Abstract
Background
Distraction osteogenesis (DO) enables gradual bone elongation through controlled osteotomy and distraction. Existing radiographic and index-based assessments face limitations. Callus pattern classification offers predictive value, but its early applicability across age groups and gap lengths remains unclear.
Methods
We retrospectively reviewed 129 patients (155 bone segments) who underwent lower limb DO between 2003 and 2023. Patients had a mean age of 37.3 years (5.6–77). The mean distraction length was 53.2 mm (11–174). Radiographs were graded monthly using modified Li classification, depicting two callus consolidation routes: Route 1 (no presence of type 5/8) and Route 2 (presence of type 5/8). Patients were stratified by age (younger: < 25 years; middle: 25–49 years; older: 
50 years) and distraction length (shorter: < 5 cm; longer: 
5 cm). Outcome measures included parameters such as distraction-consolidation time (DCT), healing index (HI), etc., compared by Mann–Whitney test and analysis of variance. Trends of radiographic progress were analyzed by Loess regression.
Results
Route 1 demonstrated favorable healing with significantly lower DCT (300.4 vs 381.6 days; p = 0.002) and HI (70.1 vs 89.1 days/cm; p = 0.021). Younger age patients showed negligible differences between two routes. However, middle and older age groups showed delayed consolidation in Route 2, particularly from the high-intermediate callus density grades, Li classification type 6 and 7, onward (p < 0.05). Long distraction gaps further amplified route-based divergence in the trajectory of consolidation.
Conclusion
Evolution of callus patterns closely correlate with outcomes of DO. Negative patterns (types 5/8) are associated with delayed healing, particularly in older patients and those with longer distraction lengths. These findings support their use as a timely marker for risk stratification and suggest heightened surveillance or protocol adjustment, such as slowdown of distraction, cyclic distraction-compression dynamization, or additional fixation methods, to optimize outcomes.
Level of evidence
Level III, retrospective cohort study.
Keywords: Distraction osteogenesis, Callus classification, Radiographic trajectory
Introduction
Distraction osteogenesis (DO) involves controlled osteotomy and application of distraction devices, external or internal, to achieve stable fixation and incremental elongation, promoting continuous callus regeneration until consolidation [1–4].
Evaluation of DO outcomes traditionally relies on various quantitative indices, including distraction consolidation time (DCT), healing index (HI), distraction index (DI), and external fixator index (EI). Each index offers specific insights on bone healing process [3, 5]. However, these methods have limitations such as their dependence on the completion of lengthy follow-up periods, variability in radiographic interpretations, and limited real-time assessment capability, potentially delaying clinical interventions when necessary. Various classifications based on radiographic features of traction callus have been developed to evaluate bone healing during DO either with shape, distribution, density or callus diameter [6–11]. A classification proposed by Li et al. [8], which categorizes 10 types of callus patterns, provides a reliable and reproducible method to evaluate callus consolidation and predict final outcome.
Several factors on DO outcomes, notably patient age, osteotomy location, distraction speed, and length of distraction had been investigated [12–15]. Age particularly stands out as a significant determinant, with older individuals generally experiencing slower consolidation, reduced regeneration capabilities, and higher complication rates compared to younger patients. However, comprehensive studies covering all age ranges simultaneously are relatively sparse [16], with most research focusing on specific age groups like children or elderly populations independently [17–19].
In this study, we focus on (1) analyzing traction callus patterns in different stages for their predictive value on final bony outcomes and (2) evaluating their strength on different population regarding age of patient and length of gap.
Patients and methods
Patients who underwent lower limb distraction osteogenesis at the authors’ institution between 2003 and 2023 were enrolled in this retrospective study approved by the ethics committee and institutional review board. Patient data were obtained from the departmental registry, medical records and radiographs. Patients who were lost to follow-up before clinical outcomes could be assessed were excluded. Additional exclusion criteria included those who experienced treatment interruption due to amputation from other causes, trauma, or poor-controlled systemic disease during distraction osteogenesis, as well as those with poor compliance, profound slowdown, or underwent the accordion maneuver that resulted in deviations from the standard distraction rate. Systemic diseases that were excluded consist of osteoporosis with Dual Energy X-ray Absorptiometry (DEXA) T-score 
-1, metabolic bone disease, rheumatoid arthritis with flare up or need of steroid use, and diabetes mellitus with glycated hemoglobin (HbA1c) 
7.0%. (Fig. 1).
Fig. 1.
Flowchart of the study design. PAOD, peripheral arterial occlusive disease; DM, diabetes mellitus
Smoking cessation was strictly required before treatment. In infected cases, radical debridement with segmental sequestrectomy was performed to eradicate infection, and an external traction device was used to maintain alignment and stability. The osteotomy for distraction was carried out after infection control was confirmed. In non-infected cases, the osteotomy was performed with external fixation surgery. Skin incision length is 1.5–2 cm for tibia and 2.5–4 cm for femur. The periosteum at the osteotomy site was carefully dissected and protected with small Hoffman retractors. A low-speed oscillating saw with narrow blades is used to cut the bone. Each sawing lasts no longer than 5 s with continuous saline irrigation for cooling. The cut is made in a rhombus path until cortices are divided circumferentially. The endosteal marrow is preserved as much as possible. A narrow osteotome is used to widen the gap, and osteotomy completeness is confirmed under C-arm.
