Percutaneous Zadek Osteotomy vs Open Haglund Resection for Insertional Achilles Tendinopathy: Early Outcomes and Complication Rates

Abstract

Background:

Insertional Achilles tendinopathy (IAT) is difficult to manage despite multiple surgical options. Two of the surgical options include an open midline Achilles tendon–splitting Haglund resection and a percutaneous Zadek osteotomy (ZO). The current study compared the patient-reported outcomes and complications of open vs percutaneous procedures in patients with IAT ± Haglund deformity.

Methods:

A retrospective comparative study of consecutive patients who underwent either an open midline Achilles tendon–splitting Haglund resection or percutaneous ZO for IAT ± Haglund deformity at a single institution with minimum 12 months’ follow-up was performed. Preoperative and postoperative Patient-Reported Outcome Measurement Information System (PROMIS) scores and complications were recorded for each patient and compared between cohorts.

Results:

Forty-three patients were treated with an open midline Achilles tendon–splitting Haglund resection and 34 patients were treated with a percutaneous ZO for IAT ± Haglund deformity. PROMIS function (P < .001), pain (P < .001), and mobility (P < .001) scores significantly improved in patients who received either procedure. In the open midline Achilles tendon–splitting Haglund resection cohort, a significant increase in wound complications (11/43, 25.6%) was observed in comparison to the percutaneous ZO group (1/34, 2.9%; P = .007).

Conclusion

Both techniques yielded clinically meaningful improvements in PROMIS scores. In this modest retrospective cohort, the percutaneous ZO was associated with fewer minor wound complications; however, this observation should be interpreted cautiously given the limited sample size, short follow-up, and potential selection bias.

Level of Evidence:

Level III, retrospective cohort study.

Introduction

Insertional Achilles tendinopathy (IAT) is a common source of retrocalcaneal pain treated by orthopaedic foot and ankle surgeons. Often, this pathology coincides with a posterosuperior calcaneal prominence known as Haglund deformity, and a retrocalcaneal bursitis. ³⁶ This triad is considered to be diagnostic of Haglund syndrome, which is a spectrum of disease. Commonly, patients will present with complaints of retrocalcaneal pain, often worsened with shoe-wear and increased activity. Nonsurgical management strategies include shoe-wear modification, eccentric loading therapy, anti-inflammatory medications, and physical therapy. However, in 20% to 40% of patients, nonoperative strategies will ultimately fail to improve symptomatology. ³⁶ After 3-6 months of persistent symptoms, despite conservative management, patients are often offered a surgical intervention.

Several surgical options exist for IAT with or without Haglund deformity, but outcomes remain variable, and no single technique is universally accepted. Surgical options commonly include an open Haglund resection, which is most often performed through an open midline Achilles tendon–splitting approach or a posterolateral approach. This procedure can be performed with or without an FHL tendon transfer, which may offer an additional benefit to patients over the age of 50 years. ⁴⁴ Nevertheless, like many open procedures in the retrocalcaneal space, the open midline Achilles tendon–splitting Haglund resection has been associated with a high rate of complications, such as wound dehiscence, infections, and persistent postoperative pain.^{7,9,13,17,31,40,41,46,47,59} In an attempt to mitigate many of these associated complications, surgeons have turned to minimally invasive surgery (MIS) techniques. MIS has been increasing in popularity in the past decade across foot and ankle orthopaedic surgery. Several studies have suggested reductions in wound problems, postoperative pain, and recovery time when MIS techniques are used for IAT, although results are heterogeneous.^{3,7,9 -13,16,23,31,33,40,41,43,50,56,60,61}

The percutaneous Zadek osteotomy (ZO) is a minimally invasive surgical option for patients with IAT with or without Haglund deformity. The ZO is a dorsal closing wedge osteotomy of the calcaneus that allows for correction of the anatomical strain on the Achilles tendon at its insertion on the calcaneus, without direct operation on the Achilles tendon itself.^{2,24,26,39,54,62} During the ZO procedure, a dorsal-based wedge of approximately 1 cm is removed from the calcaneus (Figure 1). Removing this wedge shortens and anteriorly rotates the calcaneus, so that the Achilles insertion point is moved more anterior and superior to its original station. This modification simultaneously reduces the pathological strain on the tendon, decompresses retrocalcaneal bursa, and improves the length-tendon relationship of the gastroc-soleus complex, therefore improving dorsiflexion. As described by Kaplan et al,^24,39 the ZO can be performed percutaneously with the aim of reducing wound complications, pain, and infection relative to open procedures.

