The VIDIA BONE score as a patient- and fracture-related risk score for the occurrence of fracture sequelae in proximal humerus fractures

Abstract

Purpose

The aim of the study was to determine preoperative patient- and fracture-related risk factors for estimating the risk of fracture sequelae after surgically treated proximal humerus fractures (PHF) using locking plate osteosynthesis. The purpose was to develop a fracture sequelae risk score as an additional tool to facilitate the treatment strategy for PHF.

Methods

All patients with PHF who underwent surgery with locking plate osteosynthesis were included. Inclusion criteria were complete preoperative parameters including general patient-specific information, preoperative radiographs and computed tomography (CT) scans. The patients were divided into 2 groups according to the occurrence of fracture sequelae (Group I: no fracture sequelae, Group II: fracture sequelae). Following risk factors for fracture sequelae were defined: Vascular supply interruption risk [V], Individual (female gender) [I], Diabetes [D], ISAR (Identification of Seniors at risk) [I], ASA (American Society of Anesthesiologists) score [A], Body mass index [B], Osteopenia [O], Nicotine abuse [N], Elderly > 65 [E]. The VIDIA BONE score contains 9 items. For each parameter of the VIDIA BONE score, a value between 1 and 3 was assigned based on the odds ratio (OR).

Results

This comparative case series study included 112 patients, with 82 (73.2%) in Group I and 30 (26.8%) in Group II. All defined preoperative risk factors were associated with an increased risk of fracture sequelae (FS), with odds ratios (OR) ranging from 1.3 to 3.9, although not all parameters reached statistical significance. The maximum score was 24 points, categorized as follows: scores of 1 to 8 indicating low risk, 9 to 16 indicating moderate risk, and 17 to 24 indicating high risk of fracture sequelae following PHF. Patients with a VIDIA BONE score of 1 to 8 points demonstrated favorable postoperative outcomes, with FS observed in only 6% of cases. In contrast, 84% of patients with a score of 17 to 24 developed a FS. Additionally, 90% of all patients who developed FS had a VIDIA BONE score more than 9 points.

Conclusion

The VIDIA BONE score appears to be a simple, reproducible, and valuable tool for supporting reliable treatment decisions. Locking plate osteosynthesis is an effective treatment option for patients with a low risk VIDIA BONE score. However, patients with a high risk VIDIA BONE score are more likely to experience FS and osteosynthesis failure.

Level of evidence

III.

Introduction

Proximal humerus factures (PHF) with a relevant fragment displacement are often treated with open reduction and internal fixation. Although locking plate osteosynthesis is the most common surgical procedure, high complication rates of up to 49% have been observed, particularly in geriatric cohorts [1–3]. Common complications after joint-preserving treatment were classified as fracture sequalae (FS) of the proximal humerus by Boileau et al. and categorized into 4 types including avascular head necrosis (AVN) or humeral head collapse, locked dislocation, surgical neck nonunion and tuberosity malunion [4]. These types are often associated with other implant-associated complications such as loosening, dislocation or fracture or osteochondral complications such as bone resorption and post-traumatic osteoarthrosis [5–7].

If an anatomical reduction of the PHF is not possible or is very prone to complications, primary arthroplasty should be considered. It should be noted that primary reverse shoulder arthroplasty (RSA) for multifragmentary PHF in older patients is associated with better clinical outcomes, a lower complication rate and a reduced revision rate compared to secondary RSA [8–10].

Treatment decision should always be made on an individual basis and the patient should be involved in a shared decision-making process. Several factors appear to be decisive for the choice of treatment. These include patient-specific factors such as age, gender, previous illnesses, comorbidities, general health, physical activity and mobility. In a multicenter study of around 4,000 patients, patient-related parameters like female gender, age of more than 65 years, smoking, diabetes or reduced general health were identified as risk factors [11]. In addition to patient-related factors, the fracture morphology is crucial for the functional outcome [12]. Defining these various factors and quantifying them in the form of a preoperative score could be a valuable tool for both the surgeon and the patient in selecting the most appropriate treatment strategy.

