Prevalence and risk factors for pseudarthrosis in humeral shaft fractures treated by minimally invasive plate osteosynthesis: A 10-year review

Abstract

Introduction

Fractures of the humeral shaft account for about 3% to 5% of all fractures. Although conservative treatment remains the first choice, there are a number of surgical indications. Minimally invasive plate osteosynthesis shows good functional results. However, complications have been reported, such as non-union.

Objective

To assess the prevalence and risk factors for non-union after treatment of humeral shaft fractures using minimally invasive plate osteosynthesis.

Materials and methods

This retrospective study was carried out in patients treated by minimally invasive plate osteosynthesis between 2009 and 2019. Demographic data and variables related to the fracture that could influence bone healing were analysed. The unpaired t-test and Mann–Whitney test were used for the statistical analyses. Categorical variables were analysed using the chi-square test or Fisher's exact test

Results

The study population showed a male predominance (53 of 75, 70.7%). The average time for fracture healing was 19.8 ± 15.3 weeks. Seven patients developed non-union. Only the presence of an open fracture increased the likelihood of non-union.

Conclusions

The prevalence of non-union in patients treated by minimally invasive plate osteosynthesis was 9.3%. Patients with open fractures were six times more likely to progress to non-union. Fracture characteristics (Arbeitsgemeinschaft für Osteosynthesefragen classification, fracture location and plate working length) did not influence progression to non-union.

Introduction

Fractures of the humeral shaft account for 3% to 5% of all fractures, and show a bimodal distribution, with a higher prevalence seen between the second and third decades, and after the sixth decade, of life. ¹

Most humeral shaft fractures are treated conservatively. ² Surgical treatment of humeral shaft fractures can be divided into ‘absolute’ and ‘relative’ types. ¹ The absolute type, involving the use of a dynamic compression plate (DCP), was the method of choice for many decades. Minimally invasive plate osteosynthesis (MIPO) is another viable treatment option, providing relatively good stability with biological preservation of bone; as such, this technique has been widely adopted. ¹

Even with surgical treatment of humeral shaft fractures, non-union may still occur. ³ According to the Arbeitsgemeinschaft für Osteosynthesefragen (AO) group, non-union is defined as persistent fracture without consolidation requiring surgical intervention after the initial treatment.^3,4

Data regarding consolidation of humeral shaft fractures treated by MIPO are inconsistent. Some articles reported 100% consolidation, ⁵ while meta-analyses reported non-union prevalence rates of 2% to 7%.^6,7

There are predisposing factors for progression to non-union, such as open fracture. ⁸ Although poorly studied, the characteristics and location in the diaphysis of the humerus may influence the consolidation of fractures treated by MIPO. Comminuted fractures show less deformity than simple fractures, so require a smaller plate working length. ⁹

This study was performed to assess the prevalence and associations of clinical and radiological factors with non-union in patients with humeral shaft fractures treated by MIPO.

This study was approved by our local ethics committee (approval number 30474720.9.0000.5404).

Materials and methods

In this retrospective study, we analysed the medical records of patients with humeral shaft fractures undergoing MIPO with a DCP plate between 2009 and 2019. The inclusion criteria were as follows: fractures of the humeral shaft located outside Heim’s square (proximal and distal to the humerus; number 12 fracture according to the AO group) ³ ; treated by MIPO using a DCP plate; availability of pre- and postoperative anteroposterior and lateral digital radiographs; and complete clinical records. Patients who did not provide written informed consent, and those with other fracture types, were excluded.

