- Email: [email protected]
Relationship of Cuboid Height to Plantar Ulceration and Other Radiographic Parameters in Midfoot Charcot Neuroarthropathy
The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Contents lists available at ScienceDirect
The Journal of Foot & Ankle Surgery journal homepage: www.jfas.org
Relationship of Cuboid Height to Plantar Ulceration and Other Radiographic Parameters in Midfoot Charcot Neuroarthropathy Andrew J. Meyr, DPM, FACFAS 1, Joshua A. Sebag, DPM 2 1 2
Clinical Associate Professor, Department of Podiatric Surgery, Temple University School of Podiatric Medicine, Philadelphia, PA Podiatric Student, Temple University School of Podiatric Medicine, Philadelphia, PA
a r t i c l e i n f o
a b s t r a c t
Level of Clinical Evidence: 3
The objective of the present investigation was to examine the effect of cuboid height on the presence of plantar midfoot ulceration and the relationship of cuboid height to other commonly performed radiographic parameters during evaluation of midfoot Charcot neuroarthropathy. A retrospective analysis was performed of 68 feet in 60 subjects who met the inclusion criteria. We did not observe statistically significant differences in the presence of a plantar midfoot ulceration when considering a cuboid height threshold of 0.0 mm, 2.0 mm, 2.0 mm, or 5.0 mm nor was the cuboid height a robust predictor for the presence of plantar midfoot ulceration when considering the positive predictive value, negative predictive value, sensitivity, or specificity. We observed a significant negative association between a negative cuboid height and the presence of Sanders Type 2 deformities (76.2% of those with negative height versus 50.0% of those with positive height had type 2 deformities; p ¼ .0036), the absence of radiographic visualization of the lateral tarsometatarsal joint (71.4% of those with negative height versus 26.9% of those with positive height had an absence of radiographic visualization; p ¼ .005), and lower calcaneal inclination angles (6.06 versus 15.08 ; p < .001). We further observed significant positive correlations between the cuboid height and the calcaneal–fifth metatarsal angle (0.655; p < .000), calcaneal inclination angle (0.591; p < .001), calcaneal–cuboid angle (0.254; p ¼ .038), medial column height (0.264; p ¼ .029), and first metatarsal inclination angle (0.245; p ¼ .047). We also observed negative correlations with Meary’s angle (0.475; p < .001) and the talar declination angle (0.387; p < .001). These findings showed a general trend toward a decreasing cuboid height and increasing sagittal plane deformity involving both the medial and the lateral columns. The results of the present investigation provide evidence against a single radiographic parameter being associated with the presence of plantar midfoot ulceration. Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved.
Keywords: complication diabetes mellitus lateral column height plain film radiography
Wukich et al (1) recently reported a strong association between the objective parameter of cuboid height (or lateral column height) on the lateral radiographic projection and plantar foot ulceration in patients with midfoot Charcot neuroarthropathy (CN). They observed a statistically significant difference in cuboid height between a retrospective series of midfoot CN patients with and without foot ulceration and reported that 40 of 50 patients (80%) with midfoot CN and a foot ulceration had a negative cuboid height (1). They described a negative cuboid height as one in which a structurally plantarflexed cuboid extends inferiorly beyond the plantar osseous plane of the lateral column of the foot.
Financial Disclosure: None reported. Conflict of Interest: None reported. Address correspondence to: Andrew J. Meyr, DPM, FACFAS, Department of Podiatric Surgery, Department of Surgery, Temple University School of Podiatric Medicine, 8th at Race Street, Philadelphia, PA 19107. E-mail address: [email protected] (A.J. Meyr).
Bevan and Tomlinson (2) had also previously sought to associate the radiographic angular measurements with plantar ulceration in midfoot CN. Although they did not specifically measure the cuboid height, they did find that the lateral talo–first metatarsal angle (or Meary’s angle) and calcaneal–fifth metatarsal angles were sensitive parameters associated with ulceration (2). All patients with a plantar midfoot ulceration in their cohort had a Meary’s angle more negative than 27. We identified 2 other studies that described the cuboid height in association with CN deformity. Hastings et al (3) demonstrated that the cuboid height became progressively more negative in a group of diabetic patients with CN during a 2-year period, and Schon et al (4) quantified the lateral column height in groups of CN patients with differing mid-tarsus deformities. Furthermore, Steel et al (5) provided descriptive data of the lateral column height in a group of “normal” adult feet in 1980, and Younger et al (6) also examined a group of 21 feet with adult flatfoot deformity in 2005. The latter study of flatfeet further described an intraobserver measurement (r2 values) of 0.4 and an interobserver measurement of
1067-2516/$ - see front matter Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved. http://dx.doi.org/10.1053/j.jfas.2017.02.007
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
0.96 for the lateral column height and an intraobserver measurement of 0.92 and interobserver measurement of 0.95 for the calcaneus–fifth metatarsal angle. In an earlier investigation, Hastings et al (7) also described a relatively “precise” measurement of the cuboid height between 2 orthopedic surgeons and a radiology resident (a combined rater’s least significant change of 4.7 mm). Somewhat in line with these investigations (1–7), we have also observed in our clinical practice a trend for some patients with midfoot CN to develop a cuboid that displaces plantarly (creating a more negative cuboid height) with seemingly poor clinical outcomes. In contrast, others have developed a cuboid that displaces dorsally (creating a more positive cuboid height), with seemingly better clinical outcomes. The objective of the present retrospective, observational comparative investigation was to further examine the association of cuboid height with plantar ulceration in midfoot CN. We specifically aimed to (1) describe the association of cuboid height with the presence of plantar midfoot ulceration in a retrospective cohort of patients with midfoot CN, and (2) describe the relationship of other radiographic parameters with a cuboid height in a retrospective cohort of patients with midfoot CN.
Patients and Methods After approval from our institutional review board, an “International Classification of Diseases” diagnostic code search was performed for all patients seen at the Temple University Foot and Ankle Institute during a 1-year data collection period (January 2015 to December 2015). The specific codes used in this search were 713.5 (Charcot’s arthropathy associated with diabetes mellitus), M14.671 (Charcot’s joint, right ankle and foot), M14.672 (Charcot’s joint, left ankle and foot), and M14.60 (Charcot’s joint, unspecified site). The electronic medical records found were then interrogated for the study inclusion and exclusion criteria. The inclusion criteria were the availability of at least anteroposterior (AP) and lateral weightbearing radiographic foot projections, evidence of midfoot CN (defined as the tarsometatarsal, midfoot, and/or transverse tarsal joints; Sanders/Frykberg classification types 2 and 3 [8]), no evidence of previous surgical reconstruction (defined as internal or external fixation of the midfoot, rearfoot and/or ankle), and no partial foot amputation or osseous resection involving the rearfoot, midfoot, any complete metatarsal, or >1 metatarsal head. The exclusion criteria were the lack of appropriate weightbearing or preoperative intervention radiographs, no radiographic evidence of CN, the presence of CN of the forefoot, subtalar joint, calcaneus and/or ankle, and any evidence of surgical reconstruction, partial foot amputation, or osseous resection involving the rearfoot, midfoot, complete metatarsal, or >1 metatarsal head. Thus, we included subjects with a history of digital amputations
749
and/or single metatarsal head amputations but excluded those with multiple partial or complete ray amputations, including transmetatarsal amputations. The electronic medical record was reviewed for participant age, gender, laterality of pathology, a history of lower extremity ulceration and/or amputation, and a history of eventual CN reconstruction. The radiographs were reviewed for CN classification as defined by the Sanders/Frykberg classification (8). Type 2 deformities were defined as involvement of the Lisfranc tarsometatarsal joint. Type 3 deformities were defined as involvement of the intercuneiform, cuneocuboid, naviculocuneiform, talonavicular, and calcaneocuboid joints. We also subjectively and categorically reviewed the AP radiographs to determine whether the lateral tarsometatarsal joint was visualized (Fig. 1). The lack of visualization of this joint space could have resulted from a variety of factors, including consolidation as a part of the CN process, overlapping or underlapping of the metatarsal bases on the cuboid as a part of the CN process, or severe arthritic changes. AP radiographs were further reviewed for the following radiographic parameters: AP talocalcaneal angle (Kite’s angle), calcaneal–cuboid angle, and AP talo–first metatarsal angle (1,9). The lateral radiographs were reviewed for the cuboid height, calcaneal inclination angle, talar declination angle, tibial–talar angle, first metatarsal inclination angle, fifth metatarsal inclination angle, lateral talo-first metatarsal angle (Meary’s angle), calcaneal–fifth metatarsal angle, and medial column height (1,9). The radiographic measurements were taken using computerized digital software (Opal-RAD PACS; Viztek, Garner, NC), which measured to a precision of 0.1 and 0.1 mm. The cuboid height and medial column height were measured relative to a line drawn from the plantar aspect of calcaneal tuberosity to the plantar aspect of the fifth metatarsal head (Fig. 2). The cuboid height was measured as the perpendicular distance from the most inferior aspect of cuboid to this line, and the medial column height was defined as the perpendicular distance from the most inferior aspect of the first metatarsal–cuneiform articulation to this line (1,3–7,9). The radiographic measurements were performed by 1 of us (J.S.) and confirmed by the corresponding author (A.J.M.). We attempted to be as conservative as possible regarding the radiographic measurements, given the osseous destruction that often accompanies CN. For example, we had initially planned on measuring the talonavicular coverage on the AP radiographic projections but found early in the data collection that destruction of the navicular structural integrity during the CN disease process made this measurement both difficult and unreliable; hence, we did not include it. If a given osseous landmark lacked structural integrity and made measurement of a specific angle questionable, we omitted this measurement. Furthermore, in patients with first metatarsal head resection, we did not perform measurement of the AP Meary angle, Meary’s angle, or the first metatarsal inclination angle. In patients with fifth metatarsal head resection, we did not perform measurement of the fifth metatarsal inclination angle and measured the calcaneal–fifth metatarsal angle using the plantar aspect of the fourth metatarsal head. Categorical outcomes were described in terms of frequency and were compared between groups using Fisher’s exact test. Continuous outcomes were described in terms of the mean, standard deviation, and range and compared between groups using the nonpaired Student t test. Our primary comparisons concerned the cuboid height. We first described our findings of cuboid height in terms of the mean, standard deviation, and range and graphically depicted the results on a histogram. We then divided the cohort into those with positive (0.0 mm) and negative (<0.0 mm) cuboid heights
Fig. 1. Assessment of radiographic visualization of the lateral tarsometatarsal joint. (Left) An example of what we defined as a relatively normal lateral tarsometatarsal joint with radiographically visible joint space. (Right) An example of what we defined as the lack of visualization of the lateral tarsometatarsal joint space.
