Power enough? Confidence intervals for uncertainty

Power enough? Confidence intervals for uncertainty

FERTILITY AND STERILITY威 VOL. 79, NO. 4, APRIL 2003 LETTERS TO THE EDITOR Copyright ©2003 American Society for Reproductive Medicine Published by El...

28KB Sizes 2 Downloads 120 Views

FERTILITY AND STERILITY威 VOL. 79, NO. 4, APRIL 2003

LETTERS TO THE EDITOR

Copyright ©2003 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.

Paul G. McDonough, M.D. Associate Editor

Power enough? Confidence intervals for uncertainty To the Editor: Garcia-Velasco et al. should be congratulated on the best attempt to date at comparing ultrasonography-guided ET and blind ET (1). The study design, statistical analysis, and discussion of power calculations set an example for studies of this type. However, the title of the article may better convey the findings if it said that “there was not more than a 20% chance that ultrasonography increased the pregnancy rate by 15%,” to convey the limited importance of a single study of moderate size. Comparison of these two ET techniques is difficult. As the authors pointed out, their physicians had extensive experience with the standard procedure, and ultrasonographyguided seems to improve with experience in regard to the optimal bladder fullness for each patient, experience of the person holding the ultrasonographic probe, choice of speculum, and experience with positioning of the speculum to align the uterine and vaginal canals. Ideally, the two ET would be compared at a point when the physicians are highly experienced with both; this is difficult, however, since both patients and providers tend to become convinced with time that ultrasonography-guided ET is superior. The authors did not indicate how the depth of ET was determined for the controls. The location of ET may be important to outcome, and the groups appear to not be the same in this regard. A recent report indicated that a depth of 1.5 to 2.0 cm from the top of the uterine cavity may be optimal for ultrasonography-guided ET (2). It is also difficult to relate these results to the findings results of others who use half as much ET volume and little or no air. Nevertheless, pregnancy rates were excellent in both groups. As the authors point out, meta-analysis will improve statistical power, but in the other prospective randomized trials, the control groups did not consistently benefit from mock ET or assessment of bladder fullness, and in some studies the technique used in the controls was clearly suboptimal. Hence, meta-analysis may be done, but as usual by lumping together apples and oranges. This, together with clinical experience, will promote use of ultrasonographyguided ET by most groups, and we will advance yet further in the success of IVF-ET because suboptimal technique will become less frequent and, most likely, ultrasonographyguided ET really does help.

David R. Meldrum, M.D. Reproductive Partners Medical Group Redondo Beach, California October 27, 2002

References 1. Garcia-Velasco JA, Isaza V, Martinez-Salazar J, Landazabel A, Requena A, Remohi J, et al. Transabdominal ultrasound-guided embryo transfer does not increase pregnancy rates in oocyte recipients. Fertil Steril 2002;78:534 –9. 2. Coroleu B, Barri PN, Carreras O, Martinez F, Parriego M, Hereter L, et al. The influence of the depth of embryo replacement into the uterine cavity on implantation rates after IVF: a controlled, ultrasound-guided study. Fertil Steril 2002;17:341–6.

doi:10.1016/S0015-0282(02)04932-4

To the Editor: We read with interest the article by Garcia-Velasco et al. (1). In their prospective randomized study, the authors conclude that transabdominal ultrasonography– guided ET does not increase pregnancy rates compared to a blind method. These results contradict those of many published studies, including ours (2) and a meta-analysis we recently conducted (3). For the meta-analysis, we identified 13 studies, but included only 3 prospective properly randomized (i.e., using computer-generated tables) studies. The meta-analysis showed that transabdominal ultrasonography– guided embryo transfer resulted in higher clinical pregnancy rates (odds ratio, 1.46 [95% CI, 1.23–1.74]), higher implantation rates (odds ratio, 1.42 [95% CI, 1.17–1.74]), and higher ongoing pregnancy rates (odds ratio, 1.49 [95% CI, 1.13– 1.96]) compared to the “clinical touch method.” In our meta-analysis, we calculated that to improve the clinical pregnancy rate from 23% to 30%, taking 5% as the significance level and accepting a 90% probability of finding a true difference, the least number needed to study was 786 participants in each arm. In the meta-analysis, 1005 patients undergoing ET under ultrasonographic guidance were compared to 1216 patients undergoing ET with the clinical touch method. We feel that the new study may have lacked sufficient power to elicit a statistically significant difference, as only 187 patients were included in each arm. In their effort to calculate the least number needed to study, the authors accept an 80% probability of finding a true difference and set the aim of achieving a 65% pregnancy rate (15% higher than their current pregnancy rate). Despite this, the authors reported that the pregnancy and implantation rates were higher 1039