Spontaneous Healing Of Avascular Necrosis Of The Femoral Head In Sickle Cell Disease

Abstract

Sickle cell disease (SCD) is an inherited hemoglobinopathy that affects approximately 100 000 Americans, and millions more worldwide. Hemoglobin S (HbS) is caused by a single point mutation on the beta globin gene, causing valine to be substituted for glutamic acid. HbS polymerizes when deoxygenated and deforms the red blood cell into a sickle or crescent shape. These sickled red cells cause vaso-occlusion, leading to interruption in blood flow and ischemia and clinical complications. When the bone is affected, there can be loss of the bony trabeculae, subchondral collapse, and joint destruction resulting in avascular necrosis (AVN), often affecting the femoral head. Clinical symptoms of femoral head AVN include pain, gait disturbance, and functional limitations. AVN therapies include observation, physical therapy (PT), core decompression, and in severe late stages, total hip arthroplasty. Earlier natural history studies of femoral head AVN in SCD describe universal progression to chronic skeletal and articular degenerative changes leading to decreased mobility, abnormal gait, and leg length discrepancies.1, 2 However, much of these natural history studies were performed prior to widespread hydroxyurea (HU) use or magnetic resonance imaging (MRI) screening, which is able to diagnose the disease at earlier stages. Here we present a case series of patients diagnosed with femoral head AVN, some of whom exhibited radiographic improvement of their AVN with little or no interventions. We identified all SCD patients with a diagnosis of AVN with diagnostic and follow-up imaging studies (pelvic X-rays or MRI of the hip), receiving care at Texas Children's Hospital Hematology Center from 1/2006 to 1/2018, and collected demographic, treatment, rheological and clinical laboratory data. Patient radiographic images at diagnosis and follow-up were reviewed and scored by a board certified pediatric radiologist using the Steinberg criteria.3 Longitudinal patient blood samples were collected for rheological measurements under a Baylor College of Medicine Internal Review Board approved protocol. Whole blood viscosity was measured on a cone and plate viscometer (DV3T Rheometer, AMETEK Brookfield, Middleboro, Massachusetts) at 45 and 225 seconds−1 at 37°C within 4 hours of sample collection in an EDTA vacutainer tube. Percent dense red blood cells (% DRBCs) were measured by an ADVIA 120 Hematology System (Siemens Medical Solutions USA, Inc., Malvern, Pennsylvania). Statistical analysis comparing groups with and without radiographic improvement was performed using a two-tailed Student's t-test. Sixteen patients were included in this case series with equal male and female distribution. The majority of patients had Hb SS genotype (Table 1). We divided the patients into two groups, group 1 which demonstrated radiographic improvement of AVN, and group 2, which had stable or worsening AVN during the interval between diagnostic imaging and follow-up imaging. The patients in group 1 were younger than group 2 (Table 1), P = 0.002. Both groups had comparable mean Steinberg stages at diagnosis (Table 1); P = 0.2. The mean decrease in Steinberg stage in group 1 was from IVB to IIB, while the mean increase Steinberg stage for group 2 was from IIIB to IVB (Table 1). Mean Hb, HbF, viscosity, and % DRBCs were not statistically different between the two groups. However, mean corpuscular volume (MCV) was lower in group 1 (P = 0.007). All patients were on HU treatment between diagnosis of AVN and follow-up imaging. Group 1 was at maximum tolerated dose (MTD) 48% of the observation period, and group 2 45% (P = 0.87). Sixty-nine percent of the patients were on HU prior to diagnosis of AVN. Two group 2 patients received chronic transfusion therapy; one received a short trial of this therapy years prior to AVN diagnosis while the other patient received it during the observation period of the study, beginning 1 year after AVN diagnosis to treat chronic pain. In our cohort, 5 of the 16 patients received PT for their AVN; only one showed radiographic improvement (Table 1). Five of the 16 patients received core decompression with bone marrow aspirate concentrate injection. None showed radiographic improvement. No patients underwent hip arthroplasty during the observed study period. Our retrospective study shows that significant radiographic improvement of femoral head AVN is possible, thus challenging the established natural history of this serious SCD complication. Although risk factors for the development of AVN have been described, there are no published prognostic factors to help predict which patients will experience radiographic improvement. We identified younger age as associated with radiographic improvement in AVN; regeneration/remodeling in young bone2 may be a possible mechanism but larger studies are required to confirm this association between radiographic improvement in AVN and young age. Previous studies describe a significantly higher incidence of AVN in older patients,4 however, it is possible that younger patients with earlier stage AVN were not diagnosed due to lack of sensitive imaging modalities. The advent of MRI has allowed identification of early stage AVN and thus a new assessment of the natural history of AVN. The mean Steinberg stage at diagnosis was not statistically different between the two groups; therefore, radiographic improvement observed in group 1 was not due to earlier stage at diagnosis. These findings suggest two things: that radiographic improvement is possible even with severe disease, and that severity at diagnosis may not be predictive of clinical course thus should not direct a more aggressive approach in an asymptomatic young patient. We did not find statistically significant rheological differences in whole blood viscosity, red cell deformability, and red cell density during the observation period between group 1 and group 2. Although abnormal rheology, specifically higher viscosity and DRBCs, has been described in the literature as associated with increased risk of developing AVN,5 there are no reports associating rheology with disease course. Furthermore, laboratory assessment of the two groups showed a statistically significant lower MCV in the group with radiographic improvement. MCV can be affected by co-inheritance of alpha thalassemia or the compound heterozygous state of HbS beta thalassemia; it also can be affected by HU usage. Both groups were exposed to similar amounts of effective HU therapy during the observation period when evaluated, determined by MTD (defined as the dose needed to maintain an absolute neutrophil count of 2000/μL-4000/μL per National Heart, Lung, and Blood Institute's recommended criteria), the average time on HU, and average percent of time at MTD during the observation period. Femoral head AVN is a complication of SCD that can cause significant morbidity for many patients. Controversy still exists about the optimum treatment for this condition, and to date, there are no widely accepted, evidence-based treatment guidelines. The most commonly described interventions are PT and core decompression, either alone or in combination. In our retrospective analysis, surgical management was not associated with radiographic improvement. This is important because aside from concerns about its utility, surgical interventions are associated with significant complications in SCD patients, including need for preoperative transfusion, increased intraoperative bleeding, increased infection rates, postoperative acute chest syndrome and vaso-occlusive crises, as well as risk of latent infection or chronic arthritis. Additionally, surgical intervention has not shown benefit over PT alone.6 Here we present five cases of spontaneous radiographic improvement which suggest observation may be appropriate in many cases. Moreover, this case series demonstrates the difficulty in predicting the clinical course of sickle cell-related femoral AVN. Neither radiographic severity, clinical laboratory values or rheological values, at diagnosis or during the observation period, predicted clinical course. Surgical intervention was not associated with radiographic improvement in this cohort and thus providers should closely weigh the risks and benefits on an individual patient basis before recommending surgical intervention. Further work is needed, such as larger prospective studies across multiple centers to help identify predictors of clinical course and ultimately improve outcomes for patients with SCD with femoral head AVN. The authors have no conflict of interest to report.