Figure 1.

The percutaneous Zadek osteotomy; cut guide as demonstrated by Kaplan et al. ²⁴

Percutaneous ZO has been reported as a treatment for IAT with or without Haglund deformity^{1,8,14,15,20,21,24,25,27,32,35,37,54,55}; however, direct comparisons with the traditional open approach remain scarce. ⁶ The purpose of the current study was to compare the patient-reported outcomes and complications of the open midline Achilles tendon–splitting Haglund resection with the percutaneous ZO in patients with IAT with or without Haglund deformity. We hypothesized that both percutaneous ZO and open midline Achilles tendon–splitting Haglund resection would allow for significant improvement in patient-reported outcomes, yet the percutaneous ZO would have significantly less complications in comparison to the traditional, open midline Achilles tendon–splitting Haglund resection due to reduced soft tissue trauma.

Methods

Patients of 2 fellowship-trained orthopaedic foot and ankle surgeons diagnosed with IAT ± Haglund deformity at one institution were retrospectively identified having undergone either an open midline Achilles tendon–splitting Haglund resection or percutaneous ZO. All patients had a minimum of 1-year follow-up and were analyzed retrospectively. Before either method of surgical intervention, patients were treated with nonoperative management strategies for at least 3-6 months. Nonoperative treatment of IAT included, but was not limited to, activity and shoe wear modification, immobilization, eccentric loading physical therapy, nonsteroidal antiinflammatory drugs, and oral steroids. If patients failed nonoperative management, patients were offered an open midline Achilles tendon–splitting Haglund resection, or the percutaneous ZO. Patients who were candidates for open Haglund resection were also candidates for percutaneous ZO and the decision for which surgical treatment option was performed was at the discretion of the surgeon. Preoperative workup of all patients included physical examination, PROMIS score survey, and standard weightbearing radiographs.

Patient allocation to the 2 surgical techniques was based on temporal practice patterns. Initially, both participating surgeons performed traditional open midline Achilles tendon–splitting Haglund resection for IAT. As the surgeons progressively incorporated MIS techniques into their practice, they transitioned to performing more Zadek osteotomies in cases that would have previously been treated with open midline Achilles tendon–splitting Haglund resections. This change in surgical approach created the 2 comparative groups for retrospective analysis. There was no change in indications for surgery during this period or with different surgical techniques. In patients with an acute rupture of the Achilles in the setting of IAT, an open procedure would be performed; however, these patients were excluded from the current study. The authors do not consider the presence of a Haglund deformity to be a specific requirement for intervention with either procedure; furthermore, we did not preferentially allocate patients with a Haglund deformity to one treatment over another. Similarly, the degree of tendinopathy on preoperative MRI was not a consideration in preoperative planning. Previous literature has demonstrated success in treating patients with IAT with the percutaneous ZO with or without Haglund deformity, regardless of IAT severity on preoperative MRI.^{1,18,20,21,24} Specifically, as demonstrated by Hall et al, ¹⁸ patients with grade 1-3 IAT on preoperative MRI demonstrated similar improvement in patient-reported outcomes following the percutaneous ZO. Accordingly, patients with IAT were offered either surgery, regardless of the presence of a Haglund deformity or tendinopathy severity.

Forty-three patients treated with an open midline Achilles tendon–splitting Haglund resection and 34 patients treated with a percutaneous ZO for IAT ± Haglund deformity between March of 2021 and May of 2024 met our inclusion criteria. As there were no previous comparative studies available to calculate an appropriate sample size, we included all consecutive patients to maximize the number of patients included, therefore reducing the risk of type 2 error.

Open Midline Achilles Tendon–Splitting Haglund Resection

Thirty-four females (79.1%) and 9 males (20.9%) were included in the open midline Achilles tendon–splitting Haglund resection cohort; patients had a mean age of 54.0 ± 11.1 (range, 21-78), 39.1 ± 8.6 (range, 18.5-61.1), and follow-up of 32.1 ± 12.3 (range, 12-53) months. Twenty-one open midline Achilles tendon–splitting Haglund resections were left-sided (46.5%), and 12 were right-sided (53.5%). No patients were active smokers; 9 (20.9%) had diabetes.

The patient was positioned prone. A thigh tourniquet was used. Patients received a lower extremity nerve block (popliteal and adductor canal or saphenous). Sedation was achieved by IV propofol or general anesthesia. A 6-cm incision was made midline over the heel. The Achilles was identified and split midline and elevated off the calcaneus. The Haglund deformity was identified and removed with an oscillating saw. At this time both flaps of the Achilles were debrided of all non-viable tissue and then reattached back to the calcaneus with suture anchors.

Percutaneous Zadek Osteotomy

Twenty-three females (67.8%) and 11 males (32.2%) were included in the percutaneous ZO cohort; patients had mean age of 55.0 ± 12.4 (range, 21-79) and follow up of 21.6 ± 7.3 (range, 12-36). Nineteen procedures were left-sided (55.9%), and 15 were right-sided (44.1%). Patients had a mean body mass index of 37.0 ± 8.7 (range, 18.6-59.0); 0 patients were smokers and 6 (17.6%) had diabetes.

The percutaneous ZO was performed on patients as described previously by Kaplan et al ²⁴ (Figure 1). Patients received a lower extremity nerve block (popliteal and adductor canal or saphenous). Sedation was achieved by IV propofol or general anesthesia. No tourniquet was used for the procedure. Patients were positioned in the lateral decubitus position; the operative leg was held off the end of the bed, resting on the mini-C-arm.

A 10-mm dorsal closing wedge on the lateral calcaneus was approximated with wire and lateral fluoroscopy. The safe zone over the calcaneal tuberosity, as defined by Talusan et al, ⁵² was identified. A 5-mm lateral incision was made at the apex of the Zadek osteotomy, 5 to 8 mm just anterior to the plantar fascia origin at the calcaneal tuberosity. After a blunt hemostat was used to spread down to bone, a 3 × 30-mm Shannon burr with a 4:1 reducer (Novastep, Englewood Cliffs, NJ) was advanced into the lateral calcaneus at the apex of the osteotomy. Shannon burr was set to 6000 rpm, and the burr position was confirmed on lateral fluoroscopy. Copious refrigerated (37°F) normal saline via bulb syringe was used throughout burring. ⁴² Burring was paused every 3-4 seconds to allow the burr to cool.

Once the first limb of the osteotomy was completed, the Shannon burr was run to shave a dorsal 10-mm wedge. The osteotomy was reduced with maximal ankle dorsiflexion (Figure 2). Two guidewires were placed across the osteotomy from the posterior calcaneal tuberosity into the subchondral bone. Two 7.0-mm headless compression screws were advanced across the osteotomy with the ankle kept in maximal dorsiflexion. Implant position was confirmed on lateral and Harris axial fluoroscopic views of the calcaneus. Incisions were irrigated with normal saline and closed with 2-0 nylon sutures. A postoperative soft dressing was placed.

Figure 2.