The aim of the study was to determine preoperative patient- and fracture-related risk factors for estimating the risk for FS after surgically treated PHFs using locking plate osteosynthesis. The purpose was to develop a FS risk score as an additional tool to facilitate the treatment strategy for PHF.

Methods

This is a retrospective comparative case series study. All patients signed a declaration of consent before being included. A positive vote from the responsible ethics committee is available (Ethics Committee of Heidelberg University II, Ethical Approval No.: 2024 − 824). The follow-up examination respects the ethical standards in the Helsinki Declaration of 1975, as revised in 2000, as well as the national law.

Study population

All patients with PHF who underwent surgery with locking plate osteosynthesis (PEEKPower™ Humeral Fracture Plate, Arthrex, Naples, USA) by 1 certified shoulder surgeon (L.L.) were included. To be included for the study, patients were required to have complete preoperative data, including general patient-specific information and radiographs and computed tomography (CT) scans, as well as a radiographic follow-up of at least 12 months or data from an interim revision surgery. Excluded were patients with multiple fractures in the affected limb, fracture dislocations, pathologic fractures, or those unable to attend follow-up due to medical reasons.

In postoperative radiographs and CT scans, the occurrence of fracture healing was analyzed and classified according to the Boileau classification [4]. The patients were divided into 2 groups according to the occurrence of fracture sequelae (Group I: no fracture sequelae, Group II: fracture sequelae).

Indication for locking plate osteosynthesis was made for displaced PHFs (2-/3-/4-part). Treatment decisions were made based on fracture morphology, bone quality, and patient-specific factors such as age, activity level, and comorbidities. In cases where reconstructive treatment of the PHF was deemed unfeasible or associated with a high risk of complications or in patients with painful preoperative osteoarthritis, primary joint replacement was considered.

The surgeries were performed under general or regional anesthesia with an interscalene plexus block. Patients were positioned in the beach chair position, and a deltopectoral approach was used. Based on fracture morphology, either a 3-hole or 5-hole plate with locking screws (3.5/ 4 mm) was used. A suture cerclage of the greater and/or lesser tuberosity (FiberWire^®, Arthrex^®, Naples, USA), screw fixation of the lesser tuberosity, and a tenotomy or tenodesis of the long head of the biceps tendon were performed as needed.

The VIDIA BONE score

Several clinical and radiological parameters that influence fracture healing were identified in a literature review as relevant factors for estimating the risk of failed fracture healing after PHF (Table 1) [5, 7, 11–13]. Gender, age at trauma and concomitant diseases (obesity, diabetes, nicotine abuse) were assessed. The American Society of Anesthesiologists (ASA) classification (range I to IV) was recorded as a score for the general physical health of the patients and to estimate the general preoperative risk [14]. The Identification of Seniors at risk (ISAR) test was used for geriatric screening [15]. The test comprises 6 items, whereby a total score of 6 can be achieved (range 0 to 6). A score of 2 points or higher indicates a high probability of special geriatric needs , which must be confirmed by a comprehensive geriatric assessment.

Table 1.

