Distal Oblique Osteotomy vs Arthrodesis for Advanced Hallux Rigidus: A Matched Cohort Study

Study Design and Patient Selection

This study was a retrospective review of prospectively collected registry data from a single tertiary orthopaedic centre, encompassing patients treated between 2007 and 2024. Institutional ethics approval was obtained (Ref: 2019/2778).

Only patients with advanced hallux rigidus who had failed at least 6 months of conservative management and subsequently underwent either first MTP joint arthrodesis or DOO were included. Patients were invited for scheduled clinical follow-up visits, during which objective clinical outcomes and patient-reported outcome measures (PROMs) were assessed. Patients who were lost to follow-up were contacted by telephone; those for whom contact could not be established were excluded from the analysis.

Patients were excluded if they had any of the following: (1) inflammatory arthropathies, (2) hallux varus or valgus deformity >15°, (3) prior foot or ankle surgery, (4) revision cases, or (5) previous midfoot or hindfoot fusion procedures.

In addition, matching was performed to improve comparability between the 2 cohorts. Patients were matched based on age, sex, body mass index, radiographic severity of hallux rigidus, and hallux valgus angle using a nearest-neighbour approach without replacement. This was intended to reduce baseline differences and minimise potential confounding when comparing postoperative outcomes. Covariate balance following matching was evaluated using standardised mean differences (SMD), with values below 0.10 considered indicative of adequate balance between groups. ¹²

Surgical Procedures

All procedures were performed by fellowship-trained orthopaedic foot and ankle surgeons at our institution. Two main surgical techniques were analysed—arthrodesis of the first MTP joint and DOO as described by Malerba et al. ¹³

Arthrodesis

In the arthrodesis group, a dorsal or medial approach was used depending on the surgeon’s preference. Joint preparation was performed. The joint was aligned in approximately 5° to 10° dorsiflexion and 0° to 5° valgus, ensuring neutral rotation under simulated weight-bearing. Fixation was achieved with either a dorsal plate and lag screw construct or crossed compression screws. Postoperatively, patients were placed in a rigid flat postoperative shoe and allowed immediate weight-bearing as tolerated. Stitches were removed at 2 weeks, and patients returned to regular footwear by 6-8 weeks or when there was evidence of radiographic union.

Dorsal Oblique Osteotomy

In the DOO group, the MTP joint was approached through a medial incision and the dorsal osteophytes were removed according to Mann’s technique. ¹⁴ A linear osteotomy (ie, DOO) was directed at an angle ranging from 35° to 45° in the sagittal plane, starting dorsally, adjacent to the articular cartilage and aiming proximally and plantarly (Figure 1). The osteotomy angle was subjectively assessed by the surgeon intraoperatively, with the intention to aim parallel to the plantar plane. On completion of the DOO, the metatarsal head displaced spontaneously and found its own level. Fixation was performed with either 1 or 2 headless compression screws under radiographic guidance, and the dorsal shelf of bone was shaved in line with the remaining metatarsal head. Post-surgery, patients were discharged with a heel-wedge postoperative shoe (DARCO MedSurgTM Shoe) and were allowed protected weight-bearing for 6 weeks. Stitches were removed at 2 weeks post-operatively, and patients were allowed to return to full activity by 12 weeks or when the osteotomy healed. ¹⁵

OPEN IN VIEWER

Figure 1.

Distal oblique osteotomy (DOO) procedure schematic. (A) Dorsal cheilectomy. (B) Oblique osteotomy. (C) Fixation of osteotomy with 2 headless screws.

Radiographic Grading of Hallux Rigidus

Radiographic evaluation included anteroposterior, lateral, and oblique weight-bearing radiographs of the affected forefoot taken preoperatively (Figure 2). The severity of first MTP joint degeneration was graded by 2 blinded orthopaedic surgeons using the Hattrup and Johnson Radiographic System for Grading Hallux Rigidus. ¹⁶ In cases of disagreement, a third senior orthopaedic surgeon provided the final grade. The Hattrup and Johnson classification was used as this study was based on retrospective radiographic review with limited standardised clinical data. As the Coughlin and Shurnas system incorporates clinical parameters such as pain profile and first MTP joint motion, Hattrup and Johnson was better aligned with the available data set and allowed more consistent radiographic grading while minimising misclassification. ⁷

OPEN IN VIEWER

Figure 2.

(A) Arthrodesis: postoperative radiograph. (B) Distal oblique osteotomy (DOO): pre- vs postoperative radiographs.

Outcome Assessment

Preoperative and postoperative functional data at 6 months and 24 months were collected by trained physiotherapists. First MTP joint ROM was measured using a goniometer along the longitudinal axes of the first metatarsal and proximal phalanx while the ankle was placed in neutral position. ¹⁷

The primary study outcomes were postoperative complication rates and patient satisfaction at 24 months. Wound complications were defined as any superficial skin infection or delayed healing that required medical intervention, such as the use of antibiotics. Malunion was defined as the development of a deformity resulting in a deviation from the originally intended alignment.

Secondary outcomes included visual analogue scale (VAS) for pain, ¹⁸ first metatarsophalangeal joint ROM, and American Orthopaedic Foot & Ankle Society (AOFAS) functional scores. ¹⁹ The VAS was used to evaluate pain specific to the first MTP joint on a 0 to 10 scale, where 0 indicated no pain and 10 represented the worst possible pain. To allow clearer interpretation of joint mobility, the total ROM, which includes the combined dorsiflexion and plantarflexion, of the first MTP joint was measured by an independent physiotherapist and was subsequently scored as 10 points for normal or mild restriction (≥75°), 5 points for moderate restriction (30°-74°), or 0 points for severe restriction (<30°). The AOFAS score assessed 3 components—pain (40 points), function (45 points), and alignment (15 points), with higher scores reflecting better outcomes. ²⁰ The minimal clinically important difference (MCID) for VAS was defined as 1.5, ²¹ and to date no validated MCIDs have been found for either AOFAS or ROM.

Patient satisfaction at 24 months postoperatively was evaluated using standardised questions adapted from the North American Spine Society (NASS) questionnaire, which has been previously validated in spinal surgery populations as a measure of patient-reported satisfaction.^22,23 Questions included were “How would you rate the overall results of your treatment?” and “Has the surgery met your expectations so far?” Responses were first collected on a 7-point Likert scale and subsequently converted to a binary measure of whether expectations were met. ²²

Complications recorded included wound complications, nonunion, malunion, implant prominence, and hardware failure. The re-operation rate was recorded, as well as the reasons for re-operation. Radiographic progression or secondary degenerative changes in the DOO group were not routinely assessed as part of the study outcomes.

Statistical Analysis

All statistical analyses were conducted using IBM SPSS Statistics v26.0 (IBM Corp). Continuous variables such as age, body mass index, VAS, and AOFAS scores were tested for normality using the Shapiro-Wilk test. ROM score was analysed as a continuous variable at the group level, as mean values were derived from averaged ordinal scores.

Depending on data distribution, comparisons between groups were performed using either the independent samples t-test or the Mann-Whitney U test. Within-group changes over time were analysed using the Wilcoxon signed-rank test. Categorical variables such as sex, complication rates, and revision surgeries were analysed using the χ² test or Fisher exact test when expected cell counts were low. ²⁴

All P values were 2-tailed, and a threshold of P <.05 was considered statistically significant. Data were presented as mean ± SD for continuous variables and as frequency (percentage) for categorical variables.

Although additional multivariable regression or sensitivity analyses could have been performed to further adjust for residual imbalance, these were not undertaken because of the limited sample size, which may reduce model stability and increase the risk of overfitting. ²⁵ Instead, baseline comparability was assessed using standardised mean differences, and residual imbalance was acknowledged in the interpretation of results.

Demographics and Baseline Characteristics

A total of 153 patients underwent surgery for hallux rigidus between 2007 and 2024 at a single tertiary institution. A total of 83 patients were excluded, comprising 76 patients with hallux valgus or varus deformity, 2 patients with gout, and 5 patients with mild to moderate disease. Following purposive matching, 63 patients met the inclusion criteria. Of these, 4 patients did not complete 24-month follow-up assessments and were excluded from outcome analyses, resulting in a final study cohort of 59 patients, of whom 28 underwent arthrodesis (fusion) and 31 underwent DOO (Figure 3).

OPEN IN VIEWER

Figure 3.

Patient flowchart.

Baseline demographic and preoperative clinical characteristics are summarised in Table 1. There were no significant differences between the fusion and DOO groups with respect to age, sex distribution, body mass index, preoperative VAS, preoperative ROM score, or preoperative AOFAS score. All patients included had advanced rigidus, and the mean duration of follow-up was 2.80 ± 2.50 years. Post-matching balance assessment demonstrated excellent balance for sex (SMD = 0.01) and acceptable balance for body mass index (SMD = 0.12). Residual imbalance was observed for age (SMD = 0.42) and baseline clinical severity measures, including preoperative pain (SMD = 0.38), joint motion scores (SMD = 0.47), and AOFAS (SMD = 0.34).

OPEN IN VIEWER

Table 1.

Patient Characteristics.

