Patients undergoing treatment for acute lymphoblastic leukemia (ALL) are at risk for bone morbidities such as osteonecrosis and low bone mineral density (BMD). We evaluated the incidence of hip and knee osteonecrosis (scored as ≥30% or <30% epiphyseal involvement), identified prospectively by magnetic resonance imaging (MRI) regardless of symptoms, and the associated risk factors in 334 children (aged 2–18 years at diagnosis) with ALL who were enrolled on the Total Therapy Study XV (clinicaltrials.gov identifier: NCT00137111) (Pui, et al 2009). We also studied the association of hip and knee osteonecrosis with bone mineral density (BMD) Z-scores as evaluated with quantitative computed tomography at the first and second lumbar vertebral levels.
The Total XV treatment, evaluations of osteonecrosis and BMD, and statistical analysis are described in the supporting document. MRI screening was performed at the completion of reinductions I and II and at the completion of therapy (Kaste, et al 2015). BMD was determined at diagnosis, after 120 weeks of continuation therapy, and 2 years after the completion of therapy, and was standardized to age- and sex-specific Z-scores (Gilsanz, et al 1988). We defined a BMD Z-score of –2.0 or lower as indicating profoundly low BMD.
The demographics of 334 patients are shown in Table I and Figure S1. Hip osteonecrosis was seen in 40 patients (12.0%): 24 (7.2%) with ≥30% epiphyseal involvement and 16 (4.8%) with <30% epiphyseal involvement. Knee osteonecrosis was observed in 198 patients (59.3%): 51 (15.3%) with ≥30% epiphyseal involvement and 147 (44.0%) with <30% epiphyseal involvement, which prevalence was 2.1 and 5.0 times that of the respective hip osteonecrosis (Tables I and SI). Multivariable analysis showed that older age (≥10 years) at diagnosis was significantly associated with higher incidences of osteonecrosis with ≥30% hip epiphyseal involvement (P = 0.007), any hip epiphyseal involvement (P < 0.001), and ≥30% knee epiphyseal involvement (P< 0.001) when compared to the incidences in patients aged 2–9.9 years (Table I). For knee osteonecrosis, standard/high-risk disease was significantly associated with higher incidences of ≥30% epiphyseal involvement (P = 0.001) and any epiphyseal involvement (P = 0.032) when compared to the low-risk disease.
Table I.
Patient characteristics and factors associated with hip and knee osteonecrosis
| Categorical Variables | Patients (n = 334) | Hip Osteonecrosis, >30% Epiphyseal Involvement | Hip Osteonecrosis, Any (≥30% or < 30%) Epiphyseal Involvement | Knee Osteonecrosis, >30% Epiphyseal Involvement | Knee Osteonecrosis, Any (≥30% or <30%) Epiphyseal Involvement | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Events (n = 24) | OR | 95% Cl | Univariate P |
Multi variable P |
Events (n = 40) | OR | 95% Cl | Univariate P |
Multivariable P |
Events (n = 51) | OR | 95% Cl | Univariate P |
Multivariable P |
Events (n = 198) | OR | 95% Cl | Univariate P |
Multivariable P |
||
| Age | <0.001 | 0.007 | <0.001 | <0.001 | <0.001 | <0.001 | 0.001 | 0.073 | |||||||||||||
| 2–9.9 years | 243 | 6 | 1 00 | 12 | 1.00 | 14 | 1.00 | -- | 130 | 1.00 | -- | ||||||||||
| ≥10 years | 91 | 18 | 9.74 | 3.73–25.45 | 28 | 8.56 | 4.12–17.78 | 37 | 11.21 | 5.66–22.18 | 68 | 2.57 | 1.50–4.39 | ||||||||
| Sex | 0.301 | 0.893 | 0.865 | 0.797 | |||||||||||||||||
| Female | 147 | 13 | 1.00 | 18 | 1.00 | 23 | 1.00 | -- | 86 | 1.00 | -- | ||||||||||
| Male | 187 | 11 | 0.64 | 0.28–1.48 | 22 | 0.96 | 0.49–1.86 | 28 | 0.95 | 0.52–1.73 | 112 | 1.06 | 0.68–164 | ||||||||