750
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Fig. 2. Measurement of cuboid height (lateral column height). The cuboid height was defined as the perpendicular distance between the most inferior aspect of the cuboid and a reference line between the plantar aspect of the calcaneal tuberosity and the plantar aspect of the fifth metatarsal head.
In the transverse plane, the mean standard deviation AP talocalcaneal angle was 19.06 8.73 (range 1.4 to 42 ). The mean standard deviation calcaneal–cuboid angle was 21.06 14.82 (range 14.4 to 59.3 ). The mean standard deviation AP talo–first metatarsal angle was 10.92 12.53 (25.60 to 39.3 ). In the sagittal plane, the mean standard deviation cuboid height was 2.2 8.6 (range 16.9 to 22.6) mm; 42 feet (61.76%) had a negative cuboid height. The mean standard deviation calcaneal inclination angle was 9.51 9.79 (range 17.6 to 38.3 ). The mean standard deviation talar declination angle was 38.97 8.26 (range 22 to 72.7 ). The mean standard deviation tibial–talar angle
Table 1 Cohort characteristics Characteristic
and performed a statistical comparison of all other clinical and radiographic measurements between these groups. Additionally, the individual radiographic parameters were graphically depicted against the cuboid height on a scatter plot, with calculation of a correlation line and Pearson’s correlation coefficient. We also performed a similar bivariate comparison of all clinical and radiographic parameters between those with and without history of plantar midfoot ulceration. We further defined other potential thresholds for normal and abnormal cuboid height in addition to 0.0 mm to include 2.0 mm, 2.0 mm, and 5.0 mm. We are unaware of any objective analysis providing evidence of a cuboid height of 0.0 mm as being “normal” and, therefore, also chose to analyze these other potential thresholds. The association of these thresholds with plantar midfoot ulceration was assessed using Fisher’s exact test, sensitivity, specificity, positive predictive value, and negative predictive value. Data were stored in a password-protected and encrypted personal computer for subsequent statistical analysis. All statistical analyses were performed by 1 of us (A.J.M.) using Statistical Analysis Systems software, version 9.2 (SAS Institute, Cary, NC).
Results The initial “International Classification of Diseases” diagnostic code search produced 96 feet in 81 subjects. Of these 96 feet, 8 were subsequently excluded because of a lack of appropriate radiographs, 5 were excluded because CN had occurred in an anatomic location other than the midfoot, 12 were excluded because they had no radiographic evidence of CN, and 3 were excluded because the patients had undergone multiple metatarsal head resections. This left 68 feet in 60 subjects who met the inclusion criteria (Table 1). Of these 68 feet, 6 (8.82%) had a history of first metatarsal head resection; thus, we did not measure the AP Meary angle, Meary’s angle, or first metatarsal inclination angle on those radiographs. Also, 5 feet (7.35%) had a history of fifth metatarsal head resection; thus, we did not measure the fifth metatarsal inclination angle and measured the calcaneal– fifth metatarsal axis based on the plantar aspect of the fourth metatarsal head on those radiographs. In addition, 2 feet (2.94%) had osseous destruction of the talar head; thus, we did not measure the AP talocalcaneal angle, AP talo–first metatarsal angle, talar declination angle, lateral talo–first metatarsal angle, or tibial–talar angle on those radiographs. Finally, 1 of these 68 feet (1.47%) had significant osseous destruction of the cuboid; thus, we did not measure the calcaneal– cuboid abduction angle on that radiograph. The mean standard deviation subject age was 56.74 8.57 (range 37 to 77) years. Of the 60 patients, 34 (56.67%) were male, and 29 deformities (42.65%) were right-sided. The Sanders/Frykberg classification was type 2 in 45 feet (66.18%) and type 3 in 13 feet (19.12%). Ten feet (14.71%) had evidence of CN occurring in both midfoot locations. Of the 68 feet, 37 (54.41%) lacked radiographic visualization of the lateral tarsometatarsal joint, 28 (41.18%) had a history of plantar midfoot ulceration, and 44 (64.71%) had a history of any lower extremity ulceration or partial forefoot amputation. Of the 68 feet, 17 (25.0%) subsequently underwent surgical reconstruction of the CN deformity.
Demographic Feet (n) Patients (n) Age (y) Mean SD Range Gender Male Female Laterality Right Left Sanders classification Type 2 Type 3 Both types 2 and 3 Plantar midfoot wound history Any lower extremity wound/forefoot amputation history Underwent reconstruction Radiographic AP talar–calcaneal angle ( ; n ¼ 66 feet) Mean SD Range Calcaneal–cuboid angle ( ; n ¼ 67) Mean SD Range AP talo–first metatarsal angle ( ; n ¼ 60) Mean SD Range Cuboid height (mm; n ¼ 68) Mean SD Range Calcaneal inclination angle ( ; n ¼ 68) Mean SD Range Talar declination angle ( ; n ¼ 66) Mean SD Range Lateral talar–first metatarsal angle ( ; n ¼ 60) Mean SD Range First metatarsal inclination angle ( ; n ¼ 62) Mean SD Range Fifth metatarsal inclination angle ( ; n ¼ 63) Mean SD Range Calcaneal–fifth metatarsal angle ( ; n ¼ 68) Mean SD Range Tibial–talar angle ( ; n ¼ 66) Mean SD Range Medial column height (mm; n ¼ 68) Mean SD Range Abbreviation: SD, standard deviation. Data in parentheses are percentages.
