Abstract
Context:
Pediatric Crohn's Disease (CD) is associated with deficits in trabecular bone mineral density (BMD) and cortical structure, potentially related to TNF-α effects to decrease bone formation and promote bone resorption.
Objective:
This study aimed to examine changes in bone density and structure in children and adolescents with CD following initiation of anti-TNF-α therapy.
Design and Participants:
Participants (n = 74; age 5–21 years) with CD completed a 12-month prospective cohort study.
Main Outcome Measures:
Tibia peripheral quantitative computed tomography scans were obtained at initiation of anti-TNF-α therapy and 12 months later. Musculoskeletal outcomes were expressed as sex-and race-specific z scores relative to age, based on >650 reference participants.
Results:
At baseline, CD participants had lower height, trabecular BMD, cortical area (due to smaller periosteal and larger endocortical circumferences), and muscle area z scores, compared with reference participants (all P < .01). Pediatric CD activity index decreased during the 10-week induction (P < .001), in association with subsequent gains in height, trabecular BMD, cortical area (due to recovery of endocortical bone), and muscle area z scores over 12 months (height P < .05; others P < .001). Bone-specific alkaline phosphatase levels, a biomarker of bone formation, increased a median of 75% (P < .001) during induction with associated 12-month improvements in trabecular BMD and cortical area z scores (both P < .001). Younger age was associated with greater increases in trabecular BMD z scores (P < .001) and greater linear growth with greater recovery of cortical area (P < .001).
Conclusions:
Anti-TNF-α therapy was associated with improvements in trabecular BMD and cortical structure. Improvements were greater in younger and growing participants, suggesting a window of opportunity for treatment of bone deficits.
Crohn's disease (CD) is an autoimmune condition of the gastrointestinal tract characterized by chronic inflammation and defective innate immune regulation of the gut microbiome. Children with CD have multiple risk factors for impaired bone accrual, including poor growth, delayed maturation, malnutrition, decreased muscle mass and physical activity, inflammation, and glucocorticoid therapy. The effect on bone health may be immediate, resulting in fragility fractures in childhood, or delayed, with suboptimal peak bone mass in adulthood (1, 2). Numerous studies have documented significant trabecular and cortical bone deficits and increased fracture rates in children and adults with CD (1–6). Skeletal modeling during growth is characterized by sex- and maturation-specific gains in trabecular and cortical bone mineral density (BMD) and cortical dimensions (7). TNF-α has direct adverse effects on bone metabolism and plays a pivotal role in CD pathogenesis (8). TNF-α impairs bone formation through inhibition of osteoblast differentiation maturation and activity, and promotes bone resorption through activation of osteoclasts. These adverse effects may be compounded by glucocorticoid therapy, which has similar effects on osteoblasts (9). The growing skeleton may be uniquely vulnerable to the detrimental effects of chronic inflammation and glucocorticoid therapy, potentially resulting in life-long insufficiency in bone strength.
We previously evaluated an incident cohort of 78 children and adolescents with CD examining the effect of the underlying disease on bone density and structure prior to glucocorticoid therapy (3). Trabecular volumetric BMD, cortical dimensions, and muscle area, as measured by tibia peripheral quantitative computed tomography (pQCT), were significantly reduced at diagnosis. Following initiation of CD therapy, trabecular BMD improved but periosteal dimensions failed to expand commensurate with linear growth. Despite the fact the Pediatric Crohn Disease Activity Index (PCDAI) score indicated no active disease in 85% of participants 3–4 years after diagnosis, significant deficits in trabecular BMD, cortical area, and muscle area persisted.
Infliximab, a chimeric monoclonal antibody against TNF-α, induces and maintains clinical remission in moderate-to-severe CD and is the first-line biologic agent in refractory pediatric CD (10, 11). Studies have documented improved linear growth during infliximab treatment; however, the effect on bone modeling has not been established (11, 12).
This prospective cohort study examined changes in peripheral pQCT measures of tibia volumetric BMD and cortical structure (cortical area and periosteal and endocortical circumference) over a 12-month interval following initiation of infliximab therapy in children and adolescents with CD. We hypothesized TNF-α contributes to deficits in trabecular BMD and cortical area (through impaired bone formation on the periosteal surface and increased bone resorption on the endocortical surface), and these parameters would improve during infliximab therapy. The objectives were to describe changes in pQCT measures of cortical and trabecular BMD and cortical structure (cortical area and periosteal and endocortical circumference) and to identify clinical and laboratory correlates of changes during 12 months of treatment with infliximab. This is the first study to address the potential for recovery of trabecular and cortical deficits in children and adolescents following initiation of anti-TNF-α therapy for any clinical indication with broad implications for bone health in childhood inflammatory diseases.
