When we separate studies that used dual measurements from those using aIMA, a more complete picture of the incidence of recurrence can be drawn. We reviewed all peer-reviewed published studies looking at bunion deformity correction at the time of writing this review and found deformity recurrence rates of between 4% and 73%. Those that reported measurements based on the aIMA both pre- and postoperatively were considered first and considered to provide the most accurate data on recurrence. Those who measured pre and post center of head and base (mIMA) or used dual measurements (first the aIMA and then the mIMA) were considered biased and are presented in the next section. A number of articles did not state the method used to measure the HVA or IMA, and they were omitted.

Shibuya et al. [68] compared feet that had first metatarsal osteotomies with (n = 73) and without (n = 81) an additional Akin procedure. Hallux abductus angles (HAA) were analyzed throughout this study. The HAA of the group that had the Akin procedure was significantly greater 6 months after surgery than the group that did not have Akins. The tibial sesamoid position was also significantly more laterally deviated in the Akin group when compared to their non-Akin peers. No difference in the revision rate was noted with 17.8% of the Akin group and 11.1% of the non-Akin group needing repeat surgery. Based on their results, the authors questioned the value of adding the additional hallux procedure. Bock et al. [8] reviewed 115 feet at 124 months that had undergone the scarf osteotomy procedure. ROM, VAS, HVA, IMA, DMAA, AOFAS, and sesamoid position were all significantly improved postoperative compared to preoperative. However, there was recurrence of 30% defined as an HVA of greater than 20͒ at final follow-up. The authors found correlations with recurrence to be higher HVA (preoperative and 6 weeks), higher IMA (6 weeks), sesamoid bone position, and DMAA. Iyer et al. [38] studied the proximal medial opening wedge (PMOW) osteotomy in 17 patients over an average 2.4-year follow-up. IMA improved at 6 weeks but deteriorated at final midterm follow-up. HVA also was noted to be improved at 6 weeks; however at the final midterm follow-up, the HVA was not significantly different from the preoperative value indicating a high degree of recurrence (64.7%). Interestingly the DMAA increased from 10.2 degrees to 13.6 degrees after surgery. Those patients who did have recurrence had higher HVA and DMAA preoperative scores compared to their colleagues who did not have recurrence. 23.5% of patients went on to have additional revision surgeries, and 35% had continued pain at the MTPJ at final follow-up. Fakoor et al. [28] compared the chevron, scarf, and McBride procedures in 44 feet. HVA was evaluated pre- and post-operation for each surgical group. The postoperative HVA of correction for both the chevron (16.7 degrees) and scarf (18 degrees) procedures was significantly different from the McBride (11 degrees) procedure but not from each other. The IMA of correction for the chevron (4.5 degrees) and scarf (6.3 degrees) was also significantly different from the McBride (2.6 degrees) procedure but not from each other. Osteotomy procedures had significantly better radiological outcomes than the McBride procedure. Recurrence was defined as any deformity reformation and occurred in 0% of scarf, 13% of chevron, and 27% of McBride patients. Pentikainen et al. [58] analyzed radiographic results of 100 feet at 6 weeks, 6 months, 1 year, and an average of 7.9 years (range 5.8–9.4 years) after distal chevron surgery to determine factors associated with hallux valgus recurrence. Recurrence in this study was determined to be an HVA of greater than 15 degrees and was seen in 73%. The mean HVA of patients who had recurrence was 28 degrees, and the IMA was significantly greater than in those who did not have recurrence. Every patient who had an HVA of greater than 30 degrees during the preoperative X-rays had recurrence. Along with the HVA, the position of the sesamoids, DMAA, congruence, and IMA all significantly affected recurrence rates. Choi et al. [12, 13] reviewed 24-month follow-up of 53 feet that had scarf osteotomies with soft tissue realignment. SF-36 scores had a small non-statistically significant improvement (46 pre to 52 post). It is interesting that the radiographic improvement could be reported based on the measurement technique, but the SF-36 scores pre- and post-operation did not show a statistically significant change. The complication rate was 15%, with additional operations being deemed necessary in 7.5% of feet for removal of hardware. There were no reported cases of recurrent HAV; however there was a statistically significant loss of correction of the IMA (2.2 degrees) and MSP (0.4 grades). Hallux varus occurred in 3.9%, and an additional 3.9% were noted to have first metatarsophalangeal joint arthritis. Choi et al. [12] reviewed 103 Ludloff osteotomies that were combined with other