External skeletal fixation is a treatment modality commonly used in veterinary orthopedic surgery for orthopedic ailments including fractures, angular limb deformities, and joint instability. Soft tissue inflammation at the ISI is common, particularly when ESF is maintained for > 8 weeks, with incidence approaching 100% in veterinary medicine1 and ranging from 50% to 96% in human medicine.2–7 Many authors consider this inflammation to be a minor complication; however, in human medicine, it can sometimes lead to pain, lameness, septic osteoarthritis, and chronic osteomyelitis threatening the stability of the implant-bone interface.7–9 Changes at the ISI likely cause similar outcomes in dogs as in humans. In veterinary medicine, drainage after ESF is particularly common in the proximal aspect of the antebrachium and crus, compared with results after ESF in the middle and distal aspects of the antebrachium and crus.10,11
Although complications associated with ISI inflammation during ESF are described in human2–4,7–9,12 and veterinary1,13–16 medicine, validated scoring systems that describe the condition of that interface are lacking. Development of a validated scoring system for inflammation at the ISI is an essential step to enable clinicians to reach evidence-based conclusions about outcomes and complications in clinical studies involving ESF. Without such a validated scoring system, conclusions based on defensible assessment of the ISI, and therefore treatments aimed at reducing inflammation at the ISI, cannot be made. Objectives of the study reported here were to develop an inflammation scoring system for the ISI, evaluate intraobserver repeatability and interobserver agreement of ISI inflammation scores, and evaluate in vivo and photographic agreement of ISI inflammation scores. We hypothesized that ISI inflammation scores would have high intraobserver repeatability and high interobserver agreement (Cronbach α > 0.7).17 We also hypothesized that ISI scores would have high agreement (Cronbach α > 0.7) when in vivo and photographic scores were compared.
Materials and Methods
Sample
Photographs of the ISI of canine patients treated by use of ESF from November 2012 through November 2013 were used in the study. Photographs were contributed by the Veterinary Specialty Hospital of the Carolinas, Cary, NC; Dallas Veterinary Surgical Center, Dallas, TX; and Veterinary Hospital of North Carolina State University, Raleigh, NC. Assessment of the agreement between in vivo and photographic scores involved the use of canine patients treated with ESF performed by clinicians at North Carolina State University from November 2012 through November 2013. The evaluations coincided with an ongoing, unpublished prospective clinical study involving ESF. Photographs of the ISIs of these clinical patients were obtained concurrently at the time of the in vivo evaluations. The use of vertebrate animals in this study was approved by the Institutional Animal Care and Use Committee of North Carolina State University.
Photographs
Recommended photographic settings were macro mode, 2 to 5 cm from the ISI, natural ambient light on, flash off, automatic white balance, focus centered on the ISI, superfine jpeg compression, red eye correction off, image quality size at the highest setting, and > 3 megapixels.18 Additional instructions for contributors included that the entire area of inflammation be included in the photograph, ISI be in focus, a plastic transparent ruler parallel to the camera lens and red section of a color wheela be included in the photograph, and a blue surgical drape be used as the background of the photograph. All photographs of acceptable quality were considered for inclusion. Inclusion criteria for photographs were that the ISI was in focus, there was adequate brightness, and the red section of the color wheel and the plastic transparent ruler were included in the photograph. Exclusion criteria included lack of the entire area of inflammation in the photograph, lack of the red section of the color wheel or the plastic transparent ruler in the photograph, an ISI < 1 cm from a healing incision or wound, and inclusion of another photograph of the same ISI at the same time point (ie, duplicate photographs). One author (MBML) determined which photographs were eligible for inclusion and prepared all photographs; that author did not score any of the photographs.
Each photograph was manipulated with image–processing softwareb such that the ruler in each photograph provided a 1-cm scale marker, the red section of the color wheel was cropped and moved close to the ISI without obscuring the interface, the image was rotated so that the implant was oriented vertically, and the image was cropped so that the entire area of inflammation was included with minimal extraneous details. Each photograph was printed (10 × 15-cm matte photographc).
