Radiographic and Patient-Reported Outcomes of a Low-Cost Modified Lapidus Bunion Correction Technique

Study Design

Institutional review board approval was obtained prior to initiation of the study. We performed a retrospective review of prospectively collected radiographic data and patient-reported outcomes (PROMIS-PF CAT). We identified 117 patients who underwent Lapidus by crossed lag screw technique from January 2014 to December 2020. Inclusion criteria were patients aged 14-85 years old undergoing primary bunion correction surgery by modified Lapidus procedure. Although we did not exclude anyone on the basis of medical comorbidities during our retrospective review, patients were medically optimized for any comorbid conditions prior to being offered surgery. After applying inclusion and exclusion criteria, we identified 73 feet in 68 patients eligible for inclusion. Patients provided their written consent to proceed with review of their data.

Preoperative Visit, Intraoperative Technique, and Postoperative Rehabilitation

Patients were seen in clinic preoperatively. We routinely collect PROMIS-PF scores via a CAT survey. A concerted effort was made to collect these scores on each patient, although this was based on the singular efforts of the lead author and was not consistent, especially early in the study period. Preoperative weightbearing radiographs consisting of anteroposterior, lateral, oblique, and standing axial sesamoid views were obtained prior to surgery.

For surgical technique, patients were placed in the supine position on the operating table. A thigh tourniquet was applied and regional neuraxial anesthesia was used along with general anesthesia. Patients were then prepped and draped in a sterile fashion. A scalpel was used to perform a percutaneous lateral release of the first metatarsophalangeal (MTP) joint lateral joint capsule and adductor hallucis insertion to allow for a manual varus manipulation of 20 degrees. A dorsomedial longitudinal incision was then made medial to the EHL and centered over the first tarsometatarsal (TMT) joint. Dissection was carried down to the first TMT joint. A small osteotome was placed into the joint to initially release the first tarsometatarsal capsule, particularly the medial and plantar capsule. With the soft tissue released, the bone cut from the medial cuneiform was then performed. The cut was made in the coronal plane with the same orientation of the joint, which was probed prior to the cut with the oscillating saw, and this dorsal to plantar trajectory was then marked on the medial skin of the foot. Extreme care was taken with this trajectory to avoid final dorsiflexion or plantarflexion deformity of the first ray. The cuneiform wedge cut was determined by taking a conservative cut of 2 to 3 mm from the lateral side of the medial cuneiform articular surface, with the medial side of the cut with minimal resection from the articular surface, effecting a wedge removal in the axial plane, while leaving the sagittal position of the first ray unchanged. A threaded 3.2-mm Steinmann pin was placed in the first metatarsal to allow rotational correction, with a reduction clamp also placed between the first and second metatarsals. Correction was provisionally held with a 2.0-mm threaded Kirschner wire. The fusion site was then assessed for congruity in the provisionally fixed position; if gaps were visible, the space was filled either with autograft from a separate harvest of 2-3 cm³ of calcaneal autograft. Two solid stainless steel 3.5-mm cortical screws (Stryker, Kalamazoo, MI) were placed across the first TMT in crossing fashion using a lag by technique. A third solid stainless steel 3.5-mm cortical screw was used to maintain the intermetatarsal correction and was placed from the first metatarsal through the second. If purchase of the first intermetatarsal screw was not excellent, a fourth screw was placed parallel and just distal to the third screw between the first and second metatarsals. Wounds were irrigated and skin was closed with 3-0 nylon sutures. Soft minimally compressive dressings were placed and the patient was placed in a postoperative shoe and made weightbearing as tolerated with crutch or walker assistance.

After surgery, patients were seen for 2-week, 6-week, and 3-month visits. Patients were encouraged to keep following up for a 6-month, 1-year, and 2-year follow-up, but many elected not to. Radiographs were routinely obtained at the 6-week and 3-month visits (Figure 2). Patients were radiographically followed until the 1-year postoperative visit according to their choice of returning for this visit. We obtained weightbearing anteroposterior, oblique, and lateral views of the foot as well as a standing axial sesamoid view of the first ray. Patients do not routinely receive physical therapy. Patients were allowed to transition to regular wide toe box shoe-wear as tolerated, which typically occurs around the 6-week postoperative visit.

Analysis and Statistics

Descriptive summary statistics were obtained and reported. To analyze PROMIS-PF scores, we binned postoperative scores according to postoperative date (periods: preoperative, postoperative day 1-45, 45-90, 90-180, 180-360, and 360-1400). We then used analysis of variance (ANOVA) to test the difference between groups. A post hoc comparison was implemented using Tukey method to control an overall type I error under 5%. ¹⁹ P values were reported and noted to be significant when under this value. All calculations and data operations were performed using R studio (open source, Boston, MA).

