Correlation of Preoperative Radiographic Severity With Disability and Symptom Severity in Hallux Valgus
Michael Matthews DPM, AACFAS, Erin Klein, DPM, FACFAS, Alyse Acciani, MPH,
Matthew Sorensen, DPM, FACFAS, Lowell Weil, Jr, DPM, MBA, FACFAS, Lowell S. Weil, Sr, DPM, FACFAS, Adam Fleischer, DPM, MPH, FACFAS
Recent government legislative efforts have had a profound impact on both the practice and delivery of medicine in America. As health care providers continue to be the primary drivers of resource utilization, insurance companies have taken increasingly aggressive steps in recent years to help “rein in” preference sensitive care (eg, elective surgery) through stricter coverage decisions. Private insurers continue to publish review criteria designed to restrict the types of procedures surgeons will be reimbursed for, sometimes basing these decisions almost exclusively on the need for select radiographic criteria alone being met. An example of this involves several companies that have even started to define minimum angular measurements that must be met for patients seeking operative intervention for hallux valgus before they will be willing to pay for surgery (eg, osteotomy, tarsometatarsal fusion). This logic inherently assumes that the magnitude of radiographic bunion deformity is closely related to the magnitude of symptoms and/or disability which the patient is experiencing, and this is particularly troubling as literature that correlates patient symptoms/disability with radiographic findings is specifically lacking in hallux valgus.
In this work, we hoped to better define how patient symptoms/disability (defined using a validated instrument for hallux valgus assessment—the FAOS) correlated with the magnitude of radiographic deformity in patients electing to undergo bunion surgery within a large, urban-based foot/ankle specialty practice that is located within the midwestern United States.
We conducted a retrospective analysis of patients who presented for elective hallux valgus operative correction in our practice during the last 2 years. All data were already in existence at the time of the study. Exempt determination and HIPAA waiver was obtained from our local institutional review board. Eligible subjects had to have Foot and Ankle Outcome Score (FAOS) completed preoperatively with respect to their operative foot. In our practice, FAOS data were typically obtained on many of our operative patients to help with clinical decision making and to monitor postoperative progress. The FAOS is a validated patient-based tool for evaluating postoperative hallux valgus operative outcomes that has demonstrated internal consistency, convergent validity, and structural validity.5 The survey consists of 42 items and 5 subscales: pain, symptoms, function/activities of daily living, function/sport and recreation, and foot and ankle–related quality of life. Subscale scores are reported on a scale that ranges from 0 to 100, with higher values indicating better scores (eg, less pain, fewer symptoms, higher functioning).
Included subjects also had to have preoperative weightbearing foot radiographs, and enough pain and disability with their bunion that they proceeded with hallux valgus surgery between January 1, 2016, and July 1, 2018. Similarly, surgeons had to have felt their patient’s symptoms and/or overall foot morphology required osseous realignment (via osteotomy or arthrodesis), as opposed to simple ostectomy of the first metatarsal head. In order to avoid the confounding effect of other foot/ankle deformities unrelated to the bunion, we restricted our sample by excluding patients who also complained of hammertoes, forefoot metatarsalgia requiring lesser metatarsal osteotomies, and/or required other surgery outside of hallux valgus correction. In patients who sought treatment for bilateral hallux valgus during the study, we selected the foot with the lowest FAOS total score for inclusion as the study foot, making the subject the unit of observation (rather than the foot) as has previously been recommended.
There were complete data from 107 patients (107 feet) with a mean age of 49.3 ± 13.8 years and mean body mass index of 25.2 ± 4.8 who sought treatment and subsequently underwent surgery for isolated hallux valgus during the study period. One hundred two participants were women (102/107, 95%), and 106 identified as white (106/107, 99%). Patients included in the study underwent scarf/Akin osteotomies (n=85), scarf osteotomies only (n=19), Lapidus procedures (n=2), or chevron (n=1) osteotomies in their study foot.
All radiographs were evaluated by a single rater (M.M.) using commercially available software. The following parameters were evaluated on plain films: the first and second intermetatarsal (IM) angle, the hallux valgus (HV) angle, metatarsus adductus angle via the Engels angle, metatarsal protrusion distance using Nilsonne’s method, and tibial sesamoid position using the method of Hardy and Clapham (Figure 1). Engle’s method for measuring the metatarsus angle describes the angle between the longitudinal bisection of the second metatarsal and the longitudinal bisection of the second cuneiform (as opposed to using the longitudinal axis of the lesser tarsus). Nilsonne’s method of examining metatarsal protrusion distance involves measuring the distance in millimeters between a perpendicular line to the longitudinal axis of the second metatarsal at the most distal point of the second metatarsal head to a parallel line that passes through the distalmost point of the first metatarsal head (Figure 1).
(A) 1-2 intermetatarsal (IM) angle; (B) Engel; (C) hallux valgus (HV) angle; (D) Nilsonne; (E) sesamoid position.
Spearman correlation coefficient (rho) was calculated for the 5 measured radiographic parameters and the 5 FAOS subscale scores. This was performed for the entire cohort, and the analysis was repeated after stratifying by age. The strength and direction of the correlation is given by the correlation coefficient. A coefficient greater than or equal to 0.4 was considered a moderate correlation, and coefficients less than 0.4 were considered weak to nonexistent. Correlations with P values less than .05 were considered to be statistically significant.
Mean ± SD FAOS subscale scores prior to surgery for the population were as follows: Symptoms = 79.2 ± 16.9, Pain = 73.4 ± 18.6, ADL = 82.0 ± 18.7, Sports/Rec = 70.0 ± 24.9, and QoL = 53.1. The mean ± SD radiographic measurements for the cohort were as follows: HA = 26.7 ± 7.8 degrees, IM = 12.5 ± 3.5 degrees, sesamoid position = 4.3 ± 0.9, MAA = 25.9 ± 5.7 degrees, and MPD = −3.9 ± 2.7 mm.
The correlation matrix for the study population is given in Table 1. The correlation matrices stratified by age (<56, and ≥56 years) are provided in Tables 2 and 3. No variable achieved a moderate or higher correlation with any of the FAOS subscales, with the exception being sesamoid position with FAOS Pain (rho=0.402, P = .01) in patients 56 years and older. The direction of this correlation was positive, indicating that greater sesamoid abnormalities were associated with less pain (ie, higher FAOS Pain scores). Similarly, larger fist-second IM angles were weakly associated with having less pain (rho=0.357, P = .02; Figure 2) and fewer symptoms (rho=0.316, P = .04) in patients 56 and older.