Study cohort, odds ratio and VIDIA BONE score value determination

	Group I (n=82)	Group II (n=30)	P value	Odds Ratio	95% CI	VIDIA BONE score
Vascularization supply interruption risk
Metaphyseal head extension < 8mm, n (%)	35 (43)	21 (70)	.0101	3.1	1.2, 7.7	3
Medial hinge disruption, n (%)	57 (70)	27 (90)	.0271	3.9	1.1,14.2	3
Anatomical neck, n (%)	37 (45)	22 (73)	.0081	3.3	1.3, 8.4	3
Individual (female gender), n (%)	62 (76)	25 (83)	.3851	1.6	0.5, 4.7	1
Diabetes, n (%)	9 (11)	7 (23)	.098 1	2.5	0.8, 7.4	2
ISAR score > 1, n (%)	16 (20)	10 (33)	.1251	2.1	0.8, 5.3	2
ASA score > II, n (%)	15 (18)	10 (33)	.0901	2.2	0.9, 5.7	2
BMI > 30 kg/m2, n (%)	12 (15)	11 (37)	.0111	3.4	1.3, 8.8	3
Osteopenia (DTI ≤ 1.4), n (%)	21 (26)	15 (50)	.0141	2.9	1.2, 6.9	2
Nicotine abuse, n (%)	7 (9)	5 (17)	.2181	2.1	0.6, 7.4	2
Elderly > 65, n (%)	41 (50)	17 (57)	.5321	1.3	0.6, 3.0	1
Total VIDIA BONE score, mean (SD)	8.1 (3.9)	12.6 (3.8)	<.0012	-	-	24

Group I: no fracture sequelae, Group II: fracture sequelae

95% CI 95% confidence interval, ASA American Society of Anesthesiologists, BMI Body mass index, ISAR Identification of Seniors at risk, DTI Deltoid tuberosity index, mm Millimeter, n Number, SD Standard deviation

¹Pearson’s chi-squared test, ²Wilcoxon-Mann-Whitney test

Preoperative radiographs (true anterior posterior (ap) view, lateral (Y) view) and computed tomography (CT) scans were evaluated. To evaluate the bone quality of the proximal humerus, the deltoid tuberosity index (DTI) was determined using the ap radiographs [16]. Values lower than 1.4 indicated low local bone mineral density of the proximal humerus and were defined to be a risk factor for impaired fracture healing und osteosynthesis failure. The fracture pattern was analyzed according to the Hertel criteria . The fracture patterns included in the VIDIA BONE score were metaphyseal head extension (< 8 mm), medial hinge disruption and fracture of the anatomical neck (fracture types 2, 8, 9, 10, 11, 12 according to Hertel classification) [12, 17].

Statistical evaluation

The nominal variables were summarized as percentages. The arithmetic mean and its standard deviation were used for descriptive statistics. Continuous and ordinal variables were grouped with medians and ranges. Shapiro-Wilk test was used to test the normality of the variables. Wilcoxon-Mann-Whitney test (U test) was used for continuous variables based on distribution normality. For testing the association of two ordinal or nominal variables, the Pearson’s chi-squared test was calculated. For exploratory statistics, the level of significance was set for p < 0.05.

The odds ratio (OR) was used to quantify the association of qualitative variables. Based on the OR results, specific point values were assigned to the scoring parameters. An OR between 1 and 2 was assigned 1 point, an OR between 2 and 3 was assigned 2 points, and an OR greater than 3 was assigned 3 points. The 95% confidence interval (CI) was assessed for the continuous variables.

Results

Study population

The study included 112 patients, with 82 (73.2%) patients in Group I and 30 (26.8%) patients in Group II. The mean age was 63.2 (SD 13.3) and 66.5 (SD 10.8), respectively (p = .452). The mean BMI was 25.5 (SD 4.9) and 27.1 (SD 6.3), respectively (p = .027). The DTI was 1.49 (SD 1.61) and 1.52 (SD 1.82), respectively (p = .015). Further preoperative parameters are shown in Table 1. In Group II, there were 29 (96.6%) type 1 and 1 type 4 (3.3%) FS according to the Boileau classification. In Group I, 2 revision surgeries (1.8%) involving implant removal and arthrolysis were found. In Group II, 24 (80.0%) revision surgeries were performed, resulting in arthroplasty in 21 cases and implant removal in 3 cases.

Quantification of the VIDIA BONE score parameter

Based on the OR, specific point values were assigned to each parameter of the VIDIA BONE score. Female gender (OR 1.61) and age over 65 (OR 1.31) were each assigned 1 point. Diabetes (OR 2.47), nicotine abuse (OR 2.14), an ASA score greater than II (OR 2.23), an ISAR score greater than 1 (OR 2.06), and a DTI ≤ 1.4 (OR 2.91) were each assigned 2 points. Obesity (OR 3.38) and each of the Hertel criteria (OR 3.13–3.95) were assigned 3 points. The maximum possible score was 24 points (Table 1). Figure 1 illustrates the percentage distribution of the VIDIA BONE score across the two study groups.

