Prevalence of pulmonary hypertension in hereditary spherocytosis

Vascular complications, including pulmonary hypertension (PH), have been reported to occur following splenectomy for various disorders, including hereditary spherocytosis (HS). We performed a prospective cross-sectional study of 36 adults with HS (78% with prior splenectomy) utilizing echocardiography to estimate tricuspid regurgitant jet velocity (TRV) as well as measurement of N-terminal pro-B-type natriuretic peptide (NT-proBNP) to screen for PH. No participant with HS had a significantly elevated TRV or NT-proBNP level, despite a median 25-year interval since splenectomy (95% confidence interval for point prevalence 0, 0.097). Although our study was limited by a small sample size, it appears that persons with HS, following splenectomy, appear unlikely to be at significantly increased risk of developing PH to the degree reported for thalassemia and sickle cell disease. Splenectomy is often performed in hereditary spherocytosis (HS), autoimmune hemolytic anemia, immune thrombocytopenia (ITP), and thalassemia intermedia and major. Overwhelming bacterial sepsis has long been known as a complication in persons with asplenia, but is now infrequent because of pneumococcal vaccinations, prophylactic penicillin, and prompt medical attention at the first sign of fever. During the past decade evidence has emerged that an increased risk of thrombosis, both venous and arterial, may result from splenectomy [1]. This complication has been described in diverse asplenic states including HS, thalassemia, other hemolytic anemias, and trauma [1-4]. In thalassemia and sickle cell disease, another vascular complication, pulmonary hypertension (PH), has also been described following splenectomy, with some studies reporting PH prevalence as high as 75% [5-7]. Asplenia has also been reported as an independent risk factor for PH following splenectomy for HS, ITP and other hemolytic disorders [8-10]. These observations suggest that asplenia may be a risk factor for PH irrespective of the underlying condition. However, to our knowledge no prospective studies have evaluated the prevalence of PH in asplenic patients with conditions other than thalassemia and sickle cell disease, conditions both characterized by prominent intravascular hemolysis. We, therefore, conducted a prospective study of adults with HS to evaluate the independent effects of hemolysis and splenectomy on the development of PH. We enrolled 36 adults (from 30 kindreds) with HS. Most (78%) had undergone splenectomy. Thirty-five participants had echocardiography, and one other completed the laboratory component of the study but did not have complete echocardiographic data. Twenty-six patients had NT-proBNP testing. Patient characteristics are shown in Table I. Of the 35 study participants with evaluable echocardiography data, the median TRV was 2.2 m sec−1 in the group with prior splenectomy and 2.4 m sec−1 in those with intact spleen (P = 0.05). Nine did not have measurable TR jet but with otherwise normal echocardiograms were considered not to have PH [11]. Four participants had a borderline elevated TRV in the 2.6–2.7 m sec−1 range, but none had a TRV ≥ 2.8 m sec−1 (Fig. 1a). Thus the prevalence of PH in both groups was zero (95% confidence interval for point prevalence is 0, 0.097 for n = 36 total; 0, 0.123 for splenectomy group; and 0, 0.369 for intact spleen group). Scatter plots displaying TRV (a) and NT-proBNP (b) distributions in the participants with HS with and without prior splenectomy. The dotted horizontal bars represent the median and the solid error bars represent the 25th and 75th percentiles. In panel a, nine additional participants had no measurable TRV and are therefore not included. Twenty-six participants had evaluable NT-proBNP levels available. The median value was 25.1 pg ml−1 in the splenectomy group and 33.4 pg ml−1 in the group with intact spleen (P = 0.82). No one had a level indicative of PH (>365 pg ml−1); although two women, age 41 and 80 years, had slightly elevated levels (292 and 191 pg ml−1, respectively) but both had TRV < 2.5 m sec−1 (Fig. 1b). In addition, three participants with TRV in the range 2.6–2.7 m sec−1 had normal NT-proBNP levels. Much attention has been focused recently on the role of hemolysis in the development of vascular complications, especially PH [12]. Most of the published studies have reported patients with sickle cell disease or thalassemia. The rare published cases of PH in other hemolytic anemias, such as HS, have been cited as further evidence supporting hemolysis as a risk factor for PH and related conditions, although this is controversial [1, 3, 13-16]. Several case control studies have reported an increased prevalence of prior splenectomy in patients having PH [8-10]. These findings and the fact that the majority of reported patients with thalassemia and sickle cell disease have had a prior splenectomy or functional asplenia suggests that it is not solely the hemolysis that increases the risk of PH but the absence of a spleen. We therefore further hypothesized that patients with HS and intact spleen (and ongoing primarily extravascular hemolysis) would have normal TRV but that increased prevalence of elevated TRV might exist in asplenic HS patients (without ongoing hemolysis). However, we did not find evidence to support this hypothesis. A postulated pathophysiologic mechanism for PH following splenectomy includes platelet and/or endothelial activation by abnormal red cells and red cell fragments with subsequent deposition of platelet-rich microthrombi in the pulmonary vasculature [1, 17]. Decreased nitric oxide bioavailability secondary to intravascular hemolysis resulting in endothelial dysfunction and/or platelet activation has also been proposed [12]. These putative mechanisms may contribute to vascular complications in conditions such as sickle cell disease and thalassemia with continued intravascular hemolysis following splenectomy, but not in conditions such as HS in which the splenectomy effectively eliminates the hemolysis. It has also been proposed that splenectomized individuals may have increased levels of circulating prothrombotic microparticles which have been associated with PAH perhaps as a result of decreased splenic clearance [18]. Although traditionally BNP and NT-proBNP have been utilized as markers of heart failure [19], recent studies have shown some utility in these biomarkers for predicting severity of PH and for