Materials and Methods
Study population
This prospective cohort study of bone and mineral metabolism enrolled 90 children and adolescents, ages 5–21 years, with biopsy-proven CD, at the time of initiation of infliximab therapy at the Children's Hospital of Philadelphia (CHOP). Participants were excluded for prior anti-TNF-α therapy; pregnancy; cognitive/developmental disorders that affected their ability to complete the study procedures; and medical illnesses or therapies unrelated to CD that could potentially affect bone, nutrition, or growth such as kidney disease, seizure disorder, diabetes, or liver disease. Study visits were completed at the first infliximab infusion, and 10 weeks, 6 months, and 12 months later. Short-term changes in mineral metabolism were reported, demonstrating significant increases in PTH and 1,25(OH)2 vitamin D levels in association with improvements in disease activity (13). CD participants were compared with reference data generated in greater than 650 healthy reference participants recruited from the surrounding community as part of a larger study of bone in healthy children (7). A subset of 240 reference participants, mean age (SD) of 12.7 (4.3) years, 126 females (53%), enrolled onto an ancillary study completed an additional visit at 12 months (14). The use of this longitudinal data is detailed below in statistical analysis.
The study protocol was approved by the CHOP Institutional Review Board. Informed consent was obtained from participants at least 18 years of age, and a parent or guardian of participants less than 18 years of age. Assent was obtained from participants 7–17 years of age.
Data collection
Disease characteristics, anthropometry, fracture events, and medications were recorded at each visit. Disease activity was assessed using the pediatric CD activity index (PCDAI) based on symptoms (30%), physical examination (30%), laboratory parameters (20%), and growth (20%), with scores ranging from 0–100 (15). Disease activity was categorized as none (1–10), mild (11–30), and moderate to severe (>30). Participants were provided with glucocorticoid diaries, which were reviewed at each visit, and the cumulative dose (mg/kg/d prednisone equivalents) over the 12-month study was tabulated. Height was measured using a stadiometer and weight using a digital scale. Tanner stage was ascertained by self-assessment questionnaire (16).
Laboratory measures at each visit included hematocrit, erythrocyte sedimentation rate (ESR; mm/h), serum C-reactive protein (CRP; mg/L), and albumin (g/dL) concentrations, using standard methods in the clinical laboratory. The vitamin D and cytokine assay methods were previously reported (13). Serum bone-specific alkaline phosphatase (BSAP; μg/L) was measured by two-site immunoradiometric assay (coefficient of variation [CV], 8%) and serum C-terminal telopeptide of type 1 collagen (β-CTX; pg/mL) by an electrochemiluminescent assay (CV, 5%) at Quest Diagnostics Laboratory. Cytokines were measured at enrollment, 10-week, and 12-month visits.
Left tibia pQCT scans were obtained using a Stratec XCT2000 device (Orthometrix) with a 12-detector unit, voxel size 0.4 mm, slice thickness 2.3 mm, and scan speed 25 mm/sec, as previously described (3, 4, 7, 17, 18). The reference line was placed at the proximal border of the distal tibia growth plate if open and at the distal endplate if fused. The baseline scout view was reviewed at the time of the 12-month visit to ensure consistency of reference line placement, and all source images were reviewed by one individual for adequacy of reference line placement. Trabecular volumetric BMD (mg/cm3) was measured 3% of tibia length proximal to the reference line and cortical volumetric BMD (mg/cm3), periosteal and endocortical circumference (mm), and cortical cross-sectional area (mm2) at the 38% site. Prior biomechanics studies demonstrated correlations between fracture load and pQCT trabecular BMD (R = 0.75) and diaphyseal cortical area (R = 0.93) (19). Muscle and fat cross-sectional area (mm2) were assessed at the 66% site. The manufacturer's hydroxyapatite phantom was scanned daily. Our CV in children and adolescents was <2% for test-retest scan procedures with participant repositioning for all pQCT outcomes (20).
Statistical analysis
Analyses were performed using Stata 13.0 (StataCorp.). P < .05 was considered significant and two-sided tests used throughout. Continuous variables were expressed as mean ± SD or median (range or interquartile range [IQR]). Group differences between CD and reference participants were assessed using the Student t test or Wilcoxon rank-sum test, and changes within CD participants were tested using the paired t test or Wilcoxon signed-rank test. Correlations between continuous variables were assessed by Pearson or Spearman correlations. Age- and sex-specific height and body mass index (BMI) z scores were generated using national data (21). pQCT outcomes were converted to race- and sex-specific z scores relative to age based on the reference participants using the LMS method (Chartmaker Program v2.3) that accounts for the non-linearity, heteroscedasticity, and skew of bone and body composition data (22). Cortical dimensions and muscle and fat area were highly correlated with tibia length (all P < .0001); therefore, the z scores for age were further adjusted for tibia-length-for-age z score, as described (4). Of note, the PCDAI includes points for a height velocity z score less than −1.0 and does not fully capture changes in tibia length. Sex- and Tanner stage–specific z scores for bone biomarkers were calculated using reference participant data (23).
All linear regression models for 12-month changes in pQCT z scores were adjusted for baseline z score. Glucocorticoids were assessed as a binary (any vs none) and three-category variable (none, < median, and ≥ median dose). Most laboratory parameters improved significantly during the first 10 weeks and did not improve further at 6 or 12 months; therefore, models assessed the association of the 10-week changes in these parameters with 12-month changes in pQCT outcomes. Models for change in cortical BMD z scores included changes in cortical area z scores and tibia length, as described (4, 24). Periosteal and endocortical circumferences are highly correlated during childhood and adolescents (R > 0.80) and both dimensions expand during normal growth. We used linear regression to adjust endocortical circumference z score for periosteal circumference z score and the β-coefficient for the CD vs control participant represented the adjusted z score. During growth, periosteal circumference increases in response to biomechanical loading; therefore, models for changes in periosteal circumference z scores tested the association with baseline and change in muscle area z score within the CD participants and in comparison with the subset of reference participants enrolled onto the longitudinal study (25). We also employed an alternative approach to assess changes in cortical area in CD vs controls that did not rely on z scores and incorporated the absolute change in tibia length. We generated multivariate linear regression models for the change in cortical area relative to change in tibia length, adjusted for baseline cortical area, baseline tibia length, age, sex, and sex-by-Tanner-stage interactions using longitudinal data in the CD and reference participants. The assumptions for the regression models were assessed using quantile-quantile plots to assess the normality of residuals.
