Forefoot Adduction is a Risk Factor for Jones Fracture

Abstract

Jones fractures are among the most common fractures of the foot, yet much remains unknown about their etiology. The purpose of this study was to further examine the risk factors of forefoot and hindfoot alignment on Jones fractures using an epidemiological study design. We used a retrospective, matched, case-control study design. Cases consisted of patients with acute, isolated Jones fractures confirmed on plain film radiographs that were seen at our institute from Jan. 2009 to Dec. 2013. Patients presenting with pain unrelated to metatarsal fractures served as controls. Controls were matched to cases on age (+/− 2 yrs), gender, and year of presentation. Weight bearing foot x-rays were assessed for thirteen angular relationships by a single rater. Conditional multivariable logistic regression was used to identify important risk factors. Fifty patients with acute Jones fractures and 200 controls were included. The only significant variables in the final multivariable model were metatarsus adductus angle (odds ratio 1.16 [95% CI 1.08 to 1.25]) and 4^th/5th intermetatarsal angle (odds ratio 0.69 [95% CI 0.57 to 0.83])—both measures of static forefoot adduction. The presence of metatarsus adductus (defined as >15°) on foot radiographs was associated with a 2.4 times greater risk of Jones fracture (adjusted odds ratio 2.4 [95% CI 1.2 to 4.8]). We conclude that risk of Jones fracture increases with an adducted forefoot posture. In our population which consisted primarily of patients presenting after a fall (10/50, 20%) or misstep/inversion injury (19/50, 38%), hindfoot alignment appeared to be a less important factor.

Level of Clinical Evidence

3, Prognostic Study

Introduction

Fractures of the 5^th metatarsal occur with regularity and Jones fractures, in particular, have a notoriously poor prognosis for healing (1). When treated non-operatively, nonunion rates for Jones fractures range anywhere from 11–50% (2). Although these injuries are common, much is still unknown about the etiology of Jones fractures.

Jones fractures have been reported to result after acute trauma, inversion injuries, falls, and following activities that apply repetitive low grade stress to the lateral aspect of the foot (3,4). Several authors also report that static foot posture may play an important role as well. DeLee et al. (1983) and Raikin and colleagues (2008) both found a relationship between proximal fifth metatarsal fractures and hindfoot varus deformity, implicating excess lateral column loading as a possible etiology (3,5). Similarly, certain sports (e.g., basketball and football) which involve large amounts of pivoting may also be associated with an increased risk for developing Jones fractures (6,7). This is thought to be due to the excess stress placed on the lateral side of the forefoot while the ankle is somewhat plantar flexed during pivoting maneuvers (5,8).

Metatarsus adductus, defined as “a uniplanar transverse plane deformity where the metatarsals are adducted at the Lisfranc joint” (9), may also predispose individuals to Jones fracture. Metatarsus adductus is perhaps the most common congenital foot deformity and may increase plantar pressures encountered on the lateral aspect on the foot while walking (10). Yoho et al. (2012) recently published their results of a case control (1:1) study with 60 patients examining the association between metatarsus adductus and Jones fractures (11). They found that the group with Jones fractures had a higher mean metatarsus adductus angle than those in the control group (11). While assistive in our understanding of this association, this study was limited in that it only looked only at univariate associations, did not control for the effects of confounding variables, and did not report risk estimates (e.g., odds ratios, relative risks). The purpose of this study was to expand on previous work by generating risk estimates while controlling for the effects of confounding variables (e.g., age, hindfoot alignment) through matching and multivariable analysis.

Patients and Methods

This study represents a retrospective, matched, case-control (50 cases/200 controls) study design. Consecutive patients presenting to our institution from Jan. 2009 to Dec. 2013 with acute Jones fracture of the proximal fifth metatarsal served as the cases. Control subjects were selected from the same clinic but had foot pain that was not attributable to a metatarsal fracture. Exempt status determination and waiver of consent was granted from XXXXXXXXX Institutional Review Board prior to the start of the research.

Definition of Cases

Although there are many reported definitions for Jones fractures, the most commonly used definition is a transverse fracture occurring in the proximal fifth metatarsal, typically occurring within 1.5 cm from the tuberosity (12). For the purposes of this study, “cases” then were defined as patients who presented for treatment of a transverse fracture occurring at the proximal metaphyseal-diaphyseal junction within 1.5 cm of the tuberosity and without distal extension beyond the 4^th/5^th intermetatarsal articulation (13). Those patients that would more appropriately be classified as proximal diaphyseal stress fractures were specifically excluded. All Jones fractures were radiographically confirmed by the lead author (AF).

