Increased prevalence of osteonecrosis of the femoral head in children and adolescents with sickle‐cell disease

Abstract

Osteonecrosis of the femoral head (OFH) in sickle-cell disease (SCD) has a prevalence of 6.8% on plain radiography in patients 10–24 years as reported by the Cooperative Study of SCD (CSSCD) (1). Risk factors of frequent vaso-occlusive crises (VOC) and elevated hematocrit were identified. Patients with HbSS and alpha thalassemia trait had the highest prevalence of OFH. We screened for OFH in 257 patients with SCD 10–21 years by plain radiography. We observed a prevalence of 12.4% (32 of 257) for all genotypes, SS/Sβ0 thalassemia: 14.2% (26 of 182) and SC: 9.2% (6 of 65). Bilateral disease occurred in 56% (18/32) of patients. Risk factors identified by univariate analysis for patients withSS and Sβ0 thalassemia included age (P = 0.0001), male gender (P = 0.02), decreased white blood cell count (P = 0.02), increased hematocrit (P = 0.02), reduced LDH (P = 0.0003), elevated creatinine (P = 0.037), and hydroxyurea treatment (0.009). Multivariate logistic regression analysis showed independent associations with OFH for hydroxyurea therapy [odds ratio (OR) 3.1, 95% confidence interval (CI): 1.17, 8.41, P = 0.022], gender (OR 3.1, 95% CI: 1.05, 9.02, P = 0.04, and age (OR 1.3, 95% CI: 1.07, 1.56, P = 0.008). The overall prevalence of 12.4% is 1.8-fold higher than that reported by the CSSCD and that for SS and Sβ0 thalassemia is 2.4-fold higher. Osteonecrosis of the femoral head (OFH) is a common and debilitating complication of SCD with highest prevalence in adults with Hb SS and co-existent alpha thalassemia trait as reported by the Cooperative Study of SCD (CSSCD). A report on the natural history of untreated symptomatic OFH in adult SCD demonstrated progression to collapse of the femoral head in 87% of patients within 5 years of diagnosis and the eventual need for arthroplasty in patients with advanced disease [1]. Bilateral disease is common and may occur in 40–91% of patients [3, 4]. The study reported herein screened for OFH in patients with SCD ages 10–21 years old by plain radiography as part of yearly health assessments and found an unexpectedly high prevalence at a single academic institution. The cohort of 257 subjects screened represents 83% of the 310 patients with SCD age 10–21 years who were seen from 2005 to 2010. OFH was diagnosed in 32 patients (SS: 25, SC: 6, Sβ0 thalassemia: 1) The prevalence for all genotypes is 12.4% (32/257); for HbSS and HbSβ0 thalassemia, prevalence is 14.2% (26/182) and HbSC 9.2% (6/65). Bilateral disease occurred in 56% (18/32) of patients, that is, 50 of the 64 hips were affected. The 50 affected hips were staged by the Steinberg method [5] and classified as Stage 1: 6 hips, Stage II: 26 hips, Stage III: 6 hips, Stage IV: 8 hips, Stage V: 2 hips, and Stage VI: 2 hips. Ten of 32 patients with OFH had symptomatic hip pain of which five required surgical procedures during the study period. Three HbSS patients had arthroplasty at ages 17, 19, and 20 years. Two patients, one HbSS and one HbSC, had core decompression at ages 13 and 15 years, respectively. All symptomatic patients had Steinberg Stage III or higher in the symptomatic hip. The clinical and laboratory parameters of study subjects with HbSS and HbSβ0 thalassemia are described in Table I. Analysis was confined to these genotypes, because they account for the majority of patients with OFH (26 of 32 patients). Additionally, 46% (83 of 182) of the HbSS and HbSβ0 thalassemia patients were treated with hydroxyurea and inclusion of HbSC patients would affect analysis of hydroxyurea effect on laboratory parameters. Univariate analysis showed that subjects with OFH were significantly older, more likely to be male, have lower white counts, higher hematocrit, lower LDH, higher creatinine, and more likely to be treated with hydroxuyrea. The associations for white blood cell count, hematocrit, and LDH most likely reflect hydroxyurea exposure although fetal hemoglobin percentage did not reach significance (P = 0.068). Our model for multivariate logistic regression analysis for independent associations with OFH included hydroxyurea exposure, age, gender, ethnicity, SCD genotype, body mass index (BMI), creatinine, and hospitalizations for vaso-occlusive crises (VOC). Variables with co-linearity to hydroxyurea use (hematocrit, LDH, MCV, and white blood cell count) were excluded from the model. This analysis demonstrated that the use of hydroxyurea [odds ratio (OR) 3.1, 95% confidence interval (CI): 1.17, 8.41, P = 0.022], male gender (OR 3.1, 95% CI: 1.05, 9.02, P = 0.04), and age (OR 1.3, 95% CI: 1.07, 1.56, P = 0.008) were independently associated with OFH. Compliance with