Lateral fluoroscopic images of a patient undergoing percutaneous Zadek osteotomy, demonstrating the removal of a dorsal based wedge and closing down of the osteotomy. ²⁴

Postoperative Management

Follow-up visits included radiographs and clinical evaluation. At a 2-week postoperative visit, the splint was taken down for each patient and the sutures were removed (Figure 3). Patients who underwent the open midline Achilles tendon–splitting Haglund resection were placed into a nonweightbearing, resting plantarflexion cast at 2 weeks. At the 2-week postoperative follow-up, patients were transitioned into a boot with wedges and began a progressive weightbearing protocol. Over the subsequent weeks, patients weaned out of the wedges until neutral. Patients began formal physical therapy once they had progressed to a full weightbearing status out of their boot. At 10 weeks, postoperative patients transitioned out of the boot and started physical therapy.

Figure 3.

Preoperative and postoperative lateral radiographs of a patient who underwent percutaneous Zadek osteotomy. ²⁴

For patients who received a percutaneous ZO, postoperatively they were nonweightbearing in a splint for the first 2 weeks, and then full weightbearing in a CAM boot was allowed from weeks 2-6. Patients began formal physical therapy once they had progressed to a full weightbearing status out of their boot. At this time, patients began therapy for gait training, range of motion exercises, and Achilles strengthening. Patients were not allowed to drive for 6 weeks following surgery. At 6 weeks postoperatively, patients were transitioned out of the CAM boot into supportive sneakers as tolerated. Return to low-impact activities including biking, swimming, and the elliptical was allowed at 6 weeks postoperation; high-impact activities such as running and jumping were allowed at 12 weeks postoperative as tolerated (Figure 4).

Figure 4.

Postoperative images of incisions following the (A) percutaneous ZO or (B) open midline Achilles tendon–splitting Haglund resection. ZO, Zadek osteotomy.

Data Analysis

Patient-reported Outcomes Measurement Information System (PROMIS) scores are a validated instrument for assessing patient outcomes, and the preferred patient-reported outcome metric of the American Orthopaedic Foot & Ankle Society.^4,30,48 PROMIS pain interference, physical function, and mobility scores were collected for patients at preoperative and postoperative appointments. A full 5 domain PROMIS survey was included in each patient’s preoperative and postoperative appointments. ⁵¹ For the purposes of the current study, PROMIS function, pain, and mobility scores were analyzed in each cohort. Higher PROMIS function and mobility scores are more desirable whereas lower PROMIS pain scores are more desirable (normal range: 40-60, mean 50).

To determine if postoperative PROMIS scores represented a clinically significant difference from preoperative scores, the minimum clinically important difference was conservatively considered to be 5, as previously described.^22,38 Complications, nonunion, reoperations were recorded. Cases were divided into cohorts for further analysis based on intervention (percutaneous ZO vs open midline Achilles tendon–splitting Haglund resection). For our study, any opening of the wound, of any size, was considering a wound healing complication, even if the wound healing did not change the weightbearing time, require antibiotics, or an additional surgical procedure. All data were tested for normality, which demonstrated the data was from a normal distribution. As such, preoperative and postoperative function, mobility, and pain PROMIS scores were compared using a paired t test. All patients had postoperative PROMIS scores; however 7 of 43 patients in the open midline Achilles tendon–splitting Haglund resection cohort (16.3%) underwent surgery before completing the PROMIS score survey. These patients have been excluded from the statistical analysis. All other continuous data was compared by t test; all categorical data were compared using χ² analysis. All statistical analyses were complete on IBM SPSS Statistics 25 (IBM, Armonk, NY). All P <.05 were considered significant.

Results

Cohort Demographics

There were no differences in sex (P = .256), age (P = .716), or laterality of surgery (P = .539) between patients who underwent percutaneous ZO and patients who underwent open midline Achilles tendon–splitting Haglund resection. Similarly, there was no significant difference in body mass index (P = .287) or diabetes (P = .718) between cohorts. Follow-up duration was found to be significantly longer in the open midline Achilles tendon–splitting Haglund resection cohort (P < .001) (Table 1).

Table 1.

Patient Demographics.