Definition of Controls

Patients attending the same foot and ankle specialty clinic with foot pain that was not attributable to a metatarsal fracture served as controls. Control subjects were selected from a roster of patients who received treatment for plantar fasciitis, Morton’s neuroma and/or tendonitis. Controls were matched to cases at a ratio of 4:1 on age (+/− 2 years), gender, and calendar year of presentation. Control subjects had to have weight bearing foot x-rays available for review. Control subjects were selected without any knowledge of their x-rays measurements and without knowledge of any of the study’s other covariates.

Independent Variables

Weight bearing foot x-rays were assessed for thirteen predetermined angular relationships (e.g., metatarsus adductus angle) on anterior-posterior and lateral radiographs. The method of evaluation is described below. Charts were also reviewed for body mass index (BMI), medications (e.g., benzodiazepines), and medical history including smoking status (current or non-current) and osteoporosis status (14,15).

Radiographic Measurements

All radiographic measurements were performed by a single rater (RC) using commercially available computer software (Echos, Medstrat, Downers Grove, IL). All radiographs included in this study were taken when bone healing permitted full weight bearing. Anterior-posterior radiographs were analyzed for metatarsus adductus angle (MAA), 1st and 2nd intermetatarsal angle (1/2 IMA), 4th and 5th intermetatarsal angle (4/5 IMA), Meschan’s metatarsal break angle, talo-calcaneal angle (TCA), calcaneo-cuboid abduction angle, talo-1st metatarsal angle, and hallux abductus angle. Metatarsal break angle describes the angle created by connecting the distal most aspects of the 1^st, 2^nd and 5^th metatarsals. The angle generally increases with relatively longer 1^st and/or 5^th metatarsals and decreases with short 1^st and/or 5^th metatarsals. Values less than 135 degrees are strongly indicative of a short 1^st and/or 5^th metatarsal and abnormal forefoot metatarsal parabola. (16,17). Lateral radiographs were examined for calcaneal inclination angle (CIA), talar declination angle, 1st metatarsal declination angle, 5th metatarsal declination angle, and talo-1st metatarsal (Meary’s) angle. All measurements were performed as described by Thomas et al. (2006) (17), except MAA which was assessed using the method of Dominguez and Muneura (2008) (18) as this method has been shown to be both reliable and reproducible. For the purposes of analysis, patients with angles greater than 15 degrees were classified as having metatarsus adductus (19).

X-ray Technique

A standard protocol was used to obtain foot radiographs. Anterior posterior weight bearing radiographs were obtained while patients stood erect in double support with both knees in full extension. The central beam was angled 15 degrees from vertical in the sagittal plane at a distance of 100 cm from the foot. The direction of the x-ray beam was oblique to the cassette positioned on the floor and centered over the second metatarsocuneiform joint. Similarly, lateral weight bearing radiographs were obtained in double support with both knees in full extension. The central beam was angled 90 degrees from vertical in the sagittal plane at a distance of 100 cm from the foot. The direction of the x-ray beam was perpendicular to the cassette positioned upright in the platform and centered over the fifth metatarsal base.

Statistical Analysis

The increased risk of Jones fracture for each of the study’s independent variables was estimated by calculating odds ratios with the use of conditional logistic regression. In this context, the odds ratio described the odds that a patient with a Jones fracture had been exposed to the risk factor (e.g., increased metatarsus adductus angle) divided by the odds that a control subject had been exposed to the risk factor, after adjusting for all other variables in the model. The more that the odds ratio deviated from 1, the stronger the association between the exposure variable and the condition being studied. MAA and CIA were also examined using Mantel-Hantzel test for trend to determine whether greater amounts of the exposure variable increased the risk of Jones fracture (dose-response) (20,21). A backward elimination method was used to select the final prediction model. All two-way interactions were initially included in the model. All angular measurements and BMI were initially analyzed as continuous variables. To simplify presentation of the results, these data were divided roughly into tertiles using historically important cut points; age categories were displayed by decade. Statistical analyses were conducted using SAS software version 9.4 (SAS Institute, Cary, NC, and Microsoft Corporation, Redmond, WA). All tests of significance were two-tailed with P values < 0.05 considered significant.

Results

A total of 250 subjects were included in the analysis–50 patients (35 females, 15 males) with a Jones fracture and 200 subjects (140 females, 60 males) in the control group (Table 1). There were no differences for age, gender or year of presentation among the groups which indicated good matching on the desired covariates. The average age of Jones fracture patients and control subjects was 50.4 ± 17.5yrs (range 18 – 87) and 50.5 ± 17.3yrs (range 18 – 89), respectively. The mean MAA differed significantly among cases (18.8 ± 8.2°) and controls (14.7 ± 5.1°, p<0.0001). The mechanism of injury for those presenting with Jones fractures was: inversion injury/misstep (19/50, 38%), trip/fall (10/50, 20%), walking/overuse (5/50, 10%), sport (5/50, 10%), trauma (2/50, 4%) and unclear/not reported (9/50, 18%).

Table 1.

Characteristics of Patients and Control Subjects in Relation to Risk of Jones Fracture.

