Journal of Andrology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH
Published-Ahead-of-Print May 23, 2007, DOI:10.2164/jandrol.107.002741

This Article
Right arrow Author Manuscript (PDF)
Right arrow All Versions of this Article:
28/5/638    most recent
Author Manuscript (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stokes-Riner, A.
Right arrow Articles by Swan, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stokes-Riner, A.
Right arrow Articles by Swan, S.

One Semen Sample or Two? Insights from a Study of Fertile Men

Abbie Stokes-Riner *, Sally W Thurston , Charlene Brazil , David Guzick , Fan Liu , James W Overstreet , Christina Wang , Amy Sparks , J. Bruce Redmon , and Shanna Swan

* To whom correspondence should be addressed. E-mail: abbie_stokesriner{at}urmc.rochester.edu.

Epidemiological studies that estimate semen parameters may collect a variable number of semen samples per man. Although using multiple samples can reduce measurement error, it is rarely feasible to obtain two semen samples from all participants, and this raises two statistical issues. First, do men who give two semen samples differ in any systematic way from those who give a single sample? Second, among men who give two samples, do semen parameters based on the first and second samples differ systematically? We used data from 706 fertile men participating in the Study for Future Families (SFF) who provided one (N = 82) or two (N = 624) semen samples to address these questions. Adjusted logistic regression and t-tests were used to compare characteristics of men who gave one versus two samples. Adjusted linear regression and t-tests were used to compare differences between semen parameters from the first and second visits, among men who gave two samples. No covariates significantly predicted whether a man gave one or two samples except center, which was likely a result of between-center differences in recruitment strategies. Among men who gave two samples, only mean percent motile and morphology differed between the two samples from a single man, and these differences became insignificant in the adjusted models. As long as the model for semen quality adjusts for important covariates, it makes little difference whether the analysis includes men who give one semen sample or two. Likewise, multiple semen samples should be used when available, as long as the covariate-adjusted model accounts for repeated measures.


Key words: Semen • Semen Analysis • Sperm • epidemiology • statistics






HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH
Copyright © 2007 by The American Society of Andrology.