Many studies have reported the epidemiology of C. trachomatis infection throughout the world. The prevalence of C. trachomatis infections varies from 2% to 5.6% in selected and unselected populations . In line with previous studies, the prevalence of C. trachomatis infection in our city of Sfax was found to be 4.2%. To our knowledge, C. trachomatis genovar distribution has not been reported before in our country. C. trachomatis genotyping was investigated in all urogenital samples collected in Sfax (Tunisia) found to be Cobas amplicor PCR positive since 2000. Ninety nine percent (137/138) of our samples were successfully amplified using our semi nested PCR. This sensitivity is high when compared with those reported in other studies ranging from 44 to 99% [10, 13–15].
The reverse hybridization used in our study is simple and fast. It requires low technology to be performed. It is powerful in the diagnosis of mixed infection and did not show any cross reactivity between reference strains.
In our study, genotype E was the most prevalent being detected in 70.2% of single infections. All the other genotypes were detected at percentages ranging from 0.7 to 2.2%. When considering single and mixed infections, genovar E was detected in 92.6% of the cases followed by genovar F that was detected in 10.3% of the cases. The detection of genovar E was remarkably high in our study but was in line with that of Gita et al., performed in New Delhi (India) reporting the detection of only genovar E in 22 specimens positive for C. trachomatis by PCR-RFLP. Their high percentage was explained by the small number of specimen processed which was not so in our case. According to the literature, genovars D, E, F and G were reported to be the most prevalent genovars worldwide [8, 9, 16–21]. Differences in the relative proportions of these genovars in different geographical areas were reported. For example, in Cambodia, genovar D (22.2%) was the most prevalent followed by genovar F (18.5%) . In Australia, the most common genovar was E (41%) followed by genovar F (26%) . In Sweden, the most common genovar was E (47%) followed by genovar F (17%) . Differences in genotype prevalence were found in different study populations. In Thailand, genovar F was the most prevalent genotype (60%) among non sex workers, but decreased to 29% among sex workers . The high detection rate of genovars D, E, F, and G could be related to a longer persistence time of their infections than the other genovars in the genital tract .
The high prevalence of genovar E in our study could be explained by a specific immune status of Tunisian patients and by distinctive characteristics of some genotype E strains in our country. Such strains could not be distinguished by our reverse hybridization method. Thus, the application of new typing schemes based on more discriminative method such as multilocus sequence typing or multiple loci variable number of tandem repeats analysis might provide a better understanding of our high prevalence of C. trachomatis genovar E. An interesting issue would be to further examine whether all or most of the genovar E in our patients were of the same strain or not. Furthermore, an assessment of in vitro studies of infectivity and cytotoxicity of our local strains in urogenital cell lines may also help to explain the high prevalence of genovar E in our population. Jones et al., suggested that genovar E can outcompete other genovars for nutrients and growth factors leading to a rapid expansion of genovar E when compared to the other genovars.
Mixed infections were detected in our study in 21.1% of the cases. This percentage seems to be high as most of mixed infections in the literature occur at a percentage not exceeding 15% [19, 27–30]. Mixed infections may result from two separate episodes of infection and the lack of cross protection between genotypes . In our study, single genovars were rarely detected except for genovar E. Genovars H and K were detected once in single infections but were detected at high frequencies in mixed infections (11 and 7 times, respectively). These genovars are probably relatively low grade pathogens which could only persist in the presence of other genovars .
Trachoma genovars A and B were also detected in 5 patients (3.6%). All of them were observed in mixed infections. This could not be due to cross reactivity since our technique was well studied with a mixture of reference strains. Also, these genovars were involved in urogenital infections [32–37].
The genovar detection in our study was similar among men and women in line with previous studies suggesting that specific serotypes do not preferentially infect one gender [23, 38]. However, the occurrence of mixed infections is significantly higher among men than women. Furthermore, a statistically significant difference was found for the occurrence of genovars F and K which were observed more frequently in men than in women. Men are probably sexually active with more than one partner leading to the occurrence of mixed infections and infection with another genovar than E. No statistically significant difference in distribution of genotypes and age was found such a result is in a total agreement with other studies .
In our study, no association was found between C. trachomatis genotype E and the patient’s infection status (current symptomatic infection and infertility). According to the literature, genovar E was reported at high frequencies in symptomatic adults  as well as in asymptomatic adults [16, 18, 19]. C. trachomatis genovar E was also found in our study to be the most prevalent genovar. In infertile patients from Mexico, C. trachomatis genovar F was the most prevalent genovar (54%) followed by genovars E, G, K and LGV (9%) .
Patients with N. gonorrhoeae infection were significantly more frequently infected by mixed genovars and by genovars other than the single genovar E (Table 2). According to the literature, little is known about N. gonorrhoeae coinfection and its association with C. trachomatis genovars. Papadogeorgakis et al., reported a significantly higher proportion of N. gonorrhoeae coinfections among patients with urethritis with genovariant Ja. No other associations were reported except for the significant association of N. gonorrhoeae coinfection with genovar D .
Yet, we could mention a few limitations of our study. First, the smaller sample size of infected patients could affect the relationship between C. trachomatis genotypes and clinical manifestations. Second, supplementary information about behavioral characteristics would be useful to explain the difference in genotype distribution among men and women. In addition, the reverse hybridization method used in this study could not detect new genovars and recombinant clinical strains. These could lead to cross hybridizations . The presence of trachoma genotypes and the high frequency of mixed infections have to be confirmed by cloning and complete ompA gene sequencing.