Published ahead of print on May 8, 2003, doi:10.1164/rccm.200211-1292OC
© 2003 American Thoracic Society
Pulmonary Dysfunction in Transfusion-dependent Patients with Thalassemia MajorDipartimento di Pediatria, Istituto di Fisiologia Umana I, and Dipartimento di Medicina del Lavoro, Università di Milano, Milan, Italy Correspondence and requests for reprints should be addressed to Edgardo D'Angelo, M.D., Istituto di Fisiologia Umana I, via Mangiagalli 32, 20133 Milan, Italy. E-mail: edgardo.dangelo{at}unimi.it
Pulmonary function tests were performed on 62 transfusion-dependent patients with thalassemia major, ranging in age from 8 to 33 years, and receiving chelation therapy with desferrioxamine or deferiprone. Percent predicted values for FVC, FEV1, and PEF were significantly reduced, whereas FEV1/FVC and maximal expiratory flow at 25% FVC were within normal limits, indicating a restrictive disease. Both FVC and FEV1 were negatively correlated with transfusional iron burden as indexed by age. Single-breath carbon monoxide transfer factor was reduced, even after correction for low hemoglobin concentration, and was negatively correlated with iron burden and iron overload, as indexed by serum ferritin levels. Owing to low hemoglobin concentration, blood-diffusing capacity was reduced, in spite of increased lung capillary blood volume, which was, however, adequate to normalize blood diffusing capacity when hemoglobin concentration was only partially restored by transfusion. The diffusing capacity of the alveolar–capillary membrane was substantially decreased and negatively correlated with age and serum ferritin, the fall being primarily attributed to increased membrane thickness. These findings suggest that lung fibrosis and/or interstitial edema related to iron overload are the main cause of pulmonary dysfunction observed in patients with thalassemia major.
Key Words: alveolar–capillary membrane • iron overload • lung mechanics • pulmonary capillary blood volume • pulmonary diffusing capacity Thalassemia major (TM) is a disorder characterized by ineffective erythropoiesis, leading to impaired oxygen delivery to the tissues. Although adequate, protracted transfusion programs prevent the development of the adverse effects associated with this disorder, iron accumulation eventually occurs, in spite of concomitant chelation therapy. Although the heart, liver, and pancreas are the target organs most frequently involved, and in which extensive iron-induced injury is regularly observed at necropsy, abnormalities of lung mechanics have been reported by almost all studies of patients with TM. However, there is no consensus about the nature, restrictive (1–5) or obstructive (6–8), of these defects. Moreover, the relationship between the changes of lung mechanics in transfusion-dependent patients with TM and iron burden or overload remains unclear. Indeed, substantial iron deposition in the lung has been observed on postmortem examination in some (9–11) but not in other cases (1, 2). In addition to abnormal lung mechanics, patients with TM regularly exhibit a reduced pulmonary diffusing capacity, which in most instances is only partially due to lower hemoglobin concentration. As for mechanical dysfunctions, the relation between lung diffusing capacity and iron deposition is uncertain. Moreover the mechanisms leading to the fall in lung diffusing capacity have not been investigated, except in a study on a small group of young Chinese and Malay patients (4). Indeed, reduction of lung diffusing capacity could be the consequence of the fall of either the diffusing capacity of the alveolar–capillary membrane, or the pulmonary capillary blood volume, or both (12). In the present study, we wanted to assess in a large number of transfusion-dependent, thalassemic patients the predominant type of lung mechanical abnormalities, the prevalence of the reduction in pulmonary diffusing capacity and the mechanisms involved, the relation between mechanical and diffusional alterations, and the dependence of lung dysfunctions on iron burden and overload.
The study included 62 white patients (8–33 years) with TM, whose physical characteristics are reported in Table 1 . The patients were receiving blood transfusions at 3- to 4-week intervals, and chelation therapy with desferrioxamine via subcutaneous infusion five times a week at a dose of 40–50 mg · kg–1 (39 patients), or deferiprone given orally at a daily dose of 75 mg · kg–1 for the entire week. Iron overload was estimated as the mean value of serum ferritin levels (13) of 13–17 samples obtained during the preceding year. The patients had no clinical manifestations of cardiopulmonary diseases at the time of the study, which was approved by the institutional ethics committee; informed consent was obtained from patients or parents. One patient reported asthmatic episodes, nine were receiving therapy because of modest rhythm alterations and three because of left ventricular hypokinesis without pulmonary hypertension. In these patients, echocardiographic reassessment showed subsequently that digitalis had restored normal ventricular functions. Hepatomegaly was present in 30 patients, splenomegaly was present in 4 patients, and 28 patients had undergone splenectomy. Four patients smoked 10–15 cigarettes/day.