Lengthening was routinely started 7 days after osteotomy. The distraction rate was default as 1 mm per day with four quarter turns. It may vary between 0.5 and 1.0 mm a day by doctor’s comment, depending on the patient’s tolerance of pain, location of implantation, and the adjacent joints range of motion. Patients those who had slowdown over 2 weeks were excluded from this study. Instruction regarding device operation and pin-tract care were given and confirmed by teach-back. Weight bearing as tolerable were encouraged for every patient. Visits of physical therapist and monthly radiograph follow-up were arranged.
Outcome measurement
Progress of consolidation of traction callus were classified based on a system built by Li et al. [8]. The classification was originally divided into 10 types by different densities (low, intermediate, normal) and patterns (sparse, homogenous, heterogenous and lucent). Type 1, 2, 3, and 4 were classified as low density, while type 5, 6, 7, and 8 as intermediate density and type 9 and 10 as normal density. Two main paths of bone healing, the homogenous path (type 2–6-9) and the heterogenous path (type 3–7-10), and the isolated adverse types 5 and 8 was concluded in their study. In this study, we modified the grades of density into low, low-intermediate, high-intermediate, and high density for better illustration of the consolidation progress. We also observed the interconnection of homogenous and heterogenous path. In order to investigate the influence of adverse types 5 and 8, we grouped the homogenous path and heterogenous path into route 1 (path without presence of type 5 or 8) and route 2 (path with presence of type 5 or 8). (Fig. 2) Serial radiographs of 30 patients were used to achieve inter-observer agreement on the radiologic interpretation under supervision of the senior investigator. Designation of density grades and types was assigned to every month along with the timeline for each segment. During transition, they were assigned to the month if the density and type presented in most of the days, usually over 15 days. Trends of density increase for all the segments were recorded and analyzed according to presentation of certain patterns and patient factors. Consolidation progress was defined as “uninterrupted” if no events, such as fracture, bone graft, or deformity, were noted during DO. By age, patients were divided into younger (< 25 years), middle (25–49 years) and older group (
50 years). Based on distraction length, the segments were subdivided into longer gap group (
5 cm) and shorter gap group (< 5 cm).
Fig. 2.
The two developmental routes of traction callus subdivided into four grades: low density, low-intermediate density, high-intermediate density, high density. Route 1 consisted of two main intertwined paths (type 2—type 6—type 9; type 3—type 7—type 10) with pluripotent type 1 and type 4 as proposed in the study of Li et al. Route 2 was separated by an additional presentation of low-intermediate density callus formation of type 5 and type 8
Several indices to assess overall outcomes were measured in this study. Distraction consolidation time (DCT, days) was defined as the number of days to achieve consolidation, the presence of three out of four continuous cortices in both antero-posterior and lateral projections of the radiograph, from the start of distraction. Healing index (HI, days/cm) was obtained by dividing DCT with the distance of distraction segment lengthened in centimeters. Distraction index (DI, cm/days) was obtained by dividing the length of the distraction segment in centimeters by the duration of lengthening in days. Lengthening index (LI, months/cm) was obtained by dividing the time of external fixation in months by the distance of distraction length achieved. Consolidation index (CI, days/cm) was obtained by the number of days from surgery to consolidation divided by the length of regenerate bone in centimeters.
Statistical analysis
Statistical analysis was conducted with SPSS statistics v. 22.0.0.0 and R package 4.5.1. For baseline demographic data and characteristics, continuous variables were presented in mean (standard deviation, SD) and categorical variables in counts (percentage). Counts and percentages of traction callus classification were calculated. Parameters and indices of lengthening outcomes appeared in means (SD) and were compared by analysis of variance test (ANOVA). Mann–Whitney analysis was used to compare the outcomes. The analysis was also used in the comparison of duration to reach each grade, with the time duration shown in means (SD). Statistically significant was set as p-value < 0.05. Z value and effect size were also introduced. Effect size was categorized in into negligible (
0.147), small (0.147 
0.33), medium (0.33 
0.474), and large (
0.474) [20, 21]. The trajectory of grade advancement was drawn by Loess regression [22].
Results
A total of 129 patients were enrolled in the study based on the inclusion and exclusion criteria. Demographic data are shown on Table 1. Age distribution is from 5.6 to 77 years with mean age of 37.3 years old. Overall, 155 bone segments were studied, with 53 (34.2%) located at femur and 102 (65.8%) at tibia. Six femur segments and 17 tibia segments were double-leveled. The distraction length ranged from 11 to 174 mm with a mean of 53.2 mm. The mean follow-up time was 67.4 months. There were 9.3% (12/129) diabetic patients, 0.8% (1/129) with osteoporosis, and 0.8% (1/129) with rheumatic arthritis.
Table 1.
