Saturday, June 20, 2009 S17 Method: 110 infertile women, after exclusion of male factor infertility, were evaluated for clinical and laboratory evidence of tuberculosis. Premenstrual endometrial biopsies from these women were subjected to examination of smears for presence of acid fast bacilli (AFB) and to culture for mycobacterium tuberculosis. Polymerase chain reaction and histopathology of endometrium was done in all cases. Women with tubercular endometritis were given anti-tubercular treatment. AFB staining, culture and histopathology was repeated on endometrial samples following completion of treatment. Laparoscopy was performed if spontaneous conception did not occur. Results: 40 women had raised IgG against 38 kDa M. tuberculosis antigen on ELISA. One endometrial sample was positive for AFB, four samples showed tubercular granulomas on histopathology. The PCR assay was positive for mycobacteria in 13 (10.18%) cases. Repeat endometrial samples in 9 PCR positive cases with all other tests negative for endometrial tuberculosis were again negative for tuberculosis on smear examination, culture and histopathology. Since PCR positivity may denote dead bacilli, only 4 women with tubercular endometritis on histopathology were given anti-tubercular treatment. Repeat endometrial biopsy after completion of treatment in these 4 cases showed no evidence of tuberculosis. One woman conceived but had ectopic pregnancy. Three women did not conceive during a further six month follow up period and had adhesions and blocked tubes on laparoscopy. Conclusion: Genital tuberculosis is common in infertile women in India. Investigations to exclude tuberculosis must be included in the workup of infertility. Anti-tubercular treatment is effective but fertility prognosis remains poor because of early tubal damage. O47 Risk factors associated with poor quality sputum submission in India P. Daley 1 *, A. Latha 2 , S. Suzana 2 , A. Dev 2 , W. Grandin 3 , B. Shalini 2 , L. Armstrong 2 , K. John 2 , D. Mathai 2 . 1 Dalhousie University, Halifax, Canada, 2 Christian Medical College Vellore, Vellore, India, 3 Tufts University, Boston, USA Introduction: Sputum smear remains the primary diagnostic modality in most TB endemic countries. Smear sensitivity depends on quality of sputum collection. The Q score compares sputum neutrophil and epithelial cell counts and ranks specimen quality more accurately than macroscopic description. Women have a lower smear positivity rate which is related to sputum collection, and it can be improved through instruction. Objectives: To describe factors associated with poor quality sputum specimen submission at a tertiary referral center in India. Methods: Consecutive adults with confirmed active TB were consented at a tertiary hospital DOTS clinic to submit one spot sputum. Gram stain and Q score was performed and risk factors for poor quality sputum were analyzed using logistic regression. Results: 542 patients were recruited and submitted 536 specimens, of which 181 (33.76%) were good quality (Q score 1). Mean age was 40.34 (SD 13.37), 73.8% were males, and 15.7% had HIV. 68% of good quality specimens and 40.7% of poor quality specimens came from patients who were smear positive (p < 0.001). Females were more likely to submit poor sputum (OR 1.40, 95% C.I. 0.916 2.15), as were patients from outside the local state (OR 1.60, 1.08 2.35), patients with extrapulmonary TB (OR 5.82, 2.84 11.96), patients who did not complain of cough (OR 2.29, 1.47 3.58) and patients with normal CXR (OR 2.94, 1.51 5.72). HIV, Socioeconomic category, occupation and education did not influence sputum quality. Mucoid (OR 0.36, 0.25 0.52) and bloody (OR 0.19, 0.059 0.61) specimens were less likely to be poor quality, but specimen volume did not influence quality. Conclusions: One third of specimens submitted were of good quality. Good quality was associated with smear positivity. Females, referral patients and CXR normal patients were less likely to submit good quality specimens and should be targeted for instruction before collection. O48 Evaluation of a real-time in-house methodology for Mycobacterium tuberculosis detection from patient sample-concentrated sediments M. Sharma 1 *, D. Kein 1 , A. Rendina 2 , J. Karlowsky 2 , S. Christianson 1 , J. Wolfe 1 . 1 National Microbiology Laboratory, Public Health Agency of Canada, Winnipeg, Canada, 2 Health Sciences Centre, Winnipeg, Canada PCR based methodology for direct detection of Mycobacterium tuberculosis complex (MTBC) from patient samples is utilized in mycobacteriology laboratories for rapid diagnosis of tuberculosis infection in untreated patients. A variety of commercial kits are available for MTBC detection but the majority of them are limited by strict sample acceptance criteria for processing, sensitivity and specificity issues, along with high costs. Despite being less sensitive than MTBC culture, these commercial kits have a higher sensitivity than acid fast bacilli smears alone and a rapid turn around time when compared to culture. In this study, we evaluated the sensitivity and specificity of an in-house real-time based PCR method for detection of MTBC organisms from a large variety of clinical sample types processed in a clinical laboratory for mycobacterial culture. The study tested a total of 139 samples, of which 99 were sputa and 40 were non- sputa. These samples comprised of 111 MTBCs, 22 non tuberculous mycobacteria and 6 samples negative for both mycobacterial smear and culture. Two genetic targets were utilized for detection of MTBC: the insertion sequence IS6110 and the region of deletion RD9. Detection of MTBC from culture positive samples was as follows: IS6110 detection showed 91.9% sensitivity and 100% specificity for smear positive samples and 35.1% sensitivity and 95.2% specificity for smear negative samples. RD9 detection showed 100% sensitivity and 100% specificity for smear positive samples and 50% sensitivity and 90.5% specificity for smear negative samples. The sensitivity and specificity of this test was further enhanced with the use of the combination of both targets. The benefit of this study is the development of a two-target technology that can deliver rapid, specific and accurate identification of MTBC directly from a diverse group of patient clinical samples. This can eventually result in enhanced patient care and facilitate rapid epidemiological contact tracing. Additionally, it can provide substantive cost benefits due to the reduced hospital stay and management of infection control and spread. H Virology II O49 Luminex based assay for multiplexed genotyping of 45 mucosal human papillomavirus types V. Goleski 1 *, A. Severini 2 , M. Dawood 3 , S. Ratnam 4 . 1 Public Health Agency of Canada, Winnipeg, Canada, 2 National Microbiology, Public Health Agency of Canada, Winnipeg, Canada, 3 Cadham Provincial Laboratory, Winnipeg, Canada, 4 Public Health Laboratory, St. John’s, Canada Objective: The recent introduction of effective vaccines against HPV types 6, 11, 16 and 18 is increasing the demand of type specific epidemiological studies. We have developed a Luminex-based genotyping assay that can type simultaneously 45 mucosal HPVs. This assay was evaluated in comparison with the Roche Linear Array (LA) test. Methods: Amplified single stranded HPV DNA carrying a biotin tag was generated using primers PGMY 1 and GP5+/GP6+ 2 in a nested PCR reaction. A set of 45 Luminex microspheres coupled with 45 unique HPV probes was used for detection and typing. A total of 149 cervical specimens collected in PreservCyt were utilized in the study. Results: The Luminex method identified 45 mucosal HPV types without cross hybridization. It showed a higher sensitivity than the LA test, 85 vs 73 positive samples, and 171 vs 164 total HPV types detected, respectively, with 47 multiple infections detected with both methods. On the other hand, the LA test showed slightly better sensitivity for