Bed rest after embryo transfer negatively affects in vitro fertilization a randomized controlled clinical trial

Capsule:
Immediate ambulation after embryo transfer in in vitro fertilization obtained higher live-born infant rates than bed rest, which denotes the negative effect of bed rest after embryo transfer on treatment outcome.

Authors:
Sharayu Gaikwad, M.D., Nicolas Garrido, Ph.D., Ana Cobo, Ph.D., Antonio Pellicer, M.D., José Remohi, M.D.

Volume 100, Issue 3, Pages 729-735.e2, September 2013

Abstract:

Objective:
To evaluate the influence of 10 minutes of bed rest after ET on the achievement of a live-born infant (LBI) in patients undergoing IVF treatment with oocyte donation (OD).

Design:
Prospective, randomized, parallel assignment, controlled trial.

Setting:
Private IVF center.

Patient(s):
A total of 240 patients undergoing a first IVF cycle with OD in our center.

Intervention(s):
Ten minutes of bed rest after ET or no bed rest, that is, allowing patients to ambulate immediately after the ET.

Main Outcome Measure(s):
The primary outcome was LBI rate per randomized patient. Secondary outcomes were implantation rate and biochemical and clinical miscarriage rates.

Result(s):
LBI rates (56.7% vs. 41.6%) were observed to be significantly higher in the no rest (NR) group than in the rest (R) group. And lower miscarriage rates (18.3% vs. 27.5%) were shown in the NR when compared with the R group, but the difference did not reach statistical significance. Neonatal characteristics like height, weight, and Apgar score were similar in both the groups. Comparable implantation rates were obtained with or without BR after ET.

Conclusion(s):
The statistically significant higher LBI rate shown in our NR group confirms that 10 minutes of bed rest immediately after ET has no positive effect and in fact can be negative for the outcome of IVF with OD. The anatomical/physiological or psychological reasons for this should be explored in future research.

Clinical Trial Registration Number:
NCT01343992.

  • The study aims to tackle one of the many “clinical habits” in ART that are in widespread use without significant basic scientific or clinical evidence to support them. I agree with the previous comment made about decreased LBR in both the study and the control groups compared to the preceding 5 years. It would be interesting to read the authors’ comments on this, before accepting that bedrest is actually detrimental to LBR. The authors make some interesting points in their discussion about the potential rationale against BR after ET. It would appear that gravity should not play a major role, and that “stress” levels may be increased through BR. I did not see a discussion on the role of an overdistended bladder resulting from BR, which could be a factor.
    Unfortunately this topic is so multifactorial that conclusive results are unlikely to be obtained in the future, and clinical practice is unlikely to change.
    At least based on this study patients who do not observe BR after ET for whatever reason can be reassured. And maybe patients could be given the choice of whether they want to ambulate or rest after ET. This could be an idea for a follow-up study on what patients would choose to do. My prediction: close to 100 % would choose bedrest- whether the evidence supports it or not.