Value 68 60 56.74 8.57 37–77 34 (56.67) 26 (43.33) 29 (42.65) 39 (53.35) 45 13 10 28 44 17
(66.18) (19.12) (14.71) (41.18) (64.71) (25.0)
19.06 8.73 1.4–42 21.06 14.82 14.4–59.3 10.92 12.53 25.60 to 39.3 2.2 8.6 16.9 to 22.6 9.51 9.79 17.6 to 38.3 38.97 8.26 22–72.7 21.69 9.82 1.5–51.2 17.96 6.45 5–36 5.37 4.54 3.1 to 22 11.82 10.53 19 to 43.2 133.82 10.45 111.2–162.6 9.0 10.9 28.9 to 37.3
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
751
Fig. 3. Histogram of cuboid height showing a roughly parametrically shaped data distribution.
was 133.82 10.45 (range 111.2 to 162.6 ). The mean standard deviation first metatarsal inclination angle was 17.96 6.45 (range 5 to 36 ). The mean standard deviation fifth metatarsal inclination angle was 5.37 4.54 (range 3.1 to 22 ). The mean standard deviation lateral talo–first metatarsal angle was 21.69 9.82 (range 1.5 to 51.2 ); 15 feet (25.0%) had a lateral talo–first metatarsal angle of 27. The mean standard deviation calcaneal–fifth metatarsal angle was 11.82 10.53 (range 19 to 43.2 ). The mean standard deviation medial column height was 9.0 10.9 (range 28.9 to 37.3) mm. Depiction of the cuboid height on a histogram demonstrated a roughly parametric or normal distribution (Fig. 3). Fisher’s exact tests revealed no significant differences in the rates of plantar midfoot ulceration between those with a normal versus an abnormal cuboid height, using any of the thresholds (0.0, 2.0, 2.0, and 5.0 mm). Furthermore, the positive predictive value, negative predictive value,
sensitivity, and specificity for plantar midfoot ulceration on normality/abnormality were not high (Table 2). A negative cuboid height (<0.0 mm) was also not statistically significant when isolating only the 45 subjects with Sanders type 2 CN with (n ¼ 15) and without (n ¼ 26) plantar midfoot ulceration (78.95% versus 65.38%; p ¼ .5067). The results of the comparison of clinical and radiographic variables between subjects with a positive and negative cuboid height, using 0.0 mm as the threshold, are listed in Table 3. Table 3 also includes the Pearson correlation coefficients of the other radiographic parameters measured with the cuboid height. Subjects with negative cuboid height were more likely to have a Sanders/Frykberg type 2 deformity (76.2% of those with a negative cuboid height versus 50.0% of those with a positive cuboid height had a type 2 deformity; p ¼ .036), were less likely to have a Sanders/Frykberg type 3 deformity (9.5% of those with a negative cuboid height versus 34.6% of those with a positive cuboid height had a type 3 deformity; p ¼ .023), and were more likely
Table 2 Association of cuboid height with plantar midfoot ulceration (n ¼ 68 total) CH Threshold (mm)
p Value*
Plantar Midfoot Ulceration (%) PPV
NPV
Sensitivity
Specificity
<0.0 (n ¼ 42) <2.0 (n ¼ 51) <2.0 (n ¼ 38) <5.0 (n ¼ 29)
.453 .3939 1.00 .627
34.62 29.14 40.00 38.46
54.76 54.90 57.89 55.17
32.14 17.86 42.86 53.57
57.50 70.00 55.00 40.00
Abbreviations: CH, cuboid height; NPV, negative predictive value; PPV, positive predictive value. * Fisher’s exact test comparing those with and without plantar midfoot ulceration.
752
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Table 3 Comparison of positive versus negative cuboid height and clinical and radiographic parameters Variable
Sanders classification Type 2 Type 3 Both types 2 and 3 Lack of visible lateral tarsometatarsal joint space Plantar midfoot wound history Any lower extremity wound/forefoot amputation history Underwent reconstruction AP talocalcaneal angle ( ) Calcaneal–cuboid angle ( ) AP talo-first metatarsal angle ( ) Calcaneal inclination angle ( ) Talar declination angle ( ) Tibial–talar angle ( ) First metatarsal inclination angle ( ) Fifth metatarsal inclination angle ( ) Meary’s angle ( ) Meary’s angle 27 Calcaneal–fifth metatarsal angle ( ) Medial column height (mm)
CH (mm)
p Value
0.0 (n ¼ 26)
<0.0 (n ¼ 42)
13 (50.0) 9 (34.6) 4 (15.4) 7 (26.9) 9 (34.6) 17 (65.3) 4 (15.4) 18.41 25.39 8.95 15.08 36.65 133.96 19.78 5.37 17.72 4 (15.38) 17.85 11.54
32 (76.2) 4 (9.5) 6 (14.3) 30 (71.4) 19 (45.2) 27 (64.3) 13 (31.0) 19.46 8.75 18.32 13.03 12.05 13.88 6.06 8.22 40.38 8.99 133.73 9.43 16.91 6.63 5.37 5.07 23.92 9.72 13 (30.95) 8.09 8.74 7.44 12.28
8.83 16.63 9.69 9.68 6.41 12.16 5.80 3.63 8.87 10.52 7.58
.036* .023* 1.00 .005*,y .453 1.00 .249 .6356 .0565 .3376 < .001* .0749 .9310 .0814 .998 .014* .2486 < .001* .131
Pearson Correlation Coefficient (p Value)
0.172 0.254 0.154 0.591 0.387 0.160 0.245
(.161) (.038*) (.216) (< .000*) (< .001*) (p ¼ .199) (p ¼ .047*)
0.475 (< .000*) 0.655 (< .000*) 0.264 (.029*)
Abbreviations: AP, anteroposterior; CH, cuboid height. Data presented as n (%) or mean standard deviation. Categorical outcomes compared between groups using Fisher’s exact test; continuous outcomes compared between groups using nonpaired Student’s t test. * Statistically significant. y Odds ratio 6.79 (95% confidence interval 2.01 to 23.84).
to lack visualization of the lateral tarsometatarsal joint (71.4% of those with a negative cuboid height versus 26.9% of those with a positive cuboid height had an absence of visualization of the lateral tarsometatarsal joint; p ¼ .005). Subjects with a negative cuboid height had a greater frequency of plantar midfoot wounds; however, this difference was not statistically significant (45.2% versus 34.6%; p ¼ .453). Subjects with a negative cuboid height demonstrated a lower calcaneal inclination angle (6.06 versus 15.08 ; p < .001), greater Meary’s angle (23.92 versus 17.72 ; p ¼ .014), and lower calcaneal–fifth metatarsal angle (8.09 versus 17.85 ; p < .001). A statistically significant positive Pearson correlation coefficient was observed between the cuboid height and the calcaneal–cuboid angle (0.254; p ¼ .038), calcaneal inclination angle (0.591; p < .001), first metatarsal inclination angle (0.245; p ¼ .047), calcaneal–fifth metatarsal angle (0.655; p < .001), and medial column height (0.264; p ¼ .029). A statistically significant negative Pearson correlation coefficient was observed between the cuboid height and the talar declination angle (0.387; p < .001) and Meary’s angle (0.475; p < .001). Frequency scatter plots for all variables are shown in Fig. 4. The results of the comparison of clinical and radiographic variables of the subjects with and without a history of a plantar midfoot ulceration are listed in Table 4. We did not observe any statistically significant differences between groups with respect to the clinical or radiographic variables. Discussion The objective of the present retrospective, observational investigation was to examine the association between the cuboid height and the presence of ulceration in subjects with midfoot CN. Our results have shown that the cuboid height might not be a robust measure with respect to the presence or absence of midfoot plantar foot ulceration (1). We did not observe a significant difference in cuboid height between study participants with and without a plantar midfoot ulceration (2.77 versus 1.86 mm; p ¼ .671) nor a significant difference in the presence of plantar midfoot ulceration between participants with a positive and negative cuboid height (34.6% versus
45.2%; p ¼ .453). Furthermore, we observed relatively modest positive predictive values, negative predictive values, sensitivities, and specificities when considering cuboid height and plantar midfoot ulceration at multiple cuboid height thresholds (0.0, 2.0, 2.0, and 5.0 mm; Table 2). This was against our study hypothesis and did not match the observations of Wukich et al (1). Although we have continued to use this measurement in our clinical practice and consider it an important part of the radiographic evaluation of midfoot CN, we did not find it to be associated with the presence or absence of ulceration in our cohort. This is clinically important because the presence of midfoot ulceration and associated increased risk of infection in the setting of ulceration has been cited as a major contributor in the decision-making process for conservative versus surgical treatment of midfoot CN (1–13). Second, we also did not observe any statistically significant association between the presence of plantar midfoot ulceration and other commonly performed radiographic angular measurements of the foot and ankle. We believe this emphasizes that the clinical outcomes and the presence of plantar midfoot ulcerations in CN are related to other confounding clinical factors and aspects of lower extremity function, rather than just the structural deformity as assessed with plain film radiographs. However, perhaps our assessment of midfoot CN should be more complex than simply the position of a single bone or joint. As an example, the classification of Schon et al (4) for midfoot CN describes 4 different types involving the intercuneiform and cuneocuboid articulations. Third, we believe that the present investigation has provided unique information regarding the relationship of commonly performed radiographic angular measurements at various points in the midfoot CN process and, in particular, as it relates to cuboid height. We observed a significant negative association between the cuboid height and the presence of Sanders/Frykberg type 2 deformities (76.2% of those with a negative cuboid height versus 50.0% of those with a positive cuboid height had a type 2 deformity; p ¼ .036), those lacking visualization of the lateral tarsometatarsal joint (71.4% of those with negative cuboid height versus 26.9% of those with positive cuboid height had an absence of visualization of the lateral
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
753
Fig. 4. Frequency scatter plots of radiographic parameters graphed against the cuboid height. These frequency scatter plots with corresponding linear correlation lines demonstrate the positive and negative relationships between the cuboid height and the other measured radiographic parameters. (A) represents the frequency scatter plot demonstrating a negative relationship between the AP talocalcaneal angle and cuboid height. (B) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal-cuboid angle and cuboid height. (C) represents the frequency scatter plot demonstrating a negative relationship between the AP talo-first metatarsal angel and cuboid height. (D) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal inclination angle and cuboid height. (E) represents the frequency scatter plot demonstrating a negative relationship between the talar declination angle and cuboid height. (F) represents the frequency scatter plot demonstrating a negative relationship between the lateral talo-first metatarsal angle and cuboid height. (G) represents the frequency scatter plot demonstrating a positive relationship between the first metatarsal inclination angle and cuboid height. (H) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal-fifth metatarsal angle and cuboid height. (I) represents the frequency scatter plot demonstrating a negative relationship between the tibial-talar angle and cuboid height. (J) represents the frequency scatter plot demonstrating a positive relationship between the medial column height and cuboid height.