Results
Participant characteristics
Ninety participants enrolled and 76 completed the 12-month visit. Baseline demographics, PCDAI score, height, and BMI z scores did not differ in those with vs without followup. Two participants were excluded for incomplete pQCT data, leaving 74 participants. Participant characteristics are summarized in Table 1 and notable for height and BMI z scores significantly lower than the mean (±SD) values of 0.30 ± 0.87 and 0.26 ± 0.94, respectively, in reference participants.
Table 1.
Participant Characteristics
| Characteristic | Statistic |
|---|---|
| N | 74 |
| Age, y | 14 (5–21) |
| Sex, N (% Male) | 47 (64%) |
| Race, N (%) | |
| White | 61 (82%) |
| Black | 9 (12%) |
| Other | 4 (6%) |
| Tanner, N (%) | |
| Stage I–II | 29 (39%) |
| Stage III–IV | 32 (43%) |
| Stage V | 13 (18%) |
| Height, z score | −0.57 ± 1.02 |
| BMI, z score | −0.14 ± 1.07 |
| Age at Diagnosis, y | 11.8 (5–17.6) |
| Interval since Diagnosis, y | 2.1 (0.02–9.7) |
| Disease location | |
| Ileal | 4 (5%) |
| Colonic | 20 (27%) |
| Ileocolonic | 50 (68%) |
| Disease descriptor | |
| Isolated upper | 62 (84%) |
| Perianal | 28 (38%) |
Data are presented as N (%), median (range), or mean ± SD.
Clinical course
Fifty percent of participants had severe disease activity at enrollment, with marked improvement to 6% at 10 weeks and 7% at 12 months (Table 2). Hematocrit, serum albumin, and all markers of inflammation improved significantly at 10 weeks with no difference between the 10-week and 12-month values. Serum 25(OH) vitamin D levels did not change during the 10-week induction period, as described (13). Thirty-five (47%) participants were treated with glucocorticoids during the 12 months, 11 of whom started glucocorticoids after the baseline visit. Only five received glucocorticoids over the second 6-month interval. Among those treated with glucocorticoids, the median number of days on steroids was 45 (IQR, 10–142) with a median cumulative dose of 19 (IQR, 0.97–86) mg/kg. All participants started the study on infliximab, and 67 (91%) were still receiving a TNF-α inhibitor at 12 months (62, infliximab; four, adalimumab; and one, certolizumab).
Table 2.
Disease Activity, Laboratory Results, and Medications Over the Study Interval
| Baseline | 10 Weeks | 6 Months | 12 Months | |
|---|---|---|---|---|
| Disease activity | ||||
| PCDAI, N (%) | ||||
| Not active (<10) | 9 (12) | 29 (41) | 34 (48) | 32 (46) |
| Mild (10–30) | 28 (38) | 37 (53) | 32 (46) | 33 (47) |
| Moderate-severe (>30) | 37 (50) | 4 (6)a | 4 (6) | 5 (7) |
| Laboratory results | ||||
| Albumin, g/dL | 3.9 (2.3–5.2) | 4.3 (3.2–5.0)a | 4.3 (2.5–5.0) | 4.4 (2.7–5.3) |
| ESR, mm/hr | 26 (0–115) | 9 (0–65)a | 9 (0–60) | 13 (0–64) |
| CRP, mg/L | 1.3 (0.3–19.7) | 0.5 (0.3–4.2)a | 0.5 (0.3–15) | 0.5 (0.3–5.2) |
| Hematocrit, % | 36 (17–45) | 37 (30–46)b | 38 (28–43) | 38 (33–45) |
| IL-6, pg/mL | 14.4 (0.58–208) | 4.9 (0.1–398)a | — | 5.2 (0.36–82.8) |
| TNF-α, pg/mL | 7.8 (0.4–29.8) | 1.6 (0.19–14.7)a | — | 3.3 (0.2–96.1) |
| 25(OH)D, ng/mL | 26.2 (5.4–84.9) | 27.7 (2.6–81.1) | 30.4 (8.9–84.6) | 28.3 (5.0–69.7) |
| Calcium, mg/dL | 9.1 (0.4) | 9.1 (0.3) | 9.0 (0.4) | 9.0 (0.3) |
| BSAP, z score | −1.94 (−2.87–−0.78) | −0.72 (−1.60–−0.01)a | −0.34 (−1.03–0.23) | −0.18 (−1.04–0.71) |
| β-CTX, z score | −0.46 (−1.33–0.90) | 0.33 (−0.66–1.19)c | 0.47 (−0.41–1.00) | 0.38 (−0.19–0.94) |
| Medications, N (%) | ||||
| Anti-TNF-α agent | 74 (100) | 71 (97) | 68 (93) | 67 (91) |
| Glucocorticoids | 24 (32) | 10 (14) | 4 (5) | 5 (7) |
| Aminosalicylates | 59 (80) | 56 (77) | 56 (77) | 48 (65) |
| Mecaptopurine-azathioprine | 27 (36) | 16 (22) | 10 (14) | 6 (8) |
| Methotrexate | 15 (20) | 18 (24) | 17 (23) | 18 (24) |
Data are presented as N (%), median (range), median (intraquartile range, IQR), or mean ± SD.
