Pulmonary function abnormalities in non-splenectomized and splenectomized adult hemoglobin E/β-thalassemia patients and their correlation with pulmonary hypertension

The effect of splenectomy on pulmonary function test (PFT) and pulmonary hypertension (PH) in thalassemia remains unclear. We aimed to investigate PFT and their association with PH in patients with hemoglobin E/β-thalassemia stratified by their splenic status. Thirteen splenectomized patients (SP) and 12 non-splenectomized patients (NSP) were compared regarding to the PFT abnormalities and PH (mean pulmonary artery pressure from right-heart catheterization ≥25 mmHg or estimated systolic pulmonary artery pressure from echocardiography ≥40 mmHg). Eleven (84%) SP and 9 (75%) NSP had restrictive impairment (RI). Of these, more patients having severe RI in SP than in NSP (8 vs 2, P=0.035). FVC and PaO2 were lower in SP than in NSP (66±15% vs 77±12%, P=0.043, and 79.38±1.6 mmHg vs 98.83±6.2 mmHg, P<0.001, respectively). Residual volume was higher in SP than in NSP (78±17% vs 64±15%, P=0.036). Seven (54%) SPwho developed PH had a longer time interval between splenectomy and the onset of PH than those who did not (17±4.9 years vs 9.8±6.1 years, P=0.04). In conclusion, greater severity of extrapulmonary restrictive impairment and hypoxemia were more common in SP. These patients developed PH as a late complication unrelated to hypoxemia and PFT parameters. Introduction Previous studies have reported pulmonary dysfunction in patients with β-thalassemia (Thal); mostly with restrictive impairment (RI) in patients being transfusion-dependent1-11 and also in non-transfusion dependent patients.12-14 Other pulmonary abnormalities consist of small airway obstruction,12,15 impairment of diffusion capacity,16,17 arterial hypoxemia,16,18 and pulmonary hypertension (PH).19-21 Blood transfusion and iron chelation therapy for thalassemia disease in Thailand are usually not as intensive as in Western countries due to limited resources. This could have some effects on pulmonary dysfunction in these patients. The effect of splenectomy on the pulmonary functions remains unclear and their association with PH has not yet been reported. We aimed to investigate the pulmonary function abnormalities, and their association with PH in hemoglobin E/β-Thal patients stratified by their splenic status. Materials and Methods This retrospective study was performed with the approval of our institutional ethics committee. A computer search of our hospital databases between January 2000 and December 2007 identified 42 patients with E/β-Thal, diagnosed based on hematologic data and thalassemia genotype. Out of the 42 patients, we excluded 17 patients who had no pulmonary function tests (PFT). Twenty-five patients had PFT performed approximately within 6 months apart from the echocardiography or right heart catheterization (RHC). The final study group consisted of 25 patients, whom were further classified into 2 groups, i.e., splenectomized patients (SP) and non-splenectomized patients (NSP). The medical records were reviewed for demographic data (age, gender, height, and body weight), smoking history, age at splenectomy performed, and laboratory data including hemoglobin, platelet count, nucleated red blood cells (nRBC), reticulocyte count, serum alanine aminotransferase (ALT), and serum ferritin level. An experienced chest radiologist retrospectively reviewed chest radiographs for the presence or absence of extramedullary hematopoiesis (i.e., abnormal soft-tissue masses along the paravertebral areas and/or ribs), thalassemic bone changes, and cardiomegaly. Pulmonary function tests All PFT measurements were performed using Vmax system® (Sensor Medics, Yorba Linda, CA, USA). Spirometry including forced expiratory volume at 1 second (FEV1), forced vital capacity (FVC), forced expiratory flow between 25% and 75% of vital capacity (FEF2575%), and measurement of lung volumes, i.e., total lung capacity (TLC), and lung volumes subdivision including residual volume (RV), functional residual capacity (FRC), inspiratory capacity (IC), and expiratory reserve volume (ERV) were recorded. Single-breath diffusing capacity for carbon monoxide (DLCO), DLCO value adjusted for hemoglobin concentration, and DLCO adjusted for alveolar volume (DLCO/VA) were recorded. The normal reference values for pulmonary function tests were based on the Crapo reference equation.22,23 Arterial blood gases results performed at the time of PFT measurements were also recorded. The parameters in arterial blood gases including partial pressure of oxygen (PaO2) and carbon dioxide (PaCO2) were used in the calculation of alveolar-arterial partial pressure of oxygen difference [P (A-a) O2] based on alveolar gas equation. Arterial hypoxemia was considThalassemia Reports 2013; volume 3:e5 Correspondence: Prapaporn Pornsuriyasak, Division of Pulmonary and Critical Care, Department of Medicine, 272 Ramathibodi Hospital, Rama 6 Road, Bangkok 10400, Thailand. Tel. +66.2.2011629 ext. 19/+66.8.1.6465390 Fax: +66.2.2011629. E-mail address: pprapaporn@gmail.com


Introduction
Previous studies have reported pulmonary dysfunction in patients with β-thalassemia (Thal); mostly with restrictive impairment (RI) in patients being transfusion-dependent [1][2][3][4][5][6][7][8][9][10][11] and also in non-transfusion dependent patients. [12][13][14] Other pulmonary abnormalities consist of small airway obstruction, 12,15 impairment of diffusion capacity, 16,17 arterial hypoxemia, 16,18 and pulmonary hypertension (PH). [19][20][21] Blood transfusion and iron chelation therapy for thalassemia disease in Thailand are usually not as intensive as in Western countries due to limited resources. This could have some effects on pulmonary dysfunction in these patients. The effect of splenectomy on the pulmonary functions remains unclear and their association with PH has not yet been reported. We aimed to investigate the pulmonary function abnormalities, and their association with PH in hemoglobin E/β-Thal patients stratified by their splenic status.

Materials and Methods
This retrospective study was performed with the approval of our institutional ethics committee. A computer search of our hospital databases between January 2000 and December 2007 identified 42 patients with E/β-Thal, diagnosed based on hematologic data and thalassemia genotype. Out of the 42 patients, we excluded 17 patients who had no pulmonary function tests (PFT). Twenty-five patients had PFT performed approximately within 6 months apart from the echocardiography or right heart catheterization (RHC). The final study group consisted of 25 patients, whom were further classified into 2 groups, i.e., splenectomized patients (SP) and non-splenectomized patients (NSP). The medical records were reviewed for demographic data (age, gender, height, and body weight), smoking history, age at splenectomy performed, and laboratory data including hemoglobin, platelet count, nucleated red blood cells (nRBC), reticulocyte count, serum alanine aminotransferase (ALT), and serum ferritin level. An experienced chest radiologist retrospectively reviewed chest radiographs for the presence or absence of extramedullary hematopoiesis (i.e., abnormal soft-tissue masses along the paravertebral areas and/or ribs), thalassemic bone changes, and cardiomegaly.

Pulmonary function tests
All PFT measurements were performed using Vmax system® (Sensor Medics, Yorba Linda, CA, USA). Spirometry including forced expiratory volume at 1 second (FEV 1 ), forced vital capacity (FVC), forced expiratory flow between 25% and 75% of vital capacity (FEF 25-75% ), and measurement of lung volumes, i.e., total lung capacity (TLC), and lung volumes subdivision including residual volume (RV), functional residual capacity (FRC), inspiratory capacity (IC), and expiratory reserve volume (ERV) were recorded. Single-breath diffusing capacity for carbon monoxide (DL CO ), DL CO value adjusted for hemoglobin concentration, and DL CO adjusted for alveolar volume (DL CO /VA) were recorded. The normal reference values for pulmonary function tests were based on the Crapo reference equation. 22,23 Arterial blood gases results performed at the time of PFT measurements were also recorded. The parameters in arterial blood gases including partial pressure of oxygen (PaO 2 ) and carbon dioxide (PaCO 2 ) were used in the calculation of alveolar-arterial partial pressure of oxy- ered when PaO 2 at room air was <80 mmHg. Obstructive impairment was defined as FEV 1 /FVC<75%, and FEV 1 <80%. RI was defined as FEV 1 /FVC≥75% with FVC<80% or TLC≤80%. Severity of RI was classified based on reduction in TLC (mild, TLC 60-79%; moderate, TLC 40-59%; severe, TLC<40%). Small airway involvement was based on the values of FEF 25-75% <60%, FEV 1 /FVC≥75%, and FEV 1 ≥80%. In patients having reduced FVC, reduced FEV 1 , normal FEV 1 /FVC, and reduced FEF 25-75% , restrictive impairment combined with small airway involvement was considered. DL CO was considered low when it was ≤70%, otherwise it was considered normal PFT. 24