Inflammation scoring system
A standard method for the development of an instrument to assess a subjective state was used to develop an ISI inflammation score on the basis of nonvalidated ISI scoring systems used previously for humans.19–26 Six metrics were defined: erythema, discharge amount, discharge type, swelling, hair loss or lack of hair regrowth, and granulation tissue. A scale from 1 to 10 (higher scores indicated greater severity of each of the assessed metrics of inflammation) was selected on the basis of ease of use, compared with ease of use for a visual analogue scale.27,28 Severity of discharge type was graded as serous, serosanguineous, hemorrhagic, or purulent. Hair loss or lack of hair regrowth was scored on the basis of the amount of hair at the ISI, compared with the amount of hair outside the ISI.
A photographic chart of the various severities of each metric for the scale scores was created by use of photographs collected prior to initiation of the study. Each chart consisted of 6 photographs for each of the 6 metrics; photographs depicted the range for the scale (from 1 to 10) for a specific metric (Figure 1). The photographic chart depicting the scale for each of the 6 metrics was provided to each rater for use as a reference during each scoring session.
Photograph reference chart used to grade ISIs in canine patients. The chart includes 6 photographs for each of the 6 metrics associated with ESF. Photographs show increasing severity of inflammation from left (least severe) to right (most severe). Notice the 10-point scale at the bottom of the figure.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Photograph reference chart used to grade ISIs in canine patients. The chart includes 6 photographs for each of the 6 metrics associated with ESF. Photographs show increasing severity of inflammation from left (least severe) to right (most severe). Notice the 10-point scale at the bottom of the figure.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Photograph reference chart used to grade ISIs in canine patients. The chart includes 6 photographs for each of the 6 metrics associated with ESF. Photographs show increasing severity of inflammation from left (least severe) to right (most severe). Notice the 10-point scale at the bottom of the figure.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Intraobserver repeatability and interobserver agreement
Four board-certified veterinary surgeons (DJML, SCR, BDXL, and BJT) each scored the set of photographs 5 times to assess intraobserver repeatability. Repeated assessments were made a minimum of 14 days apart. Raters were not aware of their previous scores or of the scores of the other raters. Scores of the first evaluations for the 4 raters were compared to assess interobserver agreement.
Agreement of in vivo and photographic assessment
Two board-certified veterinary surgeons (DJML and SCR) evaluated the ISI of canine patients treated by use of ESF performed by clinicians at North Carolina State University between November 2012 and November 2013. Photographs of the same ISIs were concurrently obtainedd at the time of the in vivo assessment. Photographs were subsequently scored a minimum of 14 days after the in vivo assessment. Raters were not aware of the ISI score for the in vivo assessment or of the scores of other raters.
Statistical analysis
All statistical analyses were performed with a computer statistical package.e Power analyses were performed a priori to determine minimum sample size for interobserver agreement, intraobserver repeatability, and agreement between in vivo and photographic scores. These analyses (power = 0.8) revealed that a minimum of 16 photographs scored by 4 raters were needed for interobserver agreement, 12 photographs scored 5 times by each rater were needed for intraobserver repeatability, and 24 photographs and dogs scored by 2 raters were needed for agreement between photographic and in vivo assessments.
Intraobserver repeatability was assessed by use of the Cronbach α for each rater and each metric; an ICC (a ratio of variances used to assess similarities between measurements) for each metric was calculated by combining data for all raters. Interobserver agreement was evaluated by use of the Cronbach α (cutoff value, 0.7; significance set at P < 0.05).29 Agreement between in vivo and photographic scores was evaluated by use of both a linear model and the Cronbach α for each rater and each metric. The ICC was used to evaluate the overall agreement across multiple measurements and multiple raters. Correlations between metrics were evaluated by use of the Pearson correlation coefficient between medians of each metric across raters. Principal component analysis was used to identify linearly uncorrelated metrics.