| Race | 0.823 | 0.967 | 0.045 | 0.459 | 0.183 | ||||||||||||||||
| White | 236 | 16 | 1.00 | 28 | 1.00 | 42 | 1.00 | -- | 141 | 1.00 | -- | ||||||||||
| Black | 54 | 5 | 1.40 | 0.49–4.01 | 7 | 1.11 | 0.46–2.69 | 3 | 0.27 | 0.08–0.91 | 27 | 0.67 | 0.37–1.22 | ||||||||
| Other | 44 | 3 | 1.01 | 0.28–3.61 | 5 | 0.95 | 0.35–2.62 | 6 | 0.73 | 0.29–1.84 | 30 | 1.44 | 0.73–2.87 | ||||||||
| Risk | 0.003 | 0.355 | <0.001 | 0.253 | <0.001 | 0.001 | <0.001 | 0.032 | |||||||||||||
| Low | 176 | 5 | 1.00 | 9 | 1.00 | 7 | 1.00 | -- | 88 | 1.00 | -- | ||||||||||
| High/standard | 158 | 19 | 4.67 | 1.70–12.84 | 31 | 4.53 | 2.08–9.85 | 44 | 9.32 | 4.06–21.41 | 110 | 2.29 | 1.46–3.59 | ||||||||
| Lineage | 0.192 | 0.084 | 0.854 | 0.037 | 0.727 | 0.115 | |||||||||||||||
| B | 282 | 18 | 1.00 | 30 | 1.00 | 38 | 1.00 | -- | 162 | 1.00 | -- | ||||||||||
| T | 52 | 6 | 1.91 | 0.72–5.07 | 10 | 2.00 | 0.91–4.39 | 13 | 2.14 | 1.05–4.37 | 36 | 1.67 | 0.88–3.14 | ||||||||
| WBC count at diagnosis | 0.569 | 0.378 | 0.516 | 0.802 | |||||||||||||||||
| <50 × 109/L | 248 | 19 | 1.00 | 32 | 1.00 | 36 | 1.00 | -- | 148 | 1.00 | -- | ||||||||||
| ≥50 × 109/L | 86 | 5 | 0.74 | 0.27–2.06 | 8 | 0.69 | 0.31–1.57 | 15 | 1.24 | 0.64–2.41 | 50 | 0.94 | 0.57–1.54 | ||||||||
| Physis status at the first MRI | <0.001 | 0.250 | <0.001 | 0.533 | <0.001 | 0.174 | 0.099 | 0.815 | |||||||||||||
| Open | 298 | 15 | 1.00 | 28 | 1.00 | 34 | 1.00 | -- | 172 | 1.00 | -- | ||||||||||
| Closed | 36 | 9 | 6.29 | 2.52–15.71 | 12 | 4.82 | 2.18–10.67 | 17 | 6.95 | 3.30–14.64 | 26 | 1.90 | 0.89–4.09 | ||||||||
Abbreviations: OR, odds ratio; CI, confidence interval; WBC, white blood cell; MRI, magnetic resonance imaging. The independent effect of each factor with a P value of <0.10 in univariate analysis was determined by multivariable analysis.
Table II shows the associations between degree of osteonecrosis and BMD Z-score after adjustment for BMD Z-score at diagnosis and race. The median BMD Z-score at diagnosis for patients with knee osteonecrosis involving ≥30% of the epiphysis was significantly higher than that for the combined cohorts with <30% or no epiphyseal involvement (P = 0.026). At week 120 of continuation therapy, hip and knee osteonecrosis were associated with median BMD Z-scores that were significantly lower than those of other groups, regardless of the degree of epiphyseal involvement (P ≤ 0.007 for all). At 2 years off therapy, patients with any knee epiphyseal involvement had median BMD Z-scores that were significantly lower than those of patients without osteonecrosis (P = 0.026). When the median BMD Z-score changes from week 120 of continuation therapy to 2 years off therapy were evaluated, the increases in Z-scores were significantly greater for patients with ≥30% hip epiphyseal involvement (P = 0.018), any hip epiphyseal involvement (P = 0.004), and any knee epiphyseal involvement (P = 0.031) than for other groups. Figure S2, Table SII, and the supplemental document show the association of BMD Z-scores with hip and knee osteonecrosis categorized into three groups (no, <30%, and ≥30% epiphyseal involvement). The median BMD Z-scores were decreased at week 120, relative to the Z-scores at diagnosis, but showed subsequent improvement at 2 years off therapy; these changes were more prominent in those individuals with ≥30% epiphyseal involvement. Knee osteonecrosis with ≥30% epiphyseal involvement and with any epiphyseal involvement was significantly associated with profoundly low BMD Z-scores (≤–2.0) at week 120 (P = 0.034 and P = 0.011, respectively) (Table SIII).