Characteristic and information of 155 segments from 129 patients that underwent DO
| Characteristic | All segments (N = 155) | Younger age group (age  yrs) (N = 50) |
Middle age group (25 yrs  50 yrs) (N = 63) |
Older age group (age  50 yrs) (N = 42) |
|---|---|---|---|---|
| Gender | ||||
| Patient (Male/Female)—no. (%) | 96 (74.4%) / 33 (25.6%) | 23 (53.5%) / 20 (46.5%) | 44 (86.3%) / 7 (13.7%) | 29 (82.9%) / 6 (17.1%) |
| Segment (Male/Female)—no. (%) | 118 (76.1%) / 37 (23.9%) | 28 (56.0%) / 22 (44.0%) | 56 (88.9%) / 7 (11.1%) | 34 (81.0%) / 8 (19.0%) |
| Indication—no. (%) | ||||
|
Fracture Infection Developemental Others |
69 (44.5%) 61 (39.4%) 12 (7.7%) 13 (8.3%) |
21 (42.0%) 11 (22.0%) 8 (16.0%) 10 (20.0%) |
25 (39.7%) 33 (52.4%) 4 (6.3%) 1 (1.6%) |
23 (54.8%) 17 (40.5%) 2 (4.8%) |
| Localization—no. (%) | ||||
|
Single—leveled femur Single—leveled tibia Double-leveled—upper/lower femur Double-leveled—upper/lower tibia |
41 (26.5%) 68 (43.9%) 6 (3.9%) / 6 (3.9%) 17 (11.0%) / 17 (11.0%) |
18 (36.0%) 22 (44.0%) 2 (4.0%) / 2 (4.0%) 3 (6.0%) / 3 (6.0%) |
17 (27.0%) 24 (38.1%) 3 (4.8%) / 3 (4.8%) 8 (12.7%) / 8 (12.7%) |
6 (14.3%) 22 (52.4%) 1 (2.4%) / 1 (2.4%) 6 (14.3%) / 6 (14.3%) |
| Area—no. (%) | ||||
|
Diaphysis Metaphysis |
124 (80.0%) 31 (20.0%) |
44 (88.0%) 6 (12.0%) |
49 (77.8%) 14 (22.2%) |
31 (73.8%) 11 (26.2%) |
| Type of external skeletal fixation—no. (%) | ||||
|
Circular frame Monolateral frame |
95 (61.3%) 60 (38.7%) |
16 (32.0%) 34 (68.0%) |
45 (71.4%) 18 (28.6%) |
34 (81.0%) 8 (19.0%) |
| Fixation for consolidation—no. (%) | ||||
|
Nail Plate Nail & plate External skeletal fixation Removal |
52 (33.5%) 34 (21.9%) 1 (0.6%) 3 (1.9%) 65 (41.9%) |
8 (16.0%) 21 (42.0%) 1 (2.0%) 20 (40.0%) |
29 (46.0%) 9 (14.3%) 1 (1.6%) 24 (38.1%) |
15 (35.7%) 4 (9.5%) 1 (2.4%) 1 (2.4%) 21 (50.0%) |
| Distraction length (mm)—mean (SD) | 53.2 (25.1) | 59.4 (28.3) | 50.0 (23.0) | 50.5 (23.2) |
| Following time (months)—mean (SD) | 67.4 (53.4) | 61.2 (44.6) | 81.4 (66.1) | 53.8 (34.8) |
SD, standard deviation; yrs, years
Among patients of all ages, most segments presented radiographic patterns along the two main paths of classification as shown in Table 2. In overall, type 2 outnumbered type 3 at low density level (90 vs. 54) while type 7 was more than type 6 at high-intermediate level (90 vs. 60). At high density level, number of type 9 and type 10 achieved equal (87 vs. 90). Above data revealed that patterns from homogenous path and heterogenous path take turns dominating in different stages and may exist interchange during the process, with both paths all lead to consolidation. When stratified by age, three groups manifested similar trends with the general group, while only the younger age group presented with more type 9 than type 10 callus patterns at high density level (37 > 23). As to the presence of shape, segments with shape 2 was the most (80%) in all age groups.
Table 2.
Outcome features, consolidation rates, parameters, and indices of all segments that underwent DO
| All segments (N = 155) |
Younger age group (N = 50) |
Middle age group (N = 63) |
Older age group (N = 42) |
|
|---|---|---|---|---|
| Classification—no | ||||
| Low density | ||||
| Type 1 | 104 | 31 | 45 | 28 |
| Type 2 | 90 | 27 | 41 | 22 |
| Type 3 | 54 | 17 | 20 | 17 |
| Type 4 | 10 | 3 | 4 | 3 |
| Low-intermediate density | ||||
| Type 5 | 20 | 6 | 10 | 4 |
| Type 8 | 8 | 1 | 4 | 3 |
| High-intermediate density | ||||
| Type 6 | 60 | 23 | 21 | 16 |
| Type 7 | 90 | 28 | 37 | 25 |
| High density | ||||
| Type 9 | 87 | 37 | 31 | 19 |
| Type 10 | 90 | 23 | 41 | 26 |
| Shape – no. (%) | ||||
| Shape 1 | 19 (12.3%) | 8 (16.0%) | 8 (12.7%) | 3 (7.1%) |
| Shape 2 | 123 (79.4%) | 38 (76.0%) | 50 (79.4%) | 35 (83.3%) |
| Shape 3 | 5 (3.2%) | 2 (4.0%) | 2 (3.2%) | 1 (2.4%) |
| Shape 4 | 8 (5.2%) | 2 (4.0%) | 3 (4.8%) | 3 (7.1%) |
| Uninterrupted consolidation—no. (%) | 141 (91.0%) | 44 (88.0%) | 59 (93.7%) | 38 (90.5%) |
| Parameters & indices—mean (SD) | ||||
| Distraction length (mm) | 50.8 (22.4) | 55.7 (22.7) | 47.7 (20.9) | 50.0 (23.9) |
| DCT (days) | 311.3 (112.1) | *271.6 (80.3) | 328.5 (122.0) | 330.7 (118.3) |
| DI (cm/days) | 0.71 (0.19) | 0.73 (0.14) | 0.69 (0.18) | 0.71 (0.23) |
| CI (days/cm) | 74.4 (39.4) | *56.9 (24.9) | 82.3 (41.0) | 82.3 (44.6) |
| HI (days/cm) | 72.6 (38.8) | *55.4 (24.4) | 80.5 (40.5) | 80.4 (43.7) |
| LI (months/cm) | 2.0 (1.4) | *1.5 (1.2) | 2.1 (1.3) | 2.4 (1.8) |
Parameters and indices were compared with ANOVA test. *p < 0.05
DCT, distraction-consolidation time; DI, distraction index; CI, consolidation index; HI, healing index; LI, lengthening index
Overall, the mean distraction length for consolidated cases was 50.8 mm (
22.4) with DCT 311.3 days (
112.1), DI 0.71 cm/day (
0.19), and HI 72.6 day/cm (
38.8) (Table 2). In different age groups, the mean distraction length was indifferent and also around 50 mm. The mean DI is indifferent among groups. Nevertheless, DCT, CI, HI, and LI were significantly lower in younger age group compared to the other two age groups, which means more efficient in consolidation for young patients.