Results
During the study period, 371 photographs of ISIs were submitted for possible inclusion. Often, multiple photographs of the same ISI were submitted. Of the 371 photographs, 329 were excluded from the study. Thus, 42 photographs (4 of wire-skin interfaces and 38 of pin-skin interfaces) of the ISI of dogs with ESF met inclusion criteria. These 42 photographs were of the ISI in 18 dogs. Median duration from surgery for ESF until photographs were obtained was 63 days (range, 11 to 117 days). For each dog, there was a median of 1.5 (range, 1 to 6) photographs of an ISI at various time points. Median scores for all 4 raters combined (scale, 1 to 10) were 4 for erythema, 4 for drainage amount, 5 for drainage type, 3 for swelling, 5 for no hair or lack of hair regrowth, and 4 for granulation tissue (Figure 2). Intraobserver Cronbach α for all raters and for all metrics ranged from 0.849 to 0.969 (P < 0.001; Table 1). Overall ICC for all metrics ranged from 0.922 to 0.975. Interobserver ICC ranged from 0.835 to 0.943 (P ≤ 0.001) for all metrics for pairwise comparisons (Table 2).
Photographs of representative ISIs in canine patients with low and high scores, respectively, for erythema (A [median score, 1.5] and B [median score, 8]), drainage amount (C [median score, 3] and D [median score, 8]), drainage type (E [median score, 2] and F [median score, 9]), swelling (G [median score, 2] and H [median score, 5]), no hair or lack of hair regrowth (I [median score, 1] and J [median score, 7.5]), and granulation tissue (K [median score, 1] and L [median score, 7]). Scores represent the median value for 4 raters. Each photograph contains an ISI with the area of inflammation associated with that interface. Within each photograph, the red section of a color wheel is visible (lower left). The scale marker (horizontal bar in lower right corner) differs in size among panels because of differences in magnification among the photographs; however, each bar represents 1 cm.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Photographs of representative ISIs in canine patients with low and high scores, respectively, for erythema (A [median score, 1.5] and B [median score, 8]), drainage amount (C [median score, 3] and D [median score, 8]), drainage type (E [median score, 2] and F [median score, 9]), swelling (G [median score, 2] and H [median score, 5]), no hair or lack of hair regrowth (I [median score, 1] and J [median score, 7.5]), and granulation tissue (K [median score, 1] and L [median score, 7]). Scores represent the median value for 4 raters. Each photograph contains an ISI with the area of inflammation associated with that interface. Within each photograph, the red section of a color wheel is visible (lower left). The scale marker (horizontal bar in lower right corner) differs in size among panels because of differences in magnification among the photographs; however, each bar represents 1 cm.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Photographs of representative ISIs in canine patients with low and high scores, respectively, for erythema (A [median score, 1.5] and B [median score, 8]), drainage amount (C [median score, 3] and D [median score, 8]), drainage type (E [median score, 2] and F [median score, 9]), swelling (G [median score, 2] and H [median score, 5]), no hair or lack of hair regrowth (I [median score, 1] and J [median score, 7.5]), and granulation tissue (K [median score, 1] and L [median score, 7]). Scores represent the median value for 4 raters. Each photograph contains an ISI with the area of inflammation associated with that interface. Within each photograph, the red section of a color wheel is visible (lower left). The scale marker (horizontal bar in lower right corner) differs in size among panels because of differences in magnification among the photographs; however, each bar represents 1 cm.
Citation: American Journal of Veterinary Research 76, 11; 10.2460/ajvr.76.11.931
Intraobserver repeatability (Cronbach α) for inflammation scores obtained by assessment of 42 photographs of ISIs in 18 dogs treated by use of ESF.
| Cronbach α† | ||||||
|---|---|---|---|---|---|---|
| Metric | Inflammation score (Median [range])* | Rater 1 | Rater 2 | Rater 3 | Rater 4 | Overall ICC‡ |
| Erythema | 4 (1–10) | 0.931 | 0.949 | 0.951 | 0.938 | 0.940 |
| Drainage amount | 4 (1–10) | 0.936 | 0.929 | 0.936 | 0.916 | 0.954 |
| Drainage type | 5 (1–10) | 0.969 | 0.849 | 0.932 | 0.926 | 0.948 |
| Swelling | 3 (1–10) | 0.871 | 0.931 | 0.923 | 0.921 | 0.938 |
| No hair or lack of hair regrowth | 5 (1–10) | 0.891 | 0.934 | 0.930 | 0.933 | 0.922 |
| Granulation tissue | 4 (1–10) | 0.956 | 0.948 | 0.937 | 0.900 | 0.975 |
Results represent repeatability of 5 assessments of the same photographs performed by each of 4 raters at intervals of ≥ 14 days.
