Dr Claire Philipp, Division of Hematology, MEB Rm 378, UMDNJ-Robert Wood Johnson Medical School, 1 Robert Wood Johnson Place, New Brunswick, New Jersey 08903, USA. Tel.: +732 2357682; fax: +732 2357115; e-mail: email@example.com
Summary. Menorrhagia is a common clinical problem and is unexplained in more than 50% of women. Although studies suggest that von Willebrand's Disease (VWD) is found in a substantial number of women with unexplained menorrhagia, the prevalence of platelet defects in women with menorrhagia is unknown. To determine the prevalence of platelet and other hemostatic defects, we evaluated women ages 17–55 diagnosed with unexplained menorrhagia. Seventy-four women (52 white, 16 black, six other) were studied. Bleeding time was prolonged in 23 women (31.5%). Maximal percent platelet aggregation was decreased with one or more agonists in 35 (47.3%) women. The most commonly found platelet function defects were reduced aggregation responses to ristocetin in 22 women and to epinephrine in 16 women. Sixteen of 22 women with reduced ristocetin aggregation had von Willebrand ristocetin cofactor (VWF:RCo) and von Willebrand factor antigen (VWF:Ag) > 60%. Platelet ATP release was decreased with one or more agonists in 43 (58.1%) women. Of the black women studied, 11/16 (69%) had abnormal platelet aggregation studies compared with 20/52 white women (39%) (P = 0.06). Black women with menorrhagia had a higher prevalence of decreased platelet aggregation in response to ristocetin and epinephrine than did white women (P = 0.0075, P = 0.02). Ten women (13.5%) had VWF:RCo and/or VWF:Ag < 60%. Using race and blood group specific ranges, 5 (6.8%) women had decreased VWF:RCo, VWF:Ag and/or collagen binding (VWF:CB). Mild factor XI deficiency was found in two women and one woman with mild factor V deficiency and one hemophilia A carrier were identified. We conclude that the prevalence of platelet function defects and other inherited bleeding disorders is substantial in a multiracial US population of women with unexplained menorrhagia.
Unexplained menorrhagia is a common clinical problem among women of reproductive age, frequently resulting in anemia, impairing women's daily activities, and often managed by surgery. Menorrhagia is the presenting symptom for the majority of the over 500 000 women who undergo hysterectomy yearly in the USA . More than a quarter of the USA female population undergo hysterectomy by age 60 . Approximately 20% of hysterectomies are performed for ‘dysfunctional’ uterine bleeding not attributable to uterine leiomyomas, polyps, endometrial or cervical cancer, prolapse, pregnancy, or endometriosis .
Despite the potential for excessive bleeding in women undergoing major surgical procedures, hysterectomy mortality rates ranging from six to 11 per 10 000 , and the availability of potential alternative management approaches, women in the USA with unexplained menorrhagia are currently rarely evaluated for inherited bleeding disorders. Recent reports from the USA and Europe have suggested that the prevalence of von Willebrand's disease (VWD) in white women with menorrhagia ranges from 13% to 20% [5–7], with a much lower prevalence reported in black women studied . The frequency and characteristics of qualitative platelet disorders in women with menorrhagia are unknown. Based on results of previous studies which did not systematically include evaluations for platelet disorders, the commonly held presumption has been that VWD is the major bleeding disorder in women with menorrhagia [5–7]. In this study, we evaluated women who had been diagnosed by their gynecologists with unexplained menorrhagia for underlying bleeding disorders. Our data demonstrate that the frequency of undiagnosed qualitative platelet disorders is substantial among women with menorrhagia in a USA multiracial population.
Women between the ages of 17 and 55 years seen at the UMDNJ-Robert Wood Johnson faculty gynecology practice or collaborating community gynecology practices with a physician diagnosis of menorrhagia were eligible to participate in the study. In order to determine a list of potential subjects all medical charts coded with a diagnosis relating to menorrhagia during the period between 1 January 1996 and 4 January 2002 were reviewed. Eligibility required that a pelvic exam performed by their gynecologist be documented in their medical record. Women with known bleeding or endocrine disorders, submucuous uterine fibroids or uterine polyps on pelvic exam or ultrasonography, malignancy, use of an intrauterine device, or treatment with anticoagulants within the past two months were not eligible. Women who had a hysterectomy more than 36 months prior to anticipated study entry were not eligible. Potential study subjects were contacted for participation in the study.
