Postoperatively, patients wore a wooden-soled shoe for a minimum of 6 weeks and were allowed to fully weightbear without crutches on the first postoperative day. Patients were seen at Weeks 1, 2, 4, and 6 after surgery for dressing changes. Stitches were removed 2 weeks postoperative; radiographic control films (foot in weightbearing position, AP and lateral views) were obtained 6 weeks postoperative. If these radiographs showed complete healing of the osteotomy, the patient was allowed to wear regular shoes. Return to full activity and sports was allowed 2 months postoperative.
After the 6-week followup, we followed patients at 3 months and at the final followup. The following clinical and functional variables were evaluated preoperatively and at the last followup: unsegmented VAS score for pain (0-10); ROM of the MTP1; and AOFAS score . Radiographs were taken preoperatively, at 6 weeks and at final followup.
Three of us (UL, AK, MP) evaluated the radiographs for alignment according to the guidelines described by the AOFAS  on weightbearing AP radiographs. The following variables were assessed preoperatively, 6 weeks postoperatively, and at last followup: HVA; IMA; DMAA; metatarsal index measured by a perpendicular line drawn from the distal-most point of the MT2 to the MT1; and position of the tibial sesamoid on the AP radiograph in relation to a line drawn along the center of the longitudinal axis of the MT1. A dislocation of 25% means ¼ of the tibial sesamoid bone laterally overlaps this line. The axis of the MT1 was determined as follows: a line was drawn from the center of the head of the MT1 through the center of the base of the MT1 as described by Miller . As there is extensive literature regarding interobserver variability of measurement of the HVA (0.98/0.96; preoperative/postoperative), IMA (0.94/0.86; preoperative/postoperative), and DMAA (0.09/0.002; preoperative/postoperative) [6, 26, 28, 29], only the metatarsal index and sesamoid bone position also were measured by three individuals (UL, AK, MP) not involved in the clinical evaluation. We found a Cronbach's alpha between 0.893 (sesamoid bone position preoperative/postoperative) and 0.988 (index preoperative/postoperative).
Variables were described by frequencies, mean ± SD, and range. Differences among the three subgroups regarding preoperative and postoperative variables were analyzed using the Kruskal-Wallis test; differences between times (evaluation of improvement of clinical, functional, and radiographic outcome variables) were calculated using the Wilcoxon test. Spearman's correlation was used to assess a possible relationship for the following parameters: the loss of reduction of HVA at final followup versus different preoperative and postoperative values (HVA, IMA, sesamoid bone position, loss of correction of sesamoid bone position). Statistical analysis was performed using SPSS® 15.0 (SPSS Inc, Chicago, IL).
Pain according to the mean VAS improved (p < 0.001) from 5.9 points preoperatively to 0.4 points postoperatively (Table 3).
The mean AOFAS score improved (p < 0.001) from 56 points to 90 points (Table 3). Some subscore elements (interphalangeal joint motion, metatarsophalangeal-interphalangeal stability) remained unchanged (Table 4).
The radiographic outcome measurements improved at last followup (Table 3). Shortening of the MT1 was expressed by the first metatarsal index with a preoperative value of −1.0 mm (range, −9 mm to +7 mm), a postoperative value of −2.9 mm (range, −10 mm to +4 mm), and average shortening of 1.9 mm (range, 0-4 mm).
The postoperative AOFAS score differed (p = 0.024) among the three groups: the group after previous osteotomy showed the lowest score (87 ± 3), followed by the group after resection arthroplasty (89 ± 7), and the group without primary osteotomy (92 ± 6) (Table 5). We observed no difference (p > 0.05) in the VAS for pain and the radiographic variables between the three groups.
The HVA (+2.0°; p = 0.005) and the dislocation of the sesamoid bone (+4.6%; p = 0.020) increased during the period between the 6-week and final followup radiographs (Table 6). The IMA and DMAA did not change. We found no correlation between the increase of the HVA and dislocation of the sesamoid bone and any of the following parameters: preoperative HVA, IMA, sesamoid bone position; HVA, IMA, sesamoid bone position 6 weeks postoperatively and at final followup.
Seven patients (seven feet) wore orthotic devices preoperatively (five after primary resection arthroplasty, two after bunionectomy) attributable to metatarsalgia and continued wearing them postoperatively. One patient had an asymptomatic recurrence (20° HVA, 10° IMA) after a bunionectomy; the preoperative HVA measured 35°; the preoperative IMA was 16°; no additional surgery was performed. One patient had overcorrection of 3° hallux varus after a primary resection arthroplasty; the patient complained of mild pain (VAS, 2). Three patients had intermittent paresthesias and five had pain attributable to the screw (three, head of the screw; two, tip of the screw), and the screws were removed. All the osteotomies had healed by the last followup. No patients had head necrosis, secondary displacement, or troughing (ie, when the cortices wedge into the softer cancellous bone of the metatarsal shaft, causing a functional elevation and malrotation of the first ray). No patient had increased metatarsalgia.