procedures. The patients were divided into three groups depending on the type of distal soft tissue procedure they underwent. Thirty percent had first web space releases, 34% had Akin osteotomies and trans-articular releases, and 36% had Akin osteotomies with supplementary axial K-wire fixation and trans-articular releases. AOFAS and VAS improved in all three groups, without a significant difference between groups. Recurrence, which was defined as an HVA greater than 20 degrees, occurred in 15.5%, with all three groups having similar amounts of recurrence. Sixty-eight percent of patients who had recurrence did not report any symptoms. Deveci et al. [23] reviewed 50 scarf procedures, at a follow-up of 26.2 months (range 18–36 months). Ten percent of patients reviewed were found to recurrence of the deformity which defined as an HVA of greater than 15 degrees. Incongruity of the joint, which was the authors’ hypothesized cause of recurrence, was found to be a statistically significant risk factor for recurrence. George et al. [31] examined outcomes 37.6 months following scarf osteotomies performed in 19 adolescent feet (average age of 14.3 years). IMA, HVA, and DMAA improved significantly at 6-week postoperative evaluation; however only IMA was maintained throughout the 3-year follow-up, and deterioration of the other measures was noted. 36.8% had pain and recurrence after surgery, while 9% had superficial infections. These results led the authors to the conclusion that the scarf procedure should be used with caution in adolescents. Veri et al. [74] analyzed 37 feet that had crescentic osteotomy and distal soft tissue reconstruction . A short-term follow-up was conducted at 1 year, along with a long-term follow-up at 12.2 years (31 feet). HVA and IMA values both deteriorated during the follow-up period. Ninety percent of patients were satisfied with their physical abilities during long-term follow-up, while only 80% were satisfied with the appearance. At the short-term follow-up, superficial infection occurred in 16%, 8% had delayed unions, 5% had varus after surgery, and 11% had recurrence.

To this point we have defined deviations from the normally accepted angles based on a comparison of preoperative and postoperative angles as recurrence. Looking at this issue from another perspective, we can ask, are we in reality even correcting the original deformity? Edmonds et al. [26] looked at postoperative radiographic measurements following distal, proximal, and double osteotomies in 106 juvenile feet (mean age 14.7 years). Their primary aim was to report which of the procedures returned the radiographic measurements of IMA, HVA, and DMAA to within a normal range. For the single distal first metatarsal osteotomy, the IMA was corrected to within normal limits in only 21% of the cases, HVA was within normal limits 42% of the time (however 13% of the time there was overcorrection), and the DMAA was within normal limits 46% of the time (with 4% overcorrected). The single proximal osteotomy had 36% of IMA, HVA, and DMAA within normal limits, with only DMAA having overcorrection in 7% of the cases. Finally the double osteotomy had 54% within normal limits for IMA, HVA within normal limits 40% (7% overcorrected), and DMAA was within normal limits for 56% of the cases (22% overcorrected). From these numbers it was found that the rate of HVA overcorrection was not correlated with the type of osteotomy performed, but there was a significantly higher rate of overcorrection in the double osteotomy when compared to both types of single osteotomies. This study highlights the shortcomings of these popular procedures in returning the radiographic bone segment positions to the normal range. We may in fact not be dealing with recurrence but simply our failure to correct the original deformity. If we consider the fact that the CORA for a bunion is not within this first metatarsal but at a point proximal to the deviated metatarsal, we are in fact creating a new deformity with metatarsal osteotomy. We believe that failure to correct the original deformity is likely the prime reason for poor outcomes. This is discussed in detail in Chap. 6 (Table 7.1).

As discussed above, since in the normal state the anatomic axis and the mechanical axis of the first metatarsal are collinear, using center of joint landmarks both before and after the procedure represents the exact same bias of measurement as the dual measurement technique. Alteration of reference point for the first metatarsal incorrectly reports a normal anatomic position (IMA) of the first metatarsal when in fact a new deformity has been created and the original aIMA deviation persists. The following studies used the system of dual measurements or the center of head and base technique to report on correction. Although we cannot fully critique the data, knowing the effect this convention has on the measured results, we can assume that degree of correction is overestimated and incidence of deformity recurrence is underestimated.

Jeuken et al. [40] compared 36 scarf to 37 chevron osteotomies 14 years after surgery. Patient-reported satisfaction and patient satisfaction with pain reduction ranged between 59% and 73% for each of the groups in all three of these categories based on MOXFQ, SF-36, and VAS scores. Seventy-three percent of feet in the chevron group and 78% of feet in the scarf group had recurrence based on their definition of HVA greater than 15 degrees. The high recurrence rate was determined using the center of head to center of base technique raising the question about the possibility of underestimation of true correction. As could be expected, the satisfaction was not particularly high in this study population despite the authors reporting significant improvement in AOFAS scores after surgery illustrating the potential bias between physician-rated scales and PROMs. Aiyer et al. [2] studied the recurrence of hallux valgus after 587 foot surgeries, comparing patients who have underlying metatarsus adductus (29.5%) (MA) to those who do not. Recurrence was defined as HVA of more than 20 degrees. HVA, IMA, and metatarsus adductus angle (MAA) were all measured, with MAA being considered abnormal if greater than 20 degrees. Patients with MA had greater HVA and IMA pre- and postoperative angles as compared to individuals without MA. There was a 15% recurrence rate in patients without MA, compared to a 29.6% recurrence rate in patients with MA. The rate of recurrence in the patients with MA did not vary based on procedure (Lapidus 28.5%, distal first metatarsal osteotomy 29.4%, proximal first metatarsal osteotomy 28.9%). Interestingly patients who had less severe MA (<31 degrees) were shown to have a higher rate of recurrence than those with more severe MA (82% vs 18%). In a previous study [3], reported metatarsus adductus to be associated with HAV in 30% of the cases reviewed. We have noted that the presence of metatarsus adductus clearly changes the ability to adequately and consistently correct the deformity long term and that this finding needs to be considered in the treatment algorithm. Metatarsus adductus assessment and clinical implications are discussed further in Chap. 5. Groningen et al. [34] analyzed the outcomes of 438 feet that had chevron osteotomies. The average IMA improved from 12.4 to 6.2 degrees after surgery, while the HVA improved from 28.5 to 14.8 degrees. FAOS assessments were completed at an average of 3 years for 250 of these patients, 28.3% had complications with undercorrection of the deformity occurring in 11.6% and hardware complications occurring in 9.1%. Those who had undercorrection had significantly lower FAOS assessment scores than their counterparts who had the original deformity corrected. With the use of dual measurements, we must question the true recurrence rate due to the bias imparted by the measurement technique. Agrawal et al. [1] looked at the clinical, functional, and radiological outcomes of the scarf-Akin procedure on 47 adolescent and juvenile feet. Radiological recurrence defined as IMA greater than 9 degrees and an HVA of greater than 15 degrees occurred in 29.8% with 21% needing additional or repeat surgeries. It should be noted that these high recurrence rates were probably underestimated based on center of head radiographic measurement technique and because of a very short radiographic follow-up (6-week post-radiographs used to report results). Despite the high radiographic recurrence, the AOFAS scores at short-term follow-up between the recurrent and nonrecurrent were not significantly different; therefore one has to question the validity of the outcomes scale based on these conflicting findings. Lee et al. [48] compared outcomes of proximal and distal chevron osteotomies in 92 feet in 46 female patients that were undergoing moderate to severe hallux valgus bilaterally, one foot proximal and one foot distal for comparison. The average follow-up was 40.2 months (range 24.1–80.5) at which point 6.5% of the distal group and 4.3% of the proximal group were dissatisfied. Recurrence occurred in 6.5% feet in the proximal group and 2% of feet in the distal group. Again due to measurement technique, the conclusions are biased so we cannot draw accurate conclusions based on comparison to normal. If we analyze the radiographs provided in the study, the aIMA and the TSP are quite abnormal despite the report of these measures being corrected by use of the center of head technique similar to the figures presented in this chapter (Fig. 7.5). Buciuto [9] presented a comparison of the outcomes for the Mitchell osteotomy and the chevron osteotomy. They reported satisfactory correction for both procedures (chevron