Procedure	Sex (Female), n (%)	Age, y, Mean ± SD (Range)	Laterality (Left), n (%)	BMI, Mean ± SD (Range)	Diabetes, n (%)	Follow-up, mo, Mean ± SD (Range)
Percutaneous Zadek (n = 34)	23 (67.8)	55.0 ± 12.4 (21-79)	19 (55.9)	37.0 ± 8.7 (18.6-59.0)	6 (17.6)	21.6 ± 7.3 (12-36)
Open midline Achilles tendon–splitting Haglund resection (n = 43)	34 (79.1)	54.0 ± 11.1 (21-78)	21 (48.8)	39.1 ± 8.6 (18.5-61.1)	9 (20.9)	32.1 ± 12.3 (12-53)
P	.256	.716	.539	.287	.718	<.001*

Abbreviation: BMI, body mass index.

^*Indicates statistically significant finding (P < .05).

PROMIS Scores

At a minimum of 12 months follow-up, in patients who underwent percutaneous ZO, PROMIS function scores improved from a mean 37.8 ± 6.2 (range, 21-51) to 44.0 ± 10.1 (range, 23-64) (P < .001). PROMIS pain scores improved from a mean 65.6 ± 5.2 (range, 55-75) to 52.9 ± 11.8 (range, 39-84) (P < .001). And PROMIS mobility scores improved from a mean 35.2 ± 6.8 (range, 19-46) to 42.6 ± 9.3 (range, 25-64) (P < .001) (Table 2).

Table 2.

PROMIS Scores in Patients Undergoing Percutaneous Zadek Osteotomy. ^a

	Preoperative, Mean ± SD (Range)	Postoperative, Mean ± SD (Range)	P
PROMIS function	37.8 ± 6.2 (range, 21-51)	44.0 ± 10.1 (23-64)	<.001*
PROMIS pain	65.6 ± 5.2 (55-75)	52.9 ± 11.8 (39-84)	<.001*
PROMIS mobility	35.2 ± 6.8 (19-46)	42.6 ± 9.3 (25-64)	<.001*

Abbreviation: PROMIS, Patient-reported Outcomes Measurement Information System.

^aPreoperative vs postoperative changes in PROMIS scores within each group were analyzed by paired t test. All improvements in PROMIS scores were significant (P < .001).

^*Indicates statistically significant finding (P < .05).

Similarly, in patients who underwent an open midline Achilles tendon–splitting Haglund resection, PROMIS function scores improved from a mean 37.6 ± 6.1 (range, 23-52) to 45.8 ± 10.3 (range, 29-76) (P < .001). PROMIS pain scores improved from a mean 65.4 ± 6.1 (range, 53-77) to 49.9 ± 11.2 (range, 39-72) (P < .001). PROMIS mobility scores improved from a mean 34.6 ± 5.0 (range, 28-47) to 44.6 ± 10.1 (range, 25-72) (P < .001) (Table 3).

Table 3.

PROMIS Scores in Patients Undergoing Open Midline Achilles Tendon–Splitting Haglund Resection. ^a

	Preoperative, Mean ± SD (Range)	Postoperative, Mean ± SD (Range)	P
PROMIS function	37.6 ± 6.1 (23-52)	45.8 ± 10.3 (29-76)	<.001*
PROMIS pain	65.4 ± 6.1 (53-77)	49.9 ± 11.2 (39-72)	<.001*
PROMIS mobility	34.6 ± 5.0 (28-47)	44.6 ± 10.1 (25-72)	<.001*

Abbreviation: PROMIS, Patient-reported Outcomes Measurement Information System.

^aPreoperative vs postoperative changes in PROMIS scores within each group analyzed by paired t test. All improvements in PROMIS scores were significant (P < .001).

^*Indicates statistically significant finding (P < .05).

There were no statistically significant differences detected in the change in PROMIS function, pain, or mobility score (preoperative – postoperative) between cohorts (P = .372, .334, and .313, respectively). Patients in both cohorts were found to improve in function, pain, and mobility scores larger than the minimum clinically important difference of 5 (Table 4).

Table 4.