Recurrent hallux valgus is a well-known complication. Among other reasons, insufficient lateral displacement or failure to perform an osteotomy, causing insufficient or no correction of an increased IMA, can lead to recurrence. Knowing the potential of the Scarf osteotomy to correct an increased IMA, we chose this method for revision of recurrent hallux valgus deformity for selected patients. To see whether the Scarf osteotomy successfully corrected the recurrent deformities, we asked whether (1) the Scarf osteotomy results in a reduction of the pain level; (2) the Scarf osteotomy improves the AOFAS score; (3) the Scarf osteotomy improves radiographic outcome variables; (4) results of the aforementioned outcome variables (VAS for pain, AOFAS score, radiographic outcome variables) differ among the three different study subgroups (5) the HVA, IMA, and DMAA and the dislocation rate of the sesamoid bones increase during the period between the 6-week postoperative control and final followup radiographs.
There are some limitations to our study. First, the number of patients is relatively small. This can be attributed to the generally low numbers of hallux valgus revision surgeries performed (Table 7). Second, we imposed limited indications and had some contraindications for the surgery. Not every patient with complaints after primary hallux valgus surgery was eligible for the Scarf osteotomy; however, our aim was to determine whether the Scarf osteotomy was reasonable for treating recurrent hallux valgus in a selected group of patients sharing a similar pathomechanism. We did not consider this operation in patients for whom we considered other methods to be more suitable (eg, cock-up deformity, unstable TMT1) [8, 9, 19, 21]. Third, we lacked historical or concurrent controls. Our goal was to determine the applicability of the Scarf osteotomy in a selected group of patients and not its superiority over other methods. Fourth, the majority of the patients included in this study probably are not representative of the typical patients with recurrent hallux valgus. Most recurrences were caused by failure to do an osteotomy. Bunionectomy and resection arthroplasty generally are considered simple and outdated procedures that are performed in a minority of cases, which may have resulted in a less complex recurrence, as opposed to patients with more complex cases, such as those with malunion, fracture, or undercorrection after primary osteotomy. Fifth, although our patients had various primary procedures, all had similarities in the pathomechanism: an uncorrected or undercorrected IMA. Finally, adding an Akin osteotomy could limit the success of the Scarf osteotomy. We do not think this is the case because (1) the Akin osteotomy addresses the HVA whereas the Scarf osteotomy addresses an increased IMA and the osteotomy reduced the IMA; (2) the Akin osteotomy is an established method within the concept of forefoot correction by the Scarf osteotomy [1, 2, 11, 12, 17], and (3) in revision cases the width of the first metatarsal often is reduced, thus limiting lateral displacement of the metatarsal head. An Akin osteotomy in these cases can reduce the likelihood of contact between the distal ends of the great and second toes as in five of our patients.
Pain is the main reason for undergoing a salvage procedure for a recurrent deformity. A substantial reduction was observed in our patients. Only three other studies [8, 10, 14] report on reduction of pain measured by a VAS (Table 8). Our patients had the lowest postoperative level of pain.
Comparing the AOFAS score in our study with reported scores (Table 8), our results are in the upper range, taking a lower maximum score because of loss of motion after arthrodesis in some studies into account.
All radiographic measures improved. The amounts of reduction of the HVA and IMA are similar to those of TMT1 fusion  and proximal osteotomy , differing only by up to 2° (Table 8); MTP1 arthrodesis generally achieves less reduction of the HVA and IMA (Table 8). The average shortening in our patients was 1.9 mm. Shortening of 2.7 mm for TMT1 fusion is slightly greater . Primary use of TMT1 fusion shows shortening rates of 4 mm , 4.7 mm , 4.5 mm , and 5 mm .
We found similar pain reduction and radiographic measurements in the three subgroups and only minor differences for the AOFAS score. Thus, our study identified the Scarf osteotomy as a proven method to solve recurrences in selected patients after resection arthroplasty (Figs. 5-7). Previous studies [10, 14, 21, 33] recommend arthrodesis of the MTP1 for complaints after resection arthroplasty, however, they do not differentiate between various scenarios that may occur after resection arthroplasty (such as an unstable MTP1 joint, a cock-up deformity, transfer metatarsalgia, or a recurrence). If metatarsalgia can be treated successfully nonoperatively, we see no need to fuse the MTP1 joint. Even more fusion does not guarantee pain-free lesser metatarsals . Although the joint is not normal after resection arthroplasty, there is a joint-like structure, a theory supported by de Palma et al. .
We observed an increase of the HVA by an average of 2° and in the rate of sesamoid bone dislocation by an average of 4.6%, between 6 weeks postoperatively and the latest followup. We do not, however, consider these levels of changes clinically important. Further, we identified no factors that influenced the increase in these measures between 6 weeks and the last followup.
In selected patients with recurrent hallux valgus, we found that the Scarf osteotomy reduced pain, corrected the deformity, improved the functional score, and caused only one asymptomatic recurrence. We recommend the Scarf osteotomy for salvage of recurrent hallux valgus deformity in selected patients with a stable and functional MTP1 joint, pain solely on the medial side of the MTP1, and only mild signs of osteoarthritis, and in whom the underlying pathomechanism is an uncorrected or insufficiently corrected IMA. The Scarf osteotomy should not be considered the preferred procedure if the reason for recurrence is more complex, such as a malunion or fracture. Advantages of the Scarf osteotomy include simple postoperative treatment with a wooden-soled shoe and its biomechanical stability [25, 32]. No major complications as described earlier [7, 30] were seen. For patients after resection arthroplasty, we consider only those with a stable joint showing sufficient (> 40°) ROM without the presence of a cock-up deformity. In case of a cock-up deformity, an unstable first toe, or metatarsalgia not responsive to nonoperative treatment, we recommend an arthrodesis of the MTP1 joint.
We thank B. Izay for helping with the statistical evaluation, M. Pittermann for helping with evaluation of the radiographic results, and K. Chong with English translation and wording.
1. Akin, OF. The treatment of hallux valgus: a new operative procedure and its results. Med Sentinel
1925; 33: 678-679.
2. Aminian, A., Kelikian, A. and Moen, T. Scarf osteotomy for hallux valgus deformity: an intermediate followup of clinical and radiographic outcomes. Foot Ankle Int.
2006; 27: 883-886.
3. Barouk, LS. Scarf osteotomy for hallux valgus correction: local anatomy, surgical technique, and combination with other forefoot procedures. Foot Ankle Clin.
2000; 5: 525-558.
4. Bednarz, PA. and Manoli, A II. Modified lapidus procedure for the treatment of hypermobile hallux valgus. Foot Ankle Int.
2000; 21: 816-821.
5. Catanzariti, AR., Mendicino, RW., Lee, MS. and Gallina, MR. The modified Lapidus arthrodesis: a retrospective analysis. J Foot Ankle Surg.
1999; 38: 322-332. 10.1016/S1067-2516(99)80003-9
6. Chi, TD., Davitt, J., Younger, A., Holt, S. and Sangeorzan, BJ. Intra- and inter-observer reliability of the distal metatarsal articular angle in adult hallux valgus. Foot Ankle Int.
2002; 23: 722-726.
7. Coetzee, JC. Scarf osteotomy for hallux valgus repair: the dark side. Foot Ankle Int.
2003; 24: 29-33.
8. Coetzee, JC., Resig, SG., Kuskowski, M. and Saleh, KJ. The Lapidus procedure as salvage after failed surgical treatment of hallux valgus: a prospective cohort study. J Bone Joint Surg Am.
2003; 85: 60-65.
9. Coetzee, JC. and Wickum, D. The Lapidus procedure: a prospective cohort outcome study. Foot Ankle Int.
2004; 25: 526-531.
10. Coughlin, MJ. and Mann, RA. Arthrodesis of the first metatarsophalangeal joint as salvage for the failed Keller procedure. J Bone Joint Surg Am.
1987; 69: 68-75.
11. Coughlin MJ, Mann RA, Saltzman CL, eds. Surgery of the Foot and Ankle.
Ed 8, Part I, Chapter 6: Hallux valgus. Philadelphia, PA: Mosby Elsevier; 2007:183-362.
12. Palma, L., Tulli, A. and Sabetta, SP. Histological study of the phalangeal articular side following Keller procedure for hallux valgus. J Foot Surg.
1992; 31: 355-359.
13. Dereymaeker, G. Scarf osteotomy for correction of hallux valgus: surgical technique and results as compared to distal chevron osteotomy. Foot Ankle Clin.
2000; 5: 513-524.
14. Grimes, JS. and Coughlin, MJ. First metatarsophalangeal joint arthrodesis as a treatment for failed hallux valgus surgery. Foot Ankle Int.