Consolidation following route 1 or route 2, for overall age group, exhibited significant difference in DCT, DI, CI, and HI (Table 3). A significantly shorter consolidation time and a faster consolidation rate was shown in route 1 (DCT, 300.4 days; CI, 71.8 days/cm; HI, 70.1 days/cm) than in route 2 (DCT, 381.6 days; CI, 91.0 days/cm; HI, 89.1 days/cm). Route 1 also had a higher mean distraction rate (DI, 0.722 cm/day) than route 2 (DI, 0.609 cm/day). Indices of route 1 and route 2 in younger age group and older age group was not significantly apart, whereas the DCT and DI showed difference between routes in middle age group.
Table 3.
Comparison of route 1 and route 2 with Mann–Whitney test in all ages, younger age group, middle age group, and older age group
| Route 1 | Route 2 | Z | P-value | Effect size | ||
|---|---|---|---|---|---|---|
| All segments | Age (yrs) | 37.6 | 37.4 | −0.278 | 0.781 | 0.040 |
| Distraction length (mm) | 51.1 | 49.1 | −0.248 | 0.804 | 0.035 | |
| DCT (days) | 300.4 | 381.6 | −3.074 | *0.002 | 0.439 | |
| DI (cm/days) | 0.722 | 0.609 | −2.545 | *0.011 | 0.364 | |
| CI (days/cm) | 71.8 | 91.0 | −2.276 | *0.023 | 0.325 | |
| HI (days/cm) | 70.1 | 89.1 | −2.306 | *0.021 | 0.330 | |
| LI (months/cm) | 1.98 | 2.07 | −0.054 | 0.957 | 0.008 | |
| Younger age group | Age (yrs) | 19.1 | 17.1 | −0.082 | 0.935 | 0.025 |
| Distraction length (mm) | 56.9 | 43.5 | −0.694 | 0.488 | 0.213 | |
| DCT (days) | 274.9 | 238.8 | −0.306 | 0.759 | 0.094 | |
| DI (cm/days) | 0.738 | 0.701 | −0.633 | 0.527 | 0.194 | |
| CI (days/cm) | 56.3 | 62.9 | −0.939 | 0.348 | 0.288 | |
| HI (days/cm) | 54.9 | 60.3 | −0.939 | 0.348 | 0.288 | |
| LI (months/cm) | 1.41 | 2.32 | −0.408 | 0.683 | 0.125 | |
| Middle age group | Age (yrs) | 36.6 | 36.9 | −0.010 | 0.992 | 0.002 |
| Distraction length (mm) | 46.6 | 53.2 | −0.707 | 0.479 | 0.143 | |
| DCT (days) | 307.2 | 432.6 | −2.809 | *0.005 | 0.567 | |
| DI (cm/days) | 0.714 | 0.547 | −2.667 | *0.008 | 0.539 | |
| CI (days/cm) | 79.2 | 97.5 | −1.354 | 0.176 | 0.273 | |
| HI (days/cm) | 77.3 | 95.8 | −1.394 | 0.163 | 0.282 | |
| LI (months/cm) | 2.19 | 1.88 | −0.525 | 0.599 | 0.106 | |
| Older age group | Age (yrs) | 61.7 | 54.6 | −1.944 | 0.052 | 0.545 |
| Distraction length (mm) | 50.7 | 45.4 | −0.540 | 0.589 | 0.152 | |
| DCT (days) | 321.1 | 393.8 | −1.836 | 0.066 | 0.515 | |
| DI (cm/days) | 0.715 | 0.659 | −0.583 | 0.560 | 0.164 | |
| CI (days/cm) | 79.5 | 100.6 | −1.317 | 0.188 | 0.370 | |
| HI (days/cm) | 77.6 | 98.8 | −1.317 | 0.188 | 0.370 | |
| LI (months/cm) | 2.37 | 2.26 | −0.065 | 0.948 | 0.018 |
*p < 0.05
yrs, years; DCT, distraction-consolidation time; DI, distraction index; CI, consolidation index; HI, healing index; LI, lengthening index
When subdivide by distraction length, the indices differences between 2 routes are not significant in shorter group (Table 4). In the longer group, route 1 and route 2 demonstrated significant difference in DCT, DI, CI, and HI.
Table 4.