Echocardiography and right heart catheterization
The echocardiography performed within a 6month interval apart from the date of PFT were retrospectively reviewed for the presence of PH using tricuspid regurgitant velocity to estimate the systolic pulmonary artery pressure. The hemodynamic data from RHC, including systolic pulmonary arterial pressure (SPAP) and mean pulmonary arterial pressure (mean PAP) were recorded at the time of PH diagnosis. PH was defined either by echocardiography when estimated systolic pulmonary arterial pressure (eSPAP) was ≥40 mmHg 25 or by RHC when mean pulmonary arterial pressure (PAP) was ≥25 mmHg. 26

Statistical analyses
All analyses were performed using SPSS Statistics for Windows version 17. Comparisons of normally distributed continuous data between both groups were made by using independent Student's t-test. Otherwise, Mann-Whitney U test was used. Categorical data was analyzed by Chi-square or Fisher's exact test. Results were considered statistically different when P<0.05. Pearson's correlation was used to determine the correlation between two continuous variables.

Results
The final study consisted of 25 E/β-Thal patients (11 men and 14 women). Thirteen of them were SP, and the remaining 12 patients were NSP. Table 1 shows the characteristics of patients in both groups. Of the 13 SP, the patterns of spirometric abnormalities were defined as having RI in 11 (84%) patients; 2 of 11 patients having RI combined with small airway involvement. Obstructive impairment in 1 (7.7%) patient and normal spirometry in 1 (7.7%) patient were noted. Of the 12 NSP, 9 (75%) had RI; 1 of 9 patients having RI com-bined with small airway involvement; 3 (25%) patients had normal spirometry. The severity of RI was mild in 10 patients (3 of SP and 7 of NSP), and moderate in 10 patients (8 of SP and 2 of NSP). Severe RI was not observed in any cases. Compared with the NSP, a greater proportion of SP had more severity of RI (8 of 11 in SP vs 2 of 9 in NSP, P=0.035). The hemoglobin level was not different between the two groups, but platelet count, nRBC count and reticulocyte count were higher in the SP. Table 2 shows the PFT results, PaO 2 , P(A-a)O 2 , and serum ferritin levels in both groups. Arterial hypoxemia was present in 7 of 13 SP, but was not present in the NSP (P=0.005   Serum ferritin levels in the SP group were significantly higher than those in the NSP group (3298.5 ng/mL; range 1420-5111 ng/mL vs 504.9 ng/mL; range 328-926 ng/mL, P=0.001). There were inverse correlations between FVC and serum ferritin levels (r= -0.46, P=0.02), and between PaO 2 and serum ferritin levels (r= -0.49, P=0.013). Chest radiographs were available for review in 17 patients. Of these, extramedullary hematopoiesis was found in 9 SP and 1 NSP (P=0.018). Thalassemic bone changes were found in 11 SP and 3 NSP (P=0.029). Cardiomegaly was present in 7 SP and 2 NSP (P=0.335). Echocardiographic data could be obtained from 15 patients whereas PAP from RHC data was from 9 patients. PH was found in 7 SP. The SPAP and mean PAP values from RHC were 60.3±16.30 mmHg and 37.4±10.15 mmHg, respectively. The eSPAP from echocardiography in this group was 41.5±7.6 mmHg. There were good correlations between eSPAP and SPAP (r=0.74, P=0.034) and between eSPAP and mean PAP (r=0.76, P=0.03). Factors associated with PH in SP are shown in Table 3. Seven (54%) SP who developed PH had a longer time interval between splenectomy and the onset of PH than those who did not (17±4.9 years vs 9.8±6.1 years, P=0.04). SP with PH had greater nRBC count than those without PH (1673±536.82/100 WBCs vs 782±403.93/100 WBCs, P=0.013). Of the 9 patients who had RHC performed, no evidence of emboli was reported. The eSPAP was moderately correlated with DL CO (r=-0.6, P=0.04) but was not correlated with FVC (r=-0.38, P=0.