preclinical screening of PH as well [20]. No participant in our study had a level above 365 pg ml−1, the cutoff value for predicting PH set for this study [21, 22]. Additionally, in this prospective study no participant with HS had significantly elevated TRV defined as ≥2.8 m sec−1, which is generally considered the lower limit for mild PH [23]. This was further supported by the lack of significant elevations in NT-proBNP levels in either the splenectomy or no splenectomy groups. In addition, three participants with TRV in the range 2.6–2.7 m sec−1 had normal NT-proBNP levels. We acknowledge several limitations of our present study. Its original design was to include an equal number of participants with and without prior splenectomy. However, consistent with our previous findings [24], identification and recruitment of adults with HS who have an intact spleen was difficult as most adults with HS have had a prior splenectomy during childhood. This limits our ability to assess the risk of PH in patients with HS and ongoing hemolysis (albeit extravascular). Also from this study we cannot conclude that PH in HS does not occur at an increased frequency over the general population since idiopathic PH is extremely rare (estimated prevalence less than 0.2%) [25, 26] and thus a prohibitively large number of HS patients would be required to power such a study. However, given the upper 95% confidence interval of ∼ 12% prevalence of PH for the splenectomized participants in our current study, it does not appear that splenectomy in HS confers a similar risk for PH as in thalassemia. We also acknowledge that TRV as measured by echocardiography is not the ideal diagnostic tool for PH but chose to use this measure since it has been utilized in most of the published studies of PH in sickle cell disease and thalassemia. However, the high false positive rate of elevated TRV would likely overestimate the true prevalence of PH in this population and thus is not a significant weakness in the present study [11]. In conclusion, among 36 participants with HS (28 asplenic) none had evidence of PH, despite a median 25-year interval since splenectomy. A recent abstract also reported no evidence of PH among 28 persons with splenectomy for HS, pyruvate kinase deficiency, trauma, or ITP [27]. A complete review of the literature includes only 10 definitive cases of PH following splenectomy for HS [4, 8, 9, 13, 14, 28]. Over half of these cases with clinical information were felt to be secondary to acute or chronic pulmonary embolism and may represent a distinct entity. We believe this rare occurrence should not represent significant concern for persons with HS. While this small preliminary study cannot be considered conclusive, it is informative for patients with HS considering splenectomy, especially those with moderate or severe disease who would benefit most from the procedure. A more definitive study will require greater numbers of participants obtainable only through a multiinstitutional collaboration. If splenectomy does pose a risk for vascular complications such as PH, this concern is likely most relevant when intravascular hemolysis continues following splenectomy, as in SCD and thalassemia. Further studies of the independent roles of splenectomy, intravascular and extravascular hemolysis, and anemia on the risk of vascular complications are warranted. We performed a prospective cross-sectional study of 36 participants age ≥ 18 years of age with HS recruited through our pediatric hematology clinic, most of whom were relatives of our pediatric patients with HS. We also recruited participants through our hematology colleagues in the community and had many responses from other interested individuals throughout the country who learned of our study from our website. This study was approved by the Institutional Review Board at The University of Texas Southwestern Medical Center at Dallas. A limited transthoracic echocardiogram was performed on each participant in the Children's Medical Center Dallas echocardiography laboratory to screen for increased pulmonary arterial pressure, as estimated by the velocity of tricuspid regurgitant jet [29]. Other complementary measurements compatible with elevated right ventricular pressure such as right ventricular hypertrophy, right ventricular and atrial dilation, interventricular septal curvature, and pulmonary regurgitation end diastolic velocity were also evaluated. Presence of left-sided structural abnormalities and signs of left ventricular systolic and/or diastolic dysfunction that can lead to increased pulmonary artery pressure were also evaluated and was considered an exclusion criteria if present. We defined elevated tricuspid regurgitant jet velocity (TRV) as ≥2.8 m sec−1, generally considered the lower limit for predicting even mild PH [23]. Participants without a detectable tricuspid regurgitant jet, but normal parameters for other complementary measures for PH, were considered to have systolic pulmonary artery pressure within the normal range [11]. Blood was collected from each subject for a complete blood count (CBC) and reticulocyte count. Serum was separated and stored at −70°C for subsequent analysis of N-terminal pro-B-type natriuretic peptide (NT-proBNP). NT-proBNP was assayed with Dimension RxL (Siemens, Deerfield, IL) using a “sandwich” type immunoassay [30]. A NT-proBNP level above 365 pg ml−1 has been validated to have a fair sensitivity (56–69%) and high specificity (95–100%) in predicting PH and was therefore set as the cutoff value for this study [21, 22]. Participants were excluded if they reported illness, hospitalization, or surgery within the prior 2 weeks, splenectomy within the prior 3 months, had a diagnosis of chronic lung disease or congenital heart defect, or were pregnant. Descriptive data were generated for TRV and all laboratory data and expressed as median with 25th and 75th percentiles. Point prevalence of PH expressed as percentage with 95% confidence intervals. All variables were assumed to follow nonparametric distributions, and comparisons were performed utilizing Mann–Whitney tests with significance set at P = 0.05. Shelley E. Crary SECrary@uams.edu* , Claudio Ramaciotti , George R. Buchanan* , * Division of Pediatric Hematology-Oncology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, Center for Cancer and Blood Disorders, Children's Hospital, Dallas, Texas, Division of Pediatric Cardiology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas.