Soluble transferrin receptors and tissue oxygenation in non anaemic cystic fibrosis patients

Transcription

1 Journal of Cystic Fibrosis 4 (2005) Original article Soluble transferrin receptors and tissue oxygenation in non anaemic cystic fibrosis patients Elena Christoforou a,c, Ioannis Papassotiriou b, *, Christophille Skarmoutsou a, Stavros Doudounakis d, Alexandra Stamoulakatou a, Emmanuel Kanavakis c a Haematology Laboratory, Aghia Sophia Children s Hospital, Athens, Greece b Department of Clinical Biochemistry, Aghia Sophia Children s Hospital, Athens, Greece c Medical Genetics, Athens University, Aghia Sophia Children s Hospital Athens, Greece d Cystic Fibrosis Unit, Aghia Sophia Children s Hospital, Athens, Greece Received 6 August 2004; accepted 10 May Abstract Background: Chronic pulmonary disease and progressive tissue hypoxia are major causes of morbidity and mortality in cystic fibrosis (CF). Normally the body adapts to tissue hypoxia by increasing the red cell mass and decreasing the Hb-O 2 affinity. These adaptations are commonly observed in patients with cyanotic heart disease and individuals living at high altitude. However, patients with CF not only have an impaired erythroid response to hypoxia, but also are frequently anaemic. Methods: In order to evaluate erythroid marrow activity and tissue oxygenation in 37 patients with CF we measured: the haematological and blood chemistry parameters; including red cell indices, ferritin, erythropoietin (Epo) and soluble transferrin receptors (stfr) levels; arterial blood gases, P 50 and oxygen release to the tissues (O 2 R ) and the 2,3-BPG levels. Results: The main results showed that a) patients with CF have a mild degree of tissue hypoxia which is expressed by the moderately decreased of P 50 and O 2 R values and the relative increase of Epo level, b) 2,3-BPG synthesis in patients with CF is normal and c) stfr levels are significantly increased (3-fold normal) in patients with CF compared to normal controls. Conclusions: The above observations indicate that erythroid marrow activity in patients with CF is increased. D 2005 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Cystic fibrosis; Erythropoietin; Transferrin receptors 1. Introduction * Corresponding author. Tel.: ; fax: addresses: biochaem@paidon-agiasofia.gr, jpapasotiriou@ath.forthnet.gr (I. Papassotiriou). Cystic Fibrosis (CF) is the most common life-limiting autosomal recessive genetic disorder in Caucasians, caused by mutations to the CF transmembrane conductance regulator (CFTR) gene on chromosome 7, resulting in defective regulation of chloride transport by epithelial cells [1,2]. Pulmonary disease is progressive and ultimately fatal although there are notable variations in clinical features. The major cause of morbidity and mortality in cystic fibrosis is permanent lung injury. Many proposals have been forwarded trying to explain the nature of increased bacterial infection in the lungs of persons with CF. These have ranged from suggestions that inflammation may be a primary event even before bacterial infection can be demonstrated, to virus-induced bacterial infection to CFTR mediated defects in antibacterial properties of the airways surface to innate local host response defects [3 5]. Endobronchial infection, first from Staphylococcus aureus and then colonization with Pseudomonas aeruginosa is, associated with increased airway inflammation and progressive lung injury [1,2,5]. Chronic pulmonary inflammation and tissue destruction leads to progressive tissue hypoxia [1,3]. Normally adaptation to tissue hypoxia is characterised by an increasing red cell mass and decreasing haemoglobin oxygen affinity. Significant alterations in erythropoiesis in /$ - see front matter D 2005 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. doi: /j.jcf

2 152 E. Christoforou et al. / Journal of Cystic Fibrosis 4 (2005) cystic fibrosis are suggested by the observation that polycythaemia is uncommon, even among cyanotic patients [6 12]. This is mainly due to the inadequate erythroid compensation for hypoxia, or a relative anaemia in some patients with CF who had normal haemoglobin values in the presence of chronic arterial hypoxemia. Relative anaemia should be regarded as a haemoglobin concentration that is inappropriately low for tissue oxygen demands, rather than as strictly a haemoglobin value that is below the 95th percentile limits for age and sex. Iron availability and disturbances of the regulation erythropoietin (Epo) secretion are partly responsible. Iron deficiency anaemia (IDA) and iron deficiency (ID) are common in this disease. The etiology is undoubtedly multifunctional and is normally attributed to a combination of chronic inflammation, gastrointestinal and pulmonary blood loss, pancreatic insufficiency, malnutrition and supplementary pancreatic enzymes [13 17]. ID is directly related to the severity of suppurative lung disease [19,20]. Pseudomonas aeruginosa probably contributes directly by actively acquiring iron from the host airway and indirectly by stimulating the production of circulating cytokines to divert iron away from haemoglobin synthesis and promote local iron storage as intracellular ferritin. Traditional indices used for estimating iron status can be markedly affected by the acute phase response in patients with inflammatory conditions such as patients with CF. This may hamper the investigation of ID and discrimination of IDA and anaemia of chronic disease (ACD). During the acute phase response transferrin concentration falls and ferritin concentration rises [21]. In contrast soluble transferrin receptors (stfr) levels, a new promising tool to estimate iron status is not affected by the presence of inflammation, infection or liver disease [22 26]. The transferrin receptor (TfR) is an integral membrane protein found on the cell surface of virtually all mammalian cells. Erythroid cells contain the highest number of TfR (80%) [27,28]. The number of TfR present in a cell is tightly regulated by both the iron status and the proliferation status of the cell. Regulation is effected by the binding of a protein (iron responsive element binding protein, IRE-BP) to special stem-loop structures (iron responsive elements, IRE) on the TfR mrna [29,30]. As a result of externalization of TfR during the endocyting cycle, a soluble form of TfR (stfr) can be detected in serum. Concentration of stfr is increased in iron deficiency anaemia and in disorders that are associated with expanded erythropoiesis or ineffective erythropoiesis. The level of stfr in the plasma has been shown to be closely related to the number of red cell precursors in the bone marrow and to provide a non-invasive reliable quantitative assessment of the rate of total erythropoiesis in the absence of iron deficiency. Also stfr measurement appears to be particularly useful in the differential diagnosis of IDA, (where stfrs are increased) and ACD (where stfr is regularly in normal range) [28,31,32]. Levels of circulating Epo and TfR as well as reticulocyte count may be used as a basis for functional classification of anaemia relative to defects of proliferation (hypoproliferative anaemia), maturation (ineffective erythropoiesis) or red cells survival (peripheral haemolysis). In this study we investigated the iron metabolism status, the degree of tissue hypoxia and the erythropoiesis in patients with cystic fibrosis with mild to moderate pulmonary disease. 2. Patients and methods Forty-eight randomly selected patients with CF were evaluated (mean age 11 years; age range 16 months to 33 years; 29 women). Anaemia was defined as a haemoglobin concentration that is less than lower limit of the normal value for age for each patient. Iron status was determined by measurements of iron, ferritin and transferrin and red cell indices. The cut off values used to detect abnormality for the various tests were determined according to age of every patient. Three of the CF patients had IDA, three ID, one ACD, two were carriers of b-thalassemia and two had acute infection and they were excluded. The rest of the CF patients (37 patients, mean age 11 years; range 2.5 to 33 years; 22 women) participated in the study. These patients had mild to moderate pulmonary disease. The severity of the disease was assessed by the Shwachman Kulczycki score (mean 76; range 52 to 93), standard pulmonary function tests: vital capacity (mean 85.5%; range 34 to 132), forced expiratory volume in one second (mean 78.5%; range 29% to 132%), and measurements of arterial blood gases: PaO2 (mean 97 mmhg; range 64.4 to mmhg), PaCO2 (mean 36 mmhg; range 24.8 to 49.2 mmhg) and ph (mean 7.4; range 7.34 to 7.52). All patients had CRP <3.4 mg/l, indicating the absence of an acute inflammatory process. Thirty were infected with PA, which was confirmed by a routine microbiological culture. Thirty-four received dietary supplements of pancreatic enzymes and fat-soluble vitamins, including vitamin E. Twenty normal non-anemic individuals matched for age with CF patients served as controls Analytical procedures The haematological parameters including reticulocyte counts and indices were evaluated automatically, using the ADVIA R 120 Haematology System (Bayer Corporation, Tarrytown, NY, USA). The cut off values used to detect abnormality for the various haematological tests were determined according to age of each patient (control). Reticulocyte counts were corrected according to the formula: Retics count ðpatient 0 s Hct=45ðassumed normal HctÞ to represent the Reticulocyte Production Index (RPI) in non anemics [33]. Serum ferritin and erythropoietin levels were determined by chemiluminescence using the autoanalyzer

3 E. Christoforou et al. / Journal of Cystic Fibrosis 4 (2005) Hb (g/dl) FVC and FEV1(%) Age (years) Fig. 1. Correlation between haemoglobin and age of cystic fibrosis patients studied (r =0.477 p <0.01). (Nichols Institude Diagnostics, CA, USA). The inter-and intra-assay CVs were less than 5% for both analytes. stfr and TRF assays were performed by fully mechanized latexparticle-enhanced immunonephelometric assays on the BN ProSpec nephelometer (Dade Behring, Liederbach, Germany). The inter-and intra-assay CVs were less than 7% and 6% respectively. Oxygen equilibrium curves (OEC) at the liquid phase (50 mm Bis-Tris, 140 mm NaCl, ph 7.40 at 37 -C) were performed using an automatic continuous method (Haemox-Analyzer) [13]. The P 50 values were calculated from the OEC s. Red cell 2,3-BPG levels were measured enzymatically (Boehringer Mannheim, Germany). Tissue oxygenation data were obtained either from the OEC s or from the Siggaard Andersen s Oxygen Status Algorithm (OSA), which permits computer simulations for most of data referring to oxygen delivery. Total oxygen carrying capacity parameter was calculated by multiplying the patient s haemoglobin concentration by 1.36 (quantity of O 2 bound by 1g of fully saturated Hb and expressed as vol %). Oxygen release to the tissue at the mixed venous PO 2 parameter (oxygen extraction tension) was calculated from the OECs or/and the OSA. Cardiac output parameter (Q) was calculated by simulations using the Fick equation: VO 2 ¼ 0:136 Q Hb ðsao 2 SVO 2 Þ, where: VO 2 is the amount of oxygen released per minute (l/min), Q is the blood flow (l/min), and SAO 2 and SVO 2 are the arterial and mixed venous oxygen saturations respectively Statistical analysis Data are expressed as meantsd. The correlation coefficient r between the parameters tested was computed using least squares regression analysis. The p values reported are two tailed (Fig. 1). All the statistical procedures were performed using the STATGRAFICS 5.1 for Windows program (Graphic Software System) while the regression plots have been prepared using the Sigma plot version 8.1 program (Sigma-Aldrich Chemical Co). 3. Results Erythropoietin (IU/l) Fig. 2. Correlation between FVC, FEV1 values with Epo levels (open circles FVC, closed circles FEV1) (r = 0.327, p <0.05 and r = 0.345, p <0.05 respectively). The main results of the study are summarized in Table 1 for the entire group. No significant differences were found in the haematological parameters between CF patients and controls. There was an inverse relationship between Hb concentration and PO 2 (r = 0.354, p <0.05). Hb concentration was directly related to an increase in age (r =0.477 p <0.01). Despite the fact that 2,3-BPG concentration did not differ from normal subjects 46% of CF patients had higher than normal levels and only 5.4% had lower than normal levels. The increase of 2,3-BPG concentrations was in part related and restores the increase of P 50 (r =0.586; p <0.01) and the increase of O 2 R (r = 0.546; p <0.01). The mean O 2 R Table 1 Haematological, biochemical and physiological data of patients with CF compared to normal controls Cystic fibrosis Controls Difference Hb (g/dl) 13.2T0.9 ( ) 13.5T0.8 ( ) p >0.18 RPI 1.2T0.2 ( ) 1.0 p >0.60 Ferritin (Ag/l) 30.9T16.8 ( ) 35.2T21.0 ( ) p >0.46 stfr (mg/l) 3.0T1.9 ( ) 1.2T0.2 ( ) p < Epo (IU/l) 13.9T5.3 ( ) 9.2T4.2 ( ) p < ,3-BPG (mmoles/l) 5.0T0.5 ( ) 4.9T0.2 ( ) p >0.09 P 50 (mmhg) 26.3T1.6 ( ) 26.6T0.4 ( ) p >0.4 O R 2 (vol.%) 4.7T0.7 ( ) 5.0T0.2 ( ) p <0.02

