction.

Discussion

This study presents an analysis of the incidence and type of complications after locked plating for PHFs in an active population under a worker's compensation insurance at a Level 1 trauma center. The main finding of this study is that the overall complication rate and the number of unplanned surgical reinterventions are lower than those reported in previous systematic reviews.^9,10

Study population and risk factors

Our study group was mainly composed of active middle-aged patients. The characteristics of the included studied group were mainly because the study was conducted in a dedicated workers’ accident trauma center, which provides treatment to injured workers under the national workers’ accident insurance law. During the studied period, ORIF with a PHILOS plate was chosen as the treatment of choice for displaced PHF in patients who needed to maximize their function and return to work. Reverse shoulder arthroplasty was performed in patients ≥ 65 years and a low bone mineral density (Tingart index ≤ 4 mm or DTI ≤ 1.4) with a fracture-dislocation, head split fracture, or a 4-part fracture. No hemiarthroplasty was done during the studied period. Of the included patients, one patient had low bone mineral density according to the Tingart index and three according to DTI. Age less than 65 years and a DTI > 1.4 have previously been shown to be one of the most important factors in achieving an acceptable reduction and avoiding screw penetration. ⁶

In this study, complications and reinterventions were not associated with age, presence of tabaquism, alcoholism, diabetes, BMI ≥ 30 kg/m², Tingart index ≤ 4 mm, DTI ≤ 1.4, Neer 2- and 3-part fractures, Mayo/FJD classification, preoperative or postoperative head-shaft angulation, presence of calcar comminution, medial head extension < 8 mm, medial hinge disruption > 2 mm, calcar screw position, or success of postoperative reduction. Neer 4-part fracture had a significant correlation with complications and reintervention, but as only one case was included, the Chi-square test results may not be valid.

Our results are different from previously reported factors for the failure of locked plating and poor results. Hertel et al. proposed that fractures with metaphyseal extension less than 8 mm, displacement greater than 2 mm of the medial hinge, and anatomical neck fractures have a high risk of ischemia of the humeral head. ¹⁶ Agudelo et al. found that poor reduction in varus with HSA less than 120° was the main risk factor for loss of reduction. ¹⁷ Osterhoff et al. found that fractures with calcar comminution had a greater risk of poor functional outcomes. ¹⁸ Jung et al. demonstrated through multivariate regression that osteoporosis (<−2.5 bone mineral density), varus displaced fracture (HSA < 110°), medial comminution (> 1 fragment), and insufficient medial support (no cortical or screw support) were independent risk factors for loss of reduction in ORIF with locked plates. ¹⁹ Spross et al. described that DTI (> 1.4) and metaphyseal extension > 8 mm were the most significant preoperative predictors for achieving an acceptable reduction. DTI > 1.4, age < 65 years, and a good reduction were independent factors that reduced the risk of screw penetration. ⁶

Most of the previously cited studies were published before this cohort of patients. This has allowed us to acquire the necessary knowledge to face this type of injury in an active population that still needs to maximize their function to return to work. As described by Haasters et al. in a longitudinal cohort study, complications and unplanned reinterventions have decreased over the last decade. ²⁰ The hypothesis for this observation is multifactorial. Improvements in surgical techniques and surgeons’ learning curves are known to be important parameters. Furthermore, the implementation of specialized shoulder surgery departments and the increased knowledge of alternative techniques, such as reverse shoulder arthroplasty (RSA) for complex fractures and low bone mineral density, greatly contribute to the decrease in complications and reoperations. ²⁰

Complications and reintervention

The overall complication and reintervention rates of locked plating were 13.4% and 12.6%, respectively. The main cause of reintervention was a postoperative frozen shoulder that caused disability and did not recover after six months of physical therapy. Screw penetration occurred in 4% of the patients and AVN in 1.6%, with only one of the latter requiring unplanned reintervention. This is similar to the observations made by Haan et al. They showed that within the last five years of their longitudinal study, the loss of fixation rate markedly decreased from 14.3% to 4.8%. ²⁰ In their review of 12 studies of locking plate osteosynthesis for PHF, Sproul et al. found that varus malunion due to loss of fixation occurred in 16.3% of patients. The authors indicated that this complication is of utmost importance because it may lead to screw cutout into the humeral joint and reoperations. ⁹ Complications other than postoperative frozen shoulder that required secondary surgery occurred in 11 patients (8.8%). This low complication rate may be owing to having patients younger than 65 years old who did not have a low bone mineral density rather than surgical technique or fracture type. Of the 27 (21%) patients aged 65 years and above, there were three complications (11.1%). This rate was lower than the rest of the study group, and age distribution was not statistically associated with complications. Patients who were over 65 years old were mostly gardeners, teachers, janitors, cleaning assistants, security guards, and others with administrative jobs. Demographic changes in developing countries have led to an increase in the number of older patients who need to be active and maximize their function to continue their daily activities. Spross et al. recently proposed treatment for PHF using a patient-specific, evidence-based algorithm. They showed that pre-injury quality of life could be restored as measured with the EQ-5D and approximately 90% normal shoulders as measured with the relative constant score and subjective shoulder value. Adherence to the treatment algorithm was associated with significantly better clinical outcomes and substantially reduced complication (16.3% versus 30.8%, p = 0.014) and revision rates (10.6% versus 26.9%, p < 0.001). ²¹

Surgical technique

Tension-relieving rotator cuff sutures were routinely applied to locking plate fixations to transfer deforming forces directly to the plate, thereby neutralizing their tendencies to displace. ²² Adding relieving rotator cuff sutures to locking plate fixation did not improve the stability of the construct in biomechanical models, and comparative clinical studies have not been performed. ²²

If ORIF with a locking plate is performed for a complex PHF, anatomic reduction and stable fixation are crucial to avoid complications. Reconstruction of the medial support reduces the risk of displacement and improves functional results. As medial comminution is common in complex fractures, screws for medial support are essential in the locked plating of PHFs.^19,23 In this study group, 64.6% had a correct calcar screw position, and only 8.7% were considered as malreduction, which may have contributed to the low varus collapse and screw penetration rate.

High complication rates in previous studies have made it necessary to improve the technique with augmentation methods. In 2016, Saltzman et al. conducted a systematic review of four clinical studies which used an endomedular structural bone graft for the treatment of acute PHF. Of the 136 patients, a 3.7% screw penetration and 4.4% reoperation rate were reported with an average follow-up of 20 months. ²⁴ In more recent comparative studies of displaced fractures in older patients, lower rates of complications and loss of reduction were observed in the group with plate plus structural grafts than in the group with plates only.^25,26 This augmentation technique may be used for complex fractures in patients with low bone mineral density in which you want to avoid a reverse shoulder arthroplasty.

The main characteristics of this sample were active middle-aged patients, without bone mineral density deficit and with impacted varus or valgus fractures (68.5%), medial hinge displacement (88.5%), and/or calcar comminution (76.2%). We do not believe that it is necessary to use other augmentation techniques in this population. However, there is room for improvement in the surgical technique. Achieving higher rates of anatomical reductions and calcar screws in the correct position could help further reduce complications; nevertheless, comparative studies are lacking to assess this.

Limitations

The main limitations of this study are its retrospective design and the lack of a control group. The main selection bias in the type of cohort analyzed was given by the aforementioned characteristics of the center where the study was done and complex fractures in patients over 65 with low bone mineral density have been excluded and treated with RSA. Other limitations include a heterogeneous sample of patients, small sample size, and multiple surgeons involved in treating these patients. Comparative analysis with other studies should be performed with caution and you should consider these factors. This study suggests that with improvements in the knowledge to choose the right treatment and techniques for each patient over the past 15 years, the rate of complications due to mechanical failure of locked plating for PHF in selected patients has been lowered.

Conclusions

The incidence of complications and revision surgery associated with locked plating of displaced PHF in these active working-age patients were low. This could be owing to patients’ characteristics, the improvement in the selection criteria for osteosynthesis, increased use of primary RSA for complex fracture types, and more precise surgical techniques for each patient. The main cause of reintervention was persistent stiffness. No risk factors for complications were found in this study.