Our interest in potential interactions between mood disorders of the bipolar spectrum and blood conditions typical of the Mediterranean area, such as glucose-6-phosphate dehydrogenase (G6PD) deficiency and heterozygous β-thalassaemia, was first aroused in the 1980 s. At that time, chromosomes Xq28 and 11p15, where the G6PD gene and the β-globin gene cluster are located, became very popular in psychiatric genetics following the reports of positive linkage with manic-depression 123.

We subsequently surveyed G6PD deficiency and heterozygous β-thalassaemia as phenotypic markers to be used in linkage studies 4. Unexpectedly, we found excessive proportions of the two markers among bipolar patients, especially those with psychotic symptoms 567. Among potential mechanisms underlying the observed associations, we considered linkage disequilibrium between the markers and mood disorder genes located in Xq28 and 11p15. This hypothesis, which at first was most intriguing, weakened considerably when subsequent studies reported diminished evidence of Xq28 or 11p15 linkage 89. Over the following years, the use of phenotypic markers in psychiatric genetics vanished due to the advent of molecular DNA markers. Nevertheless, with regard to heterozygous β-thalassaemia, our interest was periodically revived by the appearance of relevant reports. On the one hand, co-occurrence of the thalassaemic trait and bipolar disorder was reported in several cases from various countries, including Japan, where thalassaemia is very rare 10111213. On the other hand, a psychiatric population surveyed in Argentina revealed excessive prevalence of microcytic anaemia among patients suffering from recurrent mood disorder 14. Given the latter series of intriguing reports, an extension of our previous surveys was warranted: this is an updated account of haematological parameters regarding 1014 patients.

Outpatients consecutively admitted to the Department of Neuroscience, University of Cagliari, between 1980 and 2000 were studied. The department is one of the reference centres for the management of lithium and related treatments of mood disorders in southern Sardinia. On admission, detailed information was obtained from the patient and from other available sources (such as relatives and referring clinicians) concerning demographic characteristics and lifetime medical and psychiatric history. Moreover, all available records regarding previous prescriptions or hospitalisations were examined. Psychiatric diagnosis was made according to modified Research Diagnostic Criteria (RDC) 157. Routine laboratory tests, including blood count and serum iron concentration were available for 1014 patients. All subjects gave their informed consent to participate in the study.

Additional procedures were used aiming at the identification of heterozygous β-thalassaemia, which is ordinarily diagnosed based on reduced MCV, increased haemoglobin A₂ (HbA₂), and normal serum iron concentration. In an initial survey 5, haemoglobin electrophoresis was systematically performed in samples from 180 patients who had attended the centre between April 3rd and May 3rd, 1989. Thereafter, MCV lower than 80 μ³ was chosen as the cut-off for the performing of haemoglobin electrophoresis and/or chromatography (98% of cases with abnormally high HbA₂ concentration from the initial survey had MCV<80 μ³). Overall, electrophoresis and/or chromatography were obtained for 143 patients with MCV<80 μ³. Of the latter, 62% had increased HbA₂ (heterozygous β-thalassaemia); 22% had normal HbA₂ and normal serum iron (in Sardinia, this pattern has been characterised molecularly as heterozygous α-thalassaemia in the large majority of cases 16). In the remaining 16%, HbA₂ concentration was normal but serum iron was reduced, therefore heterozygous β-thalassaemia could not be ruled out (iron-deficiency may maintain HbA₂ within the normal range).

Overall, MCV<80 μ³ was found in 23.1% of patients (Table 1). Greater proportions were found among subgroups with bipolar course compared with the remaining subgroups (p = 0.02 at the Fisher's exact test; 183/732 = 25.0% versus 51/282 = 18.1%; relative risk = 1.38). The highest proportion was found in Manic Schizoaffective disorder, predominantly affective subtype (74/219 = 33.8%).

Heterozygous β-thalassaemia (increased HbA₂ concentration) accounted for the majority of patients with MCV<80 μ³. Rates did not vary between the initial survey (27/42 = 64%) and subsequent cases who underwent HbA₂ analysis (62/101 = 62%). If similar rates were assumed regarding the remaining 91 patients with MCV<80 μ³ but unknown HbA₂ concentration, prevalence of heterozygous β-thalassaemia in the overall sample could be estimated as at least 16.4% among bipolar and 9.9% among non-bipolar subgroups. A simulation of the Fisher's exact test provided a bipolar/non-bipolar relative risk of 1.65 with a p value of 0.01. Estimated rates of heterozygous β-thalassaemia across psychiatric diagnoses are shown in Table 1. The subgroup of patients with Manic Schizoaffective disorder, predominantly affective subtype (not shown), had the highest estimated rate of heterozygous β-thalassaemia (22%).

The distribution of patients with MCV<80 μ³ and normal HbA₂ concentration did not vary across psychiatric diagnoses (not shown).

The results from this survey and the series of intriguing literature reports are consistent with the hypothesis that heterozygous β-thalassaemia might play a role as a susceptibility factor in mood disorders, particularly of the bipolar spectrum. Since the evidence cannot be considered established, further studies are warranted. Potential approaches include: a) replication surveys of microcytic anaemia and the thalassaemic trait in psychiatric populations from high prevalence areas; b) surveys of bipolar spectrum disorders among subjects with an already established diagnosis of heterozygous β-thalassaemia (such as relatives of patients with Cooley's anaemia or participants in specific screening programs); c) family studies of patients with bipolar disorder and an established thalassaemic trait.