hydroxyurea therapy was demonstrated by the expected changes in laboratory parameters (Table II). This study reports a significantly higher prevalence of OFH in patients with SCD than that reported by the CSSCD in 1991 for a similar age group. It confirms age as a risk factor and identifies hydroxyurea exposure and male gender as additional risks. Elevated hematocrit is identified in univariate analysis and is co-linear with hydroxyurea therapy. The independent risk of frequent VOC is not identified in our patient cohort. However, the increasing use of hydroxyurea has changed the phenotype of SCD and comparison to historical trials is now problematic, because hydroxyurea would ameliorate many clinical and laboratory variables. The question of why the prevalence of OFH is higher in our clinic population and whether hydroxyurea therapy is a factor needs further exploration. When compared with the 6.8% prevalence of OFH observed in SCD patients (all genotypes) 10–24 years old in the CSSCD study, the present study detects a 1.8-fold higher rate of OFH, and for HbSS and Sβ0 thalassemia, a 2.4-fold higher rate, in a similar age group. It is difficult to account for this significant increase in OFH in our patients and for hydroxyurea as a risk factor. Hydroxyrea therapy could be a surrogate for severity, and, therefore, patients on hydroxyurea would be expected to have a higher prevalence of OFH based on disease severity. Hydroxyurea use is reported in 83 of 182 (46%) eligible HbSS or HbSβ0 thalassemia patients in this study. This does represent a bias toward inclusion of more patients treated with hydroxyurea in the study cohort, since ∼37% of eligible patients between ages 10 and 21 years followed in our clinic are treated with hydroxyurea. However, this does not account for a 2.4-fold increase for HbSS and HbSβ0 thalassemia patients. Postulating that hydroxyurea therapy is an additional risk factor for OFH can be rationalized by hydroxyurea's well-known therapeutic effect of increasing fetal hemoglobin and hematocrit possibly leading to increased blood viscosity and sickling in the microcirculation of the femoral head. Additional evidence for the association of OFH with higher hematocrit can be taken from studies of prospective screening for OFH in both pediatric and adult SCD patients in Kuwait using more sensitive magnetic resonance imaging (MRI) imaging. These studies detected a prevalence of OFH in 48.6% of adults at a mean age of 26.7 years and 26.7% in children at a mean age of 9.8 years [6, 7]. This finding is surprising, because the Kuwaiti SCD patients have milder symptoms and have elevated fetal hemoglobin levels of ∼20%, baseline hemoglobin levels of 10–11 gm/dl, and an ∼40%prevalence of alpha thalassemia [8]. There are similar findings of a higher prevalence of OFH in SCD patients from the Eastern Province of Saudi Arabia who have milder disease with higher fetal hemoglobin levels and prevalence of alpha thalassemia when compared to Western Province patients [9]. The use of hydroxyurea has changed the phenotype of more severe SCD,since it was first reported as an effective therapy for adult SCD patients with severe disease in 1995 by the Multicenter Sickle Cell Hydroxyurea Trial (MSH) [10]. Hydroxyurea use in pediatrics has demonstrated similar changes in laboratory parameters and clinical efficacy as reported intheMSH trial with a decrease in hospitalizations, VOC, and acute chestsyndrome episodes [11-13]. Likewise, there is also evidence that hydroxyurea use in children has ameliorative effects on sickle-related end-organ damage such as regeneration of splenic function, decrease in transcranial Doppler velocities, primary stroke prevention, and reversal of proteinuria [14-17]. There are two reports of long-term exposure to hydroxyurea in adults. The MSH follow-up study, after 17.5 years, reported improved survival in adult patients with greater than 10 years of hydroxyurea exposure [18], whereas stroke, organ dysfunction, infection, and malignancy were similar in patients exposed to hydroxyurea and those not. This study also reported on the prevalence of OFH in this patient cohort, which was 24.3% in patients exposed to hydroxyurea less than 5 years and 20.4% in those not exposed. The second study from a single institution in Greece examined hydroxyurea use over a 17-year period in adult patients with HbSS, HbSβ+, and HbSβ0 thalassemia [19]. This study reported on symptomatic OFH that was detected in 12% of hydroxyurea exposed patients at baseline while only 1 of 131 patients developed OFH during hydroxyurea therapy. Both these studies would suggest that hydroxyurea use is not a risk for OFH. However, both studies report on symptomatic disease and cannot be