The DSN for CT managed to identify fourteen confirmed cases and seven probable cases during the study period. No case was reported to the mandatory notification system of the HCDCP for the same time period. Surveillance of CT with the current system of the HCDCP has been in existence since 2004 based only on case definitions of symptomatic CT.
Notification in Greece is based strictly on the clinician. The heavy scheduled program of most doctors and their belief that surveillance is time-consuming are factors leading to defects in the reporting system. In previous studies in other countries, underreporting of diseases was attributed to ignorance about the requirements or methods of reporting, including not knowing which diseases are required to be reported, not knowing how a disease should be reported, concerns regarding confidentiality and perceptions that the list of reportable diseases is too extensive [10–13].
However, the DSN succeeded in collecting an adequate number of reports by using a small number of clinicians from various disciplines who collaborated under the same spectrum of cases of CT. A very small number of the reports were duplicated (sent both by neonatologists, pediatricians and ophthalmologists) also revealing the effectiveness of cross matching, hence reducing the possibility of missing any case. A laboratory surveillance network was set up, by the name ToxoSurv, in France in 2006 based on a network of specialized laboratories certified in prenatal and neonatal diagnosis of toxoplasmosis. The system provided more robust and reliable results due to the exhaustive process adopted for notification (all laboratories carrying out the diagnosis in France were invited to participate in the surveillance) . In our study, one reference laboratory (the Hellenic Pasteur Institute) was used to carry out the confirmation of suspected cases, either by serological or PCR techniques, to avoid differences in methods used by laboratories of the selected hospitals.
Response rate was quite high throughout the period of surveillance, but required reminding through telephone communication or by emails. This system of personal contact proved to be efficient and cost-effective (involving one-person part-time job) in a small country like Greece. Both the e-mail based reminding system and the web-based reporting system proved simple, flexible and minimized paperwork.
In our study the prevalence of CT was estimated without taking into consideration the number of induced abortions or terminations of pregnancy after prenatal diagnosis of CT and this could be a reason for underestimating the true birth prevalence of CT in Greece. To overcome this limitation in the future, we believe that a network of pediatricians, neonatologists and obstetricians would be an ideal combination to study the epidemiology of uncommon birth disorders in Greece. Another reason why our data may be an underestimate of the true prevalence of CT in Greece involves administration of treatment to infected pregnant women and their neonates, which could have affected both the number of confirmed cases and the proportion of symptomatic ones.
Compared to epidemiological data in other countries, the birth prevalence of CT in Greece seems much lower than in the European countries involved in the EUROTOXO project (Bénard A, Salmi LR, Mouillet E: Systematic review on the burden of congenital toxoplasmosis in Europe, unpublished), particularly Sweden. It should be noted, however, that the EUROTOXO data were largely collected before 2000. According to the EUROTOXO results, in Austria the birth prevalence of CT was estimated at approximately 7.4 per 10,000 births , in Denmark at 3 per 10,000 births , in Germany at 13 per 10,000 births , in Poland at 11.3 per 10,000 births , in Sweden at 0.73 per 10,000 births , in Switzerland at 5 per 10,000 births  and in Italy (on a regional level) 13.8 per 10,000 newborns . The incidence of symptomatic CT observed in our study was comparable to those observed in England and Ireland between 2002 and 2004 (0.16 [95% CI: 0.08-0.28] per 10,000 births)  and in France (0.34 [95% CI: 0.2-0.5] cases per 10,000 births) in 2007 .
Congenital toxoplasmosis may be characterized as a low-incidence disease in Greece, but not of low importance, due to its potentially severe complications. Epidemiological data for toxoplasmosis in general come from scarce studies throughout the country. Seroprevalence of Toxoplasma infection in the general Greek population ranges from 32.2% to 45% [23–25], whereas a 29.2% seroprevalence was reported in pregnant women  and the seroconversion rate has been estimated to be 3.3% . Variations may be due to a number of reasons, including varying eating habits in different areas, different climatic conditions etc.
Ten symptomatic CT cases at birth (47% of all confirmed and probable cases) were reported during the study. Chorioretinitis was the most frequent manifestation (50% of symptomatic cases). Although there is a low incidence of primary maternal infection in Greece, the uncertain benefit of prenatal screening to a small number of women and children should not outweigh the risk of side effects from invasive diagnosis and the high cost of diagnostic methods and treatment. Studies have shown that in the absence of prenatal treatment, the frequency of chorioretinitis and cerebral lesions was higher .
Geographical distribution of cases was variable. The higher prevalence observed in the regions of Thessaly, Macedonia and Attica may be due to different cli matic characteristics of those areas or different dietary habits of the resident populations. Further studies are needed to elucidate the reasons for such a variable geographical distribution of CT cases.
The estimated lower birth prevalence by using the data of the DSN compared to those in other countries may be attributed to different methodological approaches and the use of different case definitions and classification systems at the time of the study in each country. Additionally, epidemiological results for other countries are based on screening programmes which have been implemented and tested for years in contrast to the results of GCPSU, which were based on an experimental pilot surveillance system. This DSN study only detected lesions evident at birth. However, the burden of CT would be better assessed by long-term follow-up of cases, as congenitally infected newborns that are asymptomatic at birth are at risk of developing ocular lesions during childhood and adolescence, leading to visual impairment. Thus, greater cohorts and long-term follow-up are required to confirm the encouraging results of the present study.