1 Introduction

Rotational ankle fractures have an incidence of 187 per 100,000 and are one of the most common orthopaedic injuries seen.¹ The functional outcomes of an ankle fracture are worse when a posterior malleolus (PM) fracture is present.² Ankle fractures that involve the posterior malleolus can range from 33 to 46 %, especially in the case of Weber B and C fractures.³ The posterior inferior tibiofibular ligament, attached to the posterior malleolus, is a major contributor to ankle syndesmosis stability.⁴ A preoperative computed tomography (CT) is mandatory to delineate the posterior malleolus’ morphology, size and involvement of the incisura.⁵

The anatomy of the incisura is variable. It can vary in depth, width, shape, version and level of fibular engagement.^6,7 Boszczyk et al. reported that patients with shallow, retroverted and disengaged incisurae would be more prone to ligamentous syndesmotic injury.⁸ No study to date identifies whether there is a relationship between the anatomy of the incisura fibularis tibiae and the type of posterior malleolus fracture one may be prone to sustain. This study aimed to evaluate the relationship between the anatomy of the incisura and the pattern of PM fractures.

2 Methodology

A retrospective review of patients who sustained posterior malleolus fractures who presented and were treated at our academic hospital between 01 January 2020 and 31 December 2022. The inclusion criteria consisted of patients with posterior malleolus fractures over the age of 18 years, with ankle x-rays and pre-operative CT scans of bilateral ankles. Those who were excluded were patients who sustained pilon fractures, bilateral ankle fractures, those whose CT scans did not involve both legs and those who had missing clinical records. The data of interest included patient demographics, mechanism of injury, clinical evaluation, pre-operative radiographs, and CT scans. The relevant data was collected from the hospital admission register, outpatient files, in-patient files, picture archiving and communication system (PACS).

Pre-operative radiographs were reviewed, and the injuries were classified according to the Weber, Lauge-Hansen and AO/OTA classification systems.^4,5,9 The patients’ pre-operative CT scans involving both ankles were then reviewed. Patients were scanned supine using a multidetector scanner (Siemens SOMATOM Definition AS). The parameters for the CT examination were as follows: kV 120, mA 35, section thickness 1 mm, section collimation 0.6 mm, and reconstruction overlap of 0.5 mm. Several radiographers scanned patients in the radiology department at different times of the day.

The posterior malleolus fracture on the injured side was classified according to the Bartonicek classification system.¹⁰ The uninjured side's incisura morphology was evaluated, and this examination was performed at 5 mm proximal to the tibia plafond.¹¹ The assumption is that the uninjured incisura configuration is symmetrical to the premorbid contralateral incisura.

The documented incisura morphology included depth, version, shape, and width.^7,8 The documented fibula morphology included the shape and engagement.^7,8

Statistical analysis was done using the STATA software, version 18 (StataCorp, College Station, TX, USA). The statistical data analysed was descriptive. Categorical variables were documented as percentages and frequencies, while continuous data were documented as means and standard deviation or medians, and interquartile ranges (when it was not normally distributed). Analysis of differences was expressed using the chi-squared test or Fisher's exact test. The t-test/Mann Whitney test for more than two categories (or ANOVA/K Wallis with Benjamin Hocherberg corrections) was used to evaluate the connection between continuous and categorical variables. Statistical tests analysis were two-sided, and p-values <0.05 were statistically significant.

3 Results

The total number of patients who sustained posterior malleolus fractures enrolled in this study was 145. Females accounted for the majority, 89(61 %). The female-to-male ratio was 1.6:1. The median age for the patients overall was 40 years old (IQR 32–49 years). The median age for females was 42 years, while the median age for men was 38 years. Most PM fractures occurred on the right ankle (n = 89, 61 %).

Twisting of the ankle or fall from standing height (FFSH) was the leading cause of injury in the study population (n = 97/145, 67 %), followed by Pedestrian Vehicle Accident (PVA) (n = 34/145, 24 %) and fall from a height of more than 2 m (n = 5/145, 3 %). Motor vehicle accidents, Motorbike accidents and assault all caused 2 % (n = 3/145) of these injuries.

The most prevalent comorbidity in our patient cohort was Human Immunodeficiency Virus (HIV) positive individuals. Asthma showed a statistically significant difference in the type of PM fracture sustained among the comorbidities noted (Table 1). There was no statistically significant difference noted with employment status, alcohol and smoking use (Table 1).

