Management of abnormal uterine bleeding on anticoagulation: the patient-clinician perspective.

Our patient’s HMB is initially treated with oral progesterone, until she is able to establish with a gynecologist who evaluates her for potential structural causes of HMB and places a levonorgestrel IUD for long-term management. She also receives iron supplementation in the form of intravenous iron dextran. Three months later she is evaluated by her primary care physician, as she is no longer on anticoagulation, and reports much decreased menstrual bleeding as well as decreased fatigue and increased exercise tolerance.

As our case demonstrates, due to the fact that estrogen-containing therapies such as CHCs are a common provoking factor for VTE in reproductive age females, such therapies are often discontinued upon diagnosis. The impact of such a decision, while rarely considered by providers, can be enormous for the patient. Discontinuation of CHCs almost universally results in withdrawal bleeding, something which individuals who use such therapies continuously or regularly and are subsequently accustomed to regular or no menstrual bleeding, are rarely counseled to expect. As discussed above, this can have a dramatic, negative impact on quality of life. As CHCs are a highly effective treatment for HMB, discontinuation can be expected to result in increased menstrual blood loss in many, something which may be compounded by initiation of anticoagulation therapy. Additionally, CHCs, while associated primarily with pregnancy prevention in most providers’ minds, are effective therapies for a multitude of symptoms and conditions, including but not limited to dysmenorrhea, endometriosis, premenstrual dysphoric disorder, acne and hirsutism ( Figure 2 ). Use of CHCs can also reduce future incidence of ectopic pregnancy, benign breast disease, ovarian cysts, ovarian cancer, endometrial cancer and colorectal cancer. The importance of effective contraception is paramount, particularly in patients experiencing or at risk for VTE. Experts commonly recommend that individuals wait until at least 3 months after a VTE event to become pregnant, similar to recommendations around delaying elective surgeries. Furthermore, many anticoagulants can lead to adverse pregnancy outcomes. Warfarin has known teratogenic effects and while data on risk of direct oral anticoagulants in pregnancy are not available, use is not recommended in pregnancy.[ 26 ] Patients who subsequently suffer from pregnancy complications such as spontaneous abortion and/or ectopic pregnancy may suffer from increased risk of bleeding complications due to anticoagulant use.

While the initial focus of VTE therapy is, appropriately, prevention of progression and recurrence, it is essential also to consider the impact of treatment on uterine bleeding, particularly in menstruating individuals. This impact has been grossly underestimated thus far largely due to inadequate and inconsistent use of appropriate outcome measures in prospective studies. However, it is clear that HMB is very common among menstruating individuals on anticoagulation, that rates are increased with some anticoagulants, such as rivaroxaban, compared to others, and that HMB can have an enormous impact on patient quality of life. Opportunities for improvement in the care of these individuals abound in both clinical and research settings. In the research setting, standardization and consistent use of clinically meaningful measures of HMB and other forms of abnormal bleeding must be routinely used in studies of medications which may increase HMB. In the clinical setting there are many points of contact at which to improve care for these patients. Firstly, an up-front discussion about the increased risk of HMB on anticoagulation, taking into consideration the individual’s own menstrual history, the risks associated with the anticoagulant of choice and potential discontinuation of any therapies which may have impacted menstrual bleeding, either as a primary goal or side effect is essential. Individuals who do experience VTE in the setting of exogenous estrogen, such as CHCs, must be counseled on the risks and benefits of continuing or discontinuing these therapies and should be included in shared decision making around how to manage these important therapies. Finally, at each point of follow-up, both volume and impact of menstrual losses must be considered, extending beyond evaluation for anemia and iron deficiency to include impact on quality of life and daily functioning. In conclusion, HMB and other forms of abnormal bleeding are far more common and impactful in patients on anticoagulant therapy than previously recognized. Stigma can be a major barrier to full and open discussion about risks and concerns for both patient and provider. Risks associated with missing and failing to manage this treatment complication range from inconvenience to missed work or school, lost jobs and even increased rates of recurrent VTE. Clinicians and patients will both benefit from an increased, open, nonjudgmental communication about risks associated with anticoagulants and treatment strategies for HMB. Questions about uterine bleeding must be included in follow-up assessments for all patients at risk, particularly as patients may be less likely to report these symptoms than other, less stigmatized ones such as epistaxis or bruising. A variety of management strategies are available and appropriate management can provide great satisfaction to both patient and provider and must be an essential component of the care of anticoagulated patients.