Characteristic	Fusion (n = 28)	DOO (n = 31)	SMD
Age, y, mean ± SD	58.0 ± 4.8	55.2 ± 8.2	0.42
Female	17 (60.7)	19 (61.3)	0.01
BMI, mean ± SD	24.9 ± 3.0	24.5 ± 3.5	0.12
Preop. VAS score, mean ± SD	6.5 ± 2.4	5.7 ± 1.7	0.38
Preop. range of motion, degrees, mean ± SD	3.9 ± 2.5	5.3 ± 3.3	0.47
Preop. AOFAS score, mean ± SD	41.71 ± 15.0	46.42 ± 12.4	0.34

Abbreviations: AOFAS, American Orthopaedic Foot & Ankle Society; BMI, body mass index; DOO, distal oblique osteotomy; SMD, standardized mean difference; VAS, visual analog scale.

Comparison Between Fusion and DOO Outcomes

Primary outcomes

Patient satisfaction was high and comparable between both groups (fusion = 89.3%, DOO = 90.3%, P = 1).

In the fusion group, unsatisfied patients reported persistent pain due to prominence of the plate, limited function due to malunion, and also revision surgery required because of breakage of the plate. Meanwhile, in the DOO group, unsatisfied patients reported recurrent pain, residual stiffness and shortening of the first toe after the surgery.

The fusion group (28.6%) demonstrated a significantly higher (P = .025) overall complication rate compared with the DOO group (3.2%).

With respect to major complications (wound complications, nonunion, malunion and hardware failure), the fusion group demonstrated higher rates, including hardware failure (10.7%), wound infection (7.1%), and malunion (7.1%). In contrast, the DOO group reported only 1 case of hardware failure (3.2%). For minor complications, implant prominence was observed in the fusion group (3.6%) (Table 2).

OPEN IN VIEWER

Table 2.

Comparison Between Fusion and DOO Complications.

Complication	Fusion, n (%)(n = 28)	DOO, n (%)(n = 31)
Wound complications	2 (7.1)	0
Nonunion	0	0
Malunion	2 (7.1)	0
Implant prominence	1 (3.6)	0
Hardware failure	3 (10.7)	1 (3.2)
Total	8 (28.6)	1 (3.2)

Abbreviation: DOO, distal oblique osteotomy.

The revision surgery rate was lower in the DOO group (3.2%) than in the fusion group (17.8%), although this difference did not reach statistical significance (P = .062).

Secondary Outcomes

The majority of the patients in both fusion and DOO groups had an improvement in both VAS scores. In addition, 92.9% (n = 26) of the fusion and 96.8% (n = 30) of the DOO groups achieved the MCID for VAS improvement, with mean reductions of 5.86 points in the fusion group and 5.22 points in the DOO group. One patient from each group experienced worsening of VAS scores. The patient in the fusion group developed a hardware breakage, whereas no specific complication was identified in the DOO group.

The DOO group was associated with a significantly higher AOFAS score (P < .05) compared with the fusion group. No significant differences were noted in VAS as seen in Table 3.

OPEN IN VIEWER

Table 3.

Comparison Between Fusion and DOO Outcomes.

Outcome	Fusion (n = 28)	DOO (n = 31)	95% CI of Mean Difference	Effect Size	Significance (P Value)
Complications, n (%)	8 (28.6)	1 (3.2)	–	–	.025
Revision surgeries, n (%)	5 (17.8)	1 (3.2)	–	–	.062
Satisfied, n (%)	25 (89.3)	28 (90.3)	–	–	1
Satisfaction score, mean ± SD	2.14 ± 1.04	1.84 ± 1.07	−0.85 to 0.25	−0.28	.274
VAS, mean ± SD	0.64 ± 1.59	0.48 ± 1.55	−0.98 to 0.66	−0.10	.699
Range of motion, mean ± SD	1.43 ± 2.30	7.90 ± 2.82	5.13 to 7.81	2.47	<.05
AOFAS, mean ± SD	81.54 ± 15.44	92.23 ± 8.84	3.99 to 17.39	0.85	<.05

AOFAS, American Orthopaedic Foot & Ankle Society; DOO, distal oblique osteotomy; VAS, visual analogue scale.

The DOO group also demonstrated higher postoperative ROM scores compared with the fusion group (P < .05). Within the DOO group, ROM score improved from 5.3 ± 3.3 preoperatively to 7.9 ± 2.8 at 24 months with a mean difference of 2.6 points (P < .05), suggesting increased joint mobility at 24 months. However, the between-group ROM comparison should be interpreted with caution, as the DOO group entered surgery with a higher baseline ROM score (SMD = 0.47), and the observed postoperative difference between groups reflects both the procedural distinction and this pre-existing imbalance.

Effect Sizes

Effect sizes were calculated to quantify the magnitude of treatment differences between groups (Table 3). For postoperative VAS pain, the effect size was small (Cohen d = −0.10), suggesting minimal difference between groups. Similarly, the effect size for postoperative satisfaction score was small (Cohen d = −0.28), indicating little difference in perceived satisfaction between the 2 procedures. In contrast, the between-group effect size for postoperative AOFAS was moderate to large (Cohen d = 0.86). Moreover, the effect size for postoperative ROM was large (Cohen d = 2.43), indicating a substantial difference in joint motion between the two procedures.