Discussion
Our results confirmed that RI was commonly found in both splenectomized and non-splenectomized E/β-Thal patients. Moreover, we found a greater frequency of RI (84% of SP and 75% of NSP) than those previously reported up to 70%, 9 and a greater percentage of SP having more severe degree of RI and arterial hypoxemia, when compared to the NSP.
Although the exact mechanisms are unknown, there are possible mechanisms regarding the development of RI in E/β-Thal patients. As shown in the studies of Filosa et al. 7 and Fung et al., 8 older thalassemia patients with an average age of 19±1.4 years had more RI than younger patients with average age of 10.8±1.7 years and 15.7±1.1 years did. Nevertheless, both SP and NSP in our study were older than those in the previous studies.
The reduction in lung volumes in E/β-Thal patients may also be explained by enlarged spleen and liver or by insufficient anatomical and functional development of the lungs during early life. 5 Furthermore, an increase in vital capacity in patients following splenectomy has been reported. 18 However, the RI in this study was unlikely to be owing to the effect of organomegaly because more severity of RI was found in the SP than in the NSP.
The relation between the reduction in lung volumes and the presence of iron overload secondary to the therapy has been raised. 27 Iron deposition in the lung tissue can lead to free radical induced lung injury, regardless of serum ferritin levels. 28 However, the previous autopsy study done by Witzleban and colleague 29 showed no predilection of iron deposition to the lung (hemosiderosis). Our study showed other PFT parameters, e.g., normal or elevated DL CO adjusted for hemoglobin values and preserved FRC and RV in the lung volumes subdivision, making the intrapulmonary causes of RI unlikely. As reported previously, the elevated DL CO adjusted for hemoglobin is attributed to the effect of anemia in these patients. 30 In this study, we found a greater proportion of SP who had more severe degree of RI and higher serum ferritin levels. This suggested more iron overload in these patients that might result from more frequent red blood cell transfusion to correct their anemia. Moreover, the presence of cardiomegaly secondary to high output heart failure in consequence of chronic anemia can reduce VA and subsequently an elevated DL CO /VA can occur in these patients. 31 Interestingly, RI in SP in this study had a characteristic of preserved FRC, RV, and ERV accompanied by low IC, suggesting that the cause of RI in these patients might be related to respiratory muscle weakness or chest wall restriction rather than the intrapulmonary restriction caused by iron deposition in the lung tissue. This PFT finding was supported by the study of Hart and colleague 32 showing that patients with combined inspiratory and expiratory muscle weakness had reduced TLC, DL CO , but had increased FRC and RV. We also speculated that the greater iron overload in SP could accelerate the production of reactive oxygen species which lead to muscular fatigue and eventually to the loss of skeletal muscle mass and impaired function. Reardon and colleague 33 reported iron overloaded mice had elevated level of iron in skeletal muscle and increased serum ferritin. The level of oxidative stress product was higher compared with controls. Iron loading was found to reduce the skeletal muscle weight and function. In this study, the extramedullary hemato poiesis and thalassemic bone changes were found more frequently in the SP and might contribute to the chest wall restriction leading extrapulmonary RI. In contrast to the result of Hoyt 12 and Keens 15 in which the small airway obstruction was more common among their thalassemia patients, we found the small airway involvement to a lesser extent in only 12% of patients. All of whom with small airway involvement in our study also showed a combined restrictive impairment. This might be due to the difference in patients' age. Most thalassemia patients in both previous studies were pediatric population in which the average age was 6.8±0.6 years old and patients in the study of Keens 15 were on a hypertransfusion program. The authors proposed that the small airway obstruction could be due to iron deposition in the airways.
Arterial hypoxemia was found only in SP (28%) in our study. Hypoxemia in these patients was associated with a widening in the Article Table 3. Factors associated with development of pulmonary hypertension in splenectomized patients(data are expressed as mean±SD).  difference between alveolar and arterial partial pressure of oxygen. However, the mechanism of hypoxemia is still unclear. We speculated that hypoxemia is a consequence of alterations in pulmonary microvasculature, endothelial dysfunction known to occur in these patients and not related to other pulmonary functions related to the lung mechanics. 34 There were no significant correlations among any pulmonary function parameters, including PaO 2 and PAP values except DL CO . Atichartakarn and colleague 21 reported the PH in SP was due to increased pulmonary vascular resistance from pulmonary arteriopathy without evidence of emboli. Pathogenesis of such changes in pulmonary microvasculature was from chronic lowgrade hypercoagulable state and platelet activation after splenectomy. 35 PH was more prevalent in patients with a longer time interval between splenectomy and the onset of PH suggesting that PH was its late complication as supported by the study of Atichartakarn and colleague. 21 The main limitations of our study were a retrospective nature and a small number of patients. However, we were able to demonstrate the different patterns of PFT abnormalities between SP and NSP and also the prevalence of PH following splenectomy. In conclusion, we found that E/β-Thal patients with previous splenectomy had a greater prevalence of more severe RI, arterial hypoxemia, and PH (a wellrecognized late complication after splenectomy). The more severe RI might be attributed to an extrapulmonary causes, namely combined abnormalities related to cardiomegaly, respiratory muscle weakness, or extramedullary hemato poiesis, whereas the arterial hypoxemia might be a consequence of pulmonary arteriopathy. Nevertheless, the degree of PH was neither correlated with the degree of abnormalities of PFT nor the hypoxemia.