Women taking oral contraceptives or estrogen-based therapy were required to discontinue use for a minimum of one cycle prior to evaluation. Non-steroidal anti-inflammatory agents, aspirin and all medications and herbal agents which could potentially impair platelet function were discontinued at least 14 days prior to testing. Subjects were studied on days 3–9 of their menstrual cycle, unless they had previously undergone a hysterectomy. Participation in the study involved an in-person interview which elicited information on demographic and medical history, bleeding symptoms, family history, and a blood sample. All medical records were reviewed by study staff to confirm the diagnosis of menorrhagia.
Laboratory controls for platelet studies were normal women selected from employees, housestaff, and students at the institution who reported no non-steroidal anti-inflammatory agents, aspirin, or other medications which could potentially impair platelet function for at least 14 days and no oral contraceptives or other hormonal based medication for at least one cycle. None of these women served as donors for the establishment of any laboratory reference ranges.
Informed consent, which was approved by the Institutional Review Boards of the University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School and the Centers for Disease Control and Prevention was obtained from all study participants.
Bleeding time was performed by one of two experienced personnel using a Simplate device (Organon-Technica, Durham, NC, USA) and a modified Ivy technique. Citrated blood (3.2% sodium citrate in a 9 : 1 blood anticoagulant ratio) was obtained by clean venipuncture. Blood samples were also obtained for CBC and ABO blood typing.
Platelet aggregation and ATP release were performed at 37° in platelet-rich plasma on an optical Chronolog platelet lumi-aggregometer (Chrono-Log Corp, Haverton, PA, USA) that simultaneously monitors aggregation by turbidity and secretion of ATP by luminescence  according to standard guidelines . Ristocetin induced platelet aggregation was performed with ristocetin (Chrono-Log Corp) at three final concentrations: 0.5, 1.0 and 1.25 mg mL−1. Platelet aggregation was also performed using 5, 10 and 20 µmol L−1 of adenosine-5′diphosphate (ADP), 0.5 mmol L−1 arachidonic acid, 10 µmol L−1 epinephrine, and 2 µg mL−1 collagen (Chron-Log Corp). Aggregation tracings were performed for at least 5 min and longer, if maximal aggregation had not been reached, and the maximal percent aggregation was recorded. ATP release was measured by luminescence using the luciferin luciferase assay  with 50 µL luciferin-luciferase reagant (Chrono-Log Corp) in 450 µL platelet-rich-plasma. Luminescence was calibrated by addition of a 2 µmol L−1 ATP standard (Chrono-Log Corp). Reactions were initiated by the addition of 1 U mL−1 thrombin, 2 µg mL−1 collagen, 10 and 20 µmol L−1 ADP, and 0.5 mmol L−1 arachidonic acid (Chrono-Log Corp). ATP release was calculated by comparison of peak luminescence recorded from the patient sample with that of ATP standard. Reference ranges for platelet aggregation and ATP release were calculated as two standard deviations about the mean for 40 normal male and female subjects. Calculated maximal percentage platelet aggregation normal ranges were: arachidonic acid 74–99%; collagen 79–98%; ADP (20 µmol L−1) 72–103%; ADP (10 µmol L−1) 58–99%; ADP (5 µmol L−1) 45–105%; epinephrine 73–90%; ristocetin (1.25 mg mL−1) 87–102%; ristocetin 1.0 mg mL−1 75–98%; ristocetin (0.5 mg mL−1) 0.24–6.6. Calculated ATP release ranges (nmoles) were: arachidonic acid 0.56–1.40; collagen 0.74–1.92; ADP (20 µmol L−1) 0.56–1.16; ADP (10 µmol L−1) 0.50–1.06; thrombin >0.5. Platelet aggregation and ATP release responses were considered decreased if less than reference range.
Activated partial thromboplastin time (APTT) was performed using rabbit brain cephalin with silica as activator (PTT A, Diagnostica Stago, Parsippany, NJ, USA). Prothrombin time (PT) was performed using Neoplastine CI Plus (Diagnostica Stago). Thrombin time was performed using human thrombin (Diagnostica Stago). Factors (F)II, V, VII, IX, X, XI, and XII were measured using factor deficient plasmas. FVIII activity was measured by one-stage assay (Diagnostica Stago) using silica as activator and partial thromboplastin. All coagulation tests were performed on automated analyzers (STA or STAR, Diagnostica Stago). VWF:Ag was measured by ELISA using polyclonal antiserum (Asserachrom vWF, Diagnostica Stago, Parsippany, NJ). VWF:RCo was measured by aggregation of lyophilized normal platelets by ristocetin in an optical aggregometer (Chrono-Log Corp) as previously described  using ristocetin (American Biochemical and Pharmaceutical Corp, Marleton, NJ, USA) at a final concentration of 1 mg mL−1. Reference standards for FVIII and VWF were lyophilized commercial reference plasmas standardized against the 3rd International Standard for Factor VIII and VWF in Plasma. Blood type and race-specific reference ranges were calculated as two standard deviations about the mean for control subjects . VWF:CB was measured by ELISA using equine collagen. All samples were tested in duplicate.
Differences in means of maximal percentage aggregation between cases and controls were evaluated by t-tests. Fisher exact test for 2 × 2 tables with 95% confidence intervals and mid-P-values were used to compare prevalences between cases and controls. A P-value of less than or equal to 0.05 is considered statistically significant. Confidence intervals for odds ratios are reported where appropriate.
Pictorial chart assessment of menorrhagia
Women were provided at the time of study visit with a pictorial chart [11,12] to complete with their next menses and an explanation of how it should be completed for estimating menstrual blood loss. The chart consisted of a series of diagrams representing lightly, moderately, and heavily soiled pads and tampons. Subjects recorded each discarded item for the duration of an entire cycle. Scoring of pads and tampons was done as previously described [11,12] with tampons assigned 1, 5, and 10 and pads scored 1, 5, and 20 for lightly, moderately, and heavily soiled, respectively. Scoring of clots was performed in comparison with US coins, a modification of the previously reported comparison with Dutch coins , with one assigned to clots smaller than a quarter, and five assigned to clots a quarter or larger in size.
Characteristics of study population
Seventy-four women with unexplained menorrhagia were studied with a mean age of 40.4 ± 8.1(17–52) years. Participant characteristics are shown in Table 1. Four women (5.4%) had previously undergone hysterectomy for excessive bleeding. Fifty-two normal women with a mean age of 34.8 ± 9.7 (18–57 years) were controls for platelet studies.
Table 1. Participant characteristics (n = 74)
40.4 ± 8.1
Age at menarche (years)
12.1 ± 1.6
Pictorial blood loss assessment charts (PBAC) were returned by 59 subjects. Fifty-one (86%) had a PBAC score >100; 37 (63%) had a PBAC score >185. The mean hemoglobin of the women was 11.8 ± 1.9 g dL−1 (5.9–14.5 g dL−1). Thirty-one women had a hemoglobin < 12 g dL−1. The mean platelet count was 278 130 ± 72 702 µL−1 (102 000–439 000). Twenty-six subjects had platelet counts >300 000 µL−1. Two subjects had platelet counts between 100 000 and 150 000 µL−1. Mean corpuscular volume was < 78 fL in 18 women. Mean platelet volume was 8.7 ± 1.1 (6.9–14.1).
Twenty-five women (33.8%) with menorrhagia had one hemostatic abnormality, as defined by a prolonged bleeding time, abnormal platelet aggregation with one or more agonists, abnormal ATP release with one or more agonists, decreased VWF ristocetin cofactor or antigen, or single factor deficiency. Eighteen (24.3%) demonstrated two hemostatic abnormalities, 12 (16.2%) had three hemostatic abnormalities, 3 (4.1%) had four abnormalities, and one woman (1.3%) had five hemostatic abnormalities.
Bleeding time and platelet studies
Twenty-three women (31.5%) had prolonged bleeding times (>9 min) (Fig. 1). In only two women, the prolonged bleeding time was unassociated with any other hemostatic abnormalities. Women who were found to have prolonged bleeding times were not more likely to be anemic than women with normal bleeding times (95% CI 0.3,2.3; P = 0.9). Moreover, we could find no relationship between bleeding time and hemoglobin (P = 0.27) or hematocrit (P = 0.14) using Pearson's correlation and linear regression.
Maximal platelet aggregation was decreased with one or more agonists in 35 (47.3%) women. The prevalence odds of platelet aggregation abnormalities was 4.2-fold higher among women with menorrhagia than among control women (95% CI 1.8,9.8, P = 0.001) (Table 3). Twenty-two women (29.7%) had decreased aggregation in response to one agonist, seven women (9.5%) had decreased aggregation in response to two agonists, and six women (8.1%) had decreased aggregation in response to more than two agonists. Fifteen women (20.3%) had abnormal platelet aggregation and a prolonged bleeding time. Among controls with reduced aggregation response, 7/9 (77%) had a decreased response to one agonist.
Table 3. Platelet function testing abnormalities in women with menorrhagia and control women
Epinephrine (10 µmol L−1) induced platelet aggregation was decreased in 16 women (21.4%) with menorrhagia and in two controls (P = 0.005) (Table 4). With epinephrine as agonist, the mean percentage maximal aggregation of women with menorrhagia was significantly lower than female controls (76.33 ± 16.75 vs. 83.12 ± 13.07; P = 0.002) (Fig. 2c). Furthermore, most of the women with decreased epinephrine induced aggregation (11/16) had additional aggregation abnormalities, either with ristocetin, collagen or both agonists. All five women with isolated impaired epinephrine aggregation also had abnormalities in ATP release; three of these women had a prolonged bleeding time in addition.
Table 4. Platelet aggregation and platelet ATP release defects by agonist
Twenty-two women (29.7%) had decreased aggregation in response to ristocetin. Twenty women (27%) had decreased aggregation in response to 1.0 mg mL−1 ristocetin (Table 4); 16 women (21.6%) had decreased aggregation in response to ristocetin 1.25 mg mL−1; 14 women (18.9%) had decreased aggregation in response to both ristocetin 1.0 and 1.25 mg mL−1. Of 22 women with decreased ristocetin induced platelet aggregation, only six had VWF ristocetin cofactor and/or antigen < 60%; 12 had one or more other platelet aggregation abnormalities and an additional four had one or more ATP release abnormalities. The mean percentage maximal aggregation in response to both 1.0 and 1.25 mg mL−1 ristocetin was significantly lower in women with menorrhagia compared with control women (73.32 ± 27.93 vs. 82.96 ± 21.18; P = 0.017 and 86.31 ± 18.53 vs. 91.92 ± 4.04; P = 0.009)(Fig. 2a,b).
Arachidonic acid induced platelet aggregation was decreased in six women (8.1%) and one control (P = NS) (Table 4). Mean percent aggregation was significantly lower (P = 0.008) in cases (78.81 ± 22.51), compared with controls (85.82 ± 6.98) (Fig. 2f). Platelet aggregation was decreased in response to collagen (2 µg mL−1) in 9 (12.1%) women (Table 4). ADP induced platelet aggregation was decreased in three women (Table 4). With collagen and ADP, the mean percentage maximal aggregation was not significantly different between cases and controls (Fig. 2d,e).