Calcium levels were corrected for serum albumin.
, P < .001;
, P < .05; and
, P < .01 at 10 weeks compared with baseline.
Three children sustained fractures (one radius, one tibia, and one finger) during the 12-month study.
Anthropometric and pQCT outcomes
Table 3 and Figure 1 summarize anthropometry and pQCT outcomes in CD participants. The median (IQR) for the change in height was 4.9 (1.1–7.9) cm over the 12-month study interval. Although height and BMI z scores improved significantly over the 12-month interval, height z scores were still significantly lower at 12 months compared with reference participants. Baseline PCDAI score was not associated with baseline height or BMI z score. However, greater decreases in PCDAI scores were associated with greater increases in BMI (P < .001) and height (P < .05) z scores. None of the 10-week changes in laboratory parameters or measures of glucocorticoid exposure was associated with changes in height or BMI z score.
Table 3.
Anthropometry and pQCT Data at Baseline and 12 Months
| z Score | Visit |
Baseline Versus 12-Month P Value | ||
|---|---|---|---|---|
| Baseline | 12 Months | Change | ||
| Heighta | <.001 | |||
| SD | −0.60 ± 1.00b | −0.37 ± 0.98b | 0.23 ± 0.34 | |
| Range | −2.83–2.31 | −2.60–2.47 | −0.41–1.21 | |
| BMI | <.001 | |||
| SD | −0.14 ± 1.07c | 0.29 ± 1.03 | 0.43 ± 0.69 | |
| Range | −3.02–2.30 | −2.51–2.36 | −1.40–2.16 | |
| Trabecular BMD | <.001 | |||
| SD | −1.44 ± 1.11b | −0.99 ± 1.09b | 0.45 ± 0.76 | |
| Range | −3.90–1.24 | −4.09–2.06 | −0.73–2.27 | |
| Cortical BMD | <.001 | |||
| SD | 0.19 ± 1.08 | −0.70 ± 1.10b | −0.89 ± 1.03 | |
| Range | −2.07–3.27 | −3.32–1.65 | −3.45–0.63 | |
| Cortical area | <.001 | |||
| SD | −0.97 ± 1.35b | −0.68 ± 1.25c | 0.29 ± 0.65 | |
| Range | −4.45–1.85 | −4.10–1.38 | −0.71–2.48 | |
| Periosteal circumference | .18 | |||
| SD | −0.48 ± 1.00b | −0.44 ± 1.02c | 0.04 ± 0.28 | |
| Range | −3.13–2.23 | −3.25–1.79 | −0.63–0.93 | |
| Endocortical circumference | <.001 | |||
| SD | 0.21 ± 1.07d | 0.00 ± 1.01 | −0.21 ± 0.37 | |
| Range | −2.09–2.39 | −2.77–2.06 | −1.10–0.43 | |
| Muscle area | <.001 | |||
| SD | −0.81 ± 1.10b | −0.35 ± 1.06 | 0.46 ± 0.78 | |
| Range | −3.24–1.52 | −2.71–1.92 | −1.36–2.97 | |
Limited to those with presumed growth potential, age <20 years at baseline (71 participants).
, P < .001;
, P < .01; and
, P < .05 compared with reference participants.
Figure 1.
PCDAI scores and trabecular BMD and corical area z scores following initiation of infliximab therapy. Greater declines in PCDAI scores over the first 10 weeks were significantly associated with greater gains in trabecular BMD and cortical area z scores over the 12-month study. The dashed lines represent the median z scores in the reference participants.
Trabecular BMD
Trabecular BMD z scores were markedly lower than reference participants at the time of infliximab initiation with a mean of −1.44 (P < .001 vs reference participants; Table 3), and inversely associated with PCDAI scores (R = −0.39; P < .001). The baseline z score was not associated with age or CD duration. Trabecular BMD z scores increased an average of 0.47 during the 12-month interval; however, significant deficits remained at 12 months. Younger age (P < .001) and greater declines in PCDAI (P = .03) were associated with greater increases in trabecular BMD z score (Model R2 = 0.37). For example, the mean increase in z score was 0.75 and 0.22 among those below vs above the median age of 14 years (P < .01), respectively. Tanner stage was not associated with changes in BMD z scores independent of age.
Linear regression models demonstrated greater 10-week decreases in CRP levels (P = .03) were associated with greater increases in trabecular BMD z scores, independent of age, but not independent of change in PCDAI. Changes in serum 25(OH) vitamin D levels over the first 10 weeks were positively and independently associated with changes in trabecular BMD z scores (P = .02). Glucocorticoid exposure and changes in the remaining laboratory parameters (with the exception of CRP) were not associated with changes in trabecular BMD z score.