tarsometatarsal joint; p ¼ .005), and those with lower calcaneal inclination angles (6.06 versus 15.08 ; p < .001). Patients lacking visualization of the lateral tarsometatarsal joint had 6.79 times the odds of having a negative cuboid height. We believe that this might provide evidence that cuboid height is related to forces not only derived from the rearfoot and calcaneocuboid joint but also retrograde forces from the forefoot. It might also indicate that the structure of the fourth and fifth metatarsal bases warrants evaluation in cuboid planning procedures. We further observed significant positive correlations between the cuboid height and the calcaneal–fifth metatarsal angle (0.655; p < .000), calcaneal inclination angle (0.591; p < .001), calcaneal– cuboid angle (0.254; p ¼ .038), medial column height (0.264; p ¼ .029), and first metatarsal inclination angle (0.245; p ¼ .047). Negative correlations were observed between Meary’s angle (0.475;
p < .001) and the talar declination angle (0.387; p < .001). This, somewhat unsurprisingly, showed a trend toward decreasing cuboid heights associated with increasing sagittal plane deformity involving both the medial and the lateral columns of the foot. Just as with any scientific investigation, critical readers are encouraged to review the study design and results and reach their own conclusions. Our report represents our conclusions from the specific results. Also, just as with any study, the present study had limitations. First, the data were collected from a limited number of subjects from a single institution; thus, these results might not be representative of a broader population sampling. When considering the subject cohort, we also did not specifically account for the severity of the deformity in our inclusion and exclusion criteria. This might represent an interesting avenue for future investigation. Second, all retrospective studies have inherent limitations. We were dependent on the diagnostic codes
754
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Fig. 4. (continued).
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
755
Table 4 Comparison of patients with and without plantar midfoot ulceration history stratified by clinical and radiographic parameters Variable
Sanders classification Type 2 Type 3 Both types 2 and 3 Lack of visible lateral tarsometatarsal joint space Underwent reconstruction AP calcaneal–talar angle ( ) Calcaneal–cuboid angle ( ) AP talo–first metatarsal angle ( ) CH (mm) CH <0.0 mm Calcaneal inclination angle ( ) Talar declination angle ( ) Tibial–talar angle ( ) First metatarsal inclination angle ( ) Fifth metatarsal inclination angle ( ) Meary’s angle ( ) Meary’s angle 27 Calcaneal–fifth metatarsal angle ( ) Medial column height (mm)
History of Plantar Midfoot Ulceration
p Value
No (n ¼ 40)
Yes (n ¼ 28)
26 (65.0) 10 (25.0) 4 (10.0) 18 (45.0) 8 (20.0) 18.86 8.30 21.74 13.52 11.06 11.55 1.86 8.06 23 (57.5) 9.10 8.29 39.26 8.75 132.80 8.82 18.47 6.28 4.62 3.26 21.50 10.75 12 (30.77) 12.32 8.59 8.23 10.70
19 (67.86) 3 (10.71) 6 (21.43) 19 (67.86) 9 (32.14) 19.33 9.46 20.07 16.78 10.72 14.06 2.77 9.47 19 (67.86) 10.08 11.74 38.52 7.59 135.39 12.60 17.21 6.73 6.49 5.87 22.00 8.36 4 (16.0) 11.11 12.95 10.12 11.17
1.00 .104 .297 .085 .272 .829 .655 .915 .671 .453 .688 .724 .328 .440 .103 .846 .243 .645 .484
Abbreviations: AP, anteroposterior; CH, cuboid height. Data presented as n (%) or mean standard deviation. Categorical outcomes compared between groups using Fisher’s exact test; continuous outcomes compared between groups using nonpaired Student’s t test.
for identification of the cohort and on physician documentation and data collection of the patient characteristics. Any inaccuracies within these data would have resulted in inaccuracy of our data analysis and interpretation. Third, our primary outcome measure and secondary outcome measures were primarily derived from the radiographic analysis, dependent on consistent patient positioning and angle measurement. Patient positioning is often difficult for patients with CN and might not have been as reliable as positioning for radiographic studies examining other lower extremity pathologic entities. One might also disagree with our specific statistical analysis and plan. We believed that a direct comparison of outcomes with the Fisher exact test and nonpaired Student t test would be of most benefit and interest for readers of the present initial study; however, future investigations might consider a more complex statistical plan such as analysis of variance. Finally, the cuboid height is a relatively understudied radiographic parameter, and we are not aware of any analysis of the inter- and intrarater reliability of this measurement in a group of patients with midfoot CN (6,7). The reliability of this measurement might also represent an interesting avenue for future investigations. In conclusion, the results of the present investigation have provided evidence that it is unlikely that a single radiographic parameter used in contemporary practice is associated with plantar midfoot ulceration in patients with midfoot CN. It is our hope that these data, which provide new information on the CN deformity as it relates to cuboid height, might lead to the development of such a measure or combination of measures that are predictive of ulceration in patients with midfoot CN.
References 1. Wukich DK, Raspovic KM, Hobizal KB, Rosario B. Radiographic analysis of diabetic midfoot Charcot neuroarthropathy with and without midfoot ulceration. Foot Ankle Int 35:1108–1115, 2014. 2. Bevan WP, Tomlinson MP. Radiographic measures as a predictor of ulcer formation in diabetic Charcot midfoot. Foot Ankle Int 29:568–573, 2008. 3. Hastings MK, Johnson JE, Strube MJ, Hildebolt CF, Bohnert KL, Prior FW, Sinacore DR. Progression of foot deformity in Charcot neuropathic osteoarthropathy. J Bone Joint Surg Am 95:1206–1214, 2013. 4. Schon LC, Weinfeld SB, Horton GA, Resch S. Radiographic and clinical classification of acquired midtarsus deformities. Foot Ankle Int 19:394–404, 1998. 5. Steel MW, Johnson KA, DeQitz MA, Ilstrup MS. Radiographic measurements of the normal adult foot. Foot Ankle 1:151–158, 1980. 6. Younger AS, Sawatzky B, Dryden P. Radiographic assessment of adult flatfoot. Foot Ankle Int 26:820–825, 2005. 7. Hastings MK, Sinacore DR, Mercer-Bolton N, McCormick JJ, Hidebolt CF, Prior FW, Johnson JE. Precision of foot alignment measures in Charcot arthropathy. Foot Ankle Int 32:867–872, 2011. 8. Rogers LC, Frykberg RG, Armstrong DG, Boulton AJ, Edmonds M, Van GH, Hartemann A, Game F, Jeffcoate W, Jirkovska A, Jude E, Morbach S, Morrison WB, Pinzur M, Pitocco D, Sanders L, Wukich DK, Uccioli L. The Charcot foot in diabetes. Diabetes Care 34:2123–2129, 2011. 9. Sanner WH. Foot segmental relationships and bone morphology. In: Foot and Ankle Radiology, pp. 272–302, edited by RA Christman, Churchill Livingstone, St. Louis, MO, 2003. 10. Pinzur M. Surgical versus accommodative treatment for Charcot arthropathy of the midfoot. Foot Ankle Int 25:545–549, 2004. 11. Catanzariti AR, Medicino R, Haverstock B. Ostectomy for diabetic neuroarthropathy involving the midfoot. J Foot Ankle Surg 39:291–300, 2000. 12. Wukich DK, Sung W. Charcot arthropathy of the foot and ankle: modern concepts and management review. J Diabetes Complications 23:409–426, 2009. 13. Smith WB, Moore CA. A proposed treatment algorithm for midfoot Charcot arthropathy. Foot Ankle Spec 5:60–64, 2012.