According to the Weber classification, most patients fell into the Weber B category (n = 116, 80 %). Twenty-five (17 %) patients were classified as Weber C injuries. Two (1 %) patients were classified as Weber A injuries. Two (1 %) patients could not be categorised according to the Weber classification. Both patients suffered medial malleolus fractures (Herscovici type C)¹² with a PM fracture without the presence of a fibula fracture.

According to the Lauge-Hansen classification, 116 patients (80 %) were classified as supination external rotation (SER) injuries followed by pronation-external rotation (PER) injuries (n = 16, 11 %) and 8 (6 %) pronation-abduction (PAB) injuries. Five (3 %) patients could not be classified according to the Lauge-Hansen classification.The AO/OTA classification distribution is shown in Fig. 1.

Close

Fig. 1 AO/OTA classification frequency distribution.

Export to PPT

The frequency distribution of the Bartonicek classification in our study population is as follows: Most patients were classified as Bartonicek type 2 injuries, 83/145 (57 %). Twenty-two (15 %) patients were diagnosed with Bartonicek type 1 and type 3 injuries each, and 18/145 (12 %) patients were diagnosed with Bartonicek type 4 injuries (Fig. 2).

Close

Fig. 2 Bartonicek classification frequency distribution.

Export to PPT

Regarding incisura and fibula morphology characteristics, no statistically significant difference was noted between incisura shape, depth, version, and fibula shape with the type of posterior malleolus fracture (Tables 2 and 3). A statistically significant difference was noted with fibula engagement and incisura width with the type of PM fracture sustained.

4 Discussion

Our study showed that posterior malleolus fractures occur predominantly in females. The female-to-male ratio in this study was 1.6:1. Li et al. and Mason et al. noted slightly lower ratios of 1.15:1 and 1.27:1, respectively, of female predominance.^13,14 Stringfellow et al. conducted a multicentre cohort study, and their female-to-male ratio was 1,8:1, which was close to our study.¹⁵

The side most commonly injured in our study was the right ankle (n = 89/145, 61 %). Han et al. looked at 3952 patients over five years who were 16 years and older with ankle fractures and found the right ankle slightly more predominant (n = 2003/3952, 51 %).¹⁶ De Luna et al., in their study of ankle fracture-dislocations, similarly found the right ankle to be most commonly affected (n = 22/26, 85 %).¹⁷

The top three causes of PM fractures in our study were falls from standing height or ankle twisting, followed by pedestrian-vehicle accidents and falls from height (FFH). The top three causes of posterior malleolus fractures in the study by Li et al., low-energy fall, traffic accident, and fall from height, are consistent with those observed in our study.¹⁴

Seventy-three patients (50.3 %) in our study had no comorbidity. The comorbidities identified in the other half of the patients in our study are Human Immunodeficiency Virus (HIV), Hypertension, Diabetes Mellitus, Asthma, Epilepsy and Rheumatoid arthritis in descending order. Thirty-one percent of patients with posterior malleolus fractures in the study by Li et al. had comorbidities, while 69 % did not have a comorbidity.¹⁴ HIV was the most prevalent comorbidity in our study (n = 41, 28 %). There is a paucity of studies looking at HIV and ankle fractures specifically. However, Pramukti et al. conducted a systematic review looking at fractures in HIV-positive individuals.¹⁸ They noted that patients with HIV had a 1.9 times higher risk of fracture compared to those without HIV.¹⁸ There was no relationship between the presence of HIV infection in our patients and the pattern of PM fractures (p = 0.292)

Asthma was reported in five (3 %) patients in our study. Most asthmatic patients in our study (3/5) had Bartonicek type 4 PM fractures, which showed to be statistically significant. Corticosteroid use is known to induce osteopaenia and osteoporosis.¹⁹ Cho et al. stated that the use of inhaled corticosteroids, which are commonly used in Asthma and Chronic obstructive pulmonary disease patients, is associated with an increased risk of fractures. This could be the likely reason for the relationship between asthma and the type of PM fracture noted in our study. This result should be taken cautiously as only five Asthmatic patients were included in our study. Further studies with a larger sample size of asthmatic patients should be conducted to confirm this relationship.

