Results The analysis of the hematologic data and the clinical symptoms of 74 patients affected by beta-thal int indicate that this disease can be classified into two clinical variants referred to as sub-silent and evident beta-thal int (Table 1). The patients with sub-silent beta-thal int have a reduced total Hb content and show no or very mild clinical symptoms. The patients with evident beta-thal int show severe clinical symptoms, often associated with massive splenomegaly, and 41% of them have undergone splenectomy. The Hb content is significantly lower than that of the patient with sub-silent beta-thal int (t = 7.44; gl 72; p < 0.001). The Hb F level is higher in with evident than in patients with sub-silent beta-thal int, and the difference is statistically significant (t = 2.28; gl 72; p < 0.05). Thus, the severity of clinical symptoms of beta-thal int correlates with a decrease of the total Hb content and an increase of the Hb F content. Normally the Hb F content is positively correlated with the total Hb content, as illustrated in Fig 2. Figure 2 Epo level and Hb F content are not related in patients with β-thalassemia intermedia. The logarithm of the Epo level was plotted against the corresponding total Hb content of each patient. The patients showing an increase of Hb F higher than 40% are represented by red squares, whereas the blue squares denote the patients showing an increase of Hb F equal to or less than 40%. In the 42 iron-deficient but not thalassemic subjects, the Epo level ranging from 2.7 to 1960 mU/1 with a mean value of 160.10 ± 312.60 mU/1, was negatively correlated (r = -0.74) to the Hb content (Fig. 1). The correlation is statistically significant (n = 42; p < 0.001). Figure 1 Negative correlation between total Hb content and Epo level in iron-deficient patients. The logarithm of the Epo level plotted against the corresponding total Hb content of each patient is denoted by a blue square. The linear regression (solid line) fitting the data is described by the following equation: log [Epo] = 4.78–0.33 [Hb]. The correlation coefficient r is -0.74. The mean Epo level of beta-thal int patients is apparently lower than that of iron-deficient patients (Tab. 1). Regarding to total Hb, only between the groups of sub-silent beta-thal int patients and iron-deficient patients the difference is statistically significant (t = 2.17; gl 73; p < 0.05). The mean values of the Epo level and the Hb F content in the two groups of beta-thal int seem to be positively related, but an accurate analysis of single cases indicates (Tab. 2) that the Epo level and Hb F content are unrelated: subjects showing a low content of Hb F and an high level of Epo as well as subjects showing an high content of Hb F and a low level of Epo were frequently observed. This observation is reinforced by the analysis of groups of subjects in which the severity of anemia is similar. Within the group of evident beta tal int patients (the only adapt for this analysis) subjects showing an Hb F level higher or lower than 40% [5,6] were selected. As illustrated in Fig. 2 and Tab. 3, subjects having a very similar Epo level and Hb content fall in both classes of Hb F content, indicating that Hb F content and Epo level are not related. Table 2 Total Hb, Hb F, and Epo level in patients with evident beta thal int. Case Hb g/dl Hb F % EPO mU/l 1 10.5 87 23 4 10.6 70 55 11 7.2 21 132 19 8.2 7 51 20 7.4 24 126 21 9.0 12 233 Table 3 Epo level and Hb F content in a group of evident beta thal int patients showing a similar degree of anemia and divided for Hb F < or > 40%. Patients' group N. cases Hb F and EPO mean levels Evident beta thal int with Hb ≤ 9 g/dl 32 HbF = 13.6% EPO = 92.9 mU/l (25 cases) Hb F = 59.9% EPO = 118.8 mU/l (7 cases) Different sets of mutations of the globin genes are associated with groups of patients characterized by a different severity of the disorder. The spectrum of the beta-globin gene mutations commonly found in the Italian population can be divided in three categories, the first of which includes the -101 C→T, -92 C→T, and IVS II 844 C→G single nucleotide substitutions; the second the -87 C→G, -86 C→G or C→A, and IVS I C→T single nucleotide substitutions, and the third the beta° 39 C →T, IVS I 110 G→A, IVS II 745 C→G single nucleotide substitutions and the frameshift 6 (-A) mutation. These three categories are referred to as silent beta-thal, mild beta-thal, and severe beta-thal mutations, respectively. In the group of 33 patients with sub-silent beta-thal int the frequencies of silent beta-thal, mild beta-thal, and severe beta-thal mutations were 37%, 13% and 50%, respectively (Tab. 4). In the group of 42 patients with evident beta-thal int, the most common and equally frequent (44%) beta-globin gene defects were mild and severe beta-thal mutations. The frequency of silent beta-thal mutations was 12%. Conversely, severe beta-thal mutations were found at high frequency (83%) in thalassemia major patients. Silent beta-thal mutations were detected only in a small fraction (2.1%) of thalassemia major patients. In the 58 patients both heterozygous for a mutation of the beta-globin gene and heterozygous or rarely homozygous for the presence of three or four copies of one alpha-globin locus, the beta-globin gene defects were invariably beta° or severe beta+ thal mutations (Tab. 4 and ref. [13]); the alpha-globin gene mutation was in 54 out of 58 patients a triplicated alpha anti 3.7 I gene cluster; in 3 patients a quadruplicated alpha anti 4.2 gene cluster, and in one patient the homozygous condition for a triplicated alpha anti 3.7 I gene cluster. In these patients, the average Hb content was 10.0 ± 1.38 g/dl (Table 1), the content of Hb F was slightly higher than normal, ranging from 2.0 to 19.3 with an mean value of 5.15 ± 3.9 %, and the average Epo level was 28.45 mU/1, ranging from 7.85 to 60.72 mU/l. This value is lower than that of regularly transfused thalassemia major patients. Although the Hb F content appears to be low in all these patients, their clinical picture is highly heterogeneous, 50% of the patients having a sub-silent beta thal int, 30% an evident beta thal int and 20% a severe anemia with 5 transfusion dependent patients. Ten out of 58 patients (17.2%) have undergone splenectomy. The patients with Hb H disease have a normal Hb F level, and a Epo level lower than that of patients with beta-thal int. In the beta-thalassemia patients and in the control subjects, screened for the six mutations of the promoter region of the gamma-globin gene observed in Italy, only the Ggamma -158 C→T and AgammaT single nucleotide substitutions were detected. In all cases except 7 the AgammaT mutation was to the Agamma -225–222 bp deletion. The frequencies of the Ggamma -158 C→T and AgammaT mutations are significantly different in the various groups of beta-thalassemia patients (Tab. 5), with the lowest frequency of the Ggamma -158 C→T mutation (2.1%) in thalassemia major patients. Table 5 Distribution of the Ggamma -158 C→T and AgammaT mutations in thalassemic patients and in control subjects. GROUP OF PATIENTS Allele frequencies Ggamma-158 C→T AgammaT Sub-silent beta tal int 8.6 27.3 Evident beta tal int 18.3 26.6 beta tal int with genotype triplicated alpha + beta tal 16.6 17.6 Hb H disease 37.5 5.5 Tal major 2.1 35.4 Normal controls 25.0 18.0 In the group of patients with evident beta-thal int the frequency of the Ggamma -158 C→T mutation was 6.2% in 7 patients having a total Hb content > 9 g/dl and 1.6% in 32 patients having a total Hb content ≤ 9 g/dl.