Pulmonary function studies were performed according to recommended guidelines (14) on the day scheduled for blood transfusion. Spirometry was assessed with a computerized system (Baires-Biomedin; Biomedin, Padua, Italy). The best values of FVC, FEV1, PEF, and maximal expiratory flow at 25% FVC (MEF25) were selected from three technically acceptable maneuvers, and referred to predicted values (14). The transfer factor for carbon monoxide (TLCO) was assessed in sitting position by the single breath method using a modified, semiautomatic apparatus (Baires-Biomedin; Biomedin). In each patient, two measurements were taken first after breathing 85% oxygen for 10 minutes, and then after 15 minutes of air breathing. The test gas mixtures contained 12.5–13.5% He and 0.3% CO or 12.5–13.5% He, 0.3% CO, and 21% O2 with the balance O2 or N2, respectively. In 31 patients, the whole procedure was repeated 1–1.5 hours after the blood transfusion, hemoglobin concentration ([Hb]) and hematocrit (Hct) being measured before and after transfusion. The transfer factor of the alveolocapillary membrane for carbon monoxide (Tm) and capillary blood volume (Vc) were computed according to TLCO–1 = Tm–1 + (  · [Hb]*· Vc)–1 (12), where is the CO reaction rate and [Hb]*is [Hb]/146 g · L–1. The values of as a function of alveolar capillary oxygen tension (PCO2) were obtained from Roughton and Forster (12), PCO2 being assumed equal to PO2 measured with an oxygen analyzer (Beckman OM-11; Sensor Medics, Milan, Italy) in the last 300–500 ml exhaled after breath holding. No correction was made for carbon monoxide back pressure because it is negligible in nonsmokers (12), smokers refrained from smoking for 2–3 days before the test, and measurements under hyperoxic conditions were performed first (15). The values of TLCO, Tm, and Vc obtained in each patient and condition agreed to within 8%, and were thus averaged. The volume of red blood cells in the pulmonary capillaries (VRBC) was computed as Vc · Hct. Values of TLCO were standardized (TLCO*), that is, corrected for [Hb] = 146 g · L–1 and PCO2 = 110 mm Hg (16), using the equation given above and the experimental values of Tm and Vc. Values of TLCO*, Tm, and Vc were also expressed as the percentage of predicted normal values, which were obtained from Cotes and coworkers (17) for patients less than 17 years old, and from Bucci and coworkers (15) for older patients. The ratio of the permeability of red cell membrane to that of red cell interior was taken as infinity or 2.5 depending on whether computed Tm and Vc values were compared with those from regressions calculated by either Cotes and coworkers or Bucci and coworkers. Values of Tm and Vc assessed in 16 normal boys (8–16 years) and 28 normal adults (21–40 years) were all within ± 20% of predicted values, mean percentage predicted values being 97 and 98% for boys, and 97 and 100% for adults.
The results are presented as means ± SEM. Statistical significance was assessed by analysis of variance, the Student paired t test being used whenever appropriate. Linear regressions were computed by the least squares method and statistical assessment was made by analysis of covariance. The level of significance was taken at p
The results of spirometry are reported in Table 1. On average, FVC, FEV1, and PEF were significantly lower than predicted values, whereas FEV1/FVC and MEF25 were within normal limits. No sex-dependent differences were observed. Mean percentage predicted  FVC, FEV1, and PEF were –22.2 ± 1.4, –20 ± 1.7, and –22.6 ± 1.8% (p < 0.001), whereas (FEV1/FVC) and MEF25 were 1.3 ± 1.1 and –8.1 ± 4% (p > 0.05), respectively. On an individual basis, percentage predicted FVC, FEV1, PEF, and MEF25 were below 80% in 42, 38, 40, and 14 patients, whereas FEV1/FVC always exceeded 85%. Although all spirometric variables tended to decrease with increasing age, only percentage predicted FVC and FEV1 exhibited a significant inverse correlation with age (Figure 1)
.