4 154 E. Christoforou et al. / Journal of Cystic Fibrosis 4 (2005) stfr (mg/l) Erythropoietin (IU/l) Fig. 3. Correlation between stfr and Epo levels (r =0.373, p <0.05). was lower compared with normal subjects ( p <0.02) (19% of CF patients had higher and 51% had lower than normal values). Thus, showed that the most of the patients did not have a compensatory increase in P 50 for hypoxia and appropriate shift of the OEC to the right. No significant relationship was demonstrated between Hb, FEV1, PH, Q or 2,3-BPG. Circulating Epo levels were found 1.5 fold increased in patients with CF compared to normal controls. The increase of Epo levels correlated with the decrease of FEV1 and FVC (r = 0.327, p < 0.05 and r = 0.345, p < 0.05 respectively) (Fig. 2). None of the patients had low Epo levels. We found no significant correlation between Epo and Hb concentrations, PaO 2, oxygen saturation, P 50 or Q values. stfr levels were significantly increased compared to normal controls (2.5 fold normal). The increase in stfr levels was directly related to an increase in Epo levels (r = 0.373, p <0.05) (Fig. 3) and to a decrease in P 50 (r = 0.426, p <0.01) (Fig. 4) and O R 2 (r = 0.437, p <0.01). (increased stfr levels=60%, none had lower than normal). There was no significant relationship between stfr levels and Hb concentration, RPI, FEV1 or PO2. There was an inverse relationship between PO 2 and Q (r = 0.510, p <0.01). When 37 CF patients were divided in to three groups according to the severity of disease (S-K score group A=89, B=80, C=54) or according to FEV1 no significant correlation was found between Epo or stfr and severity of disease. The CF subjects who had elevated stfr levels (>2.5 mg/l) (59%) had significant increase in Epo. None of the patients had lower than normal stfr levels. examined, evidence of iron deficiency (ID) with or without anaemia found in 14.5%. Previous studies demonstrate a progressive increase in anaemia and iron deficiency with increasing age [11]. The incidence of iron deficiency in CF patients increase from 33% [13] in children to 62% [12] in adults. The 37 patients who were examined in detail had haemoglobin and haematocrit values that did not significantly differ from normal controls. Their serum ferritin, transferrin and red cell indices (MCV, MCH, MCHC, RDW) are within normal range and no correlation was found between haemoglobin and ferritin or transferrin. These patients did not have evidence of anaemia or ID and their Hb, Ht, MCV, MCH increased with increasing age, despte the fact that their general clinical status and pulmonary function declined with age, in keeping with the inverse relationship between age and S-K score and FEV1. From the literature the mean value of Hb is normal in CF patients but this is partly due to the large percentage of ID in the previously studied populations [10,11,35]. In this study in which we have no evidence of (ID) the mean value of haemoglobin is also at normal levels. Between CF patients and normal subjects there was no difference in the retics count and reticulocyte production index (RPI) indicating the absence of haemolysis. Although our CF patients failed to developed polycythaemia this study indicates that progressive arterial hypoxemia in our patients is accompanied by an increase in haemoglobin concentration. A negative correlation was found between PO2 and Q showing an increase in cardiac output with deterioration of arterial hypoxemia. In this study the levels of 2,3DPG and OEC S were determined in each of the CF patients. This was to determine whether a marked shift in Hb oxygen affinity stfr (mg/l) Discussion In this study the population included children, adolescents, and young adults with CF (patient s age ranged from 16mts to 33 yrs) and mild to moderate pulmonary disease. From, the group of the 48 patients who were initially P 50 (mmhg) Fig. 4. Correlation between stfr levels and P 50 values (r = 0.426, p <0.01).

5 E. Christoforou et al. / Journal of Cystic Fibrosis 4 (2005) could explain the impaired erythroid response to hypoxia. From the OEC S P50 and O2R where calculated. Our CF patients had a mild degree of tissue hypoxia, which is expressed by the moderately decreased O 2 R. In 19% of our patient s O 2 R was increased and in 51% it was decreased. The increase in 2,3-BPG is in part related and causes the increase in P 50 and the resulting increase of O 2 R levels as expected. Although this direct correlation exists, the 2,3- BPG and P 50 did not differ from normal subjects. Only 5,4% of patients had decreased 2,3 BPG levels. None of our patient s had hypophosphataemia. Thus the normal levels of Hb concentration could not be defined by the reparative decrease in haemoglobin oxygen affinity in the presence of hypoxemia. The synthesis of 2,3-BPG from red cells is influenced from a variety of factors. Such factors are anaemia, hypoxia and alkalosis. No significant correlation was found between Hb (all patients had haemoglobin concentration at normal levels), ph, FEV1 or Q and 2,3-BPG but we found a positive correlation between P 50 and PCO 2. The normal response to hypoxia includes an increase in serum Epo. Although erythropoietin levels were normal or increased (1.5 fold for the entire group) this was inadequate for the final increase of haemoglobin. No correlation is noted between Epo and ferritin, transferrin, PO 2, oxygen saturation, Hb concentration, P 50 or O 2 R. The Epo response in these patients is correlated with the severity of pulmonary function (FEV1 and FVC). Serum erythropoietin concentration is primarily related to its renal production and, under the stimulus of hypoxia, increases exponentially as haemoglobin decreases. Recent studies indicate that at any given Hb level, the higher the number of the red blood cell precursors, the lower the sepo concentration [27]. Based on these findings probably the sepo levels are regulated not only by the rate of renal production, but also by the rate of utilization by erythroid cells. Some studies in literature show that the Epo levels in CF patient s have no differences from those found in normal subjects but others show Epo levels are increase or in high normal values [10,11]. stfr are derived primarily from erythroid precursors in bone marrow, and their concentration in the absence of iron deficiency provides a reliable non invasive assessment of total erythropoiesis erythroid marrow activity [28,31]. A 2.5-fold increase of stfr levels was found in this series of patients. This response in erythroid marrow activity was associated with an increase in erythropoietin level, as demonstrated by the significant positive correlation between stfr and Epo, and is stimulated by the decrease to P 50 and O 2 R. As expected, is not related to haemoglobin level (Hb levels of the patients were within normal levels). Furthermore, no significant correlation was found between stfr levels and patient s age. None of the CF patients had low stfr levels indicating the absence of repression of myeloid activity. No significant correlation was found between stfr and RPI. Previous studies suggest that possible mechanisms for anaemia or relative anaemia (normal haemoglobin concentration in the presence of tissue hypoxia) are caused by a primary defect in red cell production or a secondary deficiency in Epo production and iron availability [11,36]. In chronic inflammatory disease Epo response is blunted and a similar observation has been made in CF patients with anaemia [11]. In this study none of the CF patients had anaemia or iron deficiency and Epo production is increased, and initiating, regulating and inducing the increase of erythroid marrow activity. Nevertheless haemoglobin concentration was normal. That reinforces the hypothesis that other factors contribute to the inadequate erythroid response to hypoxia. The data presented here demonstrate that CF patients with tissue hypoxia but with no other factors, which stimulate erythropoiesis, have increased erythroid marrow activity, which is directly related to tissue hypoxia. Moreover, the progress of arterial oxygen hypoxemia (PO 2 ) and tissue hypoxia (O 2 R ) resulted in an increase of cardiac output [11]. References [1] Welsh MJ, Tsui LC, Boat TF, Beaudet AL. Metabolic and molecular basis or inherited disease. In: Scriver CR, Beaudet AL, Sly WS, Vade D, editors. Cystic fibrosis. McGraw-Hill. [2] Welsh