Each metric was scored on a scale from 1 to 10
Values for the Cronbach α were significantly (P < 0.05) > 0.7
Values for ICC were significantly (P < 0.05) > 0.7.
Interobserver agreement (Cronbach α) for inflammation scores obtained by assessment of 42 photographs of ISIs in 18 dogs treated by use of ESF.
| Metric | Inflammation score (median [range])* | Cronbach α† | Overall ICC‡ | ||||
|---|---|---|---|---|---|---|---|
| Erythema | Rater | 0.864 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 3 (1–6) | 1 | 1.000 | 0.647 | 0.656 | 0.729 | ||
| 6 (1–10) | 2 | 0.647 | 1.000 | 0.785 | 0.569 | ||
| 4.5 (1–10) | 3 | 0.656 | 0.785 | 1.000 | 0.682 | ||
| 7 (1–10) | 4 | 0.729 | 0.569 | 0.682 | 1.000 | ||
| Drainage amount | Rater | 0.907 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 4 (1–8) | 1 | 1.000 | 0.560 | 0.692 | 0.659 | ||
| 5.5 (1–10) | 2 | 0.560 | 1.000 | 0.850 | 0.709 | ||
| 6 (1–10) | 3 | 0.692 | 0.850 | 1.000 | 0.786 | ||
| 6 (1–10) | 4 | 0.659 | 0.709 | 0.786 | 1.000 | ||
| Drainage type | Rater | 0.872 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 4.5 (1–9) | 1 | 1.000 | 0.558 | 0.623 | 0.542 | ||
| 5 (1–9) | 2 | 0.558 | 1.000 | 0.725 | 0.650 | ||
| 5.5 (1–9) | 3 | 0.623 | 0.725 | 1.000 | 0.692 | ||
| 6 (1–10) | 4 | 0.542 | 0.650 | 0.692 | 1.000 | ||
| Swelling | Rater | 0.881 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 2 (1–7) | 1 | 1.000 | 0.671 | 0.642 | 0.662 | ||
| 4 (1–10) | 2 | 0.671 | 1.000 | 0.731 | 0.472 | ||
| 4 (1–9) | 3 | 0.642 | 0.731 | 1.000 | 0.715 | ||
| 4 (2–7) | 4 | 0.662 | 0.472 | 0.715 | 1.000 | ||
| No hair or lack of hair regrowth | Rater | 0.835 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 4.5 (1–6) | 1 | 1.000 | 0.393 | 0.423 | 0.528 | ||
| 5 (1–8) | 2 | 0.393 | 1.000 | 0.787 | 0.615 | ||
| 5 (1–8) | 3 | 0.423 | 0.787 | 1.000 | 0.603 | ||
| 5 (1–9) | 4 | 0.528 | 0.615 | 0.603 | 1.000 | ||
| Granulation tissue | Rater | 0.943 | |||||
| Rater | 1 | 2 | 3 | 4 | |||
| 3 (1–9) | 1 | 1.000 | 0.873 | 0.747 | 0.770 | ||
| 5 (1–9) | 2 | 0.873 | 1.000 | 0.879 | 0.789 | ||
| 4 (1–9) | 3 | 0.747 | 0.879 | 1.000 | 0.774 | ||
| 5 (2–10) | 4 | 0.770 | 0.789 | 0.774 | 1.000 |
See Table 1 for key.
For assessment of agreement between in vivo and photographic scores, the Cronbach α for both raters ranged from 0.614 to 0.938 for individual pairwise comparisons (Table 3). Overall, the Cronbach α ranged from 0.725 to 0.932, with the Cronbach α > 0.7 for erythema, drainage amount, drainage type, swelling, and granulation tissue. Mean score for each metric differed significantly between the in vivo and photographic assessment, with the photographic score significantly (P < 0.001) higher than the in vivo score for each metric.
