Abstract

Background

Osteonecrosis (ON) is a severely disabling complication of anti-leukemic therapy, specifically long-term corticosteroid use. A hypercoagulable state is thought to underlie corticosteroid-related ON. Children with acute lymphoblastic leukemia (ALL) are also at increased risk of venous thromboembolism (VTE), indicating underlying hypercoagulability in this disease entity. Hence, we explored the relationship between ON and VTE, along with the association of ON with other variables, including age and asparaginase (ASP) therapy, in children with ALL.

Methods

Health records of children (< 18 yrs.) with de novo ALL treated at McMaster Children’s Hospital from 1992 to 2010 were reviewed. Patients were treated according to Dana-Farber Cancer Institute (DFCI) ALL Consortium Protocols. Data regarding demographics, leukemia diagnosis and therapy, development and characteristics of ON and VTE, and thrombophilia work-up, if any, were collected from computer records and chart review.

Osteonecrosis was diagnosed by plain X-ray, computed tomography (CT), magnetic resonance (MR) imaging, and/or technetium-99m (99mTc) bone scan. We included ON diagnosed during therapy and/or at any point during post-treatment follow-up.

Standard radiological measures, including venous Doppler ultrasound and/or venography (conventional, CT, MR), confirmed VTE. We included only clinically significant thromboembolic events, defined as symptomatic VTE, or asymptomatic VTE requiring anticoagulation, developing during ALL therapy.

Logistic regression analyses were performed to identify possible predictors of ON. Odds ratios (ORs) with 95% confidence intervals (CIs) and corresponding p-values were determined.

Results

Mean age of the study cohort (n = 208) was 5.4 years and male/female ratio 1.2:1. Seventy-eight (37.5%) patients had high-risk (HR) ALL and 127 (61.1%) received dexamethasone (DEX) as post-induction steroid. One hundred and sixty-two (77.9%) patients received E. coli ASP, 19 (9.1%) Erwinia ASP, and 27 (13.0%) PEG ASP.

Twenty-one (10.1%) children developed ON. Joints affected by ON included the ankle in 11 subjects, knee in 10, hip in 8, and heel in one. Fourteen of the 21 patients (66.7%) had involvement of more than one joint. All patients were diagnosed with ON during ALL treatment, with the average being 69.2 weeks following ALL diagnosis.

Forty-two (20.2%) subjects had a VTE while receiving therapy at an average of 29.4 weeks after ALL diagnosis. Nine patients had cerebral sinovenous thrombosis, 7 deep vein thrombosis (DVT), and one pulmonary embolism (PE). Twenty-six patients developed a central venous line (CVL)-related VTE.

Results of univariate logistic regression analyses for osteonecrosis are presented in Table 1. VTE strongly predicted development of ON – OR 8.85 (95% CI 3.37–23.25, p< 0.001). Thirteen (31.0%) patients with VTE developed ON compared to 8 (4.8%) of 166 subjects without VTE. In 10 of 13 (76.9%) patients who developed both VTE and ON, the diagnosis of VTE preceded that of ON.

Table 1

Univariate logistic regression for ON

FactorTotal nON, n (%)OR95% CIp-value
Age (yrs.)      
< 10 177 10 (5.6)    
>= 10 31 11 (35.5) 9.18 3.47–24.32 < 0.001 
Gender      
115 8 (7.0)    
93 13 (14.0) 2.17 0.86–5.49 0.10 
Risk-category      
SR 130 10 (7.7)    
HR 78 11 (14.1) 1.97 0.80–4.88 0.14 
Steroid      
PRED 81 5 (6.2)    
DEX 127 16 (12.6) 2.19 0.77–6.23 0.14 
ASP      
E. coli 162 13 (8.0)    
Erwinia 19 1 (5.3) 0.64 0.08–5.16 0.67 
PEG 27 7 (25.9) 4.01 1.43–11.24 < 0.01 
Cranial radiation      
113 10 (8.8)    
87 7 (8.0) 0.90 0.33–2.47 0.84 
VTE      
166 8 (4.8)    
42 13 (31.0) 8.85 3.37–23.25 < 0.001 
FactorTotal nON, n (%)OR95% CIp-value
Age (yrs.)      
< 10 177 10 (5.6)    
>= 10 31 11 (35.5) 9.18 3.47–24.32 < 0.001 
Gender      
115 8 (7.0)    
93 13 (14.0) 2.17 0.86–5.49 0.10 
Risk-category      
SR 130 10 (7.7)    
HR 78 11 (14.1) 1.97 0.80–4.88 0.14 
Steroid      
PRED 81 5 (6.2)    
DEX 127 16 (12.6) 2.19 0.77–6.23 0.14 
ASP      
E. coli 162 13 (8.0)    
Erwinia 19 1 (5.3) 0.64 0.08–5.16 0.67 
PEG 27 7 (25.9) 4.01 1.43–11.24 < 0.01 
Cranial radiation      
113 10 (8.8)    
87 7 (8.0) 0.90 0.33–2.47 0.84 
VTE      
166 8 (4.8)    
42 13 (31.0) 8.85 3.37–23.25 < 0.001 

Given that older age is a known risk factor for both VTE and ON, we conducted a multivariate analysis, which confirmed that age, ASP type, and VTE were independent, significant risk factors for ON (Table 2).

Table 2

Multivariate logistic regression for ON

FactorOR95% CIp-value
Age (>= 10 yrs.) 6.11 2.11–17.72 < 0.001 
ASP (PEG vs. E. coli) 3.83 1.19–12.31 0.02 
VTE (Y) 5.69 2.00–16.14 0.001 
FactorOR95% CIp-value
Age (>= 10 yrs.) 6.11 2.11–17.72 < 0.001 
ASP (PEG vs. E. coli) 3.83 1.19–12.31 0.02 
VTE (Y) 5.69 2.00–16.14 0.001 
Conclusion

In addition to the known impact of older age, we identified VTE and type of ASP as independent risk factors for ON in children with ALL. These observations suggest overlap in the etiopathogenesis of ON and VTE. We recommend larger, prospective studies to confirm the association of VTE and PEG ASP with ON and to assess the impact of hypercoagulability on the development of ON. This in turn may help develop preventive strategies (e.g., thromboprophylaxis) for ALL-associated ON.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.