All three classification systems (Lauge-Hansen, AO/OTA and Weber) could not entirely classify all patients included in the study. Two patients who sustained medial malleolus fractures with associated PM fractures, with intact fibulae, could not be categorised according to either of the three classification systems. Therefore, the Herscovici classification was used to assist in classifying the medial malleolus fracture but was still unable to categorise the posterior malleolus fractures and, thus, the injury entirely.¹² Both patients were females and sustained Herscovici type C injuries. One was a 56-year-old hypertensive smoker involved in a PVA, and the other was a 33-year-old with no comorbidities who fell from a standing height and twisted her ankle. One sustained a Bartonicek type 2 fracture, and the other a Barotnicek type 3 injury. In their study, McHale et al. also noted three patients who could not be classified according to the Lauge-Hansen classification.²⁰ The mean age of the three patients they mentioned was 26.7 years.²⁰ Contrastingly, the mean age of our patients who did not fit the Lauge-Hansen classification was 43 years.

The AO/OTA classification could categorise 142/145 patients (98 %), while the Lauge-Hansen classification could categorise 140/145 patients (97 %) in our study. The Weber classification could categorise the fibula fractures (143/145, 99 %). However, it did not provide any additional data regarding the medial and posterior malleoli injuries. Pfluger et al. identified the AO/OTA classification as a reliable system to classify trimalleolar ankle injuries.²¹ Patton et al. examined whether the Lauge-Hansen classification could predict the pattern of posterior malleolus fractures. No significant difference was noted between the Lauge-Hansen classification and the Bartonicek classification.

Patton et al. also noted Bartonicek type 2 injuries as the predominant diagnosis in their study (48 %).²² There was a higher percentage of Bartonicek type 3 injuries (31 %) in the study by Patton et al. than ours. Neumann et al., in their research, noted Bartonicek type 2 and 3 to be equally present, accounting for 70 % of their injuries.²³ Patton et al. and Neumann et al. noted Bartoniek type 4 injuries (9 % and 23 %, respectively) more than Bartonicek type 1 injuries (7 % and 7 %, respectively).^22,23 Our study noted more Bartonicek type 1 injuries (15 %) than Bartonicek type 4 (12 %). The overt difference between our study and the studies by Patton et al. and Neumann et al. is that our study population was younger on average.

Out of 145 patients in our study, 118 (81 %) had shallow incisurae, and 27 (19 %) were concave with a mean depth of 2.9 mm. Of note was that these measurements were performed at 5 mm proximal to the plafond. Fojtík et al. illustrated that the deepest portion of the incisura is measured in this region and that there was no rationale for measuring 10 mm proximal to the plafond.¹¹ The mean depth measured in their anatomical study was 4.5 mm.¹¹ Liu et al., Musa et al., and Taser et al. measured a mean depth of 3.29 mm, 3.44 mm and 3.6 mm, respectively.^24–26 Their results were also comparable to our study in that the percentage of those with shallow incisurae was 80.3 %, 75 % and 65 %, respectively.^24–26 Tonogai et al. measured a mean incisura depth of 4.1 mm in their study, with 64 % of their population having concave incisurae and 36 % shallow, in contrast to our study.²⁷ Elgafy et al. also documented that their study population had concave incisurae, with 67 % concave and 33 % shallow.²⁸ Incisura depth is variable, as seen in the studies mentioned above.

Interestingly, Tonogai et al. compared body mass index (BMI) between those with deep/concave and shallow incisurae.²⁷ Those with shallow incisurae had a mean BMI of 24.6 kg/m^2, whereas those with concave incisurae had a mean BMI of 27 kg/m².²⁷ BMI measurements are generally not performed in our facility in patients with lower limb fractures. The patients in our study had unilateral ankle fractures, so measuring their weight and height appropriately would have posed a challenge. As seen from our research, the local population has shallow incisurae. Shallow incisurae are prone to syndesmotic injury in the form of fracture-dislocations or recurring ankle sprains.^8,29 Wittouck et al., in their systematic review, identified that patients with unstable syndesmotic ankle fractures had shallow incisurae.³⁰ The inclusion criteria for our study involved posterior malleolus fractures, which confers syndesmotic instability. This could be a reason why shallow incisurae were predominant in our study. There was, however, no relationship between the depth of the incisurae and the type of posterior malleolus fractures the study population sustained.