had better results) based on the dual measurement technique with a loss of hallux valgus correction of 4–6 degrees. Of note they reported a 36% rate of transverse metatarsalgia pain . Recurrence rate was not specifically discussed in their analysis. Evaluation of the pre- and postoperative images provided in the paper clearly shows overestimation of the correction of all measured angles. There are a plethora of additional papers reporting individual author’s results with a wide diversity of procedures. Unfortunately the use of dual measurements and the diversity of outcomes scales make systematic analysis impossible and leave us with a lack of solid answers as to the best and most reliable methods to correct HAV. Faber et al. [27] compared the Lapidus procedure with the Hohmann distal closing wedge metatarsal osteotomy in 91 feet specifically to determine if arthrodesis procedures are necessary to control hypermobility. AOFAS scores which were taken preoperatively, 2 years postoperatively, and 10 years post-operation were compared. The AOFAS significantly increased between preoperative and both 2 and 10 years post surgery. However there was a significant decrease in AOFAS scores between 2 years and 10 years. The IMA was significantly worsened in both groups between the 2- and 10-year follow-ups. Both of these factors show the importance of long-term follow-up. The average recurrence rate in both of the groups was 8.8% with underestimation of recurrence likely secondary to measurement technique. Farrar et al. [29] assessed scarf osteotomies of 39 feet in 28 adolescent patients (mean age 14.1 years). Of the 18% that had recurrence, they defined 71% as minimal recurrence that did not desire repeat surgery and 14% as significant symptoms that required revision surgery and 14% whom did not choose to have additional surgery. Okuda et al. [55] reviewed 77 feet treated with a proximal metatarsal osteotomy omitting five cases from the result due to hallux varus at a 14–120 months final postoperative visit. Hallux valgus recurrence defined as an HVA greater than or equal to 20 degrees was found in 13.9%. Seven percent had recurrence occurring at 10 weeks after surgery. Patients who had a preoperative HVA greater than 40 degrees had an increased risk for recurrence. The authors found that patients who had an HVA of less than or equal to 15 degrees and an IMA of less than 10 degrees at 10 weeks after surgery had a decreased risk of recurrence. Kilmartin and O’Kane [45] reviewed scarf and Akin osteotomies in 73 feet at an average of 9 years post surgery. Patients reported stiffness in the first MTP joint in 8%; hallux varus occurred in 4% and recurrence of 4%. The authors considered an HVA of 20 degrees as “mild.” An additional 8% of feet had HVA greater than 15 degrees which has been defined by some authors to be abnormal. Total satisfaction rate was 88% based on their definitions. This again highlights the difficulty in comparing studies due to methodological inconsistencies. Deenik et al. [22] studied HVA in scarf and chevron osteotomies in 136 feet. Subluxation of the MTPJ occurred postoperatively in 35% of patients with preoperative HVA greater than 37 degrees which progressively lead to recurrence. Only 3% of these cases of mild HVA preoperatively lead to recurrence. Patients whose HVA was more than 37 degrees preoperatively were only 65% satisfied after surgery and had significantly more pain than those who had smaller HVAs preoperatively. Coetzee [17] investigated scarf osteotomies in 20 patients at 6 and 12 months post-operation. Preoperative AOFAS was 53, 6 months 54, and 12 months after surgery 62. Fifty-five percent were satisfied at 6 months, while only 53% were satisfied after 1 year. IMA reduced from 16 degrees preoperatively to 13 degrees at 12 months; HVA improved from a mean of 40–34 degrees. Overall satisfaction was 55%, with 45% being dissatisfied in the early postoperative and 47% still unsatisfied at 12 months. Recurrence rate was 25% defined as HVA angle greater than 32 degrees and an IMA of greater than 10 degrees. Seven patients required revision surgery. Fokter et al. [30] studied 105 feet in patients who returned for clinical exam with a mean of 21 years (range 15–24 years) after a modified Mitchell procedure. They found that pain was present either at the first MTPJ or under the lesser metatarsals in 41% of the patients. Clinical return of the hallux valgus deformity was present in 47% of their patients. First and second toe overlap was noted in 18% and rotational deformity of the hallux in 39%. They concluded that the results of the procedure could not withstand the test of time for more than a decade (Table 7.2).