Contents lists available at ScienceDirect
The Journal of Foot & Ankle Surgery journal homepage: www.jfas.org
Relationship of Cuboid Height to Plantar Ulceration and Other Radiographic Parameters in Midfoot Charcot Neuroarthropathy Andrew J. Meyr, DPM, FACFAS 1, Joshua A. Sebag, DPM 2 1 2
Clinical Associate Professor, Department of Podiatric Surgery, Temple University School of Podiatric Medicine, Philadelphia, PA Podiatric Student, Temple University School of Podiatric Medicine, Philadelphia, PA
a r t i c l e i n f o
a b s t r a c t
Level of Clinical Evidence: 3
The objective of the present investigation was to examine the effect of cuboid height on the presence of plantar midfoot ulceration and the relationship of cuboid height to other commonly performed radiographic parameters during evaluation of midfoot Charcot neuroarthropathy. A retrospective analysis was performed of 68 feet in 60 subjects who met the inclusion criteria. We did not observe statistically significant differences in the presence of a plantar midfoot ulceration when considering a cuboid height threshold of 0.0 mm, 2.0 mm, 2.0 mm, or 5.0 mm nor was the cuboid height a robust predictor for the presence of plantar midfoot ulceration when considering the positive predictive value, negative predictive value, sensitivity, or specificity. We observed a significant negative association between a negative cuboid height and the presence of Sanders Type 2 deformities (76.2% of those with negative height versus 50.0% of those with positive height had type 2 deformities; p ¼ .0036), the absence of radiographic visualization of the lateral tarsometatarsal joint (71.4% of those with negative height versus 26.9% of those with positive height had an absence of radiographic visualization; p ¼ .005), and lower calcaneal inclination angles (6.06 versus 15.08 ; p < .001). We further observed significant positive correlations between the cuboid height and the calcaneal–fifth metatarsal angle (0.655; p < .000), calcaneal inclination angle (0.591; p < .001), calcaneal–cuboid angle (0.254; p ¼ .038), medial column height (0.264; p ¼ .029), and first metatarsal inclination angle (0.245; p ¼ .047). We also observed negative correlations with Meary’s angle (0.475; p < .001) and the talar declination angle (0.387; p < .001). These findings showed a general trend toward a decreasing cuboid height and increasing sagittal plane deformity involving both the medial and the lateral columns. The results of the present investigation provide evidence against a single radiographic parameter being associated with the presence of plantar midfoot ulceration. Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved.
Keywords: complication diabetes mellitus lateral column height plain film radiography
Wukich et al (1) recently reported a strong association between the objective parameter of cuboid height (or lateral column height) on the lateral radiographic projection and plantar foot ulceration in patients with midfoot Charcot neuroarthropathy (CN). They observed a statistically significant difference in cuboid height between a retrospective series of midfoot CN patients with and without foot ulceration and reported that 40 of 50 patients (80%) with midfoot CN and a foot ulceration had a negative cuboid height (1). They described a negative cuboid height as one in which a structurally plantarflexed cuboid extends inferiorly beyond the plantar osseous plane of the lateral column of the foot.
Financial Disclosure: None reported. Conflict of Interest: None reported. Address correspondence to: Andrew J. Meyr, DPM, FACFAS, Department of Podiatric Surgery, Department of Surgery, Temple University School of Podiatric Medicine, 8th at Race Street, Philadelphia, PA 19107. E-mail address: [email protected] (A.J. Meyr).
Bevan and Tomlinson (2) had also previously sought to associate the radiographic angular measurements with plantar ulceration in midfoot CN. Although they did not specifically measure the cuboid height, they did find that the lateral talo–first metatarsal angle (or Meary’s angle) and calcaneal–fifth metatarsal angles were sensitive parameters associated with ulceration (2). All patients with a plantar midfoot ulceration in their cohort had a Meary’s angle more negative than 27. We identified 2 other studies that described the cuboid height in association with CN deformity. Hastings et al (3) demonstrated that the cuboid height became progressively more negative in a group of diabetic patients with CN during a 2-year period, and Schon et al (4) quantified the lateral column height in groups of CN patients with differing mid-tarsus deformities. Furthermore, Steel et al (5) provided descriptive data of the lateral column height in a group of “normal” adult feet in 1980, and Younger et al (6) also examined a group of 21 feet with adult flatfoot deformity in 2005. The latter study of flatfeet further described an intraobserver measurement (r2 values) of 0.4 and an interobserver measurement of
1067-2516/$ - see front matter Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved. http://dx.doi.org/10.1053/j.jfas.2017.02.007
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
0.96 for the lateral column height and an intraobserver measurement of 0.92 and interobserver measurement of 0.95 for the calcaneus–fifth metatarsal angle. In an earlier investigation, Hastings et al (7) also described a relatively “precise” measurement of the cuboid height between 2 orthopedic surgeons and a radiology resident (a combined rater’s least significant change of 4.7 mm). Somewhat in line with these investigations (1–7), we have also observed in our clinical practice a trend for some patients with midfoot CN to develop a cuboid that displaces plantarly (creating a more negative cuboid height) with seemingly poor clinical outcomes. In contrast, others have developed a cuboid that displaces dorsally (creating a more positive cuboid height), with seemingly better clinical outcomes. The objective of the present retrospective, observational comparative investigation was to further examine the association of cuboid height with plantar ulceration in midfoot CN. We specifically aimed to (1) describe the association of cuboid height with the presence of plantar midfoot ulceration in a retrospective cohort of patients with midfoot CN, and (2) describe the relationship of other radiographic parameters with a cuboid height in a retrospective cohort of patients with midfoot CN.
Patients and Methods After approval from our institutional review board, an “International Classification of Diseases” diagnostic code search was performed for all patients seen at the Temple University Foot and Ankle Institute during a 1-year data collection period (January 2015 to December 2015). The specific codes used in this search were 713.5 (Charcot’s arthropathy associated with diabetes mellitus), M14.671 (Charcot’s joint, right ankle and foot), M14.672 (Charcot’s joint, left ankle and foot), and M14.60 (Charcot’s joint, unspecified site). The electronic medical records found were then interrogated for the study inclusion and exclusion criteria. The inclusion criteria were the availability of at least anteroposterior (AP) and lateral weightbearing radiographic foot projections, evidence of midfoot CN (defined as the tarsometatarsal, midfoot, and/or transverse tarsal joints; Sanders/Frykberg classification types 2 and 3 [8]), no evidence of previous surgical reconstruction (defined as internal or external fixation of the midfoot, rearfoot and/or ankle), and no partial foot amputation or osseous resection involving the rearfoot, midfoot, any complete metatarsal, or >1 metatarsal head. The exclusion criteria were the lack of appropriate weightbearing or preoperative intervention radiographs, no radiographic evidence of CN, the presence of CN of the forefoot, subtalar joint, calcaneus and/or ankle, and any evidence of surgical reconstruction, partial foot amputation, or osseous resection involving the rearfoot, midfoot, complete metatarsal, or >1 metatarsal head. Thus, we included subjects with a history of digital amputations
749
and/or single metatarsal head amputations but excluded those with multiple partial or complete ray amputations, including transmetatarsal amputations. The electronic medical record was reviewed for participant age, gender, laterality of pathology, a history of lower extremity ulceration and/or amputation, and a history of eventual CN reconstruction. The radiographs were reviewed for CN classification as defined by the Sanders/Frykberg classification (8). Type 2 deformities were defined as involvement of the Lisfranc tarsometatarsal joint. Type 3 deformities were defined as involvement of the intercuneiform, cuneocuboid, naviculocuneiform, talonavicular, and calcaneocuboid joints. We also subjectively and categorically reviewed the AP radiographs to determine whether the lateral tarsometatarsal joint was visualized (Fig. 1). The lack of visualization of this joint space could have resulted from a variety of factors, including consolidation as a part of the CN process, overlapping or underlapping of the metatarsal bases on the cuboid as a part of the CN process, or severe arthritic changes. AP radiographs were further reviewed for the following radiographic parameters: AP talocalcaneal angle (Kite’s angle), calcaneal–cuboid angle, and AP talo–first metatarsal angle (1,9). The lateral radiographs were reviewed for the cuboid height, calcaneal inclination angle, talar declination angle, tibial–talar angle, first metatarsal inclination angle, fifth metatarsal inclination angle, lateral talo-first metatarsal angle (Meary’s angle), calcaneal–fifth metatarsal angle, and medial column height (1,9). The radiographic measurements were taken using computerized digital software (Opal-RAD PACS; Viztek, Garner, NC), which measured to a precision of 0.1 and 0.1 mm. The cuboid height and medial column height were measured relative to a line drawn from the plantar aspect of calcaneal tuberosity to the plantar aspect of the fifth metatarsal head (Fig. 2). The cuboid height was measured as the perpendicular distance from the most inferior aspect of cuboid to this line, and the medial column height was defined as the perpendicular distance from the most inferior aspect of the first metatarsal–cuneiform articulation to this line (1,3–7,9). The radiographic measurements were performed by 1 of us (J.S.) and confirmed by the corresponding author (A.J.M.). We attempted to be as conservative as possible regarding the radiographic measurements, given the osseous destruction that often accompanies CN. For example, we had initially planned on measuring the talonavicular coverage on the AP radiographic projections but found early in the data collection that destruction of the navicular structural integrity during the CN disease process made this measurement both difficult and unreliable; hence, we did not include it. If a given osseous landmark lacked structural integrity and made measurement of a specific angle questionable, we omitted this measurement. Furthermore, in patients with first metatarsal head resection, we did not perform measurement of the AP Meary angle, Meary’s angle, or the first metatarsal inclination angle. In patients with fifth metatarsal head resection, we did not perform measurement of the fifth metatarsal inclination angle and measured the calcaneal–fifth metatarsal angle using the plantar aspect of the fourth metatarsal head. Categorical outcomes were described in terms of frequency and were compared between groups using Fisher’s exact test. Continuous outcomes were described in terms of the mean, standard deviation, and range and compared between groups using the nonpaired Student t test. Our primary comparisons concerned the cuboid height. We first described our findings of cuboid height in terms of the mean, standard deviation, and range and graphically depicted the results on a histogram. We then divided the cohort into those with positive (0.0 mm) and negative (<0.0 mm) cuboid heights
Fig. 1. Assessment of radiographic visualization of the lateral tarsometatarsal joint. (Left) An example of what we defined as a relatively normal lateral tarsometatarsal joint with radiographically visible joint space. (Right) An example of what we defined as the lack of visualization of the lateral tarsometatarsal joint space.