2006; 27: 887-893.
15. Jones, S., Al Hussainy, HA., Ali, F., Betts, RP. and Flowers, MJ. Scarf osteotomy for hallux valgus: a prospective clinical and pedobarographic study. J Bone Joint Surg Br.
2004; 86: 830-836. 10.1302/0301-620X.86B6.15000
16. Kitaoka, HB., Alexander, IJ., Adelaar, RS., Nunley, JA., Myerson, MS. and Sanders, M. Clinical rating systems for the ankle-hindfoot, midfoot, hallux, and lesser toes. Foot Ankle Int.
1994; 15: 349-353.
17. Kitaoka, HB. and Patzer, GL. Salvage treatment of failed hallux valgus operations with proximal first metatarsal osteotomy and distal soft-tissue reconstruction. Foot Ankle Int.
1998; 19: 127-131.
18. Kitaoka, HB. and Patzer, GL. Arthrodesis versus resection arthroplasty for failed hallux valgus operations. Clin Orthop Relat Res.
1998; 347: 208-214. 10.1097/00003086-199802000-00025
19. Kopp, FJ., Patel, MM., Levine, DS. and Deland, JT. The modified Lapidus procedure for hallux valgus: a clinical and radiographic analysis. Foot Ankle Int.
2005; 26: 913-917.
20. Kristen, KH., Berger, C., Stelzig, S., Thalhammer, E., Posch, M. and Engel, A. The SCARF osteotomy for the correction of hallux valgus deformities. Foot Ankle Int.
2002; 23: 221-229.
21. Machacek, F Jr. Easley, ME., Gruber, F., Ritschl, P. and Trnka, HJ. Salvage of a failed Keller resection arthroplasty. J Bone Joint Surg Am.
2004; 86: 1131-1138.
22. McInnes, BD. and Bouche, RT. Critical evaluation of the modified Lapidus procedure. J Foot Ankle Surg.
2001; 40: 71-90. 10.1016/S1067-2516(01)80048-X
23. Meyer, M. Eine neue Modifikation der Hallux valgus Operation. Zbl Chir
1926; 53: 3265-3268.
24. Miller, JW. Distal first metatarsal displacement osteotomy: its place in the schema of bunion surgery. J Bone Joint Surg Am.
1974; 56: 923-931.
25. Newman, AS., Negrine, JP., Zecovic, M., Stanford, P. and Walsh, WR. A biomechanical comparison of the Z step-cut and basilar crescentic osteotomies of the first metatarsal. Foot Ankle Int.
2000; 21: 584-587.
26. Panchbhavi, VK. and Trevino, S. Comparison between manual and computer-assisted measurements of hallux valgus parameters. Foot Ankle Int.
2004; 25: 708-711.
27. Sangeorzan, BJ. and Hansen, ST Jr. Modified Lapidus procedure for hallux valgus. Foot Ankle.
1989; 9: 262-266.
28. Schneider, W., Csepan, R. and Knahr, K. Reproducibility of the radiographic metatarsophalangeal angle in hallux surgery. J Bone Joint Surg Am.
2003; 85: 494-499.
29. Shima, H., Okuda, R., Yasuda, T., Jotoku, T., Kitano, N. and Kinoshita, M. Radiographic measurements in patients with hallux valgus before and after proximal crescentic osteotomy. J Bone Joint Surg Am.
2009; 91: 1369-1376. 10.2106/JBJS.H.00483
30. Smith, AM., Alwan, T. and Davies, MS. Perioperative complications of the Scarf osteotomy. Foot Ankle Int.
2003; 24: 222-227.
31. Smith, RW., Reynolds, JC. and Stewart, MJ. Hallux valgus assessment: report of research committee of American Foot and Ankle Society. Foot Ankle.
1984; 5: 92-103.
32. Trnka, H., Parks, BG., Ivanic, G., Chu, IT., Easley, ME., Schon, LC. and Myerson, MS. Six first metatarsal shaft osteotomies: mechanical and immobilization comparisons. Clin Orthop Relat Res.
2000; 381: 256-265. 10.1097/00003086-200012000-00030
33. Vienne, P., Sukthankar, A., Favre, P., Werner, CM., Baumer, A. and Zingg, PO. Metatarsophalangeal joint arthrodesis after failed Keller-Brandes procedure. Foot Ankle Int.
2006; 27: 894-901.
© 2010 Lippincott Williams & Wilkins, Inc.
34. Wu, KK. First metatarsophalangeal fusion in the salvage of failed hallux abducto valgus operations. J Foot Ankle Surg.
1994; 33: 383-395.