Fig. 1.

Percentage distribution of the VIDIA BONE score in the two study groups (Group I: no fracture sequelae, Group II: fracture sequelae)

Determination of VIDIA BONE score risk levels

The mean VIDIA BONE score was 8.1 (SD 23.9) in Group I and 12.6 (SD 3.8) in Group II (p < .001). A VIDIA BONE score of 8 or lower was observed in 59% of Group I and 10% of Group II (Fig. 2). Scores ranging from 9 to 16 were found in 38% of Group I and 70% of Group II. A VIDIA BONE score of 17 or higher was recorded in 4% of Group I and 20% of Group II. Cohorts were categorized based on VIDIA BONE score values (Table 2). It was concluded that VIDIA BONE scores of 1 to 8 indicate a low risk, 9 to 16 indicate a moderate risk, and 17 to 24 indicate a high risk for the occurrence of FS. The overall sensitivity and specificity of the VIDIA BONE score were found to be 90% and 59%, respectively. Figure 3 provides an overview of the VIDIA BONE score, including the risk groups, which can serve as a template for clinical practice.

Fig. 2.

Bar chart of the cohort distribution by VIDIA BONE score risk groups (Group I: no fracture sequelae, Group II: fracture sequelae)

Table 2.

The VIDIA BONE score risk groups

	Group I (n=82)	Group II (n=30)	P value
VIDIA BONE score 1 to 8 (low risk)	48 (59)	3 (10)	<.0011
VIDIA BONE score 9 to 16 (moderate risk)	31 (38)	21 (70)
VIDIA BONE score 17 to 24 (high risk)	3 (4)	6 (20)

Group I: no fracture sequelae, Group II: fracture sequelae, n Number

¹Pearson’s chi-squared test

Fig. 3.

VIDIA BONE scoring system for clinical use

Discussion

The primary aim of this retrospective comparative case series study was to evaluate patient- and fracture-related risk factors for FS in a cohort of 112 patients treated with locking plate osteosynthesis for PHF. Patients were categorized into 2 groups based on the occurence of FS. The analysis demonstrated that the preoperative VIDIA BONE score is an effective prognostic marker for assessing the risk of FS. The VIDIA BONE score, designed for easy memorization and practical use in clinical settings, demonstrates high sensitivity, making it highly effective at identifying patients at risk for FS. However, its moderate specificity suggests that it is less reliable in accurately excluding patients who are not at risk.

The preoperative patient- and fracture-related variables of the VIDIA BONE score were established based on literature findings and the authors’ experience [5, 7, 11–13]. The VIDIA BONE score comprises 9 items with a total score of 24 points. All preoperative risk factors identified in this study were associated with an increased risk of FS, with OR ranging from 1.3 to 3.9. Patient age over 65 years was identified as a risk factor for FS, which is consistent with the current literature [11]. Additionally, the risk of overall complications following osteosynthesis of multifragmentary PHFs is more than three times higher in patients over 60 [18]. Moreover, in the study by Henkelmann et al., female gender was identified as a risk factor for failure of osteosynthesis (OR 1.8). This finding was corroborated by our study, which reported an OR of 1.6 for female gender. One possible explanation is the association between female gender and osteoporosis, which is more prevalent in women after the menopause. PHF is a typical osteoporosis-related fracture, commonly observed in elderly patients, particularly among women [7]. Osteoporosis can complicate surgical fixation and healing, significantly increasing the risk of loss of reduction [19]. In many patients with PHF, osteoporosis has not been investigated previously. However, it is important to assess the bone quality in patients with PHF. A conventional bone density test such as a dual-energy X-ray absorptiometry (DEXA) is usually not feasible and could unnecessarily delay the time to surgery. In this context, the DTI is a validated and simple instrument for assessing local bone quality in the proximal humerus, making it essential for clinical decision-making [16]. The smaller the DTI value, the thinner the cortex, and therefore, the greater the correlation with osteopenia. Clinical studies have shown, that the DTI has predictive value for assessing the risk of osteosynthesis failure and of screw cut out [20, 21]. In this study, the DTI was identified as a risk factor for FS (OR 2.9) and was consequently incorporated into the VIDIA BONE score.