Comparison of route 1 and route 2 with Mann–Whitney test in shorter gap group and longer gap group
| Route 1 | Route 2 | Z | P-value | Effect size | ||
|---|---|---|---|---|---|---|
| Shorter gap group | Age (yrs) | 39.5 | 35.3 | −0.489 | 0.625 | 0.097 |
| Distraction length (mm) | 33.2 | 30.1 | −0.747 | 0.455 | 0.148 | |
| DCT (days) | 282.2 | 341.2 | −1.765 | 0.078 | 0.349 | |
| DI (cm/days) | 0.675 | 0.630 | −0.682 | 0.495 | 0.135 | |
| CI (days/cm) | 93.7 | 117.3 | −1.909 | 0.056 | 0.378 | |
| HI (days/cm) | 91.3 | 114.6 | −1.877 | 0.061 | 0.371 | |
| LI (months/cm) | 2.47 | 2.71 | −0.209 | 0.835 | 0.041 | |
| Longer gap group | Age (yrs) | 35.7 | 39.7 | −0.851 | 0.395 | 0.177 |
| Distraction length (mm) | 70.1 | 70.2 | −0.308 | 0.758 | 0.064 | |
| DCT (days) | 319.7 | 426.4 | −2.715 | *0.007 | 0.565 | |
| DI (cm/days) | 0.773 | 0.586 | −3.004 | *0.003 | 0.625 | |
| CI (days/cm) | 48.4 | 61.8 | −2.543 | *0.011 | 0.529 | |
| HI (days/cm) | 47.4 | 60.8 | −2.561 | *0.010 | 0.533 | |
| LI (months/cm) | 1.46 | 1.36 | −0.027 | 0.978 | 0.006 |
*p < 0.05
yrs, years; DCT, distraction-consolidation time; DI, distraction index; CI, consolidation index; HI, healing index; LI, lengthening index
Progression of density grade along with timeline for route 1 and 2 were depicted by Loess regression line in different grouping by age and length. (Figs. 3 and 4) The average time by months to reach each density grade is provided in Tables 5 and 6.
Fig. 3.
Progression of consolidation process of route 1 and route 2 in a all ages, b younger age group, c middle age group, and d older age group. H, high; H-I, high-intermediate; L-I, low intermediate; L, low
Fig. 4.
Progression of consolidation process of route 1 and route 2 in a shorter gap group and b longer gap group. H, high; H-I, high-intermediate; L-I, low intermediate; L, low
Table 5.
Comparison of time interval from osteotomy to each density grades between route 1 and route 2 in all ages, younger age group, middle age group, and older age group
| Time to achieve target bone density (months) | Z | P-value | Effect size | ||
|---|---|---|---|---|---|
| Route 1 | Route 2 | ||||
| All segments | |||||
| Low density | 3.68 | 3.43 | −0.454 | 0.650 | 0.097 |
| High-intermediate density | 6.39 | 9.09 | −3.636 | * < 0.001 | 0.652 |
| High density | 11.02 | 13.84 | −3.008 | *0.003 | 0.428 |
| Younger age group | |||||
| Low density | 3.40 | n/a | n/a | n/a | n/a |
| High-intermediate density | 6.11 | 6.33 | −0.955 | 0.339 | 0.325 |
| High density | 10.35 | 8.75 | −0.701 | 0.484 | 0.213 |
| Middle age group | |||||
| Low density | 3.68 | 3.00 | −1.023 | 0.306 | 0.333 |
| High-intermediate density | 6.54 | 10.60 | −3.279 | *0.001 | 0.879 |
| High density | 11.41 | 15.70 | −2.840 | *0.005 | 0.571 |
| Older age group | |||||
| Low density | 4.11 | 4.00 | < 0.001 | > 0.999 | 0.000 |
| High-intermediate density | 6.50 | 9.33 | −2.255 | *0.024 | 0.789 |
| High density | 11.26 | 14.20 | −1.981 | *0.048 | 0.555 |
*p < 0.05
Table 6.
Comparison of time interval from osteotomy to each density grades between route 1 and route 2 in shorter gap group and longer gap group
| Time to achieve target bone density (months) | Z | P-value | Effect size | ||
|---|---|---|---|---|---|
| Route 1 | Route 2 | ||||
| Shorter gap group | |||||
| Low density | 3.50 | 3.50 | −0.093 | 0.926 | 0.026 |
| High-intermediate density | 5.90 | 9.40 | −2.909 | *0.004 | 0.764 |
| High density | 10.40 | 12.60 | −1.861 | 0.063 | 0.366 |
| Longer gap group | |||||
| Low density | 3.87 | 3.33 | −0.707 | 0.479 | 0.235 |
| High-intermediate density | 6.93 | 8.83 | −2.088 | *0.037 | 0.515 |
| High density | 11.67 | 15.22 | −2.485 | *0.013 | 0.515 |
*p < 0.05
In general age group, two routes share a similar trajectory before reaching high-intermediate density but curve apart afterwards significantly. Route 2 showed slowdown in consolidation. Mean time to reach low density are comparable between routes (3.68 vs. 3.43 months). On the other hand, the time to reach high-intermediate and high density is significantly longer in route 2 (9.09 vs. 6.39, 13.84 vs.11.02 months).
In middle and older age groups, curve trajectory showed similar trend, which was deviating after reaching high-intermediate density with slop of the curve significantly decreased for route 2. Nonetheless, in the younger age group, both routes demonstrated similar and close curve trajectory through all stages.