Cortical dimensions
Study participants had significantly lower cortical area z scores at enrollment with a mean of −0.97 as a consequence of the significantly lower periosteal circumference (−0.48) and greater endocortical circumference (+0.21) z scores, compared with the reference participants (Table 3). The endocortical deficits in CD were even more pronounced when adjusted for their small periosteal circumference (β-coefficient = +0.59; P < .001 vs the reference participants). Baseline z scores were not associated with age or CD duration. Cortical area z scores improved during the study interval as a result of significant decreases in endocortical circumference z scores (both P < .001). Periosteal circumference z scores did not improve and, as a result, cortical area z scores deficits only partially corrected. The multivariable models for absolute changes in cortical area, adjusted for changes in tibia length, age, sex, and Tanner stage confirmed that the gains in cortical area were significantly greater in CD, compared with reference participants (P < .001), independent of growth.
Both greater increases in tibia length (P < .001) and greater declines in PCDAI score (P = .03) were independently associated with decreases in endocortical circumference z scores (Model R2 = 0.46). Changes in endocortical circumference z scores did not differ according to sex (change in females vs males: β, −0.06;95% confidence interval [CI], −0.21–0.09; P = .44 in multivariate regression model for 12-month change in endocortical circumference). Neither change in PCDAI nor tibia length was associated with changes in periosteal circumference z scores. Glucocorticoid exposure and changes in inflammatory parameters and vitamin D levels were not associated with changes in cortical dimension z scores.
Figure 2 summarizes the associations between changes in tibia length and changes in cortical outcome z scores. The positive relation between increases in tibia length and cortical area z scores was due to greater recovery of endocortical bone in those with greater concurrent growth. The images in Figure 3 illustrate changes in trabecular BMD and endocortical dimensions.
Figure 2.
Associations between changes in cortical z scores and changes in tibia length. Only change in periosteal circumference was not significantly associated with change in tibia length (P = .86). The associations shown below remained significant in models adjusted for baseline cortical z score.
Figure 3.
Change in trabecular BMD and endocortical circumference in a 14-year-old male participant with a decrease in PCDAI from 45 to 0 and an increase in tibia length from 328 to 339 mm over the 12-month interval. The endocortical circumference decreased from 37 (A) to 35 (B) mm, corresponding with an improvement in z score from 1.35 to 0.85. Trabecular BMD increased from 174 (C) to 213 (D) gm/cm3, corresponding with an increase in trabecular BMD z score from −2.72 to −1.44.
Cortical BMD
At baseline, cortical BMD z scores were not significantly different from the reference participants, were not associated with PCDAI score, and were higher in younger participants (correlation of cortical BMD z-scores with age, R = −0.38; P < .001). During the course of the year, cortical BMD z scores decreased an average of 0.89, and at 12 months were significantly reduced, compared with the reference participants (P < .001). In linear regression models, greater increases in cortical area z scores (β, −0.95; 95% CI, −0.48–−0.1; P < .0001) and tibia length (β, −0.33; 95% CI, −0.45–−0.2; P = .001) were independently associated with greater declines in cortical BMD z scores (model R2 = 0.71). Glucocorticoid exposure, changes in 25(OH) vitamin D level, PCDAI score, and measures of inflammation were not associated with changes in cortical BMD z scores independent of changes in tibia length and cortical area z score.
Functional muscle-bone unit
At enrollment, periosteal circumference and muscle area z scores were positively associated (R = 0.51; P < .001). Greater increases in muscle area z scores were associated with greater increases (or lesser decreases) in periosteal circumference z scores (β, 0.14; 95% CI, 0.05–0.23; P < .01) during the 12 months. The gains in periosteal circumference z score for a given increase in muscle area z score, adjusted for baseline values, did not differ between CD and longitudinal reference participants (P = .82).
Bone biomarkers and pQCT outcomes
Bone biomarker (β-CTX and BSAP) z scores were low at baseline and increased significantly over 10 weeks (Table 2). The median (IQR) percent change in the absolute values over the first 10 weeks was 22% (−18; 105%) for β-CTX and 75% (35; 189%) for BSAP. Ten-week increases in BSAP and β-CTX levels were associated with 12-month increases in height (BSAP: R = 0.54; P < .0001; β-CTX: R = 0.42; P < .001), trabecular BMD (BSAP: R = 0.44; P < .001; β-CTX: R = 0.33; P < .01), cortical area (BSAP: R = 0.41; P < .001; β-CTX: R = 0.32; P < .01) z scores and with decreases in cortical BMD (BSAP: R = −0.55; P < .0001; β-CTX: R = −0.52; P < .0001) and endocortical circumference (BSAP: R = −0.50; P < .0001; β-CTX: R = −0.45; P = .0001) z scores.
The magnitude of increases in bone biomarker levels was comparable in participants who were not treated with glucocorticoids throughout induction compared with those in whom glucocorticoids were discontinued during the induction period.
Discussion
This study is the first to examine changes in volumetric BMD and cortical structure (endocortical and periosteal circumference and cortical area) following initiation of anti-TNF-α therapy in children and adolescents with CD. The data demonstrated rapid improvements in disease activity and increases in BSAP and β-CTX levels. The declines in PCDAI scores and increases in bone biomarker levels were significantly associated with subsequent improvements in height, trabecular BMD, and cortical area z scores over 12 months. Improvements in cortical area were a consequence of recovery of endocortical bone deficits. Reductions in cortical BMD z scores were consistent with our prior observation rapid bone accrual results in declines in cortical BMD, as detailed below (4, 25).