Our study's median/mean incisurae width was 26.4 mm. Yüce et al., in their study population, noted mean incisurae widths of 23.96 mm and 22.40 mm in their respective study groups.³¹ Taser et al. also noted a mean incisura width of 23.26 mm.²⁵ Their study also illustrated that males have a wider mean incisura width of 24.13 mm compared to the average female width of 21.78 mm, which was also noted in our study.²⁵ Our study shows slightly wider incisura widths compared to other studies. Taser et al. and Yüce et al. conducted their incisura measurements at 10 mm proximal to the plafond, whereas we conducted our measurements at 5 mm proximal to the plafond. This is likely why our measurements are slightly wider, as they are more distal on the tibia. Gupta et al. stipulate that wider incisurae predispose to shallower incisurae, which is a risk factor for syndesmotic injury. Yüce et al. found in their study that wider incisura widths were associated with posterior malleolus avulsion fractures.³¹ This finding can be comparable to what was found in our study. A statistical difference was noted with incisurae width and the type of posterior malleolus fracture seen in our study population (p = 0.0131). The narrower the incisura width, the higher the Bartonicek classification. Therefore, the wider widths are more prone to sustain Bartonicek type 1 fractures comparable to posterior malleolus avulsion fractures.

In our study, most fibulae were engaged (n = 128, 88 %.). Boszczyk et al., in their research, found 53/75 (71 %) of the fibulae to be disengaged in those who sustained a syndesmotic injury.⁸ Boszczyk et al. conducted their study on patients with syndesmotic injuries against an unaffected control group. In our study, the incisura morphology was examined on the unaffected leg without pathology rather than a control group. The findings of the incisura morphology on the unaffected side serve as a proxy for the pre-injury anatomy of the affected side.⁸ Statistical difference was, however, noted concerning fibula engagement and the pattern of PM fractures (p = 0.049). Seventeen (12 %) fibulae were noted to be disengaged in our study. The majority of fibula disengagement was noted to be associated with Bartonicek type 2 and 3 with 8/17 and 6/17, respectively. Boszczyk et al. noted that fibula disengagement was a notable risk factor for syndesmotic injury.⁸

Interestingly, most of our patients had shallow, flat incisurae with engaged and triangular-shaped fibulae. At face value, this appears contradictory; however, Fig. 3 illustrates the configuration. One would expect concave incisurae to be associated with engaged fibulae and shallow, flat incisurae with disengaged fibulae.

Close

Fig. 3 Axial CT scan from our study depicting a patient with a shallow, flat incisura with an engaged triangular fibula.

Export to PPT

Limitations are present in this study. This study is retrospective and was conducted at a single centre. Bias may also be introduced as the study population was not compared to a control group without injuries. Not many studies have also conducted incisura measurements at 5 mm proximal to the plafond. A standardisation of the location of incisura measurements is recommended.

5 Conclusion

Our results show that posterior malleolus fractures occur mostly in females and are noted to occur more commonly on the right ankle. Twisting of the ankle or falling from a standing height is the mechanism that most notably causes posterior malleolus fractures. HIV was the most prevalent comorbidity in our study population. The most commonly noted morphological configuration in the study population was shallow, flat incisurae with engaged triangular fibulae. Bartonicek type two was the most frequent fracture seen. Incisura width and fibular engagement showed a statistically significant association with the kind of PM fracture pattern sustained. The narrower the incisura width the higher the Bartonicek classification. Patients with disengaged fibulae are more likely to sustain a Bartonicek type 2 or 3 injury.

CRediT authorship contribution statement

Papa Kwabena Offeh Kyei: study conceptualisation, data capture, data analysis, manuscript preparation, revision and approval of final manuscript. Collen Sandile Nkosi: study conceptualisation, data analysis, manuscript preparation, manuscript revision and final manuscript approval. Richard Paterson: Manuscript revision and approval of final manuscript. Abdirashid Aden: manuscript design, manuscript preparation, revision and approval of final manuscript.

Guardian/patient's consent

No guardian/patient consent was needed for this study as this study was a retrospective review.

Before the commencement of the study, ethical approval was obtained from the Human Research Ethics Committee (Medical) of the University with reference number: M231064.

Ethics statement

All procedures performed in studies involving human participants were by the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Prior to the commencement of the study, ethical approval was obtained from the Human Research Ethics Committee (Medical) of the University with reference number: M231064.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.