Agreement (Cronbach α) of mean ± SD inflammation scores for in vivo and photographic assessment of 27 ISIs in 6 dogs treated by use of ESF.
| Inflammation score* | Cronbach α | ||||
|---|---|---|---|---|---|
| Metric | In vivo | Photographic | Rater 1 | Rater 2 | Overall score |
| Erythema | 3.1 ± 1.5 | 4.1 ± 2.0‡ | 0.898† | 0.614 | 0.855† |
| Drainage amount | 3.2 ± 2.1 | 4.4 ± 2.0‡ | 0.918† | 0.859† | 0.927† |
| Drainage type | 3.6 ± 2.5 | 4.3 ± 2.0‡ | 0.938† | 0.817 | 0.932† |
| Swelling | 2.9 ± 1.3 | 3.9 ± 1.2‡ | 0.695 | 0.720 | 0.846† |
| No hair or lack of hair regrowth | 3.8 ± 1.4 | 4.9 ± 1.5‡ | 0.726 | 0.654 | 0.725 |
| Granulation tissue | 2.9 ± 1.7 | 4.2 ± 1.8‡ | 0.860† | 0.719 | 0.904† |
Value differs significantly (P < 0.001) from the value for the in vivo assessment.
See Table 1 for remainder of key.
Pearson correlation coefficients between median scores for each metric across all raters ranged from 0.221 (P = 0.160) for the correlation between drainage type and no hair or lack of hair regrowth to 0.923 (P < 0.001) for the correlation between drainage type and drainage amount (Table 4). Erythema, swelling, and granulation tissue were correlated with all other metrics. Drainage amount and drainage type were not correlated with no hair or lack of hair regrowth. Principal component analysis did not identify linearly uncorrelated variables.
Correlation between median inflammation scores obtained by 4 raters for assessment of 42 photographs of ISIs in 18 dogs treated by use of ESF.
| Metric | ||||||
|---|---|---|---|---|---|---|
| Metric | Erythema | Drainage amount | Drainage type | Swelling | No hair or lack of hair regrowth | Granulation tissue |
| Erythema | 1.000 (—) | 0.511 (< 0.001) | 0.506 (< 0.001) | 0.575 (< 0.001) | 0.537 (< 0.001) | 0.384 (0.012) |
| Drainage amount | 0.511 (< 0.001) | 1.000 (—) | 0.923 (< 0.001) | 0.536 (< 0.001) | 0.227 (0.148) | 0.369 (0.016) |
| Drainage type | 0.506 (< 0.001) | 0.923 (< 0.001) | 1.000 (—) | 0.499 (< 0.001) | 0.221 (0.160) | 0.311 (0.045) |
| Swelling | 0.575 (< 0.001) | 0.536 (< 0.001) | 0.499 (< 0.001) | 1.000 (—) | 0.476 (0.001) | 0.794 (< 0.001) |
| No hair or lack of hair regrowth | 0.537 (< 0.001) | 0.227 (0.148) | 0.221 (0.160) | 0.476 (0.001) | 1.000 (—) | 0.388 (0.011) |
| Granulation tissue | 0.384 (0.012) | 0.369 (0.016) | 0.311 (0.045) | 0.794 (< 0.001) | 0.388 (0.011) | 1.000 (—) |
Values reported are Pearson correlation coefficient (P value).
— = Not applicable.
Discussion
The development of a validated scoring system of inflammation at an ISI is an essential step for allowing clinicians to reach evidence-based conclusions about outcomes and complications in clinical studies involving ESF for factors ranging from care of the pin site to coatings of implants. Several scoring systems of the ISI have been used for clinical studies19–26 in human medicine. The primary limitations of those scoring systems are a lack of validation and inconsistency in scoring systems among studies, which prevents meaningful comparisons. The lack of a validated scoring system weakens the conclusions of previous clinical studies in which investigators evaluated the skin response to external fixators.19–26 To our knowledge, no scoring system for inflammation at the ISI has been fully validated for human or veterinary medicine. Full validation involves face validity, intraobserver repeatability, interobserver agreement, and criterion validation. The study reported here was the first in which aspects of validity of an ISI scoring system were tested; however, criterion validity was not tested.
The scoring system investigated in the present study relied on the use of a scale (from 1 to 10) for each metric of inflammation. A reference photographic chart that illustrated these scales was used during scoring by each rater. One benefit of this scale was that the score was solely based on skin surface metrics. Because the scale did not rely on imaging modalities, it was affordable and could be repeated frequently. Although the scale did not rely on the use of other diagnostic tests, correlations with results of those tests could be made in the future. Also, the scale allowed gradations for each metric, which could be beneficial in clinical studies to enable investigators to detect subtle differences among study groups.
A potential limitation of the ISI scoring system was the lack of an objective measure for determination of validity, which is an inherent problem in evaluation of an interface. Quantitative microbial cultures of implants have been used as an objective measure of infection,30 although this does not take into account other factors that may increase inflammation at the interface (namely, thickness of soft tissues or mobility of the skin or deeper tissues relative to the implant). Future investigations could evaluate the ISI scoring system against results for objective histologic or other measures of inflammation (criterion validity testing).
An inherent difficulty of prospective clinical studies involving ESF is availability of the same clinician at each patient visit. A potential solution is to photograph the ISI at each visit and subsequently score the photographs. The present study revealed that the ISI scoring system for ESF had high intraobserver repeatability and interobserver agreement when raters evaluated photographs of an ISI. Medians differed among raters for each metric, and on the basis of lower median values, some raters were more conservative in their scoring. Cronbach α analyses control for these differences, which indicated that the scale can be reliably used by multiple observers at multiple evaluations. Therefore, the ISI scale may be used for clinical studies to begin evidence-based investigation of ISI complications.
With regard to results of photographic and in vivo assessments, the present study identified moderate agreement, with mean photographic scores being higher than in vivo scores. These differences were numerically small and ranged from 0.7 to 1.4 for the various metrics. The difference in mean scores indicated that photographic scores cannot be considered equivalent to in vivo scores when assessing an ISI. This is potentially attributable to the fact that photographs enlarge the ISI, which magnifies the images and enhances detection of the metrics of inflammation being evaluated. Subjectively, it was more challenging to evaluate some metrics versus others via photographs. This was likely the cause that led to lower values for the Cronbach α. No hair or lack of hair regrowth had the lowest agreement between in vivo and photographic scores, with a Cronbach α < 0.7. Photographic evaluation of swelling was also challenging because a 3-D metric was converted into a 2-D representation. Pending a larger study aimed at identifying the causes of these scoring differences, clinical studies should rely on photographic assessment or in vivo assessment, but the 2 assessment methods should not be used interchangeably.
A secondary goal of the present study was to determine correlations among metrics of inflammation at an ISI. Overall, all metrics had moderate positive correlates with each other. Principal component analysis did not identify key metrics that could be used to develop a simplified ISI assessment method. The 2 metrics with the highest correlation were drainage amount and drainage type. Thus, increased drainage amount was correlated with increased purulent discharge from an ISI; both of these metrics were indicative of increases in inflammation. A second pair of highly correlated metrics was granulation tissue and swelling, both of which were 3-D enlargements of tissue around the implant. Particularly in chronic situations, swelling may be granulation tissue covered with skin. Therefore, it is likely that as swelling increases, there is a greater chance of granulation tissue being visible in the wound area around the pin. Another potential cause for the high correlation between granulation tissue and swelling is that granulation tissue and swelling may be related to inflammation at the interface secondary to movement of the skin and deeper tissues relative to the implant, as opposed to inflammation secondary to infection; however, this was not investigated further in the present study. Drainage amount and drainage type were not significantly correlated with no hair or lack of hair regrowth. Difference in skin care protocols among the various institutions may have influenced these findings. The influence of skin care protocols on drainage and hair regrowth was not evaluated in this study.
The present study had several limitations. Intraobserver repeatability and interobserver agreement for in vivo assessments were not investigated because of the logistic challenge of having all 4 raters available for evaluation of clinical patients. Also, for investigation of intraobserver repeatability, making raters unaware of previous assessments would be inherently impossible in a clinical scenario of a single ISI at a single time point. Although it was not proven in this study, it is a logical extrapolation that the repeatability of in vivo assessments is equivalent to the repeatability of photographic assessments because the processes are extremely similar. A second limitation, which was mentioned previously, is that we did not test criterion validity by use of a criterion-referenced standard objective assessment method for the ISI. A third limitation was that we did not investigate the ease of use of the scale, particularly the amount of time necessary to score an ISI or obtain photographs, which would be important in clinical applications. Anecdotally, and as anticipated with all new processes, the investigators were able to use the scale more rapidly during the progression of the study because of increasing familiarity. Finally, the photographic reference chart used by the raters during scoring did not contain a reference for the red section of the color wheel or for the 1-cm scale marker because the reference chart was created before development of the photographic standardization used in the study, and this potentially could have increased variability in the ISI scores.
Several areas exist for further studies. One area is investigation into the ISI scale, such as simplifying the scale to increase ease of use, evaluating the amount of time required for use of the scale, and correlating scores of metrics of inflammation with quantitative or qualitative results for microbial cultures. A second area is use of the scale to determine the prevalence of ESF complications, the manner by which inflammation at an ISI relates to complications at the implant-bone interface, and strategies to decrease inflammation in clinical situations. Additionally, this scale may have applicability in the evaluation of other ISIs, such as transdermal osseointegrated prosthetic implants31 or long-term feeding tubes.32
Results of the present study supported the hypothesis that the ISI interface scoring system described here had high intraobserver repeatability and high interobserver agreement. However, the hypothesis that there would be high agreement between scores for in vivo and photographic assessments was not supported. Photographic scores cannot be considered equivalent to in vivo scores. Therefore, the authors concluded that this scoring system should be valid for use in clinical studies that involve the use of in vivo assessments or photographic assessments but that do not combine these assessments.
Acknowledgments
The authors thank Dr. Robert Radasch for assistance with recruitment of cases.
ABBREVIATIONS
| ESF | External skeletal fixation |
| ICC | Intraclass correlation coefficient |
| ISI | Implant-skin interface |
Footnotes
Pocket color wheel, The Color Wheel Co, Philomath, Ore.
Adobe Photoshop, version 14.1.2, Adobe Systems Inc, San Jose, Calif.
Shutterfly, Redwood City, Calif.
Canon PowerShot ELPH 320 HS, Canon USA, Melville, NY.
SAS, version 9.4, SAS Institute Inc, Cary, NC.
References
1. Anderson GM, Lewis DD, Radasch RM, et al. Circular external skeletal fixation stabilization of antebrachial and crural fractures in 25 dogs. J Am Anim Hosp Assoc 2003;39:479–498.
2. Antoci V, Ono CM, Antoci V, et al. Pin-tract infection during limb lengthening using external fixation. Am J Orthop 2008;37:E150–E154.
3. Mahan J, Seligson D, Henry SL, et al. Factors in pin tract infections. Orthopedics 1991;14:305–308.
4. Marsh JL, Mahoney CR, Steinbronn D. External fixation of open humerus fractures. Iowa Orthop J 1999;19:35–42.
5. Keating JF, Gardner E, Leach WJ, et al. Management of tibial fractures with the orthofix dynamic external fixator. J R Coll Surg Edinb 1991;36:272–277.
6. Coester LM, Nepola JV, Allen J, et al. The effects of silver coated external fixation pins. Iowa Orthop J 2006;26:48–53.
7. Parameswaran AD, Roberts CS, Seligson D, et al. Pin tract infection with contemporary external fixation: how much of a problem? J Orthop Trauma 2003;17:503–507.
8. Green SA, Ripley MJ. Chronic osteomyelitis in pin tracks. J Bone Joint Surg Am 1984;66:1092–1098.
9. Moroni A, Vannini F, Mosca M, et al. State of the art review: techniques to avoid pin loosening and infection in external fixation. J Orthop Trauma 2002;16:189–195.
10. Fox SM, Bray JC, Guerin SR, et al. Antebrachial deformities in the dog: treatment with external fixation. J Small Anim Pract 1995;36:315–320.
11. Paley D. Problems, obstacles, and complications of limb lengthening by the Ilizarov technique. Clin Orthop Relat Res 1990;250:81–104.
12. Ring D, Bruinsma WE, Jupiter JB. Complications of hinged external fixation compared with cross-pinning of the elbow for acute and subacute instability. Clin Orthop Relat Res 2014;472:2044–2048.
13. Harari J. Complications of external skeletal fixation. Vet Clin North Am Small Anim Pract 1992;22:99–107.
14. Johnson AL, Kneller SK, Weigel RM. Radial and tibial fracture repair with external skeletal fixation: effects of fracture type, reduction, and complications on healing. Vet Surg 1989;18:367–372.
15. Knudsen CS, Arthurs GI, Hayes GM, et al. Long bone fracture as a complication following external skeletal fixation: 11 cases. J Small Anim Pract 2012;53:687–692.
16. Marcellin-Little DJ, Ferretti A, Roe SC, et al. Hinged Ilizarov external fixation for correction of antebrachial deformities. Vet Surg 1998;27:231–245.
17. Bland JM, Altman DG. Statistics notes: Cronbach's alpha. Br Med J (Clin Res Ed) 1997;314:572.
18. Meijer PP, Karlsson J, LaPrade RF, et al. A guideline to medical photography: a perspective on digital photography in an orthopaedic setting. Knee Surg Sports Traumatol Arthrosc 2012;20:2606–2611.
19. Cavusoglu AT, Er MS, Inal S, et al. Pin site care during circular external fixation using two different protocols. J Orthop Trauma 2009;23:724–730.
20. Chan CK, Saw A, Kwan MK, et al. Diluted povidone-iodine versus saline for dressing metal-skin interfaces in external fixation. J Orthop Surg (Hong Kong) 2009;17:19–22.
21. Camathias C, Valderrabano V, Oberli H. Routine pin tract care in external fixation is unnecessary: a randomised, prospective, blinded controlled study. Injury 2012;43:1969–1973.
22. Camilo AM, Bongiovanni JC. Evaluation of effectiveness of 10% polyvinylpyrrolidone-iodine solution against infections in wire and pin holes for Ilizarov external fixators. Sao Paulo Med J 2005;123:58–61.
23. Egol KA, Paksima N, Puopolo S, et al. Treatment of external fixation pins about the wrist: a prospective, randomized trial. J Bone Joint Surg Am 2006;88:349–354.
24. Patterson MM. Multicenter pin care study. Orthop Nurs 2005;24:349–360.
25. Lee CK, Chua YP, Saw A. Antimicrobial gauze as a dressing reduces pin site infection: a randomized controlled trial. Clin Orthop Relat Res 2012;470:610–615.
26. Davies R, Holt N, Nayagam S. The care of pin sites with external fixation. J Bone Joint Surg Br 2005;87:716–719.
27. van Laerhoven H, Zaag-Loonen HJ, Derkx B. A comparison of Likert scale and visual analogue scales as response options in children's questionnaires. Acta Paediatr 2004;93:830–835.
28. Guyatt GH, Townsend M, Berman LB, et al. A comparison of Likert and visual analogue scales for measuring change in function. J Chronic Dis 1987;40:1129–1133.
29. Iacobucci D, Duhachek A. Advancing alpha: measuring reliability with confidence. J Consum Psychol 2003;13:478–487.
30. Holt J, Hertzberg B, Weinhold P, et al. Decreasing bacterial colonization of external fixation pins through nitric oxide release coatings. J Orthop Trauma 2011;25:432–437.
31. Brånemark R, Berlin Ö, Hagberg K, et al. A novel osseointegrated percutaneous prosthetic system for the treatment of patients with transfemoral amputation. A prospective study of 51 patients. Bone Joint J 2014;96:106–113.
32. Zopf Y, Konturek P, Nuernberger A, et al. Local infection after placement of percutaneous endoscopic gastrostomy tubes: a prospective study evaluating risk factors. Can J Gastroenterol 2008;22:987–991.