Change in PROMIS Scores Postoperation Between Percutaneous ZO and Open Midline Achilles Tendon–Splitting Haglund Resection. ^a

Procedure	Change Function, Mean ± SD (Range)	Change Pain, Mean ± SD (Range)	Change Mobility, Mean ± SD (Range)
Percutaneous Zadek (n = 34)	6.2 ± 7.4 (−12 to 28)	−12.7 ± 11.6 (−35 to 16)	7.4 ± 7.6 (−10 to 23)
Open midline Achilles tendon–splitting Haglund resection (n = 43)	7.9 ± 8.4 (−6 to 25)	−15.3 ± 11.1 (−38 to 8)	9.3 ± 8.6 (−4 to 38)
P	.372	.334	.313

Abbreviation: PROMIS, Patient-reported Outcomes Measurement Information System.

^aNo statistical differences were identified in change (∆) in PROMIS scores between the 2 cohorts.

Postoperative Complications

There was a 2.9% (1/34) rate of major complications in the percutaneous ZO group. One of 34 patients (2.9%) experienced the same or increased pain after surgery, requiring revision with an open midline Achilles tendon–splitting Haglund resection at 10 months postoperation. This patient’s persistent pain was likely due to a wedge resection of less than 10 mm during the original operation. This patient additionally experienced a wound breakdown and infection 2 months following their revision procedure with the open midline Achilles tendon–splitting Haglund resection, requiring a subsequent irrigation and debridement in the operating room. Following additional interventions, the patient is doing well.

There was a 9.3% (4/43) rate of major complications in the open midline Achilles tendon–splitting Haglund resection cohort. Two of 43 patients (4.7%) who underwent the open midline Achilles tendon–splitting Haglund resection experienced an insertional Achilles rupture following their procedure secondary to a fall; both patients were treated nonoperatively. However, one of these patients was additionally diagnosed with wound dehiscence and evidence of a wound infection that required debriding in office and a course of oral antibiotics. One of 43 patients (2.3%) had significant hypersensitivity, neuritis, pain, and swelling following their open midline Achilles tendon–splitting Haglund resection procedure. This patient was diagnosed with complex regional pain syndrome (CRPS) and elected to proceed with treatment by pain management and a CRPS specialist at an outside institution. Finally, one of 43 patients (2.3%) reported significant, persistent pain postoperatively; this patient was found to have significant heterotopic bone formation about their Achilles insertion, and a wound dehiscence at that time. This patient underwent 2 revision procedures, both with an MIS resection. This patient did also experience postoperative neuritis after their revision procedures.

In patients who underwent the percutaneous ZO, there was a 14.7% (5/34) rate of minor complications. Two of 34 patients (5.7%) underwent removal of symptomatic hardware. One of 34 patients (2.9%) experienced plantar fascia pain >6 months following percutaneous ZO that was treated successfully with a corticosteroid injection. Separately, 1 of 34 patients (2.9%) experienced sural neuritis following their percutaneous procedure, which was also successfully treated with an ultrasound guided corticosteroid injection. Finally, 1 patient (1/34, 2.9%) was diagnosed with a minor superficial breakdown of their posterior wound, which was treated with nonweightbearing immobilization and wound checks.

In the open midline Achilles tendon–splitting Haglund resection cohort, there was a 25.6% rate of minor complications (11/43). Nine of 43 patients (20.9%) were diagnosed with a postoperative wound dehiscence, which were treated with a combination of prolonged nonweightbearing immobilization in a boot, local wound care, and/or more frequent wound checks. Two of these patients (2/43, 4.7%) had a concomitant wound infection that was treated with a course of oral antibiotics. Separately, 1 of 43 patients (2.3%) developed an ipsilateral Achilles contracture and noninsertional tendinopathy throughout their postoperative recovery. This was managed with oral steroids and physical therapy. Finally, 1 of 43 patients (2.3%) experienced significant heel pain following their open procedure and was found to have a stress fracture of the calcaneus and heterotrophic ossification at 1 year postoperation. This patient was offered a revision that has not yet undergone such intervention.

