Ns were used for describing continuous variables; categorical variables were expressed in terms of frequencies and percentages. The frequency of events of interest was reported together with their corresponding 95 confidence intervals that were calculated using the bootstrap method. The association between categorical variables was evaluated with Chisquare (x2) tests, using a significance level of 0.05. Agreement between HPV results for urine and cervical JI 101 chemical information samples was evaluated using the kappa coefficient (k), and its corresponding 95 CI, classified as follows: negative values, as well as values from 0 to 0.2 = poor, from 0.21 to 0.4 = slight, from 0.41 to 0.6 = fair, from 0.61 to 0.8 = moderate, from 0.81 to 0.99 = substantial and 1.0 = perfect agreement [27]. The urine HPV tests’ operative characteristics were evaluated for determining sensitivity, specificity, positive and negative predictive values and receiver operating characteristic (ROC) area (as the average of sensitivity and specificity), taking the HPV cervical sample result as reference. STATA 9 software was used for all statistical analysis.Ethnicity (197)White Indigenous Mestizo BlackMarital status (186)Single MarriedCommon law marriage 60 (32.3) Separated Widowed Age at first intercourse (195) ,18 18 Pregnancies (192) None 1 2 3 4 Life time number of sexual partners (194) 1 2 3 4 Contraceptive method** (165) None Hormonal contraceptives Intrauterine device Surgery Condom Condom + other Smoking status (193) No Yes 21 (11.3) 29 (15.6) 118 (60.5) 77 (39.5) 9 (4.7) 45 (23.4) 51 (26.6) 47 (24.5) 40 (20.8) 30 (15.5) 48 (24.7) 43 (22.2) 73 (37.6) 33 (20.0) 5 (3.0) 11 (6.7) 35 (21.2) 54 (32.7) 27 (16.4) 163 (84.5) 30 (15.5)Results Socio-demographic dataTwo hundred and forty five women between 20 18055761 and 73 years old, were enrolled in the study (mean age: 38.1 years; SD 10.7 years) (Table 1). Two hundred and thirty nine of the 245 cervical samples (97.6 15755315 ) were positive by human b-globin amplification and 208 of the 226 urine samples (92.4 ). Fifty one women were not included in the statistical analysis due to 
their samples’ low DNA quality (negative result for b-globin) or a lack of either of the samples (cervical or urine).Human papillomavirus prevalence and type-specific distributionHPV infection frequency in cervical and urine samples was 70.6 (n = 144; 63.8?3.7 95 CI) and 63.2 (n = 129; 56.2?9.9 95 CI), respectively. Type-specific viral identification revealed that HPV-16 had the greatest prevalence in both samples, whilst HPV-31 had the second greatest prevalence in the cervical samples and HPV-58 in urine samples; the other viral types had a variable distribution in both samples (Figure 1). It was found that 55.4 (n = 113; 95 CI = 48.3?2.3) of the cervical samples had coinfection, compared to 40.2 (n = 82; 33.4?7.3 95 CI) of the urine samples. Regarding a description of the number of types of HPV simultaneously present in each sample analyzed, urine samples revealed more uninfected women or those having just one HPV-type compared to the results obtained for cervical samples where more coinfections were detected (2 to 8 types of HPV). The presence of multiple infection per sample type had a statistically significant relationship (Fisher’s exact test, p = 0.000) (Figure 2).*Categories have a size lower than 204, given that data was missing from the surveys. **Contraceptive method used at the moment of enrollment in this study. doi:10.1371/journal.pone.get BI 78D3 0056509.tCytological.Ns were used for describing continuous variables; categorical variables were expressed in terms of frequencies and percentages. The frequency of events of interest was reported together with their corresponding 95 confidence intervals that were calculated using the bootstrap method. The association between categorical variables was evaluated with Chisquare (x2) tests, using a significance level of 0.05. Agreement between HPV results for urine and cervical samples was evaluated using the kappa coefficient (k), and its corresponding 95 CI, classified as follows: negative values, as well as values from 0 to 0.2 = poor, from 0.21 to 0.4 = slight, from 0.41 to 0.6 = fair, from 0.61 to 0.8 = moderate, from 0.81 to 0.99 = substantial and 1.0 = perfect agreement [27]. The urine HPV tests’ operative characteristics were evaluated for determining sensitivity, specificity, positive and negative predictive values and receiver operating characteristic (ROC) area (as the average of sensitivity and specificity), taking the HPV cervical sample result as reference. STATA 9 software was used for all statistical analysis.Ethnicity (197)White Indigenous Mestizo BlackMarital status (186)Single MarriedCommon law marriage 60 (32.3) Separated Widowed Age at first intercourse (195) ,18 18 Pregnancies (192) None 1 2 3 4 Life time number of sexual partners (194) 1 2 3 4 Contraceptive method** (165) None Hormonal contraceptives Intrauterine device Surgery Condom Condom + other Smoking status (193) No Yes 21 (11.3) 29 (15.6) 118 (60.5) 77 (39.5) 9 (4.7) 45 (23.4) 51 (26.6) 47 (24.5) 40 (20.8) 30 (15.5) 48 (24.7) 43 (22.2) 73 (37.6) 33 (20.0) 5 (3.0) 11 (6.7) 35 (21.2) 54 (32.7) 27 (16.4) 163 (84.5) 30 (15.5)Results Socio-demographic dataTwo hundred and forty five women between 20 18055761 and 73 years old, were enrolled in the study (mean age: 38.1 years; SD 10.7 years) (Table 1). Two hundred and thirty nine of the 245 cervical samples (97.6 15755315 ) were positive by human b-globin amplification and 208 of the 226 urine samples (92.4 ). Fifty one women were not included in the statistical analysis due to their samples’ low DNA quality (negative result for b-globin) or a lack of either of the samples (cervical or urine).Human papillomavirus prevalence and type-specific distributionHPV infection frequency in cervical and 
urine samples was 70.6 (n = 144; 63.8?3.7 95 CI) and 63.2 (n = 129; 56.2?9.9 95 CI), respectively. Type-specific viral identification revealed that HPV-16 had the greatest prevalence in both samples, whilst HPV-31 had the second greatest prevalence in the cervical samples and HPV-58 in urine samples; the other viral types had a variable distribution in both samples (Figure 1). It was found that 55.4 (n = 113; 95 CI = 48.3?2.3) of the cervical samples had coinfection, compared to 40.2 (n = 82; 33.4?7.3 95 CI) of the urine samples. Regarding a description of the number of types of HPV simultaneously present in each sample analyzed, urine samples revealed more uninfected women or those having just one HPV-type compared to the results obtained for cervical samples where more coinfections were detected (2 to 8 types of HPV). The presence of multiple infection per sample type had a statistically significant relationship (Fisher’s exact test, p = 0.000) (Figure 2).*Categories have a size lower than 204, given that data was missing from the surveys. **Contraceptive method used at the moment of enrollment in this study. doi:10.1371/journal.pone.0056509.tCytological.