Improvements in trabecular BMD were greater in younger participants, and the recovery of cortical area on the endocortical surface was greatest in those with the greatest linear growth, documented by increasing tibia length. These data suggest that childhood provides a window of opportunity for recovery of trabecular and endocortical deficits. Periosteal circumference z scores did not improve. We had hypothesized that periosteal deficits in CD were related to TNF-α and that glucocorticoid effects inhibit bone formation during growth and would improve during infliximab therapy. However, the fact that periosteal circumference z scores (which were adjusted for tibia length at each visit) did not decline suggests that the participants were able to increase periosteal circumference commensurate with the increases in tibia length during the catchup growth on anti-TNF-α therapy. This was not the case during the first 6 months of CD therapy in our cohort enrolled at diagnosis, when periosteal circumference z scores decreased (only eight of 78 were treated with infliximab) (3). CD participants' gains in periosteal circumference relative to gains in muscle area were comparable to those observed in the longitudinal reference data. This suggests the functional muscle bone unit is intact during anti-TNF-α therapy, potentially related to recovery from the detrimental effects of TNF-α on osteocytes (26).
In healthy children and adolescents, cortical modeling in the diaphysis is characterized by bone formation on the periosteal surface. The primary bone laid down by periosteal osteoblasts subsequently undergoes intracortical remodeling, resulting in a transient decline in cortical BMD (27). The decline in cortical BMD z score observed following initiation of infliximab therapy in this CD cohort was likely a consequence of the rapid increases in periosteal bone formation necessary to keep pace with catchup growth. Cellular activity at the endocortical surface differs in males and females during puberty. In males, endocortical resorption enlarges the marrow cavity. In contrast, when females enter puberty, endocortical bone formation is stimulated, narrowing the medullary cavity (28). Hence, the declines (improvements) in endocortical circumference z scores may be consequences of different mechanisms depending upon sex. Although decreases in endocortical bone deficits in males may have been due to cessation of TNF-α, declines in females may have been due to restoration of expected pubertal effects. Of note, the decreases in endocortical circumference z scores did not differ according to sex.
These data illustrate the challenges of interpreting bone biomarkers in children and adolescents. We previously reported biomarkers of bone formation and resorption were positively and independently associated with height velocity and bone accrual velocity in reference participants (23). Similar associations were observed here, as greater short-term increases in BSAP and β-CTX were associated with greater catchup growth and increases in trabecular BMD and cortical area z scores. Despite the fact β-CTX is a biomarker of bone resorption, higher levels were associated with greater recovery of endocortical bone. These striking short-term increases in biomarkers of formation and resorption likely reflect recovery of the bone cell activity during accelerated linear growth and skeletal modeling, masking any effects of infliximab to decrease bone resorption in remodeling bone. These findings are in sharp contrast with results in adults with CD, reporting modest (15–26%) increases in biomarkers of bone formation and variable decreases in biomarkers of bone resorption following infliximab initiation (29–33). The study had multiple limitations. First, we did not obtain bone biopsies. Conventional pQCT has insufficient resolution to characterize disease and treatment effects on cortical and trabecular mineralization and microarchitecture. Future studies using high-resolution pQCT are needed to assess bone microarchitecture, cortical porosity, and finite element estimates of bone strength. Second, the 1-year duration was insufficient to determine whether periosteal dimensions eventually prove above and beyond the increases necessary to keep pace with catchup linear growth (24). Third, anti-TNF-α medications are the standard of care for children and adolescents with moderate-to-severe CD. Therefore, it was not possible to enroll CD patients with comparable disease severity who were not treated with anti-TNF-α as untreated controls. Fourth, we did not obtain thoracolumbar radiographs to assess for vertebral morphology (34, 35). An additional limitation is lack of data on sex steroids including changes in sex steroids during treatment and associations with changes in pQCT parameters. Finally, we are unable to determine whether the improvements in bone density and structure were a direct consequence of the blockade of TNF-α effects on bone cells. Alternative explanations include improvements in disease activity, nutrition and physical activity, decreases in IL-6 levels, and withdrawal of mercaptopurine/azathioprine and glucocorticoid therapy. However, only 32% of participants were on glucocorticoids during the first 10 weeks of therapy, and improvements in bone z scores did not differ in those who did vs did not use glucocorticoids over the study interval.
This study has many important strengths. To our knowledge, this is the first study to assess the potential for recovery of volumetric BMD and cortical structure following initiation of anti-TNF-α therapy in any pediatric disorder. Second, the use of pQCT, as opposed to duel-energy x-ray absorptiometry, provided insights into discrete changes in trabecular and cortical volumetric BMD and cortical dimensions. Third, the availability of a large and robust reference dataset (including longitudinal data) enabled us to adjust for age, sex, race, tibia length, and maturation. Fourth, the study included comprehensive assessment of concurrent glucocorticoid exposure and laboratory biomarkers.
In conclusion, anti-TNF-α therapy during growth and development is associated with rapid improvements in trabecular BMD and cortical structure. Future studies are needed to examine the effect on fracture risk.
Acknowledgments
This work was supported by National Institutes of Health Grants K23 DK082012 (to M.T.), K23 DK093556 (to M.R.D.), K24 DK076808 (to M.B.L.), and K08 HD060739 (to M.D.D.); the Clinical and Translational Science Award UL1RR024134 and UL1TR000003; and by the Penn Joint Center for Inflammatory Bowel Diseases. M.T. and M.B.L.'s research contributions were completed while they were affiliated with the Children's Hospital of Philadelphia.
Disclosure Summary: M.R.D. reports grants from NIDDK and The NephCure Foundation (American Society of Nephrology Research grant) during the conduct of the study and grants from Genentech and personal fees from Infiniti Medical, outside the submitted work. R.N.B reports personal fees from Janssen, AbbVie, and Takeda, outside the submitted work. L.A.D. reports personal fees from Imedex, Glycosyn LLC, Avaxia Biologics, outside the submitted work. In addition, L.A.D. has a patent use of Glycans and Glycosyltransferases for Diagnosing/Monitoring Inflammatory Bowel Disease licensed to Glycosyn LLC, a patent Gene Array Technique for Predicting Response in Inflammatory Bowel Diseases issued, and a patent Serological Markers of Inflammatory Bowel Disease Phenotype and Disease Progression issued. L.M.G., M.T., M.D.T., B.S.Z., J.S., R.H., J.L., and M.B.L. have nothing to disclose.
Footnotes
- β-CTX
- C-terminal telopeptide of type 1 collagen
- BMD
- bone mineral density
- BMI
- body mass index
- BSAP
- bone-specific alkaline phosphatase
- CD
- Crohn's disease
- CHOP
- Children's Hospital of Philadelphia
- CI
- confidence interval
- CV
- coefficient of variation
- ESR
- erythrocyte sedimentation rate
- IQR
- interquartile range
- PCDAI
- Pediatric Crohn Disease Activity Index
- pQCT
- peripheral quantitative computed tomography.
References
- 1. Semeao EJ, Stallings VA, Peck SN, Piccoli DA. Vertebral compression fractures in pediatric patients with Crohn's disease. Gastroenterology. 1997;112(5):1710–1713. [DOI] [PubMed] [Google Scholar]
- 2. Laakso S, Valta H, Verkasalo M, Toiviainen-Salo S, Makitie O. Compromised peak bone mass in patients with inflammatory bowel disease—A prospective study. J Pediatr. 2014;164(6):1436–1443.e1. [DOI] [PubMed] [Google Scholar]
- 3. Dubner SE, Shults J, Baldassano RN, et al. Longitudinal assessment of bone density and structure in an incident cohort of children with Crohn's disease. Gastroenterology. 2009;136(1):123–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tsampalieros A, Lam CK, Spencer JC, et al. Long-term inflammation and glucocorticoid therapy impair skeletal modeling during growth in childhood Crohn disease. J Clin Endocrinol Metab. 2013;98(8):3438–3445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Vazquez MA, Lopez E, Montoya MJ, Giner M, Perez-Temprano R, Perez-Cano R. Vertebral fractures in patients with inflammatory bowel disease compared with a healthy population: A prospective case-control study. BMC Gastroenterol. 2012;12:47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Bernstein CN, Blanchard JF, Leslie W, Wajda A, Yu BN. The incidence of fracture among patients with inflammatory bowel disease. A population-based cohort study. Ann Intern Med. 2000;133(10):795–759. [DOI] [PubMed] [Google Scholar]
- 7. Leonard MB, Elmi A, Mostoufi-Moab S, et al. Effects of sex, race, and puberty on cortical bone and the functional muscle bone unit in children, adolescents, and young adults. J Clin Endocrinol Metab. 2010;95(4):1681–1689. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Osta B, Benedetti G, Miossec P. Classical and Paradoxical Effects of TNF-alpha on Bone Homeostasis. Front Immunol. 2014;5:48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lane NE, Yao W. Glucocorticoid-induced bone fragility. Ann N Y Acad Sci. 2010;1192:81–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hyams J, Walters TD, Crandall W, et al. Safety and efficacy of maintenance infliximab therapy for moderate-to-severe Crohn's disease in children: REACH open-label extension. Curr Med Res Opin. 2011;27(3):651–662. [DOI] [PubMed] [Google Scholar]
- 11. Walters TD, Kim MO, Denson LA, et al. Increased effectiveness of early therapy with anti-tumor necrosis factor-alpha vs an immunomodulator in children with Crohn's disease. Gastroenterology. 2014;146(2):383–391. [DOI] [PubMed] [Google Scholar]
- 12. Malik S, Wong SC, Bishop J, et al. Improvement in growth of children with Crohn disease following anti-TNF-α therapy can be independent of pubertal progress and glucocorticoid reduction. J Pediatr Gastroenterol Nutr. 2011;52(1):31–37. [DOI] [PubMed] [Google Scholar]
- 13. Augustine MV, Leonard MB, Thayu M, et al. Changes in vitamin D-related mineral metabolism after induction with anti-tumor necrosis factor-alpha therapy in Crohn's disease. J Clin Endocrinol Metab. 2014;99(6):E991–E998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Thayu M, Denson LA, Shults J, et al. Determinants of changes in linear growth and body composition in incident pediatric Crohn's disease. Gastroenterology. 2010;139(2):430–438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Hyams JS, Ferry GD, Mandel FS, et al. Development and validation of a pediatric Crohn's disease activity index. J Pediatr Gastroenterol Nutr. 1991;12(4):439–447. [PubMed] [Google Scholar]
- 16. Morris NM, Udry JR. Validation of a self-administered instrument to assess stage of adolescent development. J Youth Adolesc. 1980;9(3):271–280. [DOI] [PubMed] [Google Scholar]
- 17. Lee DY, Wetzsteon RJ, Zemel BS, et al. Muscle Torque Relative to Cross-Sectional Area and the Functional Muscle-Bone Unit in Children and Adolescents with Chronic Disease. J Bone Miner Res. 2015;30(3):563–571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Leonard MB, Zemel BS, Wrotniak BH, et al. Tibia and radius bone geometry and volumetric density in obese compared to non-obese adolescents. Bone. 2015;73:69–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Kontulainen SA, Johnston JD, Liu D, Leung C, Oxland TR, McKay HA. Strength indices from pQCT imaging predict up to 85% of variance in bone failure properties at tibial epiphysis and diaphysis. J Musculoskelet Neuronal Interact. 2008;8(4):401–409. [PubMed] [Google Scholar]
- 20. Adams JE, Engelke K, Zemel BS, Ward KA. Quantitative computer tomography in children and adolescents: The 2013 ISCD Pediatric Official Positions. J Clin Densitom. 2014;17(2):258–274. [DOI] [PubMed] [Google Scholar]
- 21. Ogden CL, Kuczmarski RJ, Flegal KM, et al. Centers for Disease Control and Prevention 2000 growth charts for the United States: Improvements to the 1977 National Center for Health Statistics version. Pediatrics. 2002;109(1):45–60. [DOI] [PubMed] [Google Scholar]
- 22. Cole TJ. The LMS method for constructing normalized growth standards. Eur J Clin Nutr. 1990;44(1):45–60. [PubMed] [Google Scholar]
- 23. Tuchman S, Thayu M, Shults J, Zemel BS, Burnham JM, Leonard MB. Interpretation of biomarkers of bone metabolism in children: Impact of growth velocity and body size in healthy children and chronic disease. J Pediatr. 2008;153(4):484–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Mostoufi-Moab S, Brodsky J, Isaacoff EJ, et al. Longitudinal assessment of bone density and structure in childhood survivors of acute lymphoblastic leukemia without cranial radiation. J Clin Endocrinol Metab. 2012;97(10):3584–3592. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Wetzsteon RJ, Zemel BS, Shults J, Howard KM, Kibe LW, Leonard MB. Mechanical loads and cortical bone geometry in healthy children and young adults. Bone. 2011;48(5):1103–1108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Bonewald LF. The amazing osteocyte. J Bone Miner Res. 2011;26(2):229–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Rauch F, Travers R, Glorieux FH. Cellular activity on the seven surfaces of iliac bone: A histomorphometric study in children and adolescents. J Bone Miner Res. 2006;21(4):513–519. [DOI] [PubMed] [Google Scholar]
- 28. Seeman E. Clinical review 137: Sexual dimorphism in skeletal size, density, and strength. J Clin Endocrinol Metab. 2001;86(10):4576–4584. [DOI] [PubMed] [Google Scholar]
- 29. Abreu MT, Geller JL, Vasiliauskas EA, et al. Treatment with infliximab is associated with increased markers of bone formation in patients with Crohn's disease. J Clin Gastroenterol. 2006;40(1):55–63. [DOI] [PubMed] [Google Scholar]
- 30. Franchimont N, Putzeys V, Collette J, et al. Rapid improvement of bone metabolism after infliximab treatment in Crohn's disease. Aliment Pharmacol Ther. 2004;20(6):607–614. [DOI] [PubMed] [Google Scholar]
- 31. Miheller P, Muzes G, Rácz K, et al. Changes of OPG and RANKL concentrations in Crohn's disease after infliximab therapy. Inflamm Bowel Dis. 2007;13(11):1379–1384. [DOI] [PubMed] [Google Scholar]
- 32. Ryan BM, Russel MG, Schurgers L, et al. Effect of antitumour necrosis factor-alpha therapy on bone turnover in patients with active Crohn's disease: A prospective study. Aliment Pharmacol Ther. 2004;20(8):851–857. [DOI] [PubMed] [Google Scholar]
- 33. Veerappan SG, O'Morain CA, Daly JS, Ryan BM. Review article: The effects of antitumour necrosis factor-alpha on bone metabolism in inflammatory bowel disease. Aliment Pharmacol Ther. 2011;33(12):1261–1272. [DOI] [PubMed] [Google Scholar]
- 34. Laakso S, Valta H, Verkasalo M, Toiviainen-Salo S, Viljakainen H, Mäkitie O. Impaired bone health in inflammatory bowel disease: A case-control study in 80 pediatric patients. Calcif Tissue Int 2012;91(2):121–130. [DOI] [PubMed] [Google Scholar]
- 35. Leblanc CM, Ma J, Scuccimarri R. A154: Glucocorticoid therapy and the risk of incident vertebral fracture in children with rheumatic disorders. Arthritis Rheumatol. 2014;66 Suppl 11:S199–S200. [Google Scholar]