Strengths and Limitations

This study has several notable strengths. First, although analysed retrospectively, the data were prospectively collected within a registry, which reduces recall bias and improves data completeness. Second, the use of multiple clinically relevant outcome measures, including pain, function, patient satisfaction, and complications, allowed for a comprehensive assessment of surgical outcomes. Third, matching for key demographic and clinical variables helped minimise confounding and improved the comparability between treatment groups.

In addition, the 2-year follow-up provided meaningful midterm outcome data for both pain relief and functional recovery. The study was conducted at a single tertiary institution with procedures performed by specialist orthopaedic surgeons, which ensured consistency in surgical technique, postoperative care, and outcome assessment. The study was conducted at a single tertiary institution, with all procedures performed by specialist orthopaedic surgeons, ensuring consistency in surgical technique, postoperative care, and outcome assessment.

Despite these strengths, the long recruitment period remains an important limitation and introduces the possibility of temporal confounding. Over the study interval, foot and ankle surgery evolved considerably, with advances in implant design, broader adoption of minimally invasive techniques, and a progressive shift toward earlier postoperative weight-bearing protocols. In parallel, the organisation of orthopaedic practice has also changed. Procedures that were historically performed by general orthopaedic surgeons are now increasingly undertaken by dedicated foot and ankle subspecialists, reflecting growing procedural centralisation and specialist expertise. ⁴⁰ This evolution in surgeon experience and subspecialisation may itself influence patient selection, intraoperative decision making, technical precision, complication rates, and rehabilitation pathways. As such, differences between cohorts may not be attributable solely to the intervention under study, but may also reflect broader changes in surgical expertise and standards of care over time. Although an era-based subgroup or sensitivity analysis may have provided additional insight into temporal trends in surgical practice, the relatively small sample size limited meaningful stratification without compromising statistical validity. As such, this analysis was not performed.

As with any retrospective study, there were several limitations as well. The retrospective design introduces the inherent risk of selection bias and confounding. This was mitigated by the use of standardised data collection and matching criteria. The sample size was relatively small, which may limit the statistical power to detect differences in less frequent outcomes, such as revision surgery or rare complications. Although matching improved baseline comparability between treatment groups, several variables did not meet the predefined threshold for adequate balance (SMD < 0.10), including age, preoperative VAS, ROM, and AOFAS scores. Notably, the DOO group demonstrated more favourable baseline functional measures, which may reflect differences in disease severity and could have influenced postoperative outcomes, particularly AOFAS and ROM, potentially biasing results in favour of DOO. Additionally, patient satisfaction was assessed using an adapted NASS questionnaire, although validated in spinal surgery populations,^22,23 it has not been formally validated for use in foot and ankle conditions. As such, the interpretation of satisfaction outcomes should be approached with caution and considered exploratory. The exclusion of 4 patients due to loss to follow-up may have introduced loss-to-follow-up bias, a limitation commonly encountered in hallux rigidus series where sustained long-term surveillance is challenging. Nonetheless, every effort was made to contact patients for follow-up either physically or by telephone. Another limitation is the use of the AOFAS score. Although widely used in foot and ankle research, and with some studies demonstrating acceptable reliability, it is not a purely patient-reported outcome measure and includes clinician-assessed components.^41-43 This may introduce bias and limit its ability to fully capture patient-perceived quality of life. In particular, AOFAS places greater emphasis on pain, with comparatively less weighting on factors such as stiffness and deformity. In this study, AOFAS was therefore analysed as a secondary exploratory functional outcome rather than a primary endpoint, to allow comparison with historical literature and previously collected registry data. ⁴⁴ Additionally, the component of ROM was analysed in more detail to account for the possibility of stiffness contributing to the quality of life of a patient. The principal outcomes of this study were complication rates and patient satisfaction, which are clinically robust and less dependent on instrument validity.^{15,19,26,27,32,45,46} Nonetheless, as this study also included VAS scores and patient satisfaction rates, a more holistic evaluation of clinical outcome was obtained. Moreover, multiple clinical and radiographic outcomes were evaluated in this study. As no formal adjustment for multiple comparisons was applied, there is an increased risk of type I error. Therefore, secondary outcome analyses should be regarded as exploratory and interpreted with appropriate caution. Taken together, these factors limit the ability to draw definitive conclusions regarding functional superiority between procedures.

Future studies with prospective or randomised designs, larger sample sizes, and extended follow-up would be valuable to confirm these findings and more definitively compare the long-term roles of DOO and arthrodesis in the management of advanced hallux rigidus.