Accurate epidemiological data about the diffusion and transmission of HCV infection are necessary in order to understand the burden of the disease. Assessment of HCV infection prevalence and the identification of all of the associated risk factors are required in order to plan strategies that will interrupt the different patterns of HCV transmission. It is widely documented that the parental route is the most important route of HCV transmission: injection drug use, blood transfusions and other unsafe health-care related procedures are the most frequently mentioned risk factors. Injection drug use is generally considered to be the major risk factor of new infections in developed countries, while transfusions and unsafe therapeutic injections still represent the main source of infection transmission in developing areas .
On the other hand, data concerning sexual and intrafamilial transmission remain unclear and the role played by the family in the spread of HCV infection has yet to be defined. Some case reports have been published that clearly demonstrate sexual transmission of HCV [20, 21]. In addition some epidemiological and case-control studies have confirmed that intrafamilial transmission occurs [3, 22, 23], while others have argued that household exposure non-significantly increases the risk of HCV positivity [24–26]. This difference in the results of different studies may be a consequence of several factors such as the ELISA tests used to detect anti-HCV antibodies (first and second-generation ELISA used in previous studies have a relatively low sensitivity in relation to the newer third-generation assays), geographic area, viremia levels, sexual behaviors of the subjects studied, selection of the studied population, statistical methods used to assess within-household clusters of HCV infection and genotypes.
We conducted a cross-sectional study on the intrafamilial transmission of HCV. In our study the prevalence of anti HCV positivity in all household contacts was 8.9%, which is higher than in the general population, which is in agreement with other authors. The highest prevalence was found in sexual partners (12.1%), followed by children (2.6%). Excluding those with previous parenteral exposure, the proportion of anti-HCV antibodies in family contacts was 6.2%. These results are similar to the prevalence of anti-HCV in our control group (5.2%), supporting the hypothesis that Southern Lazio is an area with an elevated level of HCV in circulation . Moreover, the prevalence of anti-HCV in sexual contacts (12.1%) was higher than in non sexual contacts (2.5%), suggesting that HCV could be sexually transmitted. But re-analyzing data removing those contacts with known parenteral exposure, we found that stable sexual relationships cannot be considered a risk factor. Interestingly, the duration of the relationship plays a fundamental role in all of the multivariate analysis models, suggesting that the difference in the prevalence rates among sexual and non sexual contacts could be explained simply by a time factor. Beside this, we found interestingly that being retired from work is an independent factor associated to HCV positivity among contacts of HCV chronic patients. This result does not concur with other studies that have shown that sexual transmission of HCV may occur in spouses, while it does confirm that the spreading of the virus to other family members seems to be rare . However, there are conflicting data in the scientific literature regarding sexual diffusion of HCV. Evidence that supports the hypothesis for this mode of transmission is that in genotyping the HCV, some authors have detected the same genotype in the spousal couple [23, 24, 29, 30]. This is in contrast with the results obtained from other studies in which the same genotype was found in very few couples, suggesting that the risk of sexual transmission was low . Moreover, the finding that many of the cases of intrafamilial spread involved siblings suggests that shared use of domestic objects may played a more important role than sexual transmission . However, this method of transmission does not appear to be predominant when compared with the transmission of other viruses (HIV and HBV) which can be transmitted by both vertical and horizontal (sexual and non-sexual) means .
Furthermore, we investigated the duration of the relationship with the index cases between anti-HCV positive contacts and anti-HCV negative contacts. According to the distribution of HCV positive sexual contacts and the duration of the relationship, the frequency of detection for anti-HCV increased in proportionately with the duration of the relationship. This result agrees with the finding of other studies that anti-HCV positivity increases with the length of the marriage [29, 30, 34, 35].
The higher prevalence among older contacts may be due to the duration of the relationship with index cases but also to some behaviors once common in Italy, such as the re-use of medical material (syringes for intramuscle injections) administered at home .
Among the risk factors that we analyzed with multivariate analysis, those significantly associated with HCV positivity in household contacts were being a drug addict and having intercourse with drug addicts. Moreover, the length of exposure and the status of the partner could also explain the fact that the retirement status is associated with HCV positivity. As far as drug addiction is concerned, our survey confirms the findings of other Italian studies [15, 28, 37–39].
In a recent review, Ackerman and coll.  found that intrafamilial HCV transmission is associated with the severity of liver disease in the index patients, the number of family members infected with HCV, the duration of exposure and sexual contact with the index patient. Interestingly, out of 67 studies selected in this review, 25.4% were from Italy. This cross-sectional study confirms that tattoos are a means of HCV transmission .
This study has several limitations. First of all, it is a cross-sectional study and it was not possible to assess the time-relationship between types of exposure and outcome (HCV seropositivity), as in a cohort study. As an example, we are aware that independent infection among family members cannot be excluded, and temporal trends cannot be determined. Anyway, we do know that index cases are long-term patients so it is likely that family members became HCV positive after the cases did.
As far as information bias is concerned, this type of error could have been avoided as we used a very well established means of collecting information on past exposures and risk factors. With regards to selection bias, we collected information regarding the contacts of the index patients attending our Hepatology Ambulatory in Rome and from other hospitals in Southern Lazio, and we believe that our population is a random sample of patients with HCV in this Region.
Another limitation of the study refers to the lack of assessment for HCV genotype, so we were not able to address whether differing genotypes have differing rates of transmission.
In conclusion, the substantial percentage of seropositive contacts with no identifiable risk factors support the hypothesis of unapparent parenteral transmission of the virus.
Furthermore, the results of such analyses are useful in generating hypotheses about possible causes of disease, rather than drawing defined conclusions .