Strategies of HMB management are largely the same, regardless of whether or not the HMB is associated with anticoagulation. CHCs, along with tranexamic acid, are first-line therapies for HMB in the general population.[ 27 ] CHCs can be used to regulate, increase or decrease frequency of menstrual bleeding, reduce menstrual blood loss or even induce amenorrhea. Additionally, and as outlined in Figure 2 , they can be used to reduce and/or ameliorate other symptoms associated with menstruation. While the increased risk of VTE with CHCs is generally considered to outweigh the benefits in individuals with a personal history of VTE, not on anticoagulation, post-hoc data of individuals involved in registry trials of apixaban and rivaroxaban suggest that hormonal therapies, including CHCs, can be safely used in individuals on therapeutic anticoagulation.[ 14 ] This is of particular importance in disease states such as endometriosis, which may only be responsive to estrogen-based therapies, and the risk and impact of anticoagulation on quality of life may be outweighed by benefits of disease management. However, it may be relevant even in individuals who require long term anticoagulation and desire one or more of the benefits of CHCs outlined in Figure 2 . Individualized and informed risk/benefit analysis should be offered to all patients. Given the inevitability of withdrawal bleeding and high likelihood of return of or development of new HMB, many patients may prefer to continue on CHCs while anticoagulated, if given the option. Occasionally, patients on long term anticoagulation may even need to start a new therapy with CHCs, particularly in circumstances such as the development of endometriosis, for which CHCs are a major component of therapy. It is, however, important to note that the increased risk of VTE continues for 2 to 4 weeks past discontinuation of estrogen containing therapies,[ 28 ] therefore transition to an alternative method of menstrual management should be initiated a least a month before discontinuation of anticoagulation. A preferred alternative for patients with history of VTE planning to discontinue anticoagulation is the levonorgestrel intrauterine device (LNG-IUD), which has dual benefit for HMB management and highly effective contraception. In addition to reducing overall menstrual blood loss, the LNG-IUD is associated with an amenorrhea rate of 20% within the first year, increasing to ~40% between years 3–8, and a <1% contraceptive failure rate.[ 29 , 30 ] Establishing with a practitioner who is able to insert IUDs may take some time, however, and continuing CHCs (so long as anticoagulation is also continued) until placement is a very reasonable approach. The etonogestrel implant is another long acting, reversible contraceptive (LARC) method which can reduce overall menstrual blood loss, but can commonly be associated with irregular and intermenstrual bleeding which some patients find troublesome.[ 31 ] Depomedroxyprogesterone acetate (DMPA) injections boast the highest rate of amenorrhea (>55%) but have a number of side effects, including being the only progesterone-only option which is associated with increased risk of VTE and therefore is not recommended in this population outside of ongoing therapeutic anticoagulation.[ 32 ] Progesterone only pills can also be used but may also result in irregular bleeding and are less effective as contraception. Antifibrinolytics, such as tranexamic acid, are considered first line therapy for HMB outside the setting of acute thrombosis.[ 27 ] While many express theoretical concerns about the risk of new thrombosis, this theory has been disproven by large volumes of data from a variety of randomized controlled trials, including in high risk populations such as trauma victims and postpartum patients, demonstrating no increased risk of VTE.[ 33 ] It is important to note, however, that fibrinolysis is a key process to VTE resolution and therefore initiation of antifibrinolytics is best delayed at least until initial symptoms improve and likely for as long as one to three months. Data on the use of antifibrinolytics in anticoagulated patients are lacking, however tranexamic acid was one of the treatment strategies employed for HMB associated with anti-Xa use in the MEDEA study, data from which are not yet available at the time of this review.[ 34 ] In addition to the above-mentioned methods, which may be useful in patients with HMB due to any cause, in individuals on anticoagulation, consideration may be given to switching from an anticoagulant with a higher rate of HMB, such as rivaroxaban, to one with a slightly lower rate of HMB, such as apixaban. However, this approach has not been studied and in patients such as the one described here, who already suffered from HMB prior to anticoagulant initiation, this is unlikely to be adequate as a sole approach. Alternative modifications, such as interrupting or reducing the dose of anticoagulant on days of bleeding, have not been studied for safety and, given the aforementioned data regarding increased rates of VTE in individuals with HMB while on anticoagulation, may pose a serious risk.[ 35 ] Importantly, individuals who develop new HMB outside the setting of a recent initiation of anticoagulation should be evaluated by a gynecologist for potential structural causes. Menstruating individuals with heavy bleeding, and even those without, would ideally be screened for iron deficiency at the time of anticoagulation initiation. Iron supplementation can be a key component of the management of HMB and may, theoretically, even impact heaviness of bleeding itself due to impact on platelet function.[ 36 ]