Michael J, Smith Alan E. Cystic fibrosis. Scientific American. [3] Koch C, Hoiby N. Pathogenesis of cystic fibrosis. Lancet 1993; 341: [4] Hoiby N. Inflammation and infection in cystic fibrosis hen or egg? Eur Respir J 2001;17:4 5. [5] Robinson P. Cystic fibrosis. Thorax 2001;56: [6] Caplan A, Gross S. Haematologic and serologic studies in cystic fibrosis. J Pediatr 1968;73: [7] Reid L, Ryland D. The pulmonary circulation in cystic fibrosis. In: Mangos JA, Talamo RC, editors. Fundamental problems of cystic fibrosis and related diseases. New York Intercontinental Medical Book Corporation; p [8] di Sant Agnese P, Davis P. Cystic fibrosis in adults. Am J Med 1979;66: [9] Wagener JS, Corrigan JJ, McNeil GC, Lemen R, Taussing LM. Ferrokinetic and haemotologic studies in cystic fibrosis patients. Am J Pediatr Hematol Oncol 1983;5: [10] Ater JL, Herbst JJ, Landaw SA, et al. Relative anemia and iron deficiency in cystic fibrosis. Pediatrics 1983;71: [11] Vichinsky Elliott P, Pennathur-Das Rukmani, Nickerson Bruce, et al. Inadequate erythroid response to hypoxia in cystic fibrosis. J Pediatr 1984;105: [12] Pond MN, Morton AM, Conway SP. Functional iron deficiency in adults with cystic fibrosis. Respir Med 1996;90: [13] Erhardt P, Miller MG. Iron deficiency in cystic fibrosis. Arch Dis Child 1987;62: [14] Zempsky WT, Rosenstein BJ, Carroll JA, et al. Effect of pancreatic enzyme supplements on iron absorption. Am J Dis Child 1989; 143: [15] Cucchiara SF, Satamaria F, Andreotti MR, et al. Mechanisms of gastro oesophageal reflux in cystic fibrosis. Arch Dis Child 1991;66: [16] Elin RJ, Wolf SM, Finch CA. Effect of induced fever on serum iron and ferritin concentrations in man. Blood 1997;49: [17] Reid David W, Withers Nicholas J, Francis Libby, Wilson John W, Kotsimbos Thomas C. Iron deficiency in cystic fibrosis: relationship to

6 156 E. Christoforou et al. / Journal of Cystic Fibrosis 4 (2005) lung disease severity and chronic pseudomonas aeruginosa infection. Chest 2002;121: [19] Ater Joann L, Herbst John J, et al. Relative anaemia and iron deficiency in cystic fibrosis. Pediatrics 1983;71: [20] Van Asbeck BS, Verhoef J. Iron and host defence. Eur J Clin Microbiol 1983;2:6 10. [21] Gabay C, Kushner I. Acute phase proteins and other systemic responses to inflammation. N Engl J Med 1999;340: [22] Ferguson BJ, Skikne BS, Simpson KM, Baynes RD, Cook JD. Serum transferrin receptor distinguishes the anaemia of chronic disease from iron deficiency. J Lab Clin Med 1992;119: [23] Nielsen OJ, Andersen LS, Hansen NE, Hansen TM. Serum transferrin receptor in anaemic patients with rheumatoid arthritis. Scand J Clin Lab Invest 1994;54: [24] Petterson T, Kivivuori Sm, Slimes SA. Is serum transferring receptor useful for detecting iron-deficiency in anaemic patients with chronic inflammatory diseases? Br J Rheumatol 1994;33: [25] Punnonen K, Irjala K, Rajamaki A. Serrum transferrin receptor and its ratio to serum ferritin in the diagnosis of iron deficiency. Blood 1997;89: [26] Suominen P, Punnonen K, Rajamaki A, Irjala K. Evaluation of new immunoenzymometric assay for measuring soluble transferrin receptor to detect iron deficiency in anemic patients. Clin Chem 1997;43: [27] Iacopetta BJ, Morgan Eh, Yeoh GCT. Transferrin receptors and iron uptake during erythroid cell development. BBA 1982;687: [28] Cook JD, Skikne BS, Baynes RD. Serum transferrin receptor. Annu Rev Med 1999;44: [29] Müllner EW, Neupert B, Kuhn LC. A specific mrna binding factor regulates the iron dependent stability of cytoplasmic transferrin receptor mrna. Cell 1989;58: [30] Casey Jl, Hentze MW. Iron-responsive elements: regulatory RNA sequences that control mrna levels and translation. Science 1988;240: [31] Ponka Prem, Chun Nam Lok. The transferrin receptor: role in health and disease. Int J Biochem Cell Biol 1999;31: [32] Feelders Richard A, Kuiper-Kramer Ellen PA, van Elijk Henk G. Structure, function and clinical significance of transferrin receptors. Clin Chem Lab Med 1999;37:1 10. [33] Hutchinson RE, Davey FR. In: Henry JB, editor. Hematopoiesis, clinical diagnosis and management by laboratory methods, 19th edition. WB Saunders. [35] Mavromichalis J, Zannos-Mariolea L, Nicolaidou M, Georgatou Ph, Dentaki-Slovaki K, Matsaniotis N. Haematologigal findings in Greek children with cystic fibrosis. Helv Paediatr Acta 1974;29: [36] Rosenthal A, Khaw KT, Scwachman H. Haemoglobin oxygen equilibrium in cystic fibrosis. Pediatrics 1997;59: