Hum. Reprod. Advance Access originally published online on October 18, 2006
Human Reproduction 2007 22(1):311-312; doi:10.1093/humrep/del366
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letters to the editor |
Reply: WHO grade ‘a’ sperm motility and zona pellucida-binding test predict IVF outcome
1 Service d’Histologie-Embryologie-Cytogénétique, Laboratoire de Biologie de la Reproduction 2 Service de Médecine de la Reproduction, Hôpital Jean Verdier, Assistance Publique—Hôpitaux de Paris, 93140 Bondy and 3 Service de Biostatistique et d’Informatique Médicale, CHU Saint-Louis, Assistance Publique—Hôpitaux de Paris, 75475 Paris Cedex 10, France
4 To whom correspondence should be addressed at: Service d’Histologie-Embryologie-Cytogénétique, Laboratoire de Biologie de la Reproduction, Hôpital Jean Verdier, 93140 Bondy, France. E-mail: christophe.sifer{at}jvr.aphp.fr
Sir,
We thank Professor Evers for his interest in our preliminary work (Sifer et al., 2005
). We agree that a positive likelihood ratio (LR+) of 1.67 will change the likelihood of disease in a clinically not very relevant way. Indeed, we have moderated our purpose saying that an LR+ of 1.67 indicated a small impact on the post-test probability of successful IVF. However, this change was statistically significant as our study showed. Thus, we believe that in the lack of other predictive tests that could be performed routinely, this new combined test is helpful to decrease the risk of fertilization failure during IVF therapy in the case of unexplained infertility.
Concerning male factor, we have found an LR+ of 6.0, which indicated a better, though moderate, post-test impact, as we have said in our study and accordingly to Professor Evers’ letter. However, we effectively did not include the 95% confidence interval (CI) of this LR in our study, and we agree that this could lead to misinterpretation. We thought that this CI is calculated using an approximate formula, which could not be considered as valid on such a small sample. Unfortunately, we are not aware of an exact CI for the LR. Nevertheless, we admit that the result of the usual 95% CI in this case is [0.9–41], which includes 1.0. A Fisher’s exact test performed on these only 20 mild male infertilities and comparing the proportion of successful IVF yields a result near to the threshold of significance (P = 0.057). This is consistent with the comments of Professor Evers, and we cannot rule out the hypothesis that the test may not be predictive at all in mild male-factor infertility. Results were however promising in this indication and need further research. As a result, we performed, since the end of our study, 19 IVF attempts (7 with male factor and 12 with unexplained infertility) with a previous positive sperm–zona pellucida-binding test. We observed 6/7 and 10/12 fertilization rate >20%, respectively, which seems very encouraging.
We also agree that the sperm–zona pellucida-binding test needs further validation, particularly in an independent group of patients. Splitting the original data set would perhaps not have been the best solution, because recruitment of 84 patients (and among them 20 mild male infertilities) required 1 year. It is quite unlikely that we would have achieved recruitment of a sufficient sample size to allow data splitting and validation of the proposed threshold in a reasonable time. Moreover, although data splitting is quite appealing, some authors do not recommended it, because it is not as efficient as other methods such as the bootstrap approach (Harrell et al., 1996). We thus preferred to present the results obtained, waiting for further validation, possibly in multicentre, and hopefully international, data.
References
Harrell FE, Lee KL, Mark DB. (1996) Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 15:361–387.[CrossRef][Web of Science][Medline]
Sifer C, Sasportes T, Barraud V, Poncelet C, Rudant J, Porcher R, Cedrin-Durnerin I, Martin-Pont B, Hugues JN, Wolf JP. (2005) World Health Organization grade ‘a’ motility and zona-binding test accuracy predict IVF outcome for mild male factor and unexplained infertilities. Hum Reprod 20:2769–2775.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||