On χ² analysis, when directly comparing the rate of revision, removal of hardware, infection, nerve injury, plantar fascia pain, or other complications between the percutaneous ZO and open midline Achilles tendon–splitting Haglund resection cohorts, there was no statistical significance observed (P = .866, .107, .116, .866, .258, and .068, respectively). However, in comparison to the percutaneous ZO cohort (1/34, 2.9%), the open midline Achilles tendon–splitting Haglund resection cohort (11/43, 25.6%) had a statistically significantly higher rate of wound dehiscence (P = .007) (Tables 5 and 6).

Table 5.

Postoperative Complications in the Percutaneous ZO and Open Midline Achilles Tendon–Splitting Haglund Resection Cohorts. ^a

Complication	Percutaneous Zadek Osteotomy, n (%) (n = 34)	Open Midline Achilles Tendon–Splitting Haglund Resection, n (%) (n = 43)	P
Revision	1 (2.9)	1 (2.3)	.866
Removal of hardware	2 (5.9)	0 (0)	.107
Wound dehiscence	1 (2.9)	11 (25.6)	.007*
Infection	0 (0)	3 (7.0)	.116
Nerve injury	1 (2.9)	1 (2.3)	.866
Plantar fascia pain	1 (2.9)	0 (0)	.258
Other b	0 (0)	4 (9.3)	.068

Abbreviation: ZO, Zadek osteotomy.

^aData are represented as a fraction and percentage of the total N. Data were compared by χ² analysis.

^bComplications include postoperative insertional Achilles rupture (x2), stress fracture of the calcaneus, and Achilles contracture.

^*Indicates statistically significant finding (P < .05).

Table 6.

Overall Postoperative Complications in the Percutaneous ZO and Open Midline Achilles Tendon–Splitting Haglund Resection Cohorts. ^a

Procedure	Percutaneous Zadek, n (%) (n = 34)	Open Midline Achilles Tendon–Splitting Haglund Resection, n (%) (n = 43)	P
Total complications	6 (17.6)	15 (34.9)	.092
Major complications	1 (2.9)	4 (9.3)	.261
Minor complications	5 (14.7)	11 (25.6)	.243

Abbreviation: ZO, Zadek osteotomy.

^aData are represented as a fraction and percentage of the total N. Data were compared by χ² analysis.

Discussion

This retrospective series demonstrated significant 12- to 36-month improvements in PROMIS function, pain, and mobility after both percutaneous Zadek osteotomy and open midline Achilles tendon–splitting Haglund resection for insertional Achilles tendinopathy with or without Haglund deformity. The ZO cohort attained comparable symptom relief while exhibiting a lower incidence of wound-related complications (2.9 % vs 25.6 %), which is consistent with other reports of low wound complications for MIS foot and ankle surgery.^{1,5,8,19 –21,24,28,34,39,45,53,57,58,63} Nevertheless, these comparative findings should be regarded as preliminary: the study lacked an a priori power calculation and included only 34 and 43 cases per cohort, respectively. As a result, any nonsignificant between-group differences may reflect limited precision rather than clinical equivalence, and the apparent safety advantage of the minimally invasive approach awaits confirmation in larger, prospectively powered trials.

There was no significant difference between cohorts regarding the rate of overall, major, or minor complications. Nevertheless, there were a few specific trends observed in our data that may be considered potential side effects of both procedures to counsel patients on. For example, there were more instances of hardware removal and plantar fascia irritation (requiring an injection) in the percutaneous ZO cohort and more wound infections and postoperative insertional Achilles ruptures in the open midline Achilles tendon–splitting Haglund resection cohort but because of the size of our study, none of these were statistically significant.

