Continued from Part II
In my last post in this series on treatment of Asherman's syndrome I posted the actual live birth rate outcomes according to studies published peer reviewed medical journals to show why it is not the promising solution that it is often promoted as, and why prevention would be a better approach. The gold standard for treatment of intrauterine adhesions (Asherman’s syndrome) is hysteroscopic adhesiolysis (synechiolysis) and hormonal therapy. Most studies so far focus on fertility outcomes following a particular treatment method.
Yet one needs to be cautious when evaluating outcomes because some studies are presented in ways which may mislead the inexperienced or lay reader into believing that the success rates are higher than they actually are. If you have read some of the original abstracts or papers I cited in the references of my last post, you may notice that some of the results I have given differ from those apparently reported by the authors. This is because I gave live birth rates as total number of births per total number of patients treated, which gives a complete and accurate picture of the outcomes. As I explained before, the reason I did this is to include women who may never have conceived as it is very possible that surgery did not restore their fertility. Below are some other examples of how data presentation can be misleading.
'Success rates’ after surgery may not be what you had in mind. I think most patients like myself are interested in achieving a live birth. Sometimes the author is referring to menstrual outcome. Yasmin et al (1), report that 95% of women in their study resumed normal menstruation. Unfortunately, this is not closely correlated to live birth rate. Accordingly, only 1 patient out of 20 treated ie.5% went on to have a live birth (although followup was short). Restoration of menses is known to be an unreliable criteria for future fertility. In my view this is partly because it is self-reported.
Even pregnancy or conception rates per patients treated or live birth rates per total pregnancies do not necessarily reflect live birth outcomes realistically because not only do 1 in 4-5 pregnancies end in miscarriage under the best of circumstances, women with a history of Asherman’s syndrome are also prone to second trimester pregnancy loss and preterm delivery (2). Other women with residual scarring may conceive but repeatedly miscarry with no live births.
In one study (3) the abstract says that live birth rate was 86.1%. Not only was this an overall rate in a study group where mild and moderate cases greatly outnumbered severe cases (71 vs 18), it is also calculated per pregnancy (instead of per patients treated). This can be gleaned from Table 3 where a 66.6% live birth rate was given among women with severe condition at presentation. In other words, two thirds of the women with severe Asherman’s who were able to conceive after treatment went on to deliver a live baby. This is good but not that inspiring when you consider that most women in this group were not able to conceive. In fact, only 4 out of 18 (22%) women with severe AS in this study had a live birth. Another example of this occurs in Table 6 of Yu et al, 2008 (4) review where outcomes are given in terms of live births/pregnancies.
Perhaps reporting data in this way is seen as acceptable by some because doctors are all too eager to put down the inability to conceive after treatment to ‘other fertility issues’, particularly in women who have had the cruel misfortune of being inflicted with Asherman’s syndrome before having a child. The ridiculous premise is that unless women have a live birth prior to developing this iatrogenic condition, they cannot 'prove' they were ever fertile (infertile unless proven otherwise). Unless these ‘other causes of infertility’ are clearly described and (depending on the cause) pre-existing, there is no evidence to suggest that these women were infertile to begin with and should be discarded from outcomes. If the studies were conducted according to the highest standards, women with ‘other fertility issues’ would be excluded from studies to begin with and not after the fertility results are known. This brings me to…
Besides data presentation, it is also important to take into account how the study was conducted because this too has important implications on outcomes. Unfortunately, studies on Asherman’s syndrome have not been conducted to meet most rigorous scientific methodology-RCTs, making it difficult to assess the exact outcomes. I have mentioned the lack of RCTs in an earlier post. It follows that there are currently no systematic reviews or meta-analyses of the studies either since they are not conducted to stringent standards.
RCTs are clinical trials which have to meet certain criteria in order for the results to be considered unbiased, accurate and of statistical significance. In summary, studies must be done prospectively, they must be randomized with respect to treatment strategies depending on the outcome measure(s), there must be blinding of the doctors, investigators with regard to diagnoses/assessment of outcome with regards to treatment, a clear definition of exclusion and inclusion criteria for patients must be set in order to avoid outcome bias, and there must be enough participants in the/each study group(s) such that statistical analyses can be deemed to have significant value.
RCTs can be carried out to test the efficacy of a particular treatment, or to compare the efficacies of two or more different treatments. Most Asherman's studies focus on fertility outcomes following a given treatment protocol. The risk for potential bias comes mainly from unclearly defined patient inclusion/exclusion criteria, retrospective analyses and sample size and composition.