compared to either the CSSCD study or our study that reports on routine radiographic screening for OFH. Considering that approximately half of the CSSCD study patients with OFH were asymptomatic at initial diagnosis, this would suggest that the prevalence of OFH would be higher in both long-term studies if screening radiographic imaging was done in all patients. The current study has some limitations that need acknowledgment. First, it is a cross-sectional study from which a conclusion regarding causality between hydroxyurea use and OFH cannot be made. Second, data on the length of time on hydroxyurea and hospitalizations for VOC before hydroxyurea use is not known for all patients in the study. Third, hip radiographs were not obtained before starting hydroxyurea in all patients, and so there is a possibility that OFH was present before hydroxyurea therapy. Given the cross-sectional nature of this study, a prospective study was able to ascertain the incidence of OFH among patients exposed to hydroxyurea, as opposed to those not exposed would yield more robust data. In addition, a longitudinal study comparing radiographic and clinical progression of OFH in patients both exposed and not exposed to hydroxyurea is also of interest. Based on the data reported herein, we suggest that patients have screening radiographs of the hips before starting hydroxyurea therapy. From 2005 to 2010, 257 of 310 eligible children and adolescents with sickle-cell disease (SCD; SS: 177, SC: 65, Sβ+thalassemia: 7, S0thalassemia: 5, and Sother: 3) between the ages of 10–21 years were screened for OFH with radiographs of the pelvis as part of routine health maintenance at the Children's Hospital at Montefiore, Bronx, New York. Eighty-three of 182 SS or Sβ0 thalassemia were treated with hydroxyurea for a minimum of six months. Indications for hydroxyurea therapy were frequent admissions for VOC and/or multiple acute chest syndromes. The study was approved by the Institutional Review Board at Montefiore Medical Center. For the purpose of this report, all radiographs were reviewed by a single pediatric radiologist (BT) to ensure consistency in diagnosis and staging. Films were staged according to the University of Pennsylvania Steinberg classification system for OFH: Stage 0: normal plain radiographs and MRI, Stage I: normal plain radiographs, but MRI findings of OFH, Stage II, sclerosis and lucencies of the femoral head, Stage III, subchondral collapse without flattening of the femoral head, Stage IV, flattening of the femoral head, normal joint space, Stage V, joint space narrowing with acetabular changes, and Stage VI, advanced degenerative changes [5]. Patients identified as having OFH were further evaluated with MRI of the pelvis. MRI studies were performed on a 1.5 T MRI (GE) and consisted of axial T1, axial T2 fast spin echo with fat saturation, coronal T2 with fat saturation, T1 coronal of both hips, and coronal and sagital proton density of the symptomatic side, without gadolinium administration. Clinical information was collected from the hospital electronic database and included age, race/ethnicity, SCD genotype, hospitalizations for VOC in the preceding 3 years, BMI, and hydroxyurea exposure. Those who received hydroxyurea treatment for more than 6 months were considered to be exposed. The indications for hydroxyurea therapy were ≥3 admissions for VOC per year or two or more episodes of acute chest syndrome. Laboratory data abstracted from the medical record included complete blood count including red cell indices, ferritin, lactate dehydrogenase, creatinine, and fetal hemoglobin. Descriptive statistics included calculation of the mean, standard deviation, and use of t-test for continuous variables and chi-square test for categorical variables. Unconditional logistic regression analysis was used to estimate the OR and the 95% CI. All statistical analyses were conducted in STATA® (College Station, TX), and P-values less than 0.05 were considered statistically significant. Kris M. Mahadeo*, Suzette Oyeku , Benjamin Taragin , Swapnil N. Rajpathak§ ¶, Karen Moody*, Ruth Santizo*, M. Catherine Driscoll*, * Department of Pediatrics, Division of Pediatric Hematology-Oncology, Albert Einstein College of Medicine, Children's Hospital at Montefiore, Bronx, New York, Department of Pediatrics, Division of General Pediatrics, Albert Einstein College of Medicine, Children's Hospital at Montefiore, Bronx, New York, Department of Radiology, Albert Einstein College of Medicine, Children's Hospital at Montefiore, Bronx, New York, § Department of Epidemiology, Albert Einstein College of Medicine, Bronx New York, ¶ Department of Population Health and MedicineAlbert Einstein College of Medicine, Bronx, New York.