Endofound’s Sixth Annual Medical Conference: Ending Endometriosis Starts at the Beginning
The key window for diagnosing endometriosis to commence early intervention
Stacey Missmer, ScD
We have started with excellent lectures from Dr. As-Sanie focusing on clinical science, Dr. Taylor focusing on bench science and now I am going to come in with this question of key window for diagnosing endometriosis from an epidemiological perspective. This has been repeated across the last three days. We know that this being the largest study to date of 1400 women, the Global Study of Women’s Health looking at delayed diagnosis. We know that on average internationally it is about seven years. We know that the majority of women with endometriosis report that their symptom onset was in adolescence or very early adulthood. Epidemiologically this gets us to thinking about what characteristics may there be in adolescence that are influencing risk? One of my primary sources of data is a very large prospective cohort of adult women. There are almost 120,000 nurses enrolled in the Nurses’ Health Study II. They were enrolled in 1989. When they were enrolled the youngest was aged 25 and we have now been following them for more than 25 years with greater than 90 percent retention.
This part is not new news. One of the things we looked at is body size and anthropometry. To orient you to these data on the Y axis is the relative risk and on the X axis is body size increasing from left to right. This is the association with the average body size at ages five and ten years and we see that girls who are more lean in those age ranges have a higher risk. These data are consistent through to adulthood so this is the association and this was a publication from Divya Shah, at age 18, the leaner at age 18 the higher the risk of endometriosis.
It also got us to thinking about diet during adolescence. We have validated data collection methods of identifying diet during ages 14 to 18 about. For example, we see that those who had more of a dairy intake in that time period appeared to have a lower rate of diagnosis as adults. This then gets you to think about vitamin D perhaps. As Dr. Taylor pointed out on Saturday, however, 80 percent of vitamin D actually comes from sun exposure and UVB exposure. Also one of the great things about a large cohort like this is that we are able to map where they had lived at different points in their life.
When we contrast that with their UVB exposure, those who at birth lived in regions with the highest vitamin D inducing exposure had a lower risk. This is age 15. We also see in this highest group the highest vitamin D invoking exposure that they are at lower risk.
Despite my love for these types of data, despite my confidence in the validity of the questions and our approach to answering them there is a flaw when we are trying to discuss and look at adolescent endometriosis. The flaw of this research and for the majority of research that exists is that these are all studies of adult diagnosed endometriosis. The youngest woman at the start of the study was 25, and it assumes essentially a model of onset of illness in adulthood. We know that not to be true, certainly not true of adolescence presenting with symptoms.
What we really want to be capturing is this window. The illness onset in adolescence and the pertinent temporal time period for biomarker collection and evaluation, risk factor sampling and definitions. And even this model really focuses on the idea of acquired endometriosis and using the vernacular from Dr. Batt from yesterday, this importance of the window of the in utero environment and early childhood. But if we also start to think about the intriguing hypotheses presented around embryonic development then what we also need to be looking at is two generations where our important windows of perhaps markers and certainly etiology may be in this window, the multi-generational and the in utero environment for an adolescent presenting with endometriosis.
What we know for certain is that we do not know much at all for certain about adolescent endometriosis. There have not been case control and cohort studies and that it is a very exciting time period because that is currently changing. We do not know if there is specific treatment responses at Dr. Taylor just described in his answer to Dr. Griffith’s question. We do not know about this life course sampling, whether there is progression of disease or whether there is really something very distinct and unique about those who at menarche or even between thelarche and menarche are presenting with painful symptoms. We do not know what these risk factor differences are and importantly, we do not know what the biomarker differences are.
This may lead to this graphic from Christian Becker’s review paper. His excellent showing of all of the different biochemical elements that have been tested and have been tested significantly associated with endometriosis. It is out of this literal and figurative soup of biochemistry that we have yet to be able to determine a profile that is consistently diagnostic for a non-invasive manner for endometriosis.
This gets me to thinking as an epidemiologist I am going to try a new approach and we will see how it works walking through. But this how I think about some of this complexity and where a sampling in these sub-populations and timing may be coming into play.
We will walk through this together. Thinking from top to bottom and left to right if we start with the premise that we have a population of girls with endometriosis, they all truly have endometriosis. If we think about the vernacular from, well actually back up, if you look at the bottom right here our goal as clinicians for patients and their families and for scientists is to have ultimately a valid, robust, expert driven diagnostic laparoscopy and treatment with a very skilled surgeon. But in reality what we are running into in this population is, using Debra Bush’s vernacular, we first start with a sub-population of girls who have endometriosis but have discomfort and not distress. A portion of those girls will never be diagnosed and if they have no other consequences from their disease and sort of the do no harm model that may be okay. We do know that we have a sub-population of those girls who ultimately present with infertility issues. Women who are presenting otherwise asymptomatically within fertility and we have difficulty capturing them.
We also know that among those with distress those who we certainly are advocating and the training through the ME program and ENPOWR to not normalize their symptoms and seek care we know among these girls that there are many points at which they never receive a diagnosis so we are not picking them up clinically and we are not picking them up scientifically, not for large scale epidemiologic studies but also not for bench science sampling and not for animal models studies that are seeded with human tissue. We know that they can begin with distress and for the first block there never be diagnosed.