For radiographic outcomes, 2 authors (J.N.T. and C.S.P.) performed measurements of HVA, IMA, and tibial sesamoid position (TSP) as described previously in the literature. ¹⁵ These were measured and reported on routine radiographs at preoperative and 1-year postoperative visits. Continuous variables such as HVA and IMA were reported for the difference between the 2 authors. The agreement in TSP as a categorical variable was reported in Cohen kappa. Both statistics were reported with a 95% CI.

Finally, costs were obtained from the manufacturer implant costs recorded in the charge capture at the time of surgery and reported as an average.

Baseline Patient Demographics

Patients had a median age of 55.8 years (IQR 45.6-53.9), were 4 of 73 (5.5%) male, and 15 of 73 (20.6%) diabetic. Diabetic patients had a mean HbA_1c of 6.4% (IQR 6.0-7.4). Of our diabetic patients, 5 of 15 (33%) had documented neuropathy, and none were insulin dependent within 30 days of surgery or postoperatively. Current smokers comprised 11 of 73 (15.1%) of our patient population, although all were required to stop smoking for at least 1 month prior to surgery, as reported to the surgeon in the preoperative visit. These results are tabulated in Table 1.

OPEN IN VIEWER

Table 1.

Patient Demographics (Reported for 73 Feet in 68 Patients).

Demographic	Mean (IQR) or n/n (%)
Age, mean (IQR)	55.8 (45.6-53.9)
Male	4/73 (5.5)
Diabetes	15/73 (20.6)
Insulin dependent	0/15 (0)
HbA1c, %, mean (IQR)	6.4 (6.0-7.4)
Neuropathy	5/15 (33)
Smoker	11/73 (15.1)

Abbreviation: HbA_1c, glycated hemoglobin; IQR, interquartile range.

PROMIS-PF Scores

A total of 35 patients (51% of the eligible cohort) had preoperative PROMIS-PF surveys. The preoperative PROMIS-PF score was a median of 43 (IQR 37-52). Postoperative PROMIS-PF data were available for a total of 42 patients (62% of the eligible cohort) for at least 1 time point. PROMIS-PF dropped significantly at the 6-week visit to 37 (n = 11, IQR 31-39, P = .006) but rose significantly by the 3-month visit (n = 23, median 46, IQR 42-51, P = .008). At time points greater than 1 year postoperatively, patients achieved a median PROMIS-PF of 49 (n = 42, IQR 41-53). There were no significant differences between initial preoperative scores and either the 3-month or the 1-year postoperative scores (P > .05 for all comparisons). These scores are represented graphically in Figure 1.

OPEN IN VIEWER

Figure 1.

PROMIS-PF scores vs postoperative time in days. Preoperative scores fell in the first 45 days postoperatively (P = .006* for [0-1) vs [1-45)) but rose by 90 days post operatively (P = .008* for [1-45) vs [45-90)). These gains were maintained until final available follow-up (P > .05 for remaining comparisons).

OPEN IN VIEWER

Figure 2.

Representative radiographic outcome: preoperative (A) upright anteroposterior foot, (B) oblique foot, (C) sesamoid, and (D) lateral foot views shown. Final upright postoperative radiographs at 3 years and 5 months are shown as follows: (E) lateral foot, (F) anteroposterior foot, (G) oblique foot, and (H) sesamoid view are shown.

Radiographic Outcomes

Our cohort had a mode radiographic follow-up of 82 days with a median of 173 days. Preoperatively, HVA was 35 degrees (IQR 30-40 degrees), which improved to 10 degrees (IQR 5-15 degrees) at the final postoperative visit (P < .001). IMA improved significantly from 14 degrees (IQR 12-16 degrees) to 6 degrees (IQR 3-9 degrees) (P < .001). TSP was found to be 6.05 ± 1.00 preoperatively, improving to 2.22 ± 1.38 postoperatively. The agreement between the 2 authors in TSP is 0.61 (95% CI 0.70, 0.78), which falls in the range of substantial agreement justified in Landis and Koch. The authors’ HVA and IMA differences are 0.019 and −0.21, respectively. Both confidence intervals cover 1, showing no difference statistically.

Complications

Wound healing issues were defined as persistent serosanguinous drainage or wound gapping requiring leaving sutures at the 2-week follow-up or requiring local wound care and antibiotics. These occurred in 6 (8.2%) of 73 patients. Hardware complication, defined as painful and broken hardware that required removal, was seen in 9 (12.3%) of 73 feet. Three total nonunions occurred (4.1%). One (1 of 73, 1.4%) resolved after correction of hypothyroidism. One was associated with a recurrence that sought revision surgery elsewhere (1.4%), and 1 was an asymptomatic nonunion that did not desire further surgery (1.4%). Hallux varus due to overcorrection was seen in 1 patient (1.4%). There were 2 painful corrected bunions without nonunion (2 of 73, 2.7%).

Costs and Operative Time

On average, we used 3 solid stainless-steel screws and 2 Kirschner wires, a saw blade, and drapes (supplies). Our average (±SD) implant costs were $146 ± 108, and our average supply costs were $322 ± 257. The average operative time was 105.8 ± 27.7 minutes.