750
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Fig. 2. Measurement of cuboid height (lateral column height). The cuboid height was defined as the perpendicular distance between the most inferior aspect of the cuboid and a reference line between the plantar aspect of the calcaneal tuberosity and the plantar aspect of the fifth metatarsal head.
In the transverse plane, the mean standard deviation AP talocalcaneal angle was 19.06 8.73 (range 1.4 to 42 ). The mean standard deviation calcaneal–cuboid angle was 21.06 14.82 (range 14.4 to 59.3 ). The mean standard deviation AP talo–first metatarsal angle was 10.92 12.53 (25.60 to 39.3 ). In the sagittal plane, the mean standard deviation cuboid height was 2.2 8.6 (range 16.9 to 22.6) mm; 42 feet (61.76%) had a negative cuboid height. The mean standard deviation calcaneal inclination angle was 9.51 9.79 (range 17.6 to 38.3 ). The mean standard deviation talar declination angle was 38.97 8.26 (range 22 to 72.7 ). The mean standard deviation tibial–talar angle
Table 1 Cohort characteristics Characteristic
and performed a statistical comparison of all other clinical and radiographic measurements between these groups. Additionally, the individual radiographic parameters were graphically depicted against the cuboid height on a scatter plot, with calculation of a correlation line and Pearson’s correlation coefficient. We also performed a similar bivariate comparison of all clinical and radiographic parameters between those with and without history of plantar midfoot ulceration. We further defined other potential thresholds for normal and abnormal cuboid height in addition to 0.0 mm to include 2.0 mm, 2.0 mm, and 5.0 mm. We are unaware of any objective analysis providing evidence of a cuboid height of 0.0 mm as being “normal” and, therefore, also chose to analyze these other potential thresholds. The association of these thresholds with plantar midfoot ulceration was assessed using Fisher’s exact test, sensitivity, specificity, positive predictive value, and negative predictive value. Data were stored in a password-protected and encrypted personal computer for subsequent statistical analysis. All statistical analyses were performed by 1 of us (A.J.M.) using Statistical Analysis Systems software, version 9.2 (SAS Institute, Cary, NC).
Results The initial “International Classification of Diseases” diagnostic code search produced 96 feet in 81 subjects. Of these 96 feet, 8 were subsequently excluded because of a lack of appropriate radiographs, 5 were excluded because CN had occurred in an anatomic location other than the midfoot, 12 were excluded because they had no radiographic evidence of CN, and 3 were excluded because the patients had undergone multiple metatarsal head resections. This left 68 feet in 60 subjects who met the inclusion criteria (Table 1). Of these 68 feet, 6 (8.82%) had a history of first metatarsal head resection; thus, we did not measure the AP Meary angle, Meary’s angle, or first metatarsal inclination angle on those radiographs. Also, 5 feet (7.35%) had a history of fifth metatarsal head resection; thus, we did not measure the fifth metatarsal inclination angle and measured the calcaneal– fifth metatarsal axis based on the plantar aspect of the fourth metatarsal head on those radiographs. In addition, 2 feet (2.94%) had osseous destruction of the talar head; thus, we did not measure the AP talocalcaneal angle, AP talo–first metatarsal angle, talar declination angle, lateral talo–first metatarsal angle, or tibial–talar angle on those radiographs. Finally, 1 of these 68 feet (1.47%) had significant osseous destruction of the cuboid; thus, we did not measure the calcaneal– cuboid abduction angle on that radiograph. The mean standard deviation subject age was 56.74 8.57 (range 37 to 77) years. Of the 60 patients, 34 (56.67%) were male, and 29 deformities (42.65%) were right-sided. The Sanders/Frykberg classification was type 2 in 45 feet (66.18%) and type 3 in 13 feet (19.12%). Ten feet (14.71%) had evidence of CN occurring in both midfoot locations. Of the 68 feet, 37 (54.41%) lacked radiographic visualization of the lateral tarsometatarsal joint, 28 (41.18%) had a history of plantar midfoot ulceration, and 44 (64.71%) had a history of any lower extremity ulceration or partial forefoot amputation. Of the 68 feet, 17 (25.0%) subsequently underwent surgical reconstruction of the CN deformity.
Demographic Feet (n) Patients (n) Age (y) Mean SD Range Gender Male Female Laterality Right Left Sanders classification Type 2 Type 3 Both types 2 and 3 Plantar midfoot wound history Any lower extremity wound/forefoot amputation history Underwent reconstruction Radiographic AP talar–calcaneal angle ( ; n ¼ 66 feet) Mean SD Range Calcaneal–cuboid angle ( ; n ¼ 67) Mean SD Range AP talo–first metatarsal angle ( ; n ¼ 60) Mean SD Range Cuboid height (mm; n ¼ 68) Mean SD Range Calcaneal inclination angle ( ; n ¼ 68) Mean SD Range Talar declination angle ( ; n ¼ 66) Mean SD Range Lateral talar–first metatarsal angle ( ; n ¼ 60) Mean SD Range First metatarsal inclination angle ( ; n ¼ 62) Mean SD Range Fifth metatarsal inclination angle ( ; n ¼ 63) Mean SD Range Calcaneal–fifth metatarsal angle ( ; n ¼ 68) Mean SD Range Tibial–talar angle ( ; n ¼ 66) Mean SD Range Medial column height (mm; n ¼ 68) Mean SD Range Abbreviation: SD, standard deviation. Data in parentheses are percentages.
Value 68 60 56.74 8.57 37–77 34 (56.67) 26 (43.33) 29 (42.65) 39 (53.35) 45 13 10 28 44 17
(66.18) (19.12) (14.71) (41.18) (64.71) (25.0)
19.06 8.73 1.4–42 21.06 14.82 14.4–59.3 10.92 12.53 25.60 to 39.3 2.2 8.6 16.9 to 22.6 9.51 9.79 17.6 to 38.3 38.97 8.26 22–72.7 21.69 9.82 1.5–51.2 17.96 6.45 5–36 5.37 4.54 3.1 to 22 11.82 10.53 19 to 43.2 133.82 10.45 111.2–162.6 9.0 10.9 28.9 to 37.3
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
751
Fig. 3. Histogram of cuboid height showing a roughly parametrically shaped data distribution.
was 133.82 10.45 (range 111.2 to 162.6 ). The mean standard deviation first metatarsal inclination angle was 17.96 6.45 (range 5 to 36 ). The mean standard deviation fifth metatarsal inclination angle was 5.37 4.54 (range 3.1 to 22 ). The mean standard deviation lateral talo–first metatarsal angle was 21.69 9.82 (range 1.5 to 51.2 ); 15 feet (25.0%) had a lateral talo–first metatarsal angle of 27. The mean standard deviation calcaneal–fifth metatarsal angle was 11.82 10.53 (range 19 to 43.2 ). The mean standard deviation medial column height was 9.0 10.9 (range 28.9 to 37.3) mm. Depiction of the cuboid height on a histogram demonstrated a roughly parametric or normal distribution (Fig. 3). Fisher’s exact tests revealed no significant differences in the rates of plantar midfoot ulceration between those with a normal versus an abnormal cuboid height, using any of the thresholds (0.0, 2.0, 2.0, and 5.0 mm). Furthermore, the positive predictive value, negative predictive value,
sensitivity, and specificity for plantar midfoot ulceration on normality/abnormality were not high (Table 2). A negative cuboid height (<0.0 mm) was also not statistically significant when isolating only the 45 subjects with Sanders type 2 CN with (n ¼ 15) and without (n ¼ 26) plantar midfoot ulceration (78.95% versus 65.38%; p ¼ .5067). The results of the comparison of clinical and radiographic variables between subjects with a positive and negative cuboid height, using 0.0 mm as the threshold, are listed in Table 3. Table 3 also includes the Pearson correlation coefficients of the other radiographic parameters measured with the cuboid height. Subjects with negative cuboid height were more likely to have a Sanders/Frykberg type 2 deformity (76.2% of those with a negative cuboid height versus 50.0% of those with a positive cuboid height had a type 2 deformity; p ¼ .036), were less likely to have a Sanders/Frykberg type 3 deformity (9.5% of those with a negative cuboid height versus 34.6% of those with a positive cuboid height had a type 3 deformity; p ¼ .023), and were more likely
Table 2 Association of cuboid height with plantar midfoot ulceration (n ¼ 68 total) CH Threshold (mm)
p Value*
Plantar Midfoot Ulceration (%) PPV
NPV
Sensitivity
Specificity
<0.0 (n ¼ 42) <2.0 (n ¼ 51) <2.0 (n ¼ 38) <5.0 (n ¼ 29)
.453 .3939 1.00 .627
34.62 29.14 40.00 38.46
54.76 54.90 57.89 55.17
32.14 17.86 42.86 53.57
57.50 70.00 55.00 40.00
Abbreviations: CH, cuboid height; NPV, negative predictive value; PPV, positive predictive value. * Fisher’s exact test comparing those with and without plantar midfoot ulceration.
752
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Table 3 Comparison of positive versus negative cuboid height and clinical and radiographic parameters Variable
Sanders classification Type 2 Type 3 Both types 2 and 3 Lack of visible lateral tarsometatarsal joint space Plantar midfoot wound history Any lower extremity wound/forefoot amputation history Underwent reconstruction AP talocalcaneal angle ( ) Calcaneal–cuboid angle ( ) AP talo-first metatarsal angle ( ) Calcaneal inclination angle ( ) Talar declination angle ( ) Tibial–talar angle ( ) First metatarsal inclination angle ( ) Fifth metatarsal inclination angle ( ) Meary’s angle ( ) Meary’s angle 27 Calcaneal–fifth metatarsal angle ( ) Medial column height (mm)
CH (mm)
p Value
0.0 (n ¼ 26)
<0.0 (n ¼ 42)
13 (50.0) 9 (34.6) 4 (15.4) 7 (26.9) 9 (34.6) 17 (65.3) 4 (15.4) 18.41 25.39 8.95 15.08 36.65 133.96 19.78 5.37 17.72 4 (15.38) 17.85 11.54
32 (76.2) 4 (9.5) 6 (14.3) 30 (71.4) 19 (45.2) 27 (64.3) 13 (31.0) 19.46 8.75 18.32 13.03 12.05 13.88 6.06 8.22 40.38 8.99 133.73 9.43 16.91 6.63 5.37 5.07 23.92 9.72 13 (30.95) 8.09 8.74 7.44 12.28
8.83 16.63 9.69 9.68 6.41 12.16 5.80 3.63 8.87 10.52 7.58
.036* .023* 1.00 .005*,y .453 1.00 .249 .6356 .0565 .3376 < .001* .0749 .9310 .0814 .998 .014* .2486 < .001* .131
Pearson Correlation Coefficient (p Value)
0.172 0.254 0.154 0.591 0.387 0.160 0.245
(.161) (.038*) (.216) (< .000*) (< .001*) (p ¼ .199) (p ¼ .047*)
0.475 (< .000*) 0.655 (< .000*) 0.264 (.029*)
Abbreviations: AP, anteroposterior; CH, cuboid height. Data presented as n (%) or mean standard deviation. Categorical outcomes compared between groups using Fisher’s exact test; continuous outcomes compared between groups using nonpaired Student’s t test. * Statistically significant. y Odds ratio 6.79 (95% confidence interval 2.01 to 23.84).
to lack visualization of the lateral tarsometatarsal joint (71.4% of those with a negative cuboid height versus 26.9% of those with a positive cuboid height had an absence of visualization of the lateral tarsometatarsal joint; p ¼ .005). Subjects with a negative cuboid height had a greater frequency of plantar midfoot wounds; however, this difference was not statistically significant (45.2% versus 34.6%; p ¼ .453). Subjects with a negative cuboid height demonstrated a lower calcaneal inclination angle (6.06 versus 15.08 ; p < .001), greater Meary’s angle (23.92 versus 17.72 ; p ¼ .014), and lower calcaneal–fifth metatarsal angle (8.09 versus 17.85 ; p < .001). A statistically significant positive Pearson correlation coefficient was observed between the cuboid height and the calcaneal–cuboid angle (0.254; p ¼ .038), calcaneal inclination angle (0.591; p < .001), first metatarsal inclination angle (0.245; p ¼ .047), calcaneal–fifth metatarsal angle (0.655; p < .001), and medial column height (0.264; p ¼ .029). A statistically significant negative Pearson correlation coefficient was observed between the cuboid height and the talar declination angle (0.387; p < .001) and Meary’s angle (0.475; p < .001). Frequency scatter plots for all variables are shown in Fig. 4. The results of the comparison of clinical and radiographic variables of the subjects with and without a history of a plantar midfoot ulceration are listed in Table 4. We did not observe any statistically significant differences between groups with respect to the clinical or radiographic variables. Discussion The objective of the present retrospective, observational investigation was to examine the association between the cuboid height and the presence of ulceration in subjects with midfoot CN. Our results have shown that the cuboid height might not be a robust measure with respect to the presence or absence of midfoot plantar foot ulceration (1). We did not observe a significant difference in cuboid height between study participants with and without a plantar midfoot ulceration (2.77 versus 1.86 mm; p ¼ .671) nor a significant difference in the presence of plantar midfoot ulceration between participants with a positive and negative cuboid height (34.6% versus
45.2%; p ¼ .453). Furthermore, we observed relatively modest positive predictive values, negative predictive values, sensitivities, and specificities when considering cuboid height and plantar midfoot ulceration at multiple cuboid height thresholds (0.0, 2.0, 2.0, and 5.0 mm; Table 2). This was against our study hypothesis and did not match the observations of Wukich et al (1). Although we have continued to use this measurement in our clinical practice and consider it an important part of the radiographic evaluation of midfoot CN, we did not find it to be associated with the presence or absence of ulceration in our cohort. This is clinically important because the presence of midfoot ulceration and associated increased risk of infection in the setting of ulceration has been cited as a major contributor in the decision-making process for conservative versus surgical treatment of midfoot CN (1–13). Second, we also did not observe any statistically significant association between the presence of plantar midfoot ulceration and other commonly performed radiographic angular measurements of the foot and ankle. We believe this emphasizes that the clinical outcomes and the presence of plantar midfoot ulcerations in CN are related to other confounding clinical factors and aspects of lower extremity function, rather than just the structural deformity as assessed with plain film radiographs. However, perhaps our assessment of midfoot CN should be more complex than simply the position of a single bone or joint. As an example, the classification of Schon et al (4) for midfoot CN describes 4 different types involving the intercuneiform and cuneocuboid articulations. Third, we believe that the present investigation has provided unique information regarding the relationship of commonly performed radiographic angular measurements at various points in the midfoot CN process and, in particular, as it relates to cuboid height. We observed a significant negative association between the cuboid height and the presence of Sanders/Frykberg type 2 deformities (76.2% of those with a negative cuboid height versus 50.0% of those with a positive cuboid height had a type 2 deformity; p ¼ .036), those lacking visualization of the lateral tarsometatarsal joint (71.4% of those with negative cuboid height versus 26.9% of those with positive cuboid height had an absence of visualization of the lateral
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
753
Fig. 4. Frequency scatter plots of radiographic parameters graphed against the cuboid height. These frequency scatter plots with corresponding linear correlation lines demonstrate the positive and negative relationships between the cuboid height and the other measured radiographic parameters. (A) represents the frequency scatter plot demonstrating a negative relationship between the AP talocalcaneal angle and cuboid height. (B) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal-cuboid angle and cuboid height. (C) represents the frequency scatter plot demonstrating a negative relationship between the AP talo-first metatarsal angel and cuboid height. (D) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal inclination angle and cuboid height. (E) represents the frequency scatter plot demonstrating a negative relationship between the talar declination angle and cuboid height. (F) represents the frequency scatter plot demonstrating a negative relationship between the lateral talo-first metatarsal angle and cuboid height. (G) represents the frequency scatter plot demonstrating a positive relationship between the first metatarsal inclination angle and cuboid height. (H) represents the frequency scatter plot demonstrating a positive relationship between the calcaneal-fifth metatarsal angle and cuboid height. (I) represents the frequency scatter plot demonstrating a negative relationship between the tibial-talar angle and cuboid height. (J) represents the frequency scatter plot demonstrating a positive relationship between the medial column height and cuboid height.
tarsometatarsal joint; p ¼ .005), and those with lower calcaneal inclination angles (6.06 versus 15.08 ; p < .001). Patients lacking visualization of the lateral tarsometatarsal joint had 6.79 times the odds of having a negative cuboid height. We believe that this might provide evidence that cuboid height is related to forces not only derived from the rearfoot and calcaneocuboid joint but also retrograde forces from the forefoot. It might also indicate that the structure of the fourth and fifth metatarsal bases warrants evaluation in cuboid planning procedures. We further observed significant positive correlations between the cuboid height and the calcaneal–fifth metatarsal angle (0.655; p < .000), calcaneal inclination angle (0.591; p < .001), calcaneal– cuboid angle (0.254; p ¼ .038), medial column height (0.264; p ¼ .029), and first metatarsal inclination angle (0.245; p ¼ .047). Negative correlations were observed between Meary’s angle (0.475;
p < .001) and the talar declination angle (0.387; p < .001). This, somewhat unsurprisingly, showed a trend toward decreasing cuboid heights associated with increasing sagittal plane deformity involving both the medial and the lateral columns of the foot. Just as with any scientific investigation, critical readers are encouraged to review the study design and results and reach their own conclusions. Our report represents our conclusions from the specific results. Also, just as with any study, the present study had limitations. First, the data were collected from a limited number of subjects from a single institution; thus, these results might not be representative of a broader population sampling. When considering the subject cohort, we also did not specifically account for the severity of the deformity in our inclusion and exclusion criteria. This might represent an interesting avenue for future investigation. Second, all retrospective studies have inherent limitations. We were dependent on the diagnostic codes
754
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
Fig. 4. (continued).
A.J. Meyr, J.A. Sebag / The Journal of Foot & Ankle Surgery 56 (2017) 748–755
755
Table 4 Comparison of patients with and without plantar midfoot ulceration history stratified by clinical and radiographic parameters Variable
Sanders classification Type 2 Type 3 Both types 2 and 3 Lack of visible lateral tarsometatarsal joint space Underwent reconstruction AP calcaneal–talar angle ( ) Calcaneal–cuboid angle ( ) AP talo–first metatarsal angle ( ) CH (mm) CH <0.0 mm Calcaneal inclination angle ( ) Talar declination angle ( ) Tibial–talar angle ( ) First metatarsal inclination angle ( ) Fifth metatarsal inclination angle ( ) Meary’s angle ( ) Meary’s angle 27 Calcaneal–fifth metatarsal angle ( ) Medial column height (mm)
History of Plantar Midfoot Ulceration
p Value
No (n ¼ 40)
Yes (n ¼ 28)
26 (65.0) 10 (25.0) 4 (10.0) 18 (45.0) 8 (20.0) 18.86 8.30 21.74 13.52 11.06 11.55 1.86 8.06 23 (57.5) 9.10 8.29 39.26 8.75 132.80 8.82 18.47 6.28 4.62 3.26 21.50 10.75 12 (30.77) 12.32 8.59 8.23 10.70
19 (67.86) 3 (10.71) 6 (21.43) 19 (67.86) 9 (32.14) 19.33 9.46 20.07 16.78 10.72 14.06 2.77 9.47 19 (67.86) 10.08 11.74 38.52 7.59 135.39 12.60 17.21 6.73 6.49 5.87 22.00 8.36 4 (16.0) 11.11 12.95 10.12 11.17
1.00 .104 .297 .085 .272 .829 .655 .915 .671 .453 .688 .724 .328 .440 .103 .846 .243 .645 .484
Abbreviations: AP, anteroposterior; CH, cuboid height. Data presented as n (%) or mean standard deviation. Categorical outcomes compared between groups using Fisher’s exact test; continuous outcomes compared between groups using nonpaired Student’s t test.
for identification of the cohort and on physician documentation and data collection of the patient characteristics. Any inaccuracies within these data would have resulted in inaccuracy of our data analysis and interpretation. Third, our primary outcome measure and secondary outcome measures were primarily derived from the radiographic analysis, dependent on consistent patient positioning and angle measurement. Patient positioning is often difficult for patients with CN and might not have been as reliable as positioning for radiographic studies examining other lower extremity pathologic entities. One might also disagree with our specific statistical analysis and plan. We believed that a direct comparison of outcomes with the Fisher exact test and nonpaired Student t test would be of most benefit and interest for readers of the present initial study; however, future investigations might consider a more complex statistical plan such as analysis of variance. Finally, the cuboid height is a relatively understudied radiographic parameter, and we are not aware of any analysis of the inter- and intrarater reliability of this measurement in a group of patients with midfoot CN (6,7). The reliability of this measurement might also represent an interesting avenue for future investigations. In conclusion, the results of the present investigation have provided evidence that it is unlikely that a single radiographic parameter used in contemporary practice is associated with plantar midfoot ulceration in patients with midfoot CN. It is our hope that these data, which provide new information on the CN deformity as it relates to cuboid height, might lead to the development of such a measure or combination of measures that are predictive of ulceration in patients with midfoot CN.
References 1. Wukich DK, Raspovic KM, Hobizal KB, Rosario B. Radiographic analysis of diabetic midfoot Charcot neuroarthropathy with and without midfoot ulceration. Foot Ankle Int 35:1108–1115, 2014. 2. Bevan WP, Tomlinson MP. Radiographic measures as a predictor of ulcer formation in diabetic Charcot midfoot. Foot Ankle Int 29:568–573, 2008. 3. Hastings MK, Johnson JE, Strube MJ, Hildebolt CF, Bohnert KL, Prior FW, Sinacore DR. Progression of foot deformity in Charcot neuropathic osteoarthropathy. J Bone Joint Surg Am 95:1206–1214, 2013. 4. Schon LC, Weinfeld SB, Horton GA, Resch S. Radiographic and clinical classification of acquired midtarsus deformities. Foot Ankle Int 19:394–404, 1998. 5. Steel MW, Johnson KA, DeQitz MA, Ilstrup MS. Radiographic measurements of the normal adult foot. Foot Ankle 1:151–158, 1980. 6. Younger AS, Sawatzky B, Dryden P. Radiographic assessment of adult flatfoot. Foot Ankle Int 26:820–825, 2005. 7. Hastings MK, Sinacore DR, Mercer-Bolton N, McCormick JJ, Hidebolt CF, Prior FW, Johnson JE. Precision of foot alignment measures in Charcot arthropathy. Foot Ankle Int 32:867–872, 2011. 8. Rogers LC, Frykberg RG, Armstrong DG, Boulton AJ, Edmonds M, Van GH, Hartemann A, Game F, Jeffcoate W, Jirkovska A, Jude E, Morbach S, Morrison WB, Pinzur M, Pitocco D, Sanders L, Wukich DK, Uccioli L. The Charcot foot in diabetes. Diabetes Care 34:2123–2129, 2011. 9. Sanner WH. Foot segmental relationships and bone morphology. In: Foot and Ankle Radiology, pp. 272–302, edited by RA Christman, Churchill Livingstone, St. Louis, MO, 2003. 10. Pinzur M. Surgical versus accommodative treatment for Charcot arthropathy of the midfoot. Foot Ankle Int 25:545–549, 2004. 11. Catanzariti AR, Medicino R, Haverstock B. Ostectomy for diabetic neuroarthropathy involving the midfoot. J Foot Ankle Surg 39:291–300, 2000. 12. Wukich DK, Sung W. Charcot arthropathy of the foot and ankle: modern concepts and management review. J Diabetes Complications 23:409–426, 2009. 13. Smith WB, Moore CA. A proposed treatment algorithm for midfoot Charcot arthropathy. Foot Ankle Spec 5:60–64, 2012.