	No. of Patients (N=50)	No. of Controls (N=200)	Odds Ratio (95% Confidence Interval)
Gender
Male	15 (30%)	60 (30%)
Female	35 (70%)	140 (70%)
Age (in years)
≤20	3 (6%)	12 (6%)
21–30	6 (12%)	23 (11.5%)
31–40	3 (6%)	13 (6.5%)
41–50	11 (22%)	43 (21.5%)
51–60	13 (26%)	52 (26%)
61–70	9 (18%)	37 (18.5%)
>70	5 (10%)	20 (10%)
Body mass index (in kg/m2)*
≤25	17 (34%)	74 (37%)	1.0 (referent)
25 to 30	14 (28%)	72 (36%)	0.86 (0.35–2.0)
>30	19 (38%)	54 (27%)	1.2 (0.50–3.0)
Current smoker
Yes	1 (2%)	10 (5%)	0.49 (0.05–4.2)
No	49 (98%)	190 (95%)	1.0 (referent)
Metatarsus adductus angle* (in degrees)
≤15	17 (34%)	102 (51%)	1.0 (referent)
>15 to 19	13 (26%)	59 (30%)	1.5 (0.67–3.5)
>19	20 (40%)	38 (19%)	4.6 (1.88–11.4)†
Intermetatarsal angle 4th/5th* (in degrees)
>10	7 (14%)	37 (19%)	1.0 (referent)
≥7 to 10	16 (32%)	95 (47%)	1.3 (0.44–3.7)
<7	27 (54%)	68 (34%)	4.5 (1.45–13.9)¥
Calcaneal inclination angle* (in degrees)
<18	16 (32%)	63 (32%)	1.0 (referent)
≥18 to 22	11 (22%)	56 (28%)	0.73 (0.27–1.9)
>22	23 (46%)	81 (40%)	1.4 (0.60–3.4)
Meary’s angle* (in degrees)
<5	16 (32%)	65 (32%)	1.0 (referent)
≥5 to 12	18 (36%)	72 (36%)	1.0 (0.71–1.4)
>12	16 (32%)	63 (32%)	1.0 (0.70–1.5)

Odds ratios and 95% CIs are given with respect to the referent group and adjusted for other variables in the model.

^*Body mass index and radiographic measurements were analyzed as continuous variables. The data are presented in tertiles in the table.

^†p<0.001.

^¥p<0.01.

The significant variables in the final multivariable model (Table 2) were MAA (odds ratio 1.16 [95% CI 1.08 to 1.25] for each 1 degree increase) and 4/5 IMA (odds ratio 0.69 [95% CI 0.57 to 0.83] for each 1 degree increase). The Mantel-Haenszel test for trend was significant for both MAA (χ² 9.078, p=0.003) and 4/5 IMA (χ² 4.649, p=0.031) indicating that a dose-response relationship exists for both variables. Specifically, higher MAA and lower 4/5 IMA were associated with increasingly greater odds of presenting with a Jones fracture (Figure 1).

Table 2.

Final Multivariable Model (N=250).

	Parameter Estimate	Odds Ratio	95% Confidence Interval
Body mass index	0.0137	1.01	0.94 – 1.08
Current smoker (vs. non-smoker)	−0.5069	0.60	0.06 – 5.82
Metatarsus adductus angle	0.1519	1.16	1.08 – 1.25†
Intermetatarsal angle 1/2	−0.0752	0.93	0.80 – 1.07
Intermetatarsal angle 4/5	−0.3706	0.69	0.57 – 0.83†
Metatarsal break angle	−0.0244	0.98	0.94 – 1.01
Calcaneal inclination angle	0.0682	1.07	0.99 – 1.15
1st metatarsal declination angle	−0.0649	0.94	0.83 – 1.06

Odds ratios are displayed for each 1 unit (e.g., degree) increase in exposure variable.

^†p<0.0001.

Figure 1.

Larger metatarsus adductus angle (MAA) seen in (A), and smaller 4th and 5th intermetatarsal angle (4/5 IMA) seen in (B) on anterior-posterior radiograph were identified as independent risk factors for Jones fracture.

CIA showed a trend (p=0.076) in the final model, but ultimately was not significant when other covariates were also considered. There were no significant pairwise interaction terms in the final model. Smoking status and BMI were not associated with Jones fracture but were retained in the final model as potentially important confounders. Osteoporosis, number of prescription medications, and benzodiazepam use were also not associated with Jones fracture. When MAA was treated as a dichotomous variable in the final model (data not shown), the presence of metatarsus adductus (versus no metatarsus adductus) was associated with a 2.4 times greater risk of presenting with a Jones fracture (adjusted odds ratio 2.4 [95% CI 1.2 to 4.8]).