Venous thromboembolism (VTE) affects up to 900,000 individuals in the United States annually.[ 1 ] An estimated 46% of these individuals are female.[ 2 ] While risk of VTE increases with age, female individuals experience unique risks during their reproductive years, including use of hormonal therapies such as combined hormonal contraceptives (CHCs) and pregnancy. As a result, a disproportionate number of VTE events effecting reproductive-aged (15–45 years) individuals occur in females.[ 3 ] In addition to experiencing increased risk of VTE compared to male counterparts, reproductive-age females experience unique risks with regard to bleeding complications, specifically menstruation and pregnancy. Menstruation is expected to occur roughly every 28 days in virtually all individuals of female sex for an average of 40 years. This equates to up to 500 individual cycles over a lifetime. Twenty to thirty percent of the general population of menstruating individuals will experience heavy menstrual bleeding (HMB) during this time period, the current definition of which is “excessive menstrual losses which interferes with the [individual’s] physical, emotional, social and material quality of life.”[ 4 ] The combination of increased risk of VTE and the recurrent, cyclic bleeding event known as menstruation require additional consideration and attention when balancing risks and benefits of therapies for both VTE and HMB. In this review we discuss the incidence and impact of HMB in anticoagulated individuals as well as management strategies, from both the clinician perspective and the patient perspective. Accordingly, discussion will be centered around a case, highlighting patient experience in addition to data supporting diagnostic and management strategies.

As discussed above, HMB effects up to a third of the general menstruating population.[ 5 ] This clearly increases in anticoagulated individuals with an estimated prevalence of 60–70%.[ 6 ] However, estimates remain rough and highly variable due to a number of limitations in how HMB is addressed, or in many cases not addressed, in clinical studies. Virtually all registry trials of anticoagulants, including direct oral anticoagulants (DOACs) such as rivaroxaban, rely on International Society of Thrombosis and Haemostasis definitions of major bleeding (MB) and clinically relevant nonmajor bleeding (CRNMB).[ 7 , 8 ] These measures are inadequate to detect HMB for two primary reasons. The first reason MB and CRNMB criteria fail to detect HMB is that they are designed for discrete bleeding events, not recurrent ones. Many patients with HMB do develop anemia and the majority have iron deficiency with or without anemia,[ 9 ] but due to the fact that blood loss is slower, occurring over days each month, criteria such as a 2g/dL drop in hemoglobin and/or need for blood transfusion are rarely met, even though annual blood loss may very well exceed that seen in a single major bleeding event.[ 9 ] The second limitation of MB and CRNMB criteria for detecting HMB is the fact that 1) researchers rarely ask about it specifically in study questionnaires and 2) patients/research subjects are often reticent to volunteer this information. A survey of 1000 women in the UK not on hormonal birth control or hormone replacement therapy reported that 55% had had to take time off work for menstrual symptoms, 60% of whom had a disciplinary hearing for amount of time off work and 51% of whom subsequently lost their jobs. Despite this, less than 50%, and as few as 23% women in some regions of England had talked to or were willing to talk to their general practitioner about it.[ 10 ] While more research is needed to fully understand the reasons for this, there can be little doubt that the stigma associated with menstruation, treating it as something “dirty” or shameful instead of a totally normal bodily process, contributes to this hesitancy. Studies which do address HMB are most often observational, without the benefit of randomization or, in many cases, prospective data collection. Post-hoc analyses of registry trials, while offering improved oversight and comparators, are limited by criteria which grossly underestimate occurrence of HMB. Despite this recognized limitation, there is no true “gold standard” measure of HMB for studies that do address it. The most objective approach, the alkaline hematin method, can accurately quantify menstrual blood loss (MBL) but is prohibitively demanding and costly as it requires collection and processing of all menstrual hygiene products for an entire cycle.[ 11 ] Other methods, such as the Pictorial Blood Loss Assessment Chart (PBAC) are easier to administer but not universally used or adaptable to methods other than standard, disposable pads and tampons.[ 12 ] Ideally, pads and tampons should also be consistent (requiring the study to provide them) for a fully accurate comparison. For this reason, a variety of other subjective methods are often used, with inconsistent findings. While the clinical definition of HMB, based upon the perspective of the patient, has come a long way, research definitions lag behind. A number of available reviews have summarized and attempted to combine the available data on rates of HMB with various anticoagulants. Prevalence estimates are highly variable but most available data strongly suggest an increased rate of HMB with rivaroxaban (a near two-fold increase compared to low molecular weight heparin (LMWH)/warfarin), and perhaps a decreased rate with dabigatran (slightly more than half the rate compared to LMWH/warfarin) with apixaban appearing roughly equivalent to warfarin for this outcome.[ 6 , 13 – 16 ] The increased rate with rivaroxaban is presumed to be due to higher circulating doses of anticoagulant in order to facilitate daily vs. twice daily dosing. Based on previous estimates with warfarin it seems very likely that more than three-quarters of menstruating individuals treated with rivaroxaban, such as our case study, are likely to develop HMB while on anticoagulation.[ 6 ] This risk is higher if the patient has a history of HMB,[ 15 ] as our patient does, and must be balanced against the convenience of daily vs. twice daily dosing of anticoagulation, which can have a meaningful impact on adherence.[ 17 ]