  • Yalcin Yavas

    This is an interesting topic. However, I have the following comments. The authors concluded that, in their oocyte donation program, no bed rest after ET significantly improved the birth rate to 56.7% from 41.6% in patients who had bed rest after ET. However, the authors state that the birth rate in their same oocyte donation program in the 5 years preceding their study had been 60% (Statistical Analysis section, first
    sentence). It is unclear from their statement whether the patients who had 60% birth rate over the 5 years in their oocyte donation program preceding the study had bed rest or had no bed rest after their ET. Based on the nature of the study’s goal, I would presume that the patients who had 60% birth rate over the 5 years in their oocyte donation program preceding the study did have bed rest after their ET. If that is the case, then one would expect to see a birth rate close to 60% (not 41.6%) in patients who had bed rest after ET in their study. If that is the case, the authors have failed to discuss (explain) the sudden drop, in the birth rate in patients who had bed rest after their ET, from 60% over the 5 years preceding the study to 41.6% during the study. If, on the other hand, that is not the case (i.e., the patients who had 60% birth rate over the 5 years in their oocyte donation program preceding the study had no bed rest after their ET), then this should have been clarified in the article.
    My next comments are about the way the authors analyzed their data that may compromise the conclusion of their article. They simply used t-tests to compare means and chi-square test for proportions to compare some variables, including birth rates, between the 2 treatment groups: “no bed rest” (NR) vs. “best rest” (R) after ET. As we all know, besides “no bed rest” vs. “best rest” after ET, there are many other variables (factors), independent of the treatments (NR vs. R) that may affect the birth rate (the only factors that the authors controlled throughout the study are the physician performing the ETs and the type of the transfer catheter). Being aware of this, the authors simply compared those variables (listed in Tables 1 and 2) between NR and R groups, separately from (and independently of) birth rates, i.e., they compared one variable at a time. Some, if not all, of these variables (factors), although not different between the two treatment groups when compared one at a time, should have been included in a single statistical model/analysis (such as ANOVA) along with the treatment groups (NR vs. R). In other words, the effect of NR and R treatments on the birth rate should have been corrected (adjusted), in a single statistical model/analysis, for the effects of some of those variables (factors) listed in Tables 1 and 2. In other words, the effects of some of those variables (factors) should have been taken into account when comparing the effect of NR and R treatments on the birth rate. A single statistical model/analysis as such would have removed (separated) the effects of those variables (factors) on birth rates from the effects of the treatment groups (NR vs. R) on birth rates, enabling us to see the real effects of the treatment groups (NR vs. R) on birth rates.
    The authors’ rationale for not including any of those variables (factors) in a single statistical model/analysis along with the treatment groups (NR vs. R), and instead for comparing those variables between the treatment groups, separately from (and independently of) birth rates, is the fact that there were no significant differences in those variables, when compared one at a time, between the treatment groups. While this fact may well be true, some of those variables (factors), if included in a single statistical model/analysis along with the treatment groups, may (or may not) have altered the degree of (the p value for) the difference in the birth rate between the
    treatment groups. Of those factors that may affect the birth rate independently of the treatments, some are listed in Tables 1 and 2, and in Supplemental Table 1 (such as the day of ET [day 3 vs. day 5/6], type of fertilization [insemination vs. ICSI], donor’s age, recipient’s age, ease of ET), while some others are not even mentioned in the article (such as embryologist loading the ET catheter, embryologist performing the ICSI, etc.). Particularly, the percentage-point differences, within each treatment group,
    between day 3 ET and day 5/6 ET presented in Table 1 may have altered, when
    included in a single statistical model/analysis along with the treatment groups (NR vs. R), the birth rates presented in Table 3, as explained below.
    In Table 1, the authors state, “No significant differences were identified between” NR and R groups in the percentage of patients who had day 3 ET as well as in the percentage of patients who had day 5/6 ET (i.e., no difference across/between columns). That is, 40% vs. 26.7% for day 3 ET, and 60% vs. 73.3% for day 5/6 ET, i.e., 13.3 percentage-point difference for day 3 ET as well as for day 5/6 ET, which may not be significantly different as stated by the authors.
    As mentioned above, in Table 1, the authors state, “No significant differences were identified between” NR and R groups in the percentage of patients who had day 3 ET as well as in the percentage of patients who had day 5/6 ET (i.e., no difference across/between columns). However, the authors did not analyze their data to see if patients were equally distributed (or even distributed in similar proportions), within
    each treatment group, between day 3 ET and day 5/6 ET. In other words, they do