Table II.
Associations of hip and knee osteonecrosis and bone mineral density Z-scores
| Osteonecrosisa and Time Points | All Patients | ON | BMD Z-Score | No ON | BMD Z-Score | Logistic Regression | Multiple Logistic Regressionb | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| N | N | Median (Range) | N | Median (Range) | OR | 95% Cl | P | OR | 95% Cl | P | |
| Diagnosis | |||||||||||
| Hip, ≥30% epiphyseal involvement | 308 | 22 | 0.55 (−2.11 to 2.31) | 286 | 0.02(−3.27 to 3.56) | 1.51 | 0.99 to 2.30 | 0.056 | 1.49 | 0.97 to 2.28 | 0.065 |
| Hip, any epiphyseal involvement | 308 | 37 | 0.07(−2.50 to 2.73) | 271 | 0.03 (−3.27 to 3.56) | 1.16 | 0.83 to 1.62 | 0.383 | 1.16 | 0.83 to 1.62 | 0.388 |
| Knee, ≥30% epiphyseal involvement | 308 | 47 | 0.23(−1.88 to 2.59) | 261 | 0.02 (−3.27 to 3.56) | 1.37 | 1.01 to 1.87 | 0.041 | 1.44 | 1.04 to 1.98 | 0.026 |
| Knee, any epiphyseal involvement Continuation week 120 | 308 | 185 | 0.00 (−2.66 to 2.68) | 123 | 0.09 (−3.27 to 3.56) | 0.89 | 0.71 to 1.10 | 0.279 | 0.90 | 0.72 to 1.12 | 0.327 |
| Hip, ≥30% epiphyseal involvement | 292 | 19 | −1.35 (−5.62 to 1.36) | 273 | −1.08 (−5.93 to 2.05) | 0.87 | 0.62 to 1.23 | 0.442 | 0.54 | 0.34 to 0.85 | 0.007 |
| Hip, any epiphyseal involvement | 292 | 31 | −1.35 (−5.93 to 1.67) | 261 | −1.07 (−4.73 to 2.05) | 0.79 | 0.60 to 1.03 | 0.082 | 0.59 | 0.42 to 0.83 | 0.003 |
| Knee, ≥30% epiphyseal involvement | 292 | 45 | −1.45 (−5.62 to 1.80) | 247 | −1.02 (−5.93 to 2.05) | 0.76 | 0.60 to 0.96 | 0.023 | 0.61 | 0.45 to 0.82 | 0.001 |
| Knee, any epiphyseal involvement | 292 | 173 | −1.45 (−5.93 to 2.05) | 119 | −0.75 (−3.96 to 1.91) | 0.67 | 0.55 to 0.82 | <0.001 | 0.60 | 0.47 to 0.77 | <0.001 |
| 2 years off therapy | |||||||||||
| Hip, ≥30% epiphyseal involvement | 232 | 17 | 0.10 (−3.56 to 2.33) | 215 | −0.77 (−3.10 to 2.46) | 1.39 | 0.93 to 2.09 | 0.110 | 0.92 | 0.54 to 1.58 | 0.773 |
| Hip, any epiphyseal involvement | 232 | 28 | −0.01 (−3.56 to 2.33) | 204 | −0.77 (−3.10 to 2.46) | 1.19 | 0.87 to 1.64 | 0.294 | 0.96 | 0.64 to 1.46 | 0.864 |
| Knee, ≥30% epiphyseal involvement | 232 | 33 | −1.05 (−3.56 to 2.33) | 199 | −0.72 (−3.09 to 2.46) | 0.91 | 0.67 to 1.23 | 0.550 | 0.77 | 0.51 to 1.16 | 0.209 |
| Knee, any epiphyseal involvement | 232 | 140 | −0.91 (−3.56 to 2.46) | 92 | −0.52 (−3.06 to 2.45) | 0.80 | 0.64 to 1.00 | 0.046 | 0.72 | 0.54 to 0.96 | 0.026 |