In addition to age, female gender and bone density, comorbidities such as diabetes mellitus, obesity or smoking can influence the risk of osteosynthesis failure. In the multicenter study by Henkelmann et al., smoking and diabetes were identified as risk factors [11]. Interestingly, BMI was not identified as a risk factor in their analysis, which contrasts with to the findings of the present study and other studies [11]. Furthermore, validated scores such as the ASA score for general health status and the ISAR score for geriatric screening were deemed highly relevant. Key factors in determining the optimal treatment approach include the patient’s overall health, functional needs, compliance, and physical capabilities. Additionally, the feasibility of restrictive follow-up care must be considered, particularly in relation to the patient’s home care situation. For example, a patient living alone with limited mobility may find prolonged immobilization challenging and would likely benefit from early postoperative mobilization. In the context of RSA, early functional rehabilitation is often feasible [22]. In contrast, with osteosynthesis, such early mobilization may not always be advisable.

When selecting the fracture-specific parameters, the major risks defined for the occurrence of AVN, as defined by Hertel et al., were taken into account [12, 17]. These include fractures of the anatomical neck, a posteromedial metaphyseal head fragment length of less than 8 mm and disruption of the medial hinge of the proximal humerus. These criteria were used to predict the ischemia of the humeral head in PHFs [12]. A recent systematic review by Oldrini et al. found that AVN was the most common non-implant-related complication after PHF, occurring in 17.9% of cases [23]. If perfusion of the humeral head is impaired, the risk of AVN and subsequent collapse of the humeral head is increased. AVN was the most common FS in Group II of this cohort (96.6%). However, the influence of the Hertel criteria on the occurrence of FS, particularly AVN, remains controversial. Studies have reported moderate reliability of these criteria, as well as a weak predictive value [13, 24, 25]. However, fracture patterns, particularly fractures of the anatomical neck and the disruption of the medial hinge, appear to have a predictive value [11, 26]. In the study presented, all 3 selected fracture-related parameters demonstrated a high OR (3.1–3.9), and were therefore appropriately included in the VIDIA BONE score.

Various algorithms exist for the treatment of PHF, particularly for elderly patients [27–29]. Most algorithms primarily focus on patient’s age and fracture pattern, but they may not fully address individual preoperative factors that are crucial for decision-making. Developing a scoring system that integrates anamnesis, patient-specific factors, and fracture characteristics could improve the prediction of FS and thus facilitate the decision-making process. The analysis conducted in this study led to the development of the VIDIA BONE score for predicting FS risk and provided a framework for defining risk levels for classification purposes. A VIDIA BONE score of 1 to 8 indicates a low risk, 9 to 16 a moderate risk, and 17 to 24 a high risk of fracture sequelae after PHF. Patients with a low risk VIDIA BONE score experienced a successful fracture healing in 94%. In contrast, the risk of FS was 84% for patients with a high risk VIDIA BONE score had a . Among those patients with FS, 90% had a moderate or high risk VIDIA BONE score. Therefore, it can be concluded that a joint-preserving procedure is preferable for patients with a low-risk VIDIA BONE score, while those with a high risk VIDIA BONE score may benefit more from RSA. For patients with a moderate risk, the benefits and drawbacks of treatment options should be thoroughly discussed with the patient. A recommendation based solely on the VIDIA BONE score should not be made. Instead, it is important to emphasize that the score serves as a supplementary tool to aid in treatment selection and should not replace the need for individualized patient consultation and decision-making.