    not know if there exists any difference within each treatment group (within columns in Table 1) in the percentage of patients who had day 3 ET vs. day 5/6 ET, as day 3 ET and day 5 ET may result in different birth rates.
    Particularly, there seems to be a much greater, and probably significant, difference in the percentage of patients who had day 3 ET vs. day 5/6 ET within the R group than within the NR group. That is, 46.6 percentage-point difference (73.3% – 26.7%) within the R group, compared with only 20 percentage-point difference (60% – 40%) within the NR group. Are these percentage-point differences, particularly the one within the R group, significant? We don’t know but particularly the one within the R group is very likely to be significant. If so, proportions of patients who had day 3 ET and day 5/6 ET were not similar within each treatment group, which would most likely affect the birth rate independently of the treatment groups (R and NR). That is why the day of embryo transfer (day 3 vs. day 5/6) should have been included as a factor in a single statistical model/analysis along with the other factors (variables) and the treatment groups (NR vs. R).
    My last comment is about the mechanism by which bed rest after ET may decrease the birth rate, as suggested by the authors. The authors concluded that, compared to bed rest after ET (R), no bed rest after ET (NR) increased the birth rate from 41.6% to 56.7%, a 15 percentage-point increase, which was significant according to the authors’
    analysis of their data. This adverse effect of R, compared to NR, was mainly due to a higher incidence of clinical miscarriage in the R group than in the NR group (20 patients in the R vs. 10 patients in the NR had clinical miscarriage). If we calculate the clinical pregnancy by adding up the numbers of patients (shown in Table 3) who had clinical miscarriages and deliveries, and by dividing those numbers by 120, we get clinical pregnancy rates of 58.3% and 65% in the R and NR groups, respectively ([20+50]/120 and [10+68]/120). Therefore, compared to R, NR increased the clinical pregnancy from 58.3% to 65%, only a 6.7 percentage-point increase. This difference is not likely to be significant, which is neither analyzed, presented nor discussed by the authors. In other words, while NR (compared to R) significantly increased (assuming the analysis is valid) the birth rate (by 15 percentage points), both NR and R resulted in similar clinical pregnancy rates (an insignificant 6.7 percentage-point difference). In other words, while no bed rest or bed rest after ET does not affect implantation, bed rest after ET manifests its adverse effect on the birth rate not through anatomical / physiological changes at ET time but during later stages of the clinical pregnancy, by inducing clinical miscarriage.
    My question is: How can the only-10-minute bed rest after ET not affect implantation but cause clinical miscarriage much later? The authors believe that “Encouraging patients to follow their daily routine immediately after ET may help them to cope with
    anxiety during treatment and thereafter to increase their skills in maintaining relaxation throughout the treatment.”
    Then, my next question is: Could it be the other way around? Could it be that it is not the fact that the patients are told that they don’t need bed rest after ET and can follow their daily routine immediately after ET that reduces their levels of stress and anxiety during and after treatment, but it is the fact that, by some chance, patients with naturally lower levels of stress and anxiety happened to have been allocated to
    receive the treatment of no bed rest after ET (NR), resulting in lower incidence of clinical miscarriage, whereas patients with naturally higher levels of stress and anxiety happened to have been allocated to receive the treatment of bed rest after ET (R), resulting in higher incidence of clinical miscarriage? Of course, there is no way to tell that is what happened. The only way to know this was to assess the natural level of stress and anxiety of the patients prior to the study, and include this level as a variable (factor) in a single statistical model/analysis along with the other variables and the
    treatment groups (NR vs. R).

  • Lauren Johnson

    I congratulate the authors on an interesting and clinically relevant study. I was wondering if the authors could provide some additional information about the instructions that patient received after ET. Were subjects instructed to resume normal activities after leaving the clinic? Do you have a sense for patients’ activity levels in the 24 hours after the transfer?

  • Amanda N. Kallen

    The authors of this study observed a statistically significant higher LBR in patients who ambulated immediately after embryo transfer (as opposed to a period of rest). A strength of the study is the use of ET after egg donation in all patients, and the use of a single physician for all ETs. Notably, in the text, there is mention of a few difficult transfers or immediate retransfers – was there a difference in ‘difficult transfers” between groups? Additionally, the authors mention in the discussion that “gravity is unlikely to affect the position of the embryo in the uterine cavity”, which I interpret to mean that the difference in LBR is attributable to the stress associated with immediately reclining. Do the authors believe this is the case? Overall – an interesting and thought provoking study especially in light of our current practice patterns!

Translate »