I am going to move for a second to the right. Some of them will not be diagnosed from their pain presenting symptoms but will be presenting for infertility care. Those who go to primary care we also know that there are several branches of this tree that affect who we are seeing in our scientific sampling. There are those who are never diagnosed and essentially dismissed in terms of their symptoms from the health care system. There are those who are treated with hormones. We talked about this quite a bit this weekend that some of them will have improved symptoms and may never be diagnosed and perhaps again that is okay for them for their long term health but some, as the example of Lisa at 33, will present with infertility. We know that some of them, again even though they receive hormonal treatment, will not have remediation of their symptoms but will never come to a diagnosis. Some will be sent to a specialist.
Moving over to the specialist branch, we know that some of them, many of them will be sent to a specialist but not ever through a gynecology pathway. So they will be diagnosed with gastroenterologic disease, rheumatologic disease, psychiatric and psychologic disease and never come to an endometriosis diagnosis. Some of them will make it to a gynecologist but again some of them still will not receive the proper diagnosis and again, the gold standard we are trying to reach is this accurate diagnostic laparoscopy. I am approaching this a bit from the scientific questioning of, again, this heterogeneity in populations. And what we really are capturing. Here scientifically highlighted this is really where our scientific samples are coming from those who are diagnosed as infertilities, actually the majority of the evidence in literature that has been changing over about the last decade and those who have a pain presenting diagnostic laparoscopy.
It may be coming back to this soup example. That part of the reason why we are not detecting necessary and sufficient combinations of biochemistry that can lead to markers or lead to stratifying specific disease populations for targeted individualized treatment may be that we are either deluding across pertinent subgroups, we are sampling from the wrong time period and for adolescents we are not collecting samples during adolescence until very recently. It could also be that we are conflating who through that decision tree has made it into our scientific work when actually we may be missing a large proportion of the pertinent disease spectrum.
This brings me briefly back to the WERF EPHect project. I know most of you know about this and I am not going to go into great detail but again arguing that to capture this spectrum of disease both in terms of the heterogeneity of how patients present, their symptoms, the paths that they come to through that tree for clinical care and for scientific contribution. We need large, multicenter, multi-population captured information. We need from the infertility clinics, from the general practitioners, from the laparoscopy based gynecologists.
One thing that I want to focus on while I have your attention in terms of our discovery is specifically the surgical form. A very important thing that we discussed quite a bit is that we know that the phenotypic presentation varies and we do not really know in detail what that means. For some patients it may be the presence of the endometrioma that is pertinent, for some it may be the red or clear lesions, whether it is a progressive disease and we are only capturing certain cross sections across the life course or whether it actually is a distinct disease. So we have developed standardized methods for collecting information agnostically at surgery.
One complexity and great learning experience for me in this process is that we have brought together dozens of experts including Dr. Seckin and Dr. Gregersen from Endofound to help formulate these surgical forms. These are all published in Fertility and Sterility and downloadable from the WERF EPHect website. The surgical form includes information on prior imaging, procedures undertaken during this surgery to try to equate with symptom and treatment response, residual disease at close, intraoperative complications if there are any, the pathology that was observed in great detail, deep endometriosis which currently is not included in the ASRM staging information, endometriomas, endometriotic nodules, the locations and sizes and then a lot of detail about the peritoneal lesions.
What I will say is that there has been a lot of discussion during and since about how much information is being collected and what is necessary. What I will say has been fascinating is that in the many surgeons who contributed to this form every single item on the form was felt to be critically important by a certain handful of the surgeons. On almost every single question there was at least one surgeon who was certain that it was completely irrelevant. So we have taken a very agnostic approach arguing that we do not know what information is important there are not data that definitively establishes what details of this surgical presentation really is helping us to define what the appropriate treatment should be and what the long term prognosis for these patients are in terms of ideological discovery with really an eye ultimately to prevention of disease.
So our hope and goal is that after a couple of years of large scale collection of this surgical information we will have a very large analytic process of correlating very detailed patient data with these surgical presentations and hopefully we will learn from this that there is a small subset of details at the surgical visualization point that actually are pertinent. What we may discover is that there is a small subset but it differs based on age of presentation or symptom presentation. That is critically important because at the moment there is a lot of certainty in individual patients but what this process has really taught us is that there is not a lot of overlap in terms of what absolutely is important. We just do not know.
Again, I encourage as many surgeons as possible to contribute to this and start collecting information from the surgical forms. Also, this is not, in any way, a static process. WERF EPHect is designed so that from feedback, from the clinical community, from the patient community, from the scientific community, we will be revising and updating and reposting information and that contribution is absolutely invaluable.
I will conclude with once again thanking WERF EPHect and highlighting the ROSE study which is already implementing many of these methods and you will hear more about that from Dr. Gregersen. Our colleagues at Oxford Krina Zondervan and Christian Becker and my team at Boston Center for Endometriosis – thanks.
Audience Member: Thank you very much for your presentation. Before I ask my question I would like to give some background to set the framework of why I am asking the question. About ten years ago the American Medical Association established a commission to look at health care disparities. As you are aware, by 2042 it is estimated that the population will be a majority minority – more than 50 percent will be a Black and Hispanic population. What we are trying to do on this National Commission is get sensitized both for medical students as well as practicing clinicians and researchers to look at health care disparities in the work that is being done. My question to you therefore I looked up to your form and I am wondering if race and ethnicity is included in that form and if in the data being collected you are looking at the disparities or the differences that may exist along racial and ethnic lines as it relates to endometriosis?
Stacey Missmer, ScD: It is an excellent question. The data are very unclear about race, ethnicity and risk for endometriosis. Fairly consistently African-American women, Hispanic women have a lower proportion of diagnosis. There appears to be a higher proportion of diagnosis in Japanese women for example. But I think we need to take a very conservative, interpretive approach here that the disparity may not be biologic or environmental exposure basis in terms of race and ethnicity. It may be simply diagnostic bias. So where women fall out, girls and women fall out in that tree of who ultimately makes it to a diagnosis at the point of infertility or at the point of the diagnostic laparoscopy. And so, there absolutely is information collected. I have only reviewed here very briefly the surgical form, but there is a, what we refer to as the clinical form that is a self-administered questionnaire from patients. It takes about 30 to 40 minutes to complete. We, and others, have collected now those data now on more than 600 participants and that does include information about race and ethnicity in many other potentially important demographic and other factors.
- Very nice Stacey. That was really excellent, thanks so much. One of the things that comes to mind and I know how challenging these prospective kinds of studies can be in terms of data collection but focusing on the adolescents really also will give you the opportunity unparalleled to watch the disease progress. My question really then is, how well do you think you are tapping into, and you have mentioned a couple of times and I know that your center is probably one of the leading ones on the planet for identifying these young women, but it seems to me that that is really an important thing to try to capture and I do not know how easy it is going to be but it would be great to get that.
Stacey Missmer, ScD: Thank you. I should have paid you to ask that question! Absolutely, so we talk a lot about disease progression, symptom recurrence, early diagnosis and intervention and preventing sub-fertility and we really cannot get at those questions without longitudinal follow up. As Dr. Taylor described I was lucky enough to begin my training and remain in my scientific discovery in groups that are established and all about long term follow up of patients and participants. The fact of the matter is that this is a critical element but it is unrealistic to expect surgeons, even the most highly clinically investigative surgeons to establish their own robust, long term follow up systems with very little loss to follow up over time. That is truly the critical part. If you are only capturing the patients who come back to you in the future, whether it is because they have not had long term…mediation or for other reasons, if that is all you are capturing you are missing the outcomes of those patients who either had very successful treatment at one, two, three, five years on. You know cancer, our criteria is survival at five years. We have national cancer registries for tracking that information. There is no such thing for endometriosis and really for other diseases as well. I would argue, and I am obviously biased about this, but we have tools and techniques that are very robust and fairly straightforward to implement but not always completely obvious for those of us who do prospective cohort research that is multidisciplinary approach of I can only follow patients when I am in collaboration with those who are seeing and treating and diagnosing patients. What our expertise brings is how to do this long term tracking to ensure that we are capturing everyone.
Audience Member: Thank you very much for this very nice presentation. In your follow up do you include those women with endometriosis who become pregnant and having a look at this pregnancy outcome? I think it will be important in the future because there are many, many women sent to IVF programs with endometriosis who are becoming pregnant and at this very moment there are no data at all in the literature what the outcome of these pregnancies are and what after the pregnancy will be the evolution of these women?
Stacey Missmer, ScD: Another excellent question. Absolutely, it is incorporated into these data collection methods to do exactly that follow up. But that brings the point of the importance of within my own enrollment team at The Boston Center for Endometriosis we have more than 500 participants. That is huge! It is not close to large enough to look at all of these very important long term questions. It is why it is so critical that we have a lot of clinical investigators, a lot of multidisciplinary teams collecting these data. So as we pose these questions can combine them and have the sample size and the power to address these important questions.
The other important thing about particularly IVF treatment and outcomes – you give the example of the specific pregnancy outcome, which is an important and not well addressed question. Another important question is that there have been really excellent randomized controlled trials evaluating various points of treatment in the IVF population whether it is excision of endometriomas, whether it is stimulation protocols and other things comparing and contrasting women with endometriosis versus without. Those are really nice robust trials for the conception and live birth outcome for which they were designed. The issue becomes that some of those data are now being used longitudinally to say among those patients what proportion of patients came back two and three years later for pain treatment. The difficulty in directly applying those data as a trial is that a randomized trial is just a prospective cohort where the clinician investigator chose the exposure. You still have to treat it like a prospective study. For example, the goal of that original IVF randomized control trial was for women to become pregnant. Some will have successfully become pregnant and some will not. We know that pregnancy has at least a short term impact on pain. What happens in the interim between enrolling in that trial focused on IVF very short term outcomes and looking three or four years later at pain outcomes is that everything that happened in between, the adjuvant hormone therapy, the pregnancy, breastfeeding, those type of things, weight change, are not being captured. Again, it is another opportunity to apply these long term collection methods so we can get at that more cleanly.