The study has several limitations. Although the study was homogenous, with all patients treated via locked plate osteosynthesis, it does not include patients who received conservative treatment or other joint-preserving procedures apart from locking plate osteosynthesis. Therefore, the choice of technique might have a significant impact on the occurrence of FS. Furthermore, the selection of VIDIA BONE parameters was based solely on a literature review and the author’s experience, meaning they were subjectively chosen. Apart from the definition of the FS, specific complications such as failure of osteosynthesis (screw cut-out, screw displacement, malposition, non-anatomical reduction, etc.) were not explicitly considered as independent complications. HoweverNevertheless, all patients with a screw cut-out, for example, were assigned to Group II, as they all also presented with FS.

The VIDIA BONE score appears to be a valuable tool, providing a simple and objective method without the need for complex diagnostic procedures. It is currently the first scoring system to integrate various patient- and fracture-specific parameters to aid in treatment decisions. Nevertheless, further validation is necessary. Prospective studies involving larger patient cohorts and machine learning models may be valuable for this purpose. In addition, the risk of FS associated with other joint-preserving treatments, aside from locking plate osteosynthesis, including conservative approaches, should be further investigated. Future research should also investigate the inclusion of additional items in the VIDIA BONE score to enhance the assessment of moderate and high risk cases.

Conclusion

The VIDIA BONE score appears to be a simple, reproducible, and valuable tool for supporting reliable treatment decisions. Locking plate osteosynthesis is an effective treatment option for patients with a low risk VIDIA BONE score. However, patients with a high risk VIDIA BONE score are more likely to experience FS and osteosynthesis failure.

Abbreviations

ap

Anterior posterior

ASA

American Society of Anesthesiologists

BMI

Body mass index

CT

Computed tomography

CI

Confidence interval

DTI

Deltoid tuberosity index

ISAR

Identification of Seniors at risk

FS

Fracture sequela

n

Number

OR

Odds ratio

PHF

Proximal humerus fracture

RSA

Reverse shoulder arthroplasty

SD

Standard deviation

Authors’ contributions

Conceptualization: MK, NL, JS, VR, CG, LL; Data curation: MK, NL; Formal Analysis: MK; Funding acquisition: MK, LL; Investigation: MK, NL, VR, JS; Methodology: MK, NL, JS, VR; Project administration: MK, LL; Resources: MK; Software: MK; Supervision: CG, LL; Validation: CG, LL; Visualization: MK; Writing – original draft: MK; Writing – review & editing: MK, NL, JS, VR, CG, LL.

Funding

The research foundation received financial payments of Fa. Arthrex (Munich, Germany) for the conduct of this study. The authors, their immediate family did not receive any financial payments or other benefits from any commercial entity related to the subject of this article.

Data availability

Patient consent does not include publication of the data, so publication of the raw data is not possible. The datasets generated and/or analysed during the current study are not publicly available due not written consent of the patients but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the local Ethics committee of the University of Mannheim/Heidelberg, Germany. Written informed consent was obtained from all patients. Ethics approval number is 2024-824. The conducted experiments respect the ethical standards in the Helsinki Declaration of 1975, as revised in 2000, as well as the national law.

Consent for publication

All subjects have been informed about the open access publication. A declaration of the informed consent form for publication was signed by the subject.

Competing interests

Lars-Johannes Lehmann is a consultant for Arthrex and DJO. Christian Gerhardt is consultant for DJO. Michael Kimmeyer, Norbert Liebherr, Jonas Schmalzl and Verena Rentschler have no competing interest to disclose. The research foundation received financial payments of Fa. Arthrex (Munich, Germany) for the conduct of this study.

Funding

The research foundation received financial payments of Fa. Arthrex (Munich, Germany) for the conduct of this study. The authors, their immediate family did not receive any financial payments or other benefits from any commercial entity related to the subject of this article.