In the systematic review by Bakaes et al, ¹ a mean complication rate of 5.8% was reported among the percutaneous ZO studies. This was echoed by the recent retrospective analysis by Hall et al, ²¹ who reported a complication rate of 3.8% with the percutaneous ZO in more than 100 patients at a minimum 2 years’ follow-up. Meanwhile, in the retrospective study by Kilic et al, ²⁹ the open midline Achilles tendon–splitting Haglund resection had a 20.6% overall complication rate, most commonly involving wound healing. In their retrospective study, So et al ⁴⁹ reported complications in 50% of their open Haglund resection cohort; most commonly their patients experienced sural nerve entrapment or wound complications. Interestingly, So et al reported a significantly higher (P = .01) complication rate in women (58%) in comparison to men (30%) in their population.

In the current study, the major complication rate for the percutaneous ZO and open midline Achilles tendon–splitting Haglund resection was found to be 2.9% (1/34) and 9.3% (4/43), respectively, aligning with previous literature. However, when additionally considering the minor complications observed, our overall complication rates were 17.6% (6/34) and 34.9% (15/43) among the percutaneous ZO and open midline Achilles tendon–splitting Haglund resection cohorts, respectively. Although comparable literature remains limited, the overall complication rates reported in the current study are on the higher side of what has been previously reported, possibly in the result of differences in complication classification systems and our potentially stricter definition of a wound healing complication. Additionally, in contrast to the findings of So et al, there was no difference in the complications observed between men and women in either cohort of the current study.

Because there are limited studies directly comparing the percutaneous ZO to open midline Achilles tendon–splitting Haglund resection, there remains much we do not yet know regarding best surgical management for patients with IAT ± Haglund deformity. However, the similar improvement in patient-reported outcomes observed between cohorts in the current study is congruent with findings of the very few retrospective studies in prior literature. In a smaller retrospective cohort study, Choi and Suh ⁶ compared the outcomes of patients undergoing the percutaneous ZO (n = 11) and the open midline Achilles tendon–splitting Haglund resection (n = 14). Patient-reported outcomes were collected in the form of visual analog scale (VAS) and Victorian Institute of Sports Assessment–Achilles (VISA-A) scores. They found no significant difference in postoperative VAS and VISA-A scores between cohorts. ⁶

The present study has several important limitations, including its retrospective design, unmatched baseline confounders, and the lack of a prestudy power analysis First, this was a retrospective study. Patients were not age, sex, or body mass index matched between cohorts. Cases were not prospectively randomized; therefore, cohorts did not have equal distribution of potentially confounding factors, such as comorbidities. To determine the degree of influence each of these variables had over our results, we studied them each in our retrospective analysis. There was no statistical difference in the distribution of confounding variables observed. Additionally, the senior authors perform ZO on all patients who would be candidates for open midline Achilles tendon–splitting Haglund resection, therefore limiting selection bias. This retrospective study was limited to the patients of 2 fellowship-trained foot and ankle surgeons at a single academic institution. Each surgeon involved in the study has extensive experience with MIS foot and ankle surgery. Nevertheless, the percutaneous ZO has been associated with a learning curve only relevant regarding surgical time. ¹⁹ This learning curve is resolved around case 14 and has demonstrated no influence over patient-reported outcomes or complications following percutaneous ZO. Still, the relatively small size of this study may have limited the generalizability of these data, in addition to the statistical differences we were able to observe. Finally, sample-size imprecision and the exploratory nature of our analyses preclude definitive comparative conclusions.

Despite these limitations, this study was able to demonstrate improvement in patient-reported pain, function, and mobility following both percutaneous ZO and the open midline Achilles tendon–splitting Haglund resection. We observed significantly fewer wound complications in patients who underwent a percutaneous ZO in comparison to the open midline Achilles tendon–splitting Haglund resection. Findings in the current study may assist surgeons when deciding on an appropriate operative plan for patients with IAT.

Conclusion

Both techniques yielded clinically meaningful improvements in PROMIS scores. In this modest retrospective cohort, the percutaneous Zadek osteotomy was associated with fewer minor wound complications; however, this observation should be interpreted cautiously given the limited sample size, short follow-up, and potential selection bias. Larger, prospective multicenter studies are needed to confirm these preliminary findings.