The regression curve of route 1 and route 2 in shorter and longer gap groups both separated at high intermediate density grade, but with some difference. (Fig. 4) The curves in the shorter gap group went closer at high density grade, but in longer gap group they split further when approached high density grade. In shorter gap group, route 2 presented slowdown in consolidation with mean duration to reach high-intermediate density level were 9.40 months, compared to 5.90 months for route 1 (p-value = 0.004, effect size = 0.764). Nevertheless, the time to reach high density grade in two routes was 10.40 and 12.60 months with no significant difference (p-value = 0.063, effect size = 0.366). The longer gap group, on the other hand, demonstrated significant difference (p-value = 0.037, effect size = 0.515) between routes at high-intermediate density grade and even more significant at high density grade (p-value = 0.013, effect size = 0.515).
Discussion
Factors correlated with the outcome of DO had been widely analyzed in previous studies [12–15]. Age and the distraction length were the two most well-established factors among them [12–15]. Patterns of consolidation progression had been proposed to be associated with clinical outcomes, but continuous trending analysis for their clinical impacts has not been discussed. Based on Li’s pattern classification [8], current study proposed two distinct routes of consolidation which departs at high intermediate density and show differences in trends. The route featured with Li’s type 5 or 8 showed slowdown in consolidation. In addition, elder patients and patients with distraction length longer than 5 cm in our study showed higher subjectivity to the presence of this feature. However, most of the callus from both routes achieved fully consolidation even with slowdown in progression. Periosteal-preserving oscillating saw osteotomy used in current study yielded comparable overall results to previous studies with multiple drilling technique [23].
Healing index (HI) has been used widely to assess the quality of DO over decades. Early studies reported HI between 35 and 45 days/cm in children and adolescents with distraction length from 4.5 to 7 cm [3, 7, 15, 16, 24]. However, HI elevated as clinical settings became more complicated, such as patients with more difficult surgical indications and older age. Studies performed on older patients with indications of osteomyelitis, distraction length ranging from 4.4 to 8.4 cm, reported HI between 45 and 57 days/cm [25–27]. Current study obtained an overall mean HI of 72.6 days/cm (median = 63.5 days/cm). Although higher than prior studies, this data may not indicate inferior outcomes. Firstly, a lower median HI compared to mean HI was observed in all age groups. This could be due to the presence of extreme cases with clinical challenges which ended up with very high HI in our series. Secondly, studies have proven a strong negative hyperbolic relationship between distraction length and HI, DCI or maturation index. The indices rise sharply as the distraction length decreases and gradually level off as the distraction length increases [12, 13, 15]. The conversion of this tendency fell roughly at distraction length of 6 cm, just above the current study, so a drastic rise in HI was expected. Lastly, this study enrolled patients with a wider spectrum in age and a mean age larger than many of the prior studies, which may contribute to the elevation of HI [16]. The original study of Li et al., which reported similar distraction length (6.5 cm) and patient age (24.7 years), shared a close HI (72 days/cm) [8].
In this study, we explored the potential of specific callus patterns as a prognostic factor for outcome of DO. By far, there were several studies that took similar approaches. Initially, classifications based on the contour and features of traction callus were proposed independently [6, 7]. Two-dimensional classification of traction callus, by its density and feature, was proposed by Li et al. Their study was the first to observe two main callus formation paths: homogenous and heterogenous path. The former led to a better outcome and the latter progressed to a worse outcome. Other types that did not land on the two paths were either multipotent (type 1 and 4) or adverse (type 5 and 8) [8]. Tirawanish et al. built a grading system of traction callus to predict its subsidence and scored them by their width and density [28]. However, no further evidence to recognize specific patterns as pathognomonic in clinical usage. Based on Li’s classification, we observed the profound interconnection between homogenous and heterogenous paths and suggested to combine them together as a non-pathognomonic feature. We proposed type 5 and type 8, types without continuity at the central part of distraction gap, as pathognomonic features may be attributed to previous studies on the biological and biomechanical mechanism of new bone formation. Although intramembranous and endochondral ossification occur simultaneously during bone healing [4], blood supply and progenitors originate at intact bone ends; mechanical strain gradients favor bone at the fracture edges [29–31]. Thus, the center of the fracture gap is prone to interference of micro-movement and may serve as early manifestation of imperfectly stable healing environment. In addition, the “creeping substitution” phenomenon raised by Prof. McKibbin also supported the resistibility of callus formation at fracture ends [32].
Our study was consonant with previous studies that concluded aging poses a deleterious effect on bone healing [12–14, 16, 33]. A rat model study reported that older rats have relative deficit in endosteal bone formation and most likely related to cell organization than to cell proliferation [14]. Fifty to 100 percent increase in lengthening index of 12 adults compared to 48 pediatric patients was described in Paley’s study [33]. In our study, we identified the negative effect of the presence of type 5 and 8 on consolidation generally and for different age groups, which bear different subjectivity to this deleterious effect. Consolidation was not affected for those under 25 years old. On the other hand, consolidation in patients from middle age and older age group revealed significant slowdown when presented with type 5 or 8. The higher regenerative potential in younger population could be attributed to more potent progenitor (mesenchymal stem cells) and osteoblast activity, stronger angiogenic response, thicker highly osteogenic vascularized periosteum, higher growth factor and hormonal milieu, and the lesser inflammation/cellular senescence that slowdown fracture healing [34–37].
Influence of distraction length on the DO outcome was observed in previous studies [12, 15], but not specifically on the radiographic features. In the current study, both the shorter and the longer gap group showed being under the negative effect of type 5 or 8 and the effect was profound for longer group.
