Misreporting published data is not the way forward for a constructive scientific debate

Dear Editor,

We have carefully read the comment by Dr. Orvieto to our recent publication [1] and we believe the JARG readership would benefit from a clarification about the Dr. Orvieto’s misinterpretation of our clinical research study [2]. Moreover, we are seriously concerned that the publication of commentaries inconsistent from the data originally published is becoming a common speculative trend in our field, aimed at supporting personal assumptions and positions.

First, Orvieto states that a closer analysis was necessary to reveal that only patients with < 12 mIU/mL of FSH and > 0.5 ng/mL of AMH were enrolled in the study. This information was not hard to find as clearly stated in the material and method, results, and Fig. 1, where we present the study design.

Second, his assumption that only patients with > 10 oocytes retrieved were included is false. In fact, average number of oocytes retrieved was 10.9 ± 5.8 for the PGT-A group and 10.2 ± 5.0 for the control group (P = 0.021), and the average number of mature oocytes was 8.3 ± 4.1 and 7.9 ± 3.7, respectively (P = NS). We are convinced that Dr. Orvieto knows how to interpret means and standard deviations as highly experienced reproductive specialists with significant publication records. In addition, as clearly stated on Table 1, the mean female age was significantly lower and AMH was higher for the control group, potentially favoring the control group rather than the PGT-A group. We believe that these two populations have been well matched as the number of usable oocytes is similar. On the contrary, there are significant differences between the dropout group and the PGT-A and control groups. In fact, as we clearly report in our article, the dropout group is composed of patients that had signed the consent for PGT-A but achieved less than 5 oocytes fertilized and were counseled regarding the low chances of having a euploid embryo to transfer. The decision on whether to continue with PGT-A analysis or switch to a normal cycle was theirs.

This decision point resulted in the formation of two branches of study groups, i.e., one that continued its IVF/PGT-A treatment (n cases = 308) and one that reverted to conventional IVF (dropout, n cases = 106). This second group did not have any PGT-A-related intervention, and therefore cannot be accounted as part of the PGT-A group, as Orvieto suggests. Accordingly, the PGT-A and control groups were well matched and could be reliably compared, while the dropout group was a clearly distinct category of patients which was appropriately evaluated separately. Therefore, our colleague’s additional calculations are not pertinent.

Third and most concerning, Orvieto is suggesting higher ongoing pregnancy rate (OPR) per patient in the control group, in a setting where patients from the control group have performed significantly higher number of IVF cycles compared with the PGT-A group (1.3 vs 1.1, P < 0.001). It is not surprising that OPR per patient is higher if more IVF cycles are performed, and the dropout group is also included in the PGT-A arm. This seems a truly misleading way to support the author’s own perspective. Indeed, as reported, the delivery rate per cycle (non-biased by the number of cycles performed per patient) was similar between PGT-A and control group. In this more reliable scenario, even improperly including the dropout patients in the PGT-A group, no significant differences are detected for cumulative live birth rate per cycle (90/414 for PGT-A + dropout vs 698/2905; two-tailed Fisher’s exact test, P = 0.3236).

Fourth, the ensuing comments about the deleterious application of PGT-A in patients with poor ovarian response are yet not supported by evidence. As every scientific statement, Orvieto’s professional viewpoint should be backed by properly designed studies that compare patients with low numbers of oocytes available/fertilized undergoing PGT-A vs conventional IVF treatment. This type of study has yet to be performed. Nonetheless, in this study, we show that patients with a low number of oocytes available for treatment (i.e., < 5 fertilized) that refused PGT-A intervention despite matching eligibility criteria suffered a high incidence of miscarriage (> 40%). These findings support one of the PGT-A objectives which is to reduce miscarriage occurrence by identifying embryos with chromosomal abnormalities that, if implanted, would inevitably result in miscarriage or pregnancy loss. Based on these results, poor ovarian responders may benefit from PGT-A not by increasing pregnancy rate per cycle started but by minimizing their risk of miscarriage.

Finally, Orvieto states “Based on Sacchi et al. study, the 370 couples enrolled were probably chosen from a larger cohort of patients, probably 500 couples, if taking 25% as an average prevalence of POR in this age group.” As we stated in our original manuscript and confirmed at the beginning of this letter, all filters applied have been communicated in the article. All consecutive patients with eligibility criteria (age between 38 and 44, < 12 mIU/mL of FSH and > 0.5 ng/mL of AMH) and willing to undergo PGT-A at our center during the period of the study were enrolled, and their data were analyzed and reported in the publication. Furthermore, the rate of POR patients in PGT-A and control group is reported in Table 1 and not significantly different between the two groups.

We would thus recommend to our colleague a more careful reading of the clinical studies and publications with the eyes of an independent thinker rather than misreporting the published data to support predefined assumptions and perspectives.