With regards to study design, the most blatant weakness is when studies are done retrospectively. This means that the doctor/investigator chooses the data to present after they already know the outcomes. This approach introduces the possibility of bias in outcomes because they may select, for example, all of the patients who went on to have live births after treatment, but leave out a portion of those who did not regain their fertility from the study without ever mentioning it. This would obviously make the success rate of treatment appear to be higher than in actuality. Most studies are now carried out prospectively.
Even if a study is done prospectively there are ways in which the outcome can be premeditatingly skewed to enhance outcomes. This is why in proper RCTs everything about the study (except for the results and conclusions) is decided in detail in advance and recorded in a repository. This leaves little space for modifications which can incorporate bias.
If exclusion/inclusion criteria are not stringent from the outset, some of the women who did not achieve a pregnancy could still be discounted on account that they had ‘other' fertility problems. Some authors write their study before allowing a reasonable followup time (1,5) and suggest that the success rate is potentially higher because at the time of writing it was too soon for X number of patients to try to conceive. Others may include patients with ongoing pregnancies (4,6,7,8) which may later not eventuate in live birth on complete followup. In one study, (9) live birth rates were reported as 83% a figure that sticks out like a sore thumb compared to other studies. It turns out that the study, which was done on 365 women, only included 186 patients in the live birth rate outcome. The justification for this was the 179 patients who were left out did not desire a pregnancy. Call me skeptical but I find it hard to believe that 179 women would undergo surgery just to get their periods back. If you include all 365 women in the study the live birth rates don’t look quite as remarkable -42.7%- blending in with the rates recorded by other studies.
Another thing to keep in mind is patient composition: how many have mild, moderate and severe disease? Since an increase in disease severity is associated with a poorer prognosis in terms of live births, a study group consisting of proportionally fewer severe cases could artificially bump up the live birth rate. Therefore it is important to either indicate this by giving a breakdown of outcomes according to disease severity, or to carry out studies where there are equal numbers of patients of the different disease classifications. Many studies will report overall birth rates among patients without giving a breakdown according to disease severity at presentation. It is up to the reader to do their own calculations.
I think the above examples show why proper RCTs are needed to not just to compare different methods of treatment, but to give accurate data regarding live birth outcomes. It would also be of help if there was a more standard way of reporting outcomes such as live births per total patients treated.
1. Yasmin H, Nasir A, Noorani KJ. Hystroscopic management of Asherman’s syndromeJ Pak Med Assoc. 2007 Nov;57(11):553-5.
2. Capella-Allouc, S, Morsad, F, Rongieres-Bertrand, C, Taylor, S, and Fernandez, H. Hysteroscopic treatment of severe Asherman's syndrome and subsequent fertility. Hum Reprod 1999;14(5):1230-3.
3. Roy KK, Baruah J, Sharma JB, Kumar S, Kachawa G, Singh N.Reproductive outcome following hysteroscopic adhesiolysis in patients with infertility due to Asherman's syndrome. Arch Gynecol Obstet. 2009
4. Yu, D, Li, TC, Xia, E, Huang, X, Liu, Y, and Peng, X. Factors affecting reproductive outcome of hysteroscopic adhesiolysis for Asherman's syndrome. Fertil Steril 2008;89(3):715-22.
5. A. Thomson, J. Abbott, A. Kingston, M. Lenart, T. Vancaillie. Fluoroscopically guided synechiolysis for patients with Asherman’s syndrome: menstrual and fertility outcomes. Fertil Steril, 2007; Volume 87(2):405-410
6. Goldenberg, M, Sivan, E, Sharabi, Z, Mashiach, S, Lipitz, S, and Seidman, DS. Reproductive outcome following hysteroscopic management of intrauterine septum and adhesions. Hum Reprod 1995;10(10):2663-5.
7. Robinson, JK, Colimon, LM, and Isaacson, KB. Postoperative adhesiolysis therapy for intrauterine adhesions (Asherman's syndrome). Fertil Steril 2008;90(2):409-14.
8. Protopapas, A, Shushan, A, and Magos, A. Myometrial scoring: a new technique for the management of severe Asherman's syndrome. Fertil Steril 1998;69(5):860-4.
9. Feng, Z, Yang, B, Shao, J, and Liu, S. Diagnostic and therapeutic hysteroscopy for traumatic intrauterine adhesions after induced abortions: clinical analysis of 365 cases. Gynaecol Endosc 1999;8(2):95-98.
- FAQs on Asherman’s syndrome
- How Asherman's syndrome causes infertility or misc...
- Frequency of intrauterine adhesions after D+C
- Management of Intrauterine Adhesions.
- Publications: Etiology, Incidence, Prevention
- Publications: Diagnosis, Classification and Treatm...
- Publications: Reproductive Outcomes, Obstetric Com...
- Introduction: why blog about Ashermans syndrome?
- About me