Despite the novel findings of this study, several limitations should be acknowledged. We are well aware of the retrospective nature and the heterogeneity of the study. Thus, the results of the study cannot be applied directly to the general population and may not be the sole evidence for guideline development. Additionally, the number of operations done on each patient before index procedure was not fully weighted. Possible confounding factors, including osteoporosis or systemic comorbidities (e.g. rheumatologic disorders, diabetes mellitus), were not entirely excluded. However, patients enrolled to the study with these factors went through strict standards of disease status control and only constituted a small fraction. Nonetheless, to our knowledge, our study represents the largest retrospective series on the radiographic change of DO and the first on the relationship between age, length, and radiographic outcome of DO.
Conclusion
We concluded that age and distraction length remained as two of the major factors that affect consolidation in DO and certain callus patterns would pose greater risk of delay healing. While middle and older age group were significantly subjected to this impact, younger patients showed more resistance. Similarly, longer distraction length (> 5 cm) bears higher subjectivity to it. Therefore, an active and concentrated surveillance is suggested to the presence of those negative features during follow-up. If prolonged negative features are observed, interventions such as slowdown of distraction, cyclic distraction-compression dynamization, or additional fixation methods can be adopted. However, further prospective interventional studies are required to explore their roles as therapeutic target to improve treatment outcomes.
Acknowledgements
Not applicable.
Abbreviations
- DO
Distraction Osteogenesis
- DCT
Distraction Consolidation Time
- HI
Healing Index
- DI
Distraction Index
- EI
External Fixator Index
- LI
Lengthening Index
- CI
Consolidation Index
- SD
Standard Deviation
- yrs
Years
- ANOVA
Analysis of Variance
- n/a
Not Applicable
Author contributions
KCYK and WCC wrote the manuscript and collected the data. WLC designed the study and helped revise the manuscript. KCYK, WLC, and SAK prepared Tables 1, 2, 3, 4, 5 and 6. KCYK, KHH, and HYH analyzed the data and prepared Figs. 1, 2, 3 and 4. YPS and HYH supervised the project, administered it, and performed validation.
Funding
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was approved by the institutional review board of Taipei Veterans General Hospital (IRB number: 2025–09-004AC). We confirmed that all methods were carried out in accordance with relevant guidelines and regulations.
Consent for publication
Not applicable.
Competing interests
The authors have no relevant financial or non-financial interests to disclose.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Goldstein RY, Jordan CJ, McLaurin TM, Grant A. The evolution of the Ilizarov technique: part 2: the principles of distraction osteosynthesis. Bull Hosp Jt Dis. 2013;2013(71):96–103. [PubMed] [Google Scholar]
- 2.Jordan CJ, Goldstein RY, McLaurin TM, Grant A. The evolution of the Ilizarov technique: part 1: the history of limb lengthening. Bull Hosp Jt Dis. 2013;2013(71):89–95. [PubMed] [Google Scholar]
- 3.De Bastiani G, Aldegheri R, Renzi-Brivio L, Trivella G. Limb lengthening by callus distraction (callotasis). J Pediatr Orthop. 1987;7:129–34. [DOI] [PubMed] [Google Scholar]
- 4.Forriol F, Denaro L, Longo UG, Taira H, Maffulli N, Denaro V. Bone lengthening osteogenesis, a combination of intramembranous and endochondral ossification: an experimental study in sheep. Strateg Trauma Limb Reconstr. 2010;5:71–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Aldegheri R, Trivella G, Lavini F. Epiphyseal distraction. Hemichondrodiatasis. Clin Orthop Relat Res. 1989;128–136. [PubMed]
- 6.Hamanishi C, Yasuwaki Y, Kikuchi H, Tanaka S, Tamura K. Classification of the callus in limb lengthening. Radiographic study of 35 limbs. Acta Orthop Scand. 1992;63:430–3. [DOI] [PubMed] [Google Scholar]
- 7.Minty I, Maffulli N, Hughes TH, Shaw DG, Fixsen JA. Radiographic features of limb lengthening in children. Acta Radiol. 1994;35:555–9. [PubMed] [Google Scholar]
- 8.Li R, Saleh M, Yang L, Coulton L. Radiographic classification of osteogenesis during bone distraction. J Orthop Res. 2006;24:339–47. [DOI] [PubMed] [Google Scholar]
- 9.Hughes TH, Maffulli N, Green V, Fixsen JA. Imaging in bone lengthening. A review. Clin Orthop Relat Res. 1994;308:50–3. [PubMed] [Google Scholar]
- 10.Maffulli N, Thornton A. Ultrasonographic appearance of external callus in long-bone fractures. Injury. 1995;26:5–12. [DOI] [PubMed] [Google Scholar]
- 11.Maffulli N, Cheng JC, Sher A, Ng BK, Ng E. Bone mineralization at the callotasis site after completion of lengthening. Bone. 1999;25:333–8. [DOI] [PubMed] [Google Scholar]
- 12.Fischgrund J, Paley D, Suter C. Variables affecting time to bone healing during limb lengthening. Clin Orthop Relat Res. 1994;301:31–7. [PubMed] [Google Scholar]
