Materials and Methods

Consent was obtained to conduct this study from the John R. Wolfe Human Subjects Committee and from the Research Committee of the Los Angeles BioMedical Research Institute at Harbor-UCLA Medical Center. A waiver for informed consent was granted because no personal identifying information was collected.

This is a retrospective cohort study of all premenopausal women treated at Harbor-UCLA during the 6 years of 2008–2013, with a hemoglobin value of less than 5 g/dL because of heavy menstrual bleeding. A list of potential subjects was developed from the International Classification of Diseases , ninth revision, codes for excessive menstruation (285.9, 626.2) and hemoglobin values less than 5 g/dL. The diagnosis of excessive men-struation was made clinically by the woman’s gynecologists, based on the patient’s description of the duration and/or flow supported by laboratory findings and the absence of any other etiology for her anemia.

Harbor-UCLA Medical Center is part of the Los Angeles Department of Health Services and serves primarily indigent, uninsured minority (predominantly Hispanic) residents. The medical records at Harbor-UCLA are in transition from paper charts to electronic versions. Once the computerized list of names of potential subjects was obtained, electronic records were consulted to verify that the patient’s anemia was attributed by her clinician to her menstrual bleeding and to obtain computerized laboratory results.

Individual patient paper charts were reviewed for additional information from any of the written records including the emergency room sheets, nursing notes, and daily notes and follow-up clinic notes. Demographic information, menstrual bleeding records, medical and surgical history, and records at the time of presentation were all collected as were the physical findings, laboratory test results, and discharge treatments provided. Data were entered into Excel and calculations made using standard statistical programs. Statistical significance was set at P < .05 when χ ² comparisons were made.

Results

The computer search identified 271 premenopausal women who experienced severe anemia, (hemoglobin <5 g/dL) and excessive menstrual blood loss; 122 women were excluded because their anemia was not related to heavy menses; the most common nongynecological causes for severe anemia in these women were end-stage renal disease, leukemia, and trauma. Interestingly, the admission notes in these cases did not include any information about menstrual bleeding patterns of the women. Other cases were excluded because the low hemoglobin did not occur at the time that heavy bleeding was noted to be a problem.

After appropriate exclusions were made, the study population included 149 women who experienced 168 episodes of hemoglobin less than 5 g/dL, thought at the time of the presentation to be due only to heavy menstrual bleeding treated at this facility during those years.

Table 1 displays the demographic characteristics of the study population. The mean age at the time of diagnosis was 41.6 years, with a range of 19–55 years; 77% of the women were in their 40s. The gravidity of the study population ranged from 0 to 10, with the mean value of 2. Parity ranged from 0 to 7, with the mean of 1. The majority of the study population was ethnically split between African-American (40%) and Hispanic (44%), with white non-Hispanic, Asian, Pacific Islanders, and others each contributing 1–6% of the population. For women, information about ethnicity was not available.

Many women reported medical comorbidities, but not all the histories were complete. Hypertension was reported in 33 of the admissions, diabetes in 21, thyroid dysfunction in 11, hepatic abnormalities in 6, idiopathic thrombocytopenic purpura in 3, sickle cell trait in 6, thalassemia in 3, and von Willebrand disease in 1 instance.

The menstrual histories provided by women were varied and are displayed in Table 2 . At one extreme, 7.8% of women considered their menses to be normal, and 7% were referred to the hospital because of an incidental finding of severe anemia discovered during routine screening elsewhere. At the other extreme, almost 90% reported heavy or prolonged menses; 66% admitted to having heavy bleeding for more than 6 months and another 13% said their problem had existed for 3–6 months. Nearly 40% had previously received blood transfusions. Fewer than 55% of women presented with active bleeding.

The body mass indices averaged 28.9 kg/m ² ; 43.8% were obese and 23.8% were overweight. Pelvic examinations were documented on admission in 69.6% of women; 42.9% were found to be actively bleeding.

Most women were admitted to the gynecology service, but a substantial minority of the time (25%), the woman was treated only in the emergency room. One woman was treated for a few days on the gynecology service for her bleeding but was then admitted to the psychiatry service; her total hospitalization lasted 22 days. The mean number of units of packed red blood cells given to each woman was 3.75 U, with a median of 4 U and a range of 0–9 U. Of note, 6% of the women were Jehovah’s Witnesses; all but 1 of these women declined transfusions.

Table 3 displays the laboratory findings. The mean hemoglobin was 4.2 g/dL, with a range of 1.6–6.4 g/dL. The lowest hemoglobin for individual women during hospitalization ranged from 1.6 to 4.9 g/dL. Mean corpuscular volume (MCV) was 62.2 fL; 93.4% were abnormally low. Similarly, the mean value for mean corpuscular hemoglobin concentration (MCHC) was 29.2 g/dL; 99.4% were abnormally low.

Ferritin is the most accurate test to identify iron deficiency anemia. It was tested in 56.6% of cases; 86.3% of those values were abnormally low. Coagulation factors were tested in slightly more than 80% of patients; prothrombin time and partial thromboplastin time were elevated in more than 80% of tested women. Other related tests that might have been helpful in diagnosing underlying medical causes of excessive bleeding were obtained in just more than half of the cases. Endometrial biopsies results were available for 79% of the cases. Only 13% revealed any endometrial pathology (2.6% endometrial carcinoma). Table 4 displays pathology reports for each of those specimens.

Very interestingly, 41 women were found to have platelet counts exceeding 420,000/mm ³ , although no episodes of venous thromboembolism occurred, and most platelet counts resolved with transfusion and later iron therapy; 37 of them (25% of the total population) clearly had severe anemia-induced reactive thrombocytosis. The mean platelet count of that subgroup was 726.6 × 10 ³ /mL, ranging from 462 to 1891 × 10 ³ /mL). Six percent of the women had thrombocytopenia.

The red blood cell indices were used to confirm the reliability of each woman’s estimate of the duration of her excessive bleeding. In virtually all of the women who reported excessive bleeding for at least 3 months, the indices indicated iron deficiency. However, only 1 of the women who reported normal menses had normal indices.

Ultimately, radiological imaging (primarily ultrasound) results were available for 101 of women (68%). Eleven were normal, and 8 suggested endometrial pathology (thickening, polyps, complexities consistent with cancer). Myometrial abnormalities (leiomyoma, adenomyosis, enlarged uterus) were reported in 75 women with or without other pelvic pathology. Isolated adnexal pathology was reported in 7 other women. After all the test results became available, excessive bleeding was attributed to leiomyoma in 47.9% of the women; cancer (both gynecological and other) was detected in another 4.8%.

At the time of hospital discharge, 57 women (35.1%) were offered no therapy to prevent further episodes, although they were given follow-up appointments. Hysterectomy was done within 3 months in 15 women; in 3–7 months, another 19 women underwent hysterectomy. Ablation was performed in 4 women, polypectomy in 2 women, and myomectomy in 4 cases. Medical therapies included GnRH agonist (n = 4); depot medroxyprogesterone acetate (n = 8), progestin-only pills (n = 2), and levonorgestrel intrauterine system (n = 4). Overall, 35.1% of the women were lost to follow-up prior to initiating therapy and 41% were lost shortly thereafter. Another 26.8% of the women had multiple subsequent transfusions, including 17 more episodes for hemoglobin levels again below 5 g/dL.

Two cases are presented in detail to illustrate the complexities of this problem. One challenging case involved a 44 year old, slender nulligravida woman who reported at her admission in 2012 that she had required transfusion in 1995 for a hemoglobin level of 7.9 g/dL and again later that year when her hemoglobin dropped to 4.2 g/dL. She was also transfused in 2003 and 2006. She had undergone uterine artery embolization and myomectomy. She was bleeding when she presented with complaints of weakness and fatigue, dizziness, dyspnea on exertion, chest pain, palpitations, nausea, tenderness, and pain. Her sister told her it was “all in her head” and that she “should be able to walk on stairs without fear.”

On admission, her initial hemoglobin was 1.8 g/dL with an MCV of 59.0 fL and an MCHC of 27.9 g/dL. She required care in the intensive care unit for 27 days, with respiratory failure requiring multiple intubations, metabolic acidosis, congestive heart failure, elevated troponin levels, sepsis (methicillin-resistant Staphylococcus aureus ), and pulmonary edema. In total, she received 9 U of packed red blood cells and 2 U of fresh frozen plasma. Once she was stabilized, she agreed to and underwent a hysterectomy.

The second case was a 47 year old nulligravida woman with a body mass index of 35 kg/m ² who denied any abnormal vaginal bleeding. She reported having once-monthly menses lasting 7 days with the use of 2–3 pads per day. In the emergency room, she said she had experienced only 1 day of heavy bleeding, but her mother reported she had bled for 1 week. She had sought care for weakness and fatigue, dyspnea on exertion, and cough. Her hemoglobin was 1.6 g/dL, with a depressed MCV (66.4 fL), MCHC (27.4 g/dL), and ferritin levels (4.5 ng/mL). She suffered a seizure in the emergency room and required treatment for pulmonary edema. Initially her bleeding was attributed to her large fibroid uterus (one myoma measured 8–9 cm). Hysterectomy was delayed because the patient still was seeking fertility. A month later she developed hepatic dysfunction and aplastic anemia because of leukemia.