OBJECTIVES: Recent studies have hypothesized that laboratory contamination may influence interpretation of Hybrid Capture II (HCII) human papillomavirus (HPV) detection assay values. STUDY DESIGN: To test this hypothesis, 572 consecutive HCII samples were statistically evaluated to test the null hypothesis that cross-well contamination was not present. In addition, 874 consecutive paired samples from patients followed by both HCII and polymerase chain reaction (PCR) analysis were compared. RESULTS: A Kendall's tau measure of association among adjacent wells yielded a P value of .016, rejecting the null hypothesis of no contamination. Analysis of relative light unit values between 0.8 and 1.5 rejected the null hypothesis at P=.077. Moreover, PCR positivity was significantly higher for samples with HCII values above 1.5 vs 0.8 to 1.5 (P=.001). CONCLUSION: Cross-well contamination of samples occurs during processing, and may influence interpretation of some borderline positives. The proportion of cases at risk is low (<3%). Nevertheless, this information may be germane to the interpretation and reporting of marginally positive HCII test values.
OBJECTIVES: Recent studies have hypothesized that laboratory contamination may influence interpretation of Hybrid Capture II (HCII) human papillomavirus (HPV) detection assay values. STUDY DESIGN: To test this hypothesis, 572 consecutive HCII samples were statistically evaluated to test the null hypothesis that cross-well contamination was not present. In addition, 874 consecutive paired samples from patients followed by both HCII and polymerase chain reaction (PCR) analysis were compared. RESULTS: A Kendall's tau measure of association among adjacent wells yielded a P value of .016, rejecting the null hypothesis of no contamination. Analysis of relative light unit values between 0.8 and 1.5 rejected the null hypothesis at P=.077. Moreover, PCR positivity was significantly higher for samples with HCII values above 1.5 vs 0.8 to 1.5 (P=.001). CONCLUSION: Cross-well contamination of samples occurs during processing, and may influence interpretation of some borderline positives. The proportion of cases at risk is low (<3%). Nevertheless, this information may be germane to the interpretation and reporting of marginally positive HCII test values.
Authors: Jill Koshiol; Wen-Qiang Wei; Aimee R Kreimer; Wen Chen; Patti Gravitt; Jian-Song Ren; Christian C Abnet; Jian-Bing Wang; Farin Kamangar; Dong-Mei Lin; Magnus von Knebel-Doeberitz; Yu Zhang; Raphael Viscidi; Guo-Qing Wang; Maura L Gillison; Mark J Roth; Zhi-Wei Dong; Esther Kim; Philip R Taylor; You-Lin Qiao; Sanford M Dawsey Journal: Int J Cancer Date: 2010-07-01 Impact factor: 7.396