- 13.Noonan KJ, Leyes M, Forriol F, Cañadell J. Distraction osteogenesis of the lower extremity with use of monolateral external fixation. A study of two hundred and sixty-one femora and tibiae. J Bone Joint Surg Am. 1998;80(6):793–806. [DOI] [PubMed] [Google Scholar]
- 14.Aronson J, Gao GG, Shen XC, McLaren SG, Skinner RA, Badger TM, et al. The effect of aging on distraction osteogenesis in the rat. J Orthop Res. 2001;19:421–7. [DOI] [PubMed] [Google Scholar]
- 15.Sakurakichi K, Tsuchiya H, Uehara K, Kabata T, Tomita K. The relationship between distraction length and treatment indices during distraction osteogenesis. J Orthop Sci. 2002;7:298–303. [DOI] [PubMed] [Google Scholar]
- 16.Maffulli N, Lombari C, Matarazzo L, Nele U, Pagnotta G, Fixsen JA. A review of 240 patients undergoing distraction osteogenesis for congenital post-traumatic or postinfective lower limb length discrepancy. J Am Coll Surg. 1996;182:394–402. [PubMed] [Google Scholar]
- 17.Castelein S, Docquier PL. Complications associated with bone lengthening of the lower limb by callotasis. Acta Orthop Belg. 2016;82:806–13. [PubMed] [Google Scholar]
- 18.Bonnard C, Favard L, Sollogoub I, Glorion B. Limb lengthening in children using the Ilizarov method. Clin Orthop Relat Res. 1993;293:83–8. [PubMed] [Google Scholar]
- 19.Iliopoulos E, Morrissey N, Cho S, Khaleel A. Outcomes of the Ilizarov frame use in elderly patients. J Orthop Sci. 2017;22:783–6. [DOI] [PubMed] [Google Scholar]
- 20.Cliff N. Dominance statistics: ordinal analyses to answer ordinal questions. Psychol Bull. 1993;114:494–509. [Google Scholar]
- 21.Feng D, Cliff N. JMASM29: Dominance Analysis of Independent Data (Fortran). J Mod Appl Stat Methods. 2009;8(2):32. [Google Scholar]
- 22.Cleveland WS, Devlin SJ. Locally weighted regression: an approach to regression analysis by local fitting. J Am Stat Assoc. 1988;83(403):596–610. [Google Scholar]
- 23.Corona PS, Ramirez-Nunez LJ, Amat C, Carrera L. Outcome of oscillating saw open osteotomy in two-stage lower extremity bone transport with monolateral frame. Injury. 2017;48:2285–91. [DOI] [PubMed] [Google Scholar]
- 24.Blondel B, Launay F, Glard Y, Jacopin S, Jouve J, Bollini G. Limb lengthening and deformity correction in children using hexapodal external fixation: preliminary results for 36 cases. Orthop Traumatol Surg Res. 2009;95:425–30. [DOI] [PubMed] [Google Scholar]
- 25.Spiegl U, Patzold R, Friederichs J, Hungerer S, Militz M, Buhren V. Clinical course, complication rate and outcome of segmental resection and distraction osteogenesis after chronic tibial osteitis. Injury. 2013;44:1049–56. [DOI] [PubMed] [Google Scholar]
- 26.Salcedo Canovas C, Martinez Ros J, Ondono Navarro A, Molina Gonzalez J, Hernandez Torres A, Moral Escudero E, et al. Infected bone defects in the lower limb. Management by means of a two-stage distraction osteogenesis protocol. Eur J Orthop Surg Traumatol. 2021;31(7):1375–86. [DOI] [PubMed] [Google Scholar]
- 27.Lavini F, Dall’Oca C, Bartolozzi P. Bone transport and compression-distraction in the treatment of bone loss of the lower limbs. Injury. 2010;41:1191–5. [DOI] [PubMed] [Google Scholar]
- 28.Tirawanish P, Eamsobhana P. Prediction of callus subsidence in distraction osteogenesis using callus formation scoring system: preliminary study. Orthop Surg. 2018;10:121–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Trueta J, Morgan JD. The vascular contribution to osteogenesis. I. Studies by the injection method. J Bone Joint Surg Br. 1960;42-b:97–109. [DOI] [PubMed] [Google Scholar]
- 30.Colnot C. Skeletal cell fate decisions within periosteum and bone marrow during bone regeneration. J Bone Miner Res. 2009;24:274–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Claes LE, Heigele CA. Magnitudes of local stress and strain along bony surfaces predict the course and type of fracture healing. J Biomech. 1999;32:255–66. [DOI] [PubMed] [Google Scholar]
- 32.McKibbin B. The biology of fracture healing in long bones. J Bone Joint Surg Br. 1978;60-b:150–62. [DOI] [PubMed] [Google Scholar]
- 33.Paley D. Problems, obstacles, and complications of limb lengthening by the Ilizarov technique. Clin Orthop Relat Res. 1990. 10.1097/00003086-199001000-00011. [PubMed] [Google Scholar]
- 34.Malizos KN, Papatheodorou LK. The healing potential of the periosteum molecular aspects. Injury. 2005;36(Suppl 3):S13-19. [DOI] [PubMed] [Google Scholar]
- 35.Ekbote VH, Khadilkar VV, Khadilkar AV, Mughal Z, Chiplonkar SA, Palande SA, et al. Relationship of insulin-like growth factor 1 and bone parameters in 7–15 years old apparently, healthy Indian children. Indian J Endocrinol Metab. 2015;19:770–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Clark D, Nakamura M, Miclau T, Marcucio R. Effects of aging on fracture healing. Curr Osteoporos Rep. 2017;15:601–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Menger MM, Manuschewski R, Hans S, Braun BJ, Kayali M, Ehnert S, et al. Age-related alterations of angiogenesis, inflammation and bone microarchitecture during fracture healing in mice. Geroscience. 2025;47:4321–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
No datasets were generated or analysed during the current study.