Re: Intra‐amniotic Doppler measurement of blood flow in placental vascular anastomoses in twin–twin transfusion syndrome

Abstract

Nakata et al. recently reported a novel, invasive intra-amniotic Doppler technique to measure blood flow rates in arteriovenous anastomoses (AVAs) in twin–twin transfusion syndrome (TTTS) of up to 25 mL/min or 36 L/day1. Velocities of this magnitude underpin previous reports that AVAs are detectable non-invasively by transabdominal ultrasound2. However, we are concerned that the biological implausibility of the huge total intertwin net flows they derive appears to have escaped the attention of both the authors and the reviewers. The net transfusional blood flow of up to 11.6 mL/min that they report equates to nearly 17 L per day. No adult would survive this depletion or addition to its circulation for more than a few hours, let alone a mid-trimester fetus. For instance, we estimate that at 20 weeks' gestation, when fetal placental volume is approximately 40 mL, such a net loss of blood volume would take only 3 min to result in fetal exsanguination. In contrast, we have recently derived more biologically plausible net flows (0.004 mL/min at 28 weeks or 60 mL/day) with data obtained non-invasively by color Doppler insonation of superficial anastomoses3. Nakata et al. fail to acknowledge that their assessment of flow is instantaneous and therefore unreliable if flow rates fluctuate, as seems likely. Indeed, support for temporal variation comes from their somewhat counter-intuitive result showing that the gradient of net flow in TTTS is from recipient to donor. There are several possibilities for the authors' seemingly erroneous calculation of net flow. Firstly, they made no attempt to exclude false-negative anastomoses that they may have failed to identify at endoscopy. Da Paepe et al.4 recently showed that on average more than two AVAs are missed per laser procedure. In addition, our group has shown using placental cast studies that in only half the shared cotyledons in monochorionic placentae are AVAs characterized by cotermination of an artery and the vein on the chorionic plate surface5. Instead, large numbers of deep anastomoses lie beneath the chorionic plate and therefore cannot be visualized directly by chorionic plate inspection. Second, although the authors imply that their technique is more accurate than transabdominal Doppler, they made no attempt to validate their new technique or establish its reproducibility, such as with flow phantoms, or in larger in-vivo vessels like the umbilical artery with known flow rates, or by ex-vivo perfusion experiments. Nor do they acknowledge the limitations of vessel diameter measurement using color Doppler, or of the equations used to derive blood flow. Small inaccuracies in diameter will be exaggerated by calculations dependent on the square of the radius to calculate blood flow velocity. Finally, patients with anterior placentae were selected in this study. This may be an additional explanation for missed anastomoses, as the same group have previously described the difficulties with visualization in such cases6. We congratulate the authors on their creativity, but suggest that both the technique and endoscopic visualization will need validation before useful pathophysiological inferences can be drawn from such results. While appreciating that this was only a preliminary report in correspondence, we do not feel this exonerates the Journal from peer review, nor the authors from acknowledging the limitations of their method. M. J. O. Taylor*, L. Y. Wee*, M. L. Denbow*, N. M. Fisk*, * Royal Devon Exeter Hospital (Heavitree), Gladstone Road, Exeter, EX1 2ED, UK