Steroid-Induced Hypertension during Induction Chemotherapy for Acute Lymphoblastic Leukemia in US Children's Hospitals

Background: Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer. The induction phase of chemotherapy, including high-dose corticosteroids, is critical and effective as 98% of patients achieve remission after induction. It is generally understood that most non-life threatening side effects of chemotherapy during induction should be tolerated and treated with additional medication (if necessary) rather than reducing or discontinuing chemotherapeutic medication. Although hypertension (HTN) is a known possible complication of high-dose corticosteroids, the actual incidence and risk factors for this adverse event are poorly understood. Our objective was to utilize a national administrative data source to begin to answer these uncertainties.

Methods: The Pediatric Health Information System (PHIS) database consists of inpatient information from 43 US free-standing pediatric hospitals representing 25% of US pediatric centers and the majority of tertiary care pediatric hospitals. The PHIS database includes de-identified demographic data, discharge diagnoses, and medications prescribed among other data. Our study examined all new cases of ALL from age 0 to 28 years from Jan 1, 2009 to Dec 31, 2013. We defined our population of ALL patients receiving induction chemotherapy as those who: 1) had a non-relapse ALL ICD-9 code (204.0, 204.00, 204.01) and 2) received induction chemotherapy medications within the first 14 days of admission. Patients receiving antihypertensive medication before or on the date of chemotherapy initiation were excluded from our population. Due to the unknown sensitivity and specificity of ICD-9 codes for hypertension in this patient population, we elected to define hypertension based on recorded use of anti-hypertensive medications rather than diagnostic codes.

Results: We identified 4,917 unique patients who received induction chemotherapy for ALL. Of these patients, 13.0% received anti-hypertensive drugs during their admission. Adjusted logistic regression demonstrated increased odds of developing steroid-induced hypertension in obese patients [2.3; 95% CI: (1.42-3.71)] and those receiving anthracycline drugs [1.3; 95% CI: (1.004-1.67)]. Decreased odds of developing steroid-induced hypertension were independently associated with the 5 to 9 year age group [0.61; 95% CI: (0.49-0.76)] as compared to the 0 to 4 year age group.

To assess the reliability of using ICD-9 codes to identify patients with hypertension, sensitivity and specificity were calculated. Using receipt of anti-hypertensive drugs as the gold standard, the sensitivity of ICD-9 codes to detect hypertension was 0.60 [95% CI: (0.56-0.64)] and specificity was 0.96 [95% CI: (0.96-0.97)].

Conclusion: To our knowledge, this was the first study of steroid-induced HTN in pediatric ALL patients using a national data source. We found that HTN is a common complication of induction ALL therapy. Increased risk for HTN was independently associated with obesity and anthracycline use. Although ICD-9 codes are very useful in ruling out hypertension, they are significantly less accurate in detecting those patients with hypertension. The lack of sensitivity of the HTN ICD-9 codes was notable and suggests underreporting of HTN by physicians. This study is limited as patient information is from a single inpatient admission. As cure rates of pediatric ALL approach 85%, the long-term effects of steroid-induced HTN deserves further study.