We performed a standard approach using the MIPO technique with two incisions. The proximal incision was a distal deltopectoral approach. The cephalic vein was retracted laterally. Deeply, the pectoral major muscle and the biceps tendon were retracted medially. The insertion of the deltoid was retracted laterally to reach the anterior and proximal aspect of the humerus. The decision of a distal incision was based on the location of the fracture line. Standard humeral shaft fractures with no distal fracture, an anterior incision was made on the arm. The biceps muscle is retracted medially. The lateral antebrachial cutaneous nerve was identified, and the brachialis muscle had undergone longitudinal splitting. The plate was placed on the anterior aspect of the humerus (Figure 1). The radial nerve was protected by the lateral portion of the brachialis muscles. Humeral shaft fractures with distal extension, a slightly proximal incision described by Kocher was made. The plate was placed on the anterior face of the lateral column of the distal humeral shaft (Figure 2) and the radial nerve was protected by a subperiosteal dissection and retraction of the brachioradialis and extensors radialis carpi muscles origin. We used only retractors such as Farabeuf and Langenbeck. We did not use Hohmann retractors to avoid pressing radial nerve against the bone. Livani et al. ¹⁰ published this technique in 2004. We only add a third oblique incision between the brachialis and the brachioradialis muscles to explore the radial nerve in patients with palsy. This additional incision gives the surgeon the possibility to perform the MIPO technique and explore the radial nerve. ¹¹

Figure 1.

The plate placed on the anterior aspect of the humerus.

Figure 2.

The plate placed on the anterior face of the lateral column of the distal humeral shaft.

The radiographs were interpreted by an orthopaedist who did not participate in the surgeries. Consolidation was considered complete when bone callus formation was observed in at least three cortices. The AO group’s definition of non-union was used, that is, non-union is defined as persistent fracture without consolidation requiring surgical intervention after the initial treatment. ⁴

Data including sex, age, fracture type (closed or open) and presence of radial nerve injury at the time of trauma were analysed.

On anteroposterior radiographs, the distance between the fracture line and Heim’s square was measured (Figure 3). More distal fractures had shorter distances than proximal ones. On lateral radiographs, the plate working length was also calculated, defined as the distance between the screws closest to the fracture (proximally and distally) on anteroposterior radiographs (Figure 4). ¹²

Figure 3.

The distance (mm) between the fracture line and Heim's square was measured.

Figure 4.

The plate working length: distance (mm) between the screws closest to the fracture (proximally and distally).

The associations of demographic and radiological data with non-union were examined. Non-categorical variables were assessed for normality using the Kolmogorov–Smirnov test. The unpaired t-test (for parametric variables) and Mann–Whitney test (for non-parametric variables) were used for the statistical analyses. Categorical variables were analysed using the chi-square test or Fisher’s exact test. Statistical analyses were performed using SPSS software (version 27.0; SPSS Inc., Chicago, IL, USA). In all analyses, p < 0.05 was taken to indicate statistical significance.

Results

During the 10-year study period, 146 patients with humeral shaft fractures underwent MIPO at our institute. Of these, 75 were included in the study because they fulfilled the inclusion criteria. Some reasons have forced us to exclude patients from the study. Mainly, patients did not return and were lost to follow-up. Other reasons were fractures with extension to Heim’s square and lesions treated with a locking plate.

The study population showed a male predominance (53 of 75, 70.7%) with a sex ratio of 2.4 men for every woman. The average age of the patients was 35.7 ± 14.4 years, and 58.7% of fractures were on the left side.

There was a predominance of simple compared to complex fractures, especially with the transverse pattern (33.3%). All demographic data are presented in Table 1.

Table 1.

Demographic data of all patients.

Variable	Value
Age (mean ± SD) (years)	35.7 ± 14.4
Time to heal (mean ± SD) (weeks)	19.8 ± 15.3
Sex (no. (%))
Men	53 (70.7)
Women	22 (29.3)
AO Classification (no. (%))
A012A3b	2 (2.67)
AO12A1a	1 (1.33)
AO12A1b	3 (4.0)
AO12A1c	5 (6.67)
AO12A2b	4 (5.33)
AO12A2c	1 (1.33)
AO12A3b	25 (33.3)
AO12A3c	6 (8.0)
AO12B1b	3 (4.0)
AO12B1c	1 (1.33)
AO12B2b	15 (20.0)
AO12B2c	1 (1.33)
AO12B3b	1 (1.33)
AO12C1b	2 (2.67)
AO12C2b	1 (1.33)
AO12C3b	1 (1.33)
AO12C3c	1 (1.33)
AO13A1c	1 (1.33)
AO13A2c	1 (1.33)
Affected side (no. (%))
Right	31 (41.3)
Left	44 (58.7)

SD: standard deviation; AO: Arbeitsgemeinschaft für Osteosynthesefragen.

A total of 10 patients (13.1%) had open fractures. Eleven (14.4%) patients had a radial nerve injury before the surgery, and they were subjected to MIPO osteosynthesis associated with a nerve exploration (Table 2). All of them recovered the radial nerve function.

Table 2.

Rate of radial nerve palsy and fracture classifications.

Variable	Number (%)
Radial nerve palsy	11 (14.4)
Fracture classification
Closed	65 (86.8)
Open	10 (13.1)

Two patients had radial palsy after the MIPO procedure. They were excluded from the study. Although, they were subjected to postoperative ultrasonographic evaluation as described by Livani et al. ¹³ The exam excluded that the radial nerve was entrapped in the fracture and confirmed that it was above the plate. For this reason, they were treated conservatively. One patient recovered all radial nerve function after 3 months and the other after 4 months.

The average time for fracture healing was 19.8 ± 15.3 weeks. Seven (9.33%) patients progressed to non-union. There were no cases of non-union associated with infection. The average time between the first and second surgeries was 10 ± 3 months (range: 6–14 months). Non-union was treated by plate removal. An iliac graft was used in four cases due to hypotrophic non-union, and compression-only of the focus was performed in three cases of hypertrophic non-union. All cases of non-union were treated with a DCP plate.

Factors that could influence the prevalence of non-union were evaluated. There were no significant differences in sex, fracture classification or presence of radial nerve injury between cases that did and did not progress to non-union. However, open fractures showed a 6.5-fold higher likelihood of progressing to non-union than closed fractures (odds ratio (OR) 6.53; p = 0.046). The data are shown in Table 3.

Table 3.

Incidence of non-union according to clinical variables.

		Non-union (7)	Healed (68)	P	OR	CI
Radial nerve palsy (11)		1	10	1.0a	0.96	1.0–8.9
Open fracture (10)		3	7	0.046a	6.53	1.2–35.3
AO classification	A (49)	4	45
	B (21)	2	19	0.68a
	C (5)	1	4
Sex	Male (53)	6	47	0.66a	0.37	0.4–3.2
	Female (22)	1	21

P-values were calculated using Fisher's exact test. Data are numbers of patients. AO: Arbeitsgemeinschaft für Osteosynthesefragen; CI: confidence interval; OR: odds ratio. ^aFisher's exact test.

The average distance between the fracture and Heim’s square was 85.7 ± 45.8 mm. The distance was not associated with the fracture healing time (p = 0.91). The median plate working length was 110.3 mm (range: 5–170 mm), which was also not significantly associated with fracture healing time (p = 0.24).

The mean distances from Heim’s square to the fracture and plate working length were not significantly different between the cases that did and did not progress to non-union (Table 4).

Table 4.

Incidence of non-union according to radiological variables.

	Non-union	Healed	P
Distance from Heim’s square to fracture line (mm)	51.3 ± 26.2	85.7 ± 45.8	0.10a
Plate working length (mm)	90.5 (50–150)	110.3 (5–170)	0.51b

P-values were calculated using the unpaired t-test for parametric variables and the Mann–Whitney test for non-parametric variables. ^aUnpaired t test. ^bMann-Whitney test.

Discussion

Humeral shaft fractures in young males are generally associated with high-energy trauma. ¹⁴ Consistent with the literature of developing countries, our study showed a male predominance. ¹⁵ The most prevalent fracture classification was AO12A3, similar to a previous report by Tsai et al. ¹⁶ Although transverse fractures are considered to be simple fractures, they are normally associated with direct high-energy trauma. ³

Open fractures of the humeral shaft are rare, accounting for only 2% to 5% of humeral shaft fracture cases.^15,17 However, the incidence of bone exposure as a result of traffic accidents is higher in developing countries. ¹⁸ Barbieri et al. ¹⁹ reported an incidence of open fracture of 20.3%. In the present study, the rate of open fractures was 13.1%, similar to data from developing countries. Epidemiologically, our results were similar to studies conducted in the developing countries that did not find a bimodal age distribution. ¹⁸

Conservative treatment is still considered the main method for treating humeral shaft fractures. ²⁰ Surgical treatment is indicated in cases of open fractures, fractures with unacceptable deviations, and cases with associated vascular injuries requiring repair. ²¹ Although the anatomical reduction is still the standard surgical technique for humeral shaft fractures, a number of studies reported good results with MIPO.^22,23 MIPO was developed for secondary consolidation of fractures and bone callus formation, as well as intramedullary nailing. ²²

Despite the benefits of MIPO, complications may still occur, including non-union. In our study, the prevalence of non-union was 9.3%. In a survey of 34 cases, Kulkarni et al. ²⁴ reported that all patients achieved consolidation. In a study of 32 cases, Esmailiejah et al. ²⁵ reported that non-union occurred in 3% of patients. Davies et al. ²⁶ reported a non-union rate of 7% among 15 cases of humeral shaft fracture treated with MIPO. A meta-analysis reported a non-union rate of 2.03% among 197 patients treated using MIPO. ²⁷ However, most studies excluded open fractures, which may explain the higher prevalence of non-union in our study.

Several factors can lead to non-union. Open fractures promote bone devascularisation, affect healing and favour progression to non-union. The rate of progression to non-union of open fractures in our study was six-fold higher compared to closed fractures reported in the literature.^8,28,29

Fractures of the humeral shaft lead to radial nerve damage in 8% to 12% of cases.^11,30 In the present study, radial nerve palsy was found in 14.4% of cases. Despite the higher prevalence of this injury compared to the literature, neurological damage to the radial nerve was not associated with non-union in our cohort. This was consistent with the findings of Livani and Belangero, ¹⁰ who did not observe a greater risk of non-union in cases treated with MIPO and radial nerve exploration. Several studies since 2004 have published about the MIPO. Some systematic reviews showed that the MIPO is safe for the radial nerve. Furthermore, the chance of iatrogenic radial nerve lesion after the MIPO is the same or lower than open reduction and internal fixation or intramedullary nailing.^7,10,30

The plate working length is important for MIPO. Giodarno et al. ¹² reported that a larger plate working length was associated with lower local tension, which could facilitate consolidation through fixation. In the present study, the plate working length was not associated with non-union. We hypothesise that there is a tendency to increase the plate working length in cases of a simple fracture, and vice versa for complex fractures.

The plate working length is important for stabilisation and controlled, microscopic movements in the fracture focus. Theoretically, when the fracture line is closer to Heim’s square, increasing the distance between the proximal and distal screws is less feasible, thus limiting the enlargement of the plate’s working length. We postulated that this could affect fracture consolidation, especially for simple fractures requiring a larger plate working length. However, this hypothesis was not tested. Regardless of the distance between the fracture site and Heim’s square, it is likely that adequate stabilisation was achieved in our cases. Even in cases with more distal fractures, the plate was supported by the lateral column of the humerus. This modification, first reported by Livani et al., ¹¹ allows the establishment of an adequate plate working length favouring fracture consolidation.

This study had some limitations, including its retrospective nature and the lack of evaluation of other clinical factors, such as diabetes, hypertension and smoking. However, to our knowledge, this is the first study to assess the associations’ plate working length and fracture location with the development of non-union in patients with humeral shaft fractures that underwent MIPO.

Conclusions

The prevalence of non-union in patients with humeral shaft fractures treated by MIPO was 9.3%. Patients with open fractures were six times more likely to progress to non-union than those with closed fractures. Fracture characteristics (AO classification, fracture location and plate working length) did not influence progression to non-union.