| Change from diagnosis to week 120 | |||||||||||
| Hip, ≥30% epiphyseal involvement | 273 | 17 | −2.21(−5.19 to-0.19) | 256 | −1.10 (−4.39 to 1.83) | 0.51 | 0.34 to 0.77 | 0.001 | NA | ||
| Hip, any epiphyseal involvement | 273 | 29 | −1.86 (−5.19 to-0.19) | 244 | −1.12 (−4.39 to 1.83) | 0.58 | 0.42 to 0.80 | 0.001 | NA | ||
| Knee, ≥30% epiphyseal involvement | 273 | 42 | −1.72 (−5.19 to 0.40) | 231 | −1.06 (−4.13 to 1.83) | 0.58 | 0.44 to 0.78 | <0.001 | NA | ||
| Knee, any epiphyseal involvement | 273 | 163 | −1.30 (−5.19 to 1.08) | 110 | −0.96 (−3.97 to 1.83) | 0.65 | 0.52 to 0.82 | <0.001 | NA | ||
| Change from week 120 to 2 years off therapy | |||||||||||
| Hip, ≥30% epiphyseal involvement | 219 | 15 | 1.09 (−0.68 to 2.45) | 204 | 0.42 (−2.23 to 4.23) | 1.91 | 1.13 to 3.25 | 0.016 | 1.99 | 1.13 to 3.50 | 0.018 |
| Hip, any epiphyseal involvement | 219 | 26 | 0.97 (−0.68 to 2.84) | 193 | 0.40 (−2.23 to 4.23) | 1.96 | 1.27 to 3.03 | 0.002 | 1.92 | 1.23 to 3.00 | 0.004 |
| Knee, ≥30% epiphyseal involvement | 219 | 32 | 0.94 (−1.36 to 2.93) | 187 | 0.44 (−2.23 to 4.23) | 1.41 | 0.96 to 2.09 | 0.083 | 1.47 | 0.98 to 2.22 | 0.064 |
| Knee, any epiphyseal involvement | 219 | 132 | 0.68 (−1.67 to 4.23) | 87 | 0.37 (−2.23 to 2.33) | 1.43 | 1.06 to 1.93 | 0.021 | 142 | 1.03 to 1.94 | 0.031 |
Abbreviations: pts, patients; ON, osteonecrosis; BMD, bone mineral density; OR, odds ratio; CI, confidence interval; NA, not applicable
Any epiphyseal involvement includes both ≥30% and <30% involvement.
P values were adjusted for bone mineral density Z-score at diagnosis and race. In this analysis, the results of continuation week 120 and change from diagnosis to week 120 are identical.
To our knowledge, this is the first prospective study to evaluate both hip and knee osteonecrosis in patients with ALL. Hip osteonecrosis and knee osteonecrosis (with ≥30% epiphyseal involvement) were significantly associated with older age at diagnosis (≥10 years). Knee osteonecrosis was more prevalent than hip osteonecrosis. Therefore, although knee involvement was seen in 68 of the 91 older patients (74.7%), it was also seen in 130 of the 243 younger patients (53.5%), and the standard/high-risk groups, which required more intensive chemotherapy, were significant for knee osteonecrosis.