Discussion

Forefoot posture, and in particular metatarsus adductus, has been implicated previously in the development of Jones fractures of the 5th metatarsal base (11); however, our study provides some of the strongest evidence to date for a causal relationship. The current study is an improvement on prior studies (5,11) as we employed a multivariate analytical approach and controlled for select covariates through matching. We also describe, for the first time, that a narrower 4th and 5th intermetarsal angle (another distinct measurement of forefoot adduction posture) is independently associated Jones fractures, even after controlling for the presence/absence of metatarsus adductus. Having an adducted forefoot posture and having a heightened risk of developing a Jones fracture also appears to be biologically reasonable/plausable. Individuals with an adducted forefoot posture may be more likely to experience increased lateral column pressures at the more prominent metaphyseal-diaphyseal junction of the fifth metatarsal. Fishco and colleagues (10), for example, recently found that peak pressures and pressure-time integral were both significantly greater in the lateral mid- and forefoot during barefoot walking in people with metatarsus adductus compared to those without metatarsus adductus. The relatively strong association we observed between an adducted forefoot posture and Jones fracture injuries (as evidenced by the fairly large odds ratios), the presence of a dose-response relationship (i.e., biologic gradient), and the inherent biologic plausibility, collectively, make a strong argument for causality.

Identifying causative factors for developing a Jones fracture is important as these injuries are common and can be particularly challenging to treat. A compromised blood supply after injury (22) and an already tenuous blood supply to the metaphyseal-diaphyseal region of the bone many times leads to prolong healing times and a relatively high prevalence of nonunions and refractures (23,24). Therefore, recognizing risk factors for this injury may have important implications for prevention and rehabilitation after injury.

Raikin et al. (5) previously observed a high incidence of hindfoot varus in patients presenting with Jones fractures. This study concluded a varus heel posture increased forces placed on the fifth metatarsal, thus predisposing these patients to sustaining a Jones fracture. In an effort to reduce refracture rates after intramedullary screw fixation, the authors used a polypropylene orthotic with a corrective lateral hindfoot wedge extended out to a lateral forefoot post once healing was complete (5). Other studies have similarly indicated there may be a benefit of using corrective orthotics to reduce lateral column loads in preventing and reducing the incidence of Jones refractures (25,26).

Like past work (5), we also assessed hindfoot varus using the calcaneal inclination angle and talar-1^st metatarsal angle (Meary’s angle). However, we did not find hindfoot position to be as important a risk factor as forefoot posture. In fact, we found that the effect of hindfoot varus was significantly lessened when also considering the effects of the forefoot. One of the possible explanations as to why others found an association, while we did not, is the mechanism of Jones fracture injury reported by many of our patients. While most of the patients in our work presented after an inversion injury, misstep, or a fall, many of those in Raikin et al.’s study were playing sports at the time of injury (which involves cutting and repetitive loading) where the heel position may have a greater impact on lateral load bearing. In contrast, the heel does not typically contact the ground first in missteps and falls and therefore would have less influence on mid- and forefoot loading. Furthermore, nearly fifty percent (10/21) of the participants in Raikin et al.’s work reported prodromal symptoms, so it is also possible that some of the fractures in their study more closely resembled proximal diaphyseal stress fractures, as opposed to true Jones fractures.

There are several limitations in this study, most notably the possibility of chance or residual confounding as an explanation of the findings. If we had a larger pool of patients it is possible that other covariates (e.g., smoking and osteoporosis) for which we had only a few observations might have found their way into the final model. Also, given the retrospective nature of the work, the majority of the covariates which we examined were restricted to radiographic measurements. Finally, the lack of true frontal plane radiographs (e.g., Saltzman views or long leg axial views) and absence of a structured standing and walking clinical exam make it difficult to be certain that hindfoot alignment is not a more important factor in Jones fracture injuries. Lateral x-ray measurements are, at best, a proxy as to what is occurring in the heel in the frontal plane. Because previous authors (5) found a strong correlation between CIA and heel position, and also used CIA in a similar way to examine the association with heel position and Jones fractures, we were comfortable using CIA as an indirect measure of frontal plane heel position in our work; however, this does not take away from the fact that true frontal plane views and/or clinical exam findings would help make a stronger case for any conclusions reached regarding heel position.

The strength of the work is that it provides good evidence (strong association, dose-response, and biologic plausibility) that an adducted forefoot posture occupies a place in the causal pathway in the development of acute Jones fractures. In fact, we found that patients with metatarsus adductus alone were at 2.4 times greater risk of presenting with a Jones fracture than those without this risk factor.

In summary, the risk of Jones fracture increases with adducted forefoot posture. In our population which consisted primarily of patients presenting after fall or after inversion/misstep injuries, hindfoot alignment appeared to be a less important risk factor. Methods of shielding the lateral forefoot with orthoses and braces may help to prevent fractures and reduce the likelihood of recurrent injuries.