Serum ferritin levels were not significantly different between male and female patients (Table 2) . They were positively correlated with age (Figure 1), but not with any of the spirometric variables.
The mean group values of TLCO and TLCO*on air, Tm, and Vc for male and female patients are reported in Table 2, together with blood hemoglobin and hematocrit. No variable differed significantly between male and female patients. The average values of TLCO* and Tm were significantly lower than normal, whereas those of Vc were larger, percentage predicted TLCO*, Tm, and Vc amounting to –24.8 ± 2.5, –40.8 ± 1.7, and 29.4 ± 3.9% (p < 0.001), respectively. On an individual basis, percentage predicted TLCO* and Tm were less than 80% in 38 and 57 patients, respectively, whereas Vc was greater than 120% in 37 patients. The ratio of blood to alveolar–capillary membrane resistance (Tm/ · Vc = 0.62 ± 0.03) was lower than the average values (0.7–1.12) of normal subjects (15–18), an even lower value (0.40 ± 0.02) being obtained after correction for [Hb]. A proportional relationship between Tm and the volume of red blood cells in the pulmonary capillaries (VRBC = Vc · Hct) was demonstrated at a satisfactory level of significance in the present patients (Figure 2)
. The slope of this relationship was substantially lower than normal (Table 2), percentage predicted (Tm/VRBC) being –27 ± 3% (p < 0.001).
Percentage predicted TLCO*, Tm, and Tm/VRBC were inversely correlated with both age and serum ferritin levels, whereas no significant correlations were found with Vc (Figures 3 and 4) .
The mean group values of TLCO, Tm, and Vc obtained from 15 female and 16 male patients before and after transfusion are reported in Table 3 . With transfusion there was a significant increase in both TLCO and Tm, which remained, however, significantly lower than normal ( Tm = –20 ± 3%; TLCO = –13 ± 4%; p < 0.001), whereas Vc was unchanged. On an individual basis, percentage predicted TLCO and Tm were less than 80% in 25 and 28 patients, respectively, before transfusion, and in 14 and 16 patients after transfusion, whereas percentage predicted Vc was more than 120% in the same 16 patients both before and after transfusion.
Blood diffusing resistance decreased with transfusion only because of the larger [Hb]. Because a relatively greater fall occurred for the alveolar–capillary membrane resistance, the ratio of blood to alveolar–capillary membrane resistance computed for standard PCO2 increased with transfusion, changing from 0.41 ± 0.02 to 0.55 ± 0.02 (p < 0.001). The increased conductance of the alveolar–capillary membrane was due mainly to the increase in its effective surface, as indexed by VRBC, because the Tm/VRBC ratio did not change with transfusion (Table 3).
Abnormalities of lung mechanics have been reported by almost all investigators (1–8), but uncertainty persists about their nature and pathogenesis. The present spirometric data collected from the largest number of patients with TM evaluated to date have shown that the only observed lung mechanical abnormality is restrictive (Table 1 and Figure 1). Indeed, whereas FVC, PEF, and FEV1 were reduced, indices of large (FEV1/FVC) and small airway resistance (MEF25) did not differ from predicted normal values. Although iron deposition has been demonstrated in the lungs of patients with TM (9–11), a clear relation between lung hemosiderosis and restrictive disease has not been established. In line with previous studies (2, 3, 5), we found no significant correlation between spirometric variables and serum ferritin levels. On the other hand, both these variables decreased significantly with increasing iron burden, as indexed by age (Figure 1), in agreement with the finding by Factor and coworkers (3) of a strong inverse correlation between total lung capacity or FVC and directly measured iron burden. During normal growth, the air spaces increase disproportionately more than the airway system, this "dysynaptic growth" usually ending after 8–12 years of life (19–21). It has been thus suggested that in children with TM the restrictive disease is the consequence of limited growth of the alveolar compartment (1). However, percentage predicted FVC continued to decline substantially well beyond 12 years of age (Figure 1), in line with previous observations (3, 4). Hence, abnormal development cannot be the only cause of the restrictive defect. On the contrary, this finding suggests that the protracted transfusional therapy and iron overload eventually cause, or at least worsen, the restrictive disease. In all patients, TLCO was substantially reduced (Table 2), as expected on the basis of the low hemoglobin concentration. On the other hand, there is no general consensus concerning whether standardized TLCO is also lower than normal. It should be noticed that in previous studies neither PCO2 nor Tm and Vc were actually measured, standardization having been made according to the Cotes and coworkers (16) equation, which is based on assumptions that might not always be fulfilled. The