Pandora's box of endometriosis therapy
Pandora’s box of endometriosis therapy
Evidence based medicine
Therapy in medicine should be based on evidence. This is important to ascertain that a therapy is useful without side effects. It protects the patient from practices without a proven benefit and reduces the cost of medicine. Evidence based medicine has developed a ranking of evidence known as the pyramid of evidence. This ranking is mathematically correct with the Randomised Controlled Trial on top. Randomisation avoid an allocation bias and ascertains that the 2 groups are identical.
The limitations of a RCT
The effect of a treatment should be evaluated without bias. This is obvious If the effect can be objectively measured as height or weight. For endpoints as pain or well being, there is the well know placebo effect and observer bias. RCT on pain and well being thus have to be double blinded to be valid.
The results of a RCT are rue only for the group of women investigated and cannot be extrapolated. What can be a valid conclusion for 20 year old is not necessarily valid for a 60 year old. For this reason RCT are not suited for multi-morbidities.
A RCT evaluates a group as a whole and cannot detect or exclude a (hidden) small subgroup with an opposite effect. Only after detection by clinical observation this can be evaluated by another RCT.
The player Bias
Endometriosis is treated by specialists either in infertility, or in medical therapy or in surgery. The all round gynecologic specialist in endometriosis no longer exists. Each of them have their bias This is also caused by the absence of adequate non-invasive diagnosis. Only a laparoscopy by an experienced surgeon gives a 100% diagnosis. A diagnosis which is correct in 90% and wrong in 10% indeed can be considered as ‘sufficient to avoid surgery’ or ‘insufficient to start medical therapy’.
Industry also is an important player. Industry assist individuals and societies with research and travel grants, and provide support to congresses. The budgets of surgical companies are relatively small and they are rarely directly involved in trials. On the contrary all major trials of medical treatment of endometriosis were organized by the pharmaceutical industry.
For infertility medical therapy of endometriosis is not useful
The TNFa trial was the only double blind RCT that evaluated pain associated with endometriosis. Remarkably, in this trial there was a very strong placebo effect. Indeed with placebo infusions some women who previously needed monthly morphine injections became almost pain free. All trials affecting menstruation unfortunately were not blinded since the women were aware of their menstruation.
Another specific bias of all major endometriosis trials is the inclusion criteria, which used to be pain and ‘laparoscopic and/or histologically proven endometriosis in the last 1, 2 or 3 years’. However, during this diagnostic laparoscopy necessary to confirm endometriosis all visible superficial and cystic lesions of endometriosis are generally excised or coagulated. In such cases it would be questionable whether the remaining pain is still caused by endometriosis.
A laparoscopy is needed to diagnose superficial endometriosis, during which typical lesions can be excised, vaporized or coagulated. The proof of efficacy of this treatment is limited to one double blind randomized trial for pain. This trial moreover demonstrates a huge placebo effect for several months whereas the magnitude of effect is highly variable. Cystic ovarian endometriosis
Cystic ovarian endometriosis is diagnosed by ultrasound. After surgery, - excision or vaporization- spontaneous cumulative pregnancy rates are 50% to 60% and recurrence rates vary from 5% to 20%. To undertake a trial designed to evaluate whether surgical treatment and adhesiolysis affects progression, would ethically be questionable. The outcome of a surgical intervention seems to vary with the surgeon. A special problem is the small cystic ovarian endometrioma especially in young girls. Indeed, since both an endometrioma and surgery can damage oocyte reserve, and considering a recurrence rates between 5% and 20% the decision to perform surgery balances between the risk that the cyst become bigger and the risk of repetitive surgery. Deep endometriosis
Deep endometriosis is associated with severe pain in most women and severe bowel and ureter problems in some. The natural history is not known but clinical observation suggests that most lesions are no longer progressive, when the diagnosis is made. However, some such lesions may progress rapidly. The added value of ultrasound and MRI for the diagnosis and the radicality of surgical excision or the need of bowel resection remains debated. Notwithstanding the reported postoperative 20% to 50% spontaneous cumulative pregnancy rates it is unclear whether surgery improves fertility.
Pain relief is well documented albeit not in randomized controlled trials the feasibility and ethical aspects of which would obviously be questionable. The recurrence rate of deep endometriosis nodules is rare (less than 1%). Deep endometriosis surgery is difficult and complication prone. Besides the technical aspects the ‘experience’ of the surgeon and his knowledge of the disease endometriosis are important.
It would be useful if we could focus on what we agree upon. This could become the basis of the information given to women with pelvic pain, infertility or endometriosis. Available evidence permits us to make following statements. 1. A woman with pain and/or infertility has a 50% probability of having typical endometriosis or worse. 2. Medical treatment of endometriosis over long periods without a diagnosis is not recommended. 3. Superficial endometriosis can only be diagnosed by laparoscopy. 4. Medical therapy of endometriosis can reduce pain but is ineffective for infertility and for cystic ovarian endometriosis. 5. It is unknown whether medical therapy prevents progression of deep endometriosis in all women 6. Diagnostic laparoscopy should be recorded to permit subsequent confirmation of diagnosis and of completeness of diagnosis. 7. It is preferable to have the possibility to treat the disease as part of the diagnostic laparoscopy. 8. Quality control of surgery is only possible with video-registration of the entire intervention. 9. Informed consent requires the patient to be given correct information on the indication, planned intervention and level of experience of the surgeon. 10. EBM should be based upon the best evidence available. This includes rare events and the limitation of RCT’s. 11. Our actual clinical management based on experience should be kept unless proven otherwise.
What is needed
- Adequate referral
- no medical therapy without a diagnosis - risk of making endometriosis worse in some
- no IVF without a laparoscopy – certainly not in deep endometriosis
- quality control of surgery by video-registration
- more stringent control that correct information was given
For complete information and references : see recent articles in JMIG