The decrease in BMD Z-score was more pronounced in those individuals with osteonecrosis, especially in those with ≥30% epiphyseal involvement. The decrease in BMD during ALL treatment can be aggravated by complications resulting from osteonecrotic lesions, such as restriction from performing weight-bearing activities and destruction of bone architecture (den Hoed, et al 2015). Profoundly low BMD (BMD Z-score ≤ −2.0) was associated only with knee osteonecrosis. In our cohort, older patients had significantly higher BMD Z-scores at diagnosis than did younger patients (Inaba, et al 2018). However, knee osteonecrosis was seen not only in patients aged 10 years or older but also in younger patients. Therefore, younger patients with knee osteonecrosis can develop profoundly low BMD. There were improvements in BMD Z-scores at 2 years off therapy, but the scores did not return to the baseline values at diagnosis and additional follow-up is necessary.
Although BMD tends to improve after therapy, large osteonecrosis lesions typically do not resolve, especially in older patients (Kaste, et al 2015). Prospective MRI monitoring for osteonecrosis and BMD evaluation can help the incorporation of early interventions such as enhanced physical activities, pharmaceutical approaches (e.g., with bisphosphonates or vitamin D supplementation), and the adjustment of chemotherapeutic agents, although the effects of such interventions need to be confirmed in prospective studies. As prospective MRI evaluation may not be feasible, patients should undergo BMD assessment when they develop symptomatic osteonecrosis because of its strong associations with BMD, and MRI can be used to evaluate osteonecrosis in patients with reduced BMD.
Our prospective monitoring of osteonecrosis identified a higher incidence of osteonecrosis in the knee than in the hip, the clinical risk factors, and associations of knee and hip osteonecrosis with decreased BMD Z-scores. Further identification of clinical and pharmacogenomic risk factors for osteonecrosis and low BMD may help to better identify high-risk patients who would benefit from supportive care.
Supplementary Material
Acknowledgments
The authors thank Keith A. Laycock, PhD, ELS, for scientific editing of the manuscript.
Funding sources: This work was supported by grants CA21765, CA02394, and P50 M115279 from the National Institutes of Health and by ALSAC. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Conflict of interest disclosures: Hiroto Inaba and Mary V. Relling receive research funding from Servier. The authors have declared no other conflicts of interest.
REFERENCES
- den Hoed MA, Pluijm SM, te Winkel ML, de Groot-Kruseman HA, Fiocco M, Hoogerbrugge P, Leeuw JA, Bruin MC, van der Sluis IM, Bresters D, Lequin MH, Roos JC, Veerman AJ, Pieters R & van den Heuvel-Eibrink MM (2015) Aggravated bone density decline following symptomatic osteonecrosis in children with acute lymphoblastic leukemia. Haematologica, 100, 1564–1570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gilsanz V, Gibbens DT, Roe TF, Carlson M, Senac MO, Boechat MI, Huang HK, Schulz EE, Libanati CR & Cann CC (1988) Vertebral bone density in children: effect of puberty. Radiology, 166, 847–850. [DOI] [PubMed] [Google Scholar]
- Inaba H, Cao X, Han AQ, Panetta JC, Ness KK, Metzger ML, Rubnitz JE, Ribeiro RC, Sandlund JT, Jeha S, Cheng C, Pui CH, Relling MV & Kaste SC (2018) Bone mineral density in children with acute lymphoblastic leukemia. Cancer, 124, 1025–1035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaste SC, Pei D, Cheng C, Neel MD, Bowman WP, Ribeiro RC, Metzger ML, Bhojwani D, Inaba H, Campbell P, Rubnitz JE, Jeha S, Sandlund JT, Downing JR, Relling MV, Pui CH & Howard SC (2015) Utility of early screening magnetic resonance imaging for extensive hip osteonecrosis in pediatric patients treated with glucocorticoids. J Clin Oncol, 33, 610–615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pui CH, Campana D, Pei D, Bowman WP, Sandlund JT, Kaste SC, Ribeiro RC, Rubnitz JE, Raimondi SC, Onciu M, Coustan-Smith E, Kun LE, Jeha S, Cheng C, Howard SC, Simmons V, Bayles A, Metzger ML, Boyett JM, Leung W, Handgretinger R, Downing JR, Evans WE & Relling MV (2009) Treating childhood acute lymphoblastic leukemia without cranial irradiation. N Engl J Med, 360, 2730–2741. [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.