present TLCO* values (Table 2) concur with those of previous studies (1, 2, 4, 5) showing that lung diffusing capacity is lower than normal. This is consistent with the presence of a restrictive disease that has reduced the surface of the alveolar–capillary membrane because of limited expansion of the airspaces, and/or has increased the thickness of the alveolar–capillary membrane because of fibrosis, interstitial edema, or both. Percentage predicted TLCO* and Tm were significantly correlated with age and serum ferritin levels (Figures 3 and 4). Stepwise linear regression analysis showed that both age and serum ferritin levels were independent, significant predictors of TLCO* and Tm. The inverse correlation with serum ferritin levels suggests that hemosiderosis could play an important role in promoting the observed pulmonary abnormalities. The relationship between serum ferritin and percentage predicted TLCO* has been specifically addressed in only one previous study (22), which, in contrast with the present results, did not find any significant correlation. This discrepancy is most likely the consequence of smaller number of patients (21 versus 62 in the present study), nearly normal percentage predicted TLCO* (89 versus 75%), and markedly lower mean serum ferritin concentration (0.03 versus 2.7 ng · µl–1). Indeed, no significant correlation between serum ferritin levels and TLCO* or Tm was found in the present population when the data from the 15 patients with serum ferritin levels greater than 3 ng · µl–1 were discarded. Moreover, in all these patients percentage predicted TLCO* and Tm (56 ± 2 and 47 ± 3%, respectively) were markedly lower than normal (Figure 4). It could be, therefore, suggested that a serum ferritin concentration of 3 ng · µl–1 represents a threshold above which the probability of lung injury becomes high. In this connection, it is interesting to note that the prognosis for survival without overt cardiac diseases is excellent for transfusion-dependent, thalassemic patients whose serum ferritin concentration is less than or equal to 2.5 ng · µl–1 (23). In the present patients, both the blood diffusing capacity and that of the alveolar–capillary membrane were decreased. The fall of the former was due to decreased [Hb], which was only partly compensated by the concomitant increase in Vc. The latter was, however, sufficient to normalize the blood diffusing capacity with partial restoration of [Hb] by transfusion. Augmentation of the capillary blood volume has often been demonstrated in patients with various cardiopulmonary diseases (18, 24), but there was no evidence of such events in the present patients. Alternatively, the expansion of the pulmonary vascular bed could have been the consequence of repeated transfusions (25). Finally, chronic and marked anemia could have also played a role. Indeed, in patients with sickle cell anemia Femi-Pearse and coworkers (26) found an increase in Vc, similar to the present one, only when hemoglobin content was chronically low (less than or equal to 95 g · L–1). The fall in the diffusing capacity of the alveolar–capillary membrane could be related to the decrease in its surface area and to the increase in its thickness. The surface dependence was demonstrated by the significant positive correlation between Tm and the volume of red blood cells in lung capillaries (Figure 2), whereas increased thickness was suggested by the slope of this relationship being significantly lower than average predicted normal values (Table 2). Its constancy with transfusion (Table 3) supports the notion that this slope represents an index inversely proportional to the thickness of the alveolar–capillary membrane.
Assessment of Tm and Vc in TM patients has been previously attempted in only one study (4) performed on a small number of young Chinese and Malay subjects. The results concur with the present results, showing a reduction of Tm, although less prominent (30 versus 44%), but contrast as far as Vc is concerned, which was found to be normal. This discrepancy could be explained by differences in intrinsic characteristics of the two populations, evolution of the disease, In conclusion, the present study has shown that restrictive disease and reduced lung diffusing capacity are the predominant abnormalities of pulmonary function in patients with TM. Apart from the contribution due to low hemoglobin concentration, the latter defect is determined by a fall in the diffusing capacity of the alveolar–capillary membrane, attributed primarily to an increase in its thickness. This, together with the dependence of the reduced pulmonary diffusing capacity on age and serum ferritin levels, as well as of the entity of restrictive disease on age, suggests that pulmonary dysfunctions in patients with TM are due mainly to lung fibrosis and/or interstitial edema related to iron overload. Received in original form November 7, 2002; accepted in final form May 6, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
0.05. 
A/
distribution, or methodology. 
