2.1

Inclusion criteria were: (1) acute closed fracture; (2) patients with ankle fractures involving the PMFs diagnosed with medical history, symptoms, signs and imaging examinations; (3) agreement to sign right of informed consents, and consent for arthroscopic surgery on the PMFs; (4) PMFs fixed with absorbable screws; (5) complete medical records, and follow-up time ≥ 12 months.

Exclusion criteria were: (1) severe osteoporosis or secondary pathological fractures; (2) combination with severe neurovascular injuries; (3) open or non-acute fractures; (4) combination with other serious injuries or bad compliance that affected postoperative rehabilitation training; (5) incomplete information or missed follow-up.

2.2

Based on the above inclusion and exclusion criteria, the data of 33 patients who underwent treatment for PMFs using posterior ankle arthroscopy with absorbable screws in our hospital between January 2022 and December 2023 were retrospectively analyzed, including 20 males (60.61 %) and 13 females (39.39 %) (43.85 ± 8.90 (25−63) years old). The causes of injuries included traffic injuries (10, 30.30 %), fall injuries (16, 48.48 %), fall-down injuries (4, 12.12 %), and squeeze injuries (3, 9.10 %). According to Haraguchi classification , there were 18 cases (54.55 %) of type I and 15 cases (45.45 %) of type II. According to Bartoníček and Rammelt (B&R) classification , there were 18 cases (54.55 %) of type 2, 11 cases (33.33 %) of type 3, and 4 cases (12.12 %) of type 4 ( Table 1 ).

3.1

Following admission, manipulative reduction with plaster fixation or bone traction therapy was administered based on the specific type of ankle fracture. The three-dimensional (3D) CT of the ankle joint was conducted to characterize PMFs. Low-molecular-weight heparin was given preoperatively, if necessary, to prevent thrombosis, and the vascular ultrasound of lower limbs was performed to determine whether thrombosis occurred. Moreover, surgical contraindications were excluded. When the swelling of the affected limb subsided and the skin tension was moderate, surgical treatment was performed after about 5–7 days after the injury occurrence.

3.2

(1) Following successful anesthesia (general anesthesia with endotracheal intubation or continuous epidural anesthesia), the patient took a prone position, and a pneumatic tourniquet was applied to the proximal thigh of the affected limb.

(2) Fixation of the lateral malleolar : A 6–8 cm longitudinal incision was made over the fibular or lateral malleolar fracture site. The fracture was exposed through the anterior interval between the peroneus longus and brevis muscles. After elevating the periosteum and evacuating the hematoma, the fracture was anatomically reduced and stabilized using an anatomic locking plate with multiple screws (Tianjin Zhengtian Medical Equipment Co., Ltd., Tianjin, China). The deep interval between the peroneus longus and brevis tendons was not incised.

(3) The posteromedial (PM) and posterolateral (PL) portal sites were identified as follows: First, a horizontal line was drawn at the level of the tip of the lateral malleolus. Then the PM and PL portals were located at the points where this line intersects the medial and lateral aspects of the Achilles tendon. When the ankle joint was in the dorsal extension position, 11# scalpel was used to incise the skin by about 0.5 cm. The PM and PL portals were established using incision and expansion techniques. A 4.0-mm 30 º arthroscope was introduced through the PL portal, while a 4.0-mm shaver was inserted via the PM portal. After meeting, soft tissues were debrided and the posterior ankle working space was established. The flexor hallucis longus (FHL) was identified, thus establishing the medial safe zone. ( Fig. 1 A–C). The posterior joint capsules were debrided, hematomas and any small loose bodies were removed. The distal articular surface of the tibia, the posterior tibiofibular ligament, and the syndesmotic congruence were assessed. The posterior talofibular and the posterior tibiofibular ligaments were protected.

A: Surface positioning of posterior ankle arthroscopy; B-C: in vitro and in vivo operational perspectives of posterior ankle arthroscopy respectively; D: intercalary fragment in PMFs (die punch fragment) embedded on sagittal plane under ankle joint 3D CT; E: die punch fragments observed under posterior ankle arthroscopy; F: reduced die punch fragments under posterior ankle arthroscopy with talus as a template; G: broken end of posterior malleolus fracture exposed; (H): articular surface of posterior malleolus fracture after anatomical reduction; (I): placement of guide wires under posterior ankle arthroscopy; (J): placement of absorbable screws under posterior ankle arthroscopy.

(4) A proximal posterolateral (PPL) portal was established approximately 2–4 cm proximal to the posterolateral (PL) portal and served as an auxiliary portal. A 0.5-cm skin incision was made with 11# scalpel, after which the tract was dilated with a vascular clamp until the bone surface was reached. The resulting portal served three purposes: (A) to push the posterior malleolar fragment to the distal end of the joint for anatomical reduction; (B) to enable temporary fixation or placement of a more proximal anti-sliding screw for large fragments; and (C) to retract the flexor hallucis longus tendon, thereby protecting the medial neurovascular structures and expanding the surgical field of view.

(5) Reduction techniques: The articular reduction was monitored arthroscopically through the PM portal. The talar articular surface served as a template for reduction, assisted by dorsiflexion traction, distal pushing of the fragment via the PPL portal, and joystick manipulation using a Kirschner wire inserted into the fragment. If a small impacted osteochondral fragment (“die punch fragment”) was present ( Fig. 1 D, E), it was first reduced using the talus as a template ( Fig. 1 F). If necessary, temporary fixation with a Kirschner wire could be performed, followed by reduction of the posterior malleolar fragment using the previously described techniques.

(6) Fixation: A 1.0 mm guide wire was inserted into the posterior malleolar fragment via the PL portal, aiming at the center of the ankle joint and positioning the posterior tibia tubercle. Through the PPL portal, vascular forceps retracted the FHL tendon medially to enhance access to the posterior tibial margin. A second 1.0 mm guide wire was inserted higher than the first guide wire via the PM portal. If the fracture fragment was large, a third 1.0 mm guide wire can be inserted via the PPL portal. The fracture reduction status and the position of the Kirschner wire were assessed with fluoroscopy. The anatomical reduction of the articular surface was confirmed again under arthroscopy. After drilling, depth measurement, and tapping, a 3.5 mm cannulated absorbable screws (INION, Finland) in an appropriate length were inserted ( Fig. 1 G–H).

(7) Final steps: After arthroscopic and fluoroscopic confirmation, guide wires were removed, and portal incisions were closed with 3–0 sutures. If a medial malleolus fracture was present, the operative table was tilted slightly toward the affected side with the knee flexed, and an anteromedial arc incision was made for reduction and internal fixation. For certain fracture types (e.g., B&R type 3: posteromedial two-part fragment involving the medial malleolus, cases with a large medial malleolar fragment), temporary fixation of the medial malleolus with two or three Kirschner wires was performed after lateral malleolus fixation. Following reduction and fixation of the posterior malleolus, the Kirschner wires were replaced with two 3.5-mm cannulated screws. The distal tibiofibular syndesmotic congruence was probed via posterior ankle arthroscopy, and the syndesmotic stability was assessed by performing the Cotton test. If necessary, syndesmotic screw or elastic fixation could be applied. All incisions were closed layer-by-layer, and after dressing, a neutral-position plaster splint was applied to immobilize the ankle.

3.3

Antibiotics were routinely used to prevent infection until 24 h after surgery, and the affected limb and the ankle joint were raised and fixed with neutral plaster. Anterior and lateral X-rays and 3D CT of the ankle joint were re-examined. When there was no contraindication for anticoagulants, subcutaneous injection of enoxaparin sodium (4000 U/d) was started at 12 h postoperatively until discharge. After discharge, oral rivaroxaban (10 mg/d) was continued until 30 days after the operation. The plaster was removed 4 weeks postoperatively, and non-weight-bearing extension and flexion exercises were performed. Partial weight-bearing exercises were started 6–8 weeks postoperatively, and gradually transited to complete weight bearing.

4.1

The surgical time and the posterior ankle arthroscopic surgery time were recorded.

4.2

The articular surface reduction (sagittal displacement > 2 mm is bad reduction) and the syndesmotic congruence were postoperatively re-examined via 3D CT .

4.3

The ankle joint flexion mobility at the last follow-up (before take-out of internal fixation) was recorded. Pains were assessed using Visual Analogue Scale (VAS) (0–2: excellent, 3–5: good, 6–8: acceptable, >8: bad). Functional efficacy was evaluated using American Orthopeadic Foot and Ankle Society (AOFAS) score (90–100: excellent, 75–89: good, 50–74: acceptable, <50: bad) , and Olerud-Molander Ankle Score (OMAS) (91–100: excellent, 61–90: good, 31–60: acceptable, 0–30: bad) .

4.4

After the surgery, secondary injuries (e.g., neurovascular injuries), and soft tissue complications (e.g., wound infection and skin necrosis) were recorded. The ankle joints were reexamined with regular X-rays (monthly) until complete healing to evaluate fracture healing, and any reduction loss was recorded. At the last follow-up (before internal fixation removal), whether big toe flexion contracture occurred was recorded.