Platelet ATP release was decreased in 43 women (58.1%) in response to one or more agonists (OR 4.3, 95% CI 1.9,9.4, P = 0.0002) (Table 3). Of the 43 women with abnormal ATP release, 19 women also had a prolonged bleeding time, 16 had abnormal platelet aggregation in response to one or more agonists, and 13 had both a prolonged bleeding time and abnormality in platelet aggregation. Thirteen women with abnormal ATP release, had neither an abnormality in platelet aggregation or bleeding time. Thirty women (35.2%) had decreased ADP-induced ATP release; seven of these women had an isolated reduction in ADP-induced ATP release without other abnormalities in release, aggregation or bleeding time.
Of the black women with menorrhagia studied, 11/16 (69%) had abnormal platelet aggregation in response to one or more agonists compared with 20/52 white women (39%) with menorrhagia (OR 3.1, 95% CI 0.9,10.4, P = 0.06). Ristocetin induced platelet aggregation was decreased in 9/16 (56.3%) black women compared with 11/52 (21%) white women with 1.0 mg mL−1 concentration of ristocetin (P = 0.0075) (Table 5) and 8/16 black women compared with 6/52 white women with 1.25 mg mL−1 concentration of ristocetin (P = 0.001) in women with menorrhagia. Seven of nine black women with a ristocetin induced platelet aggregation defect had other platelet aggregation defects and/or a prolonged bleeding time. The mean percentage maximal aggregation was lower in black women compared with white women with menorrhagia using 1.0 mg mL−1 ristocetin (P = 0.015) and 1.25 mg mL−1 ristocetin (P = 0.006) as agonists. Defects in ristocetin induced platelet aggregation were also more prevalent in black women with menorrhagia compared with black controls (1.25 mg mL−1 ristocetin, P = 0.03; 1.0 mg mL−1 ristocetin, P = 0.05). Three of 15 black controls were identified with decreased ristocetin aggregation. Epinephrine-induced platelet aggregation was decreased in 7/16 (44%) black women compared with 8/52 (15%) white women (P < 0.02). Collagen induced platelet aggregation was decreased in 5/16 (31%) black women compared with 4/52 (8%) white women (P < 0.02) (Table 5). Abnormalities in epinephrine induced platelet aggregation were not seen in black controls; one black control had decreased collagen, ADP, and ristocetin aggregation. Bleeding time was prolonged in 5/16 (31%) black women compared with 16/52 (31%) white women. There were no significant differences between black and white women with menorrhagia in abnormalities of ATP release.
Table 5. Platelet aggregation and ATP release defects by race *
Women with menorrhagia
White (n = 52)
Black (n = 16)
* Data for races other than white or black not shown (n = 6). ** Ristocetin concentration 1.0 mg mL−1. ***ADP concentration 20 µmol L−1.
Ten women (13.5%) were identified with VWF:RCo less than 60%. Using race and blood type specific reference ranges, 5/74 (6.7%) women (four white, one black) were below range for VWF:RCo, VWF:Ag, and/or VWF:CB. One woman had absent high and intermediate multimers and was classified as a type IIA VWD. There was a significant relationship between VWF:CB and VWF:RCo (P < 0.0001, by linear regression) regardless of blood type but not with ristocetin induced platelet aggregation (P > 0.5).
Two women had FXI deficiency in the heterozygote range. One woman with FV activity in the heterozygote range was identified. She was also found to have a 14-min BT and abnormal platelet aggregation in response to arachidonic acid, epinephrine, and ristocetin. One woman was found to have a mild factor VIII deficiency with normal VWF:RCo, VWF:Ag, VWF:CB, normal factor VIII binding, and family history consistent with hemophilia A carrier status.
Our results demonstrate that underlying hemostatic defects, in particular qualitative platelet abnormalities, occur in the majority of women with unexplained menorrhagia. We found that 49% of women we studied with menorrhagia had two or more hemostatic defects as previously defined. Our results are consistent with prior studies which have also found that a substantial number of women with unexplained menorrhagia have bleeding disorders [5–7]. However, in contrast to previous studies, we found that in this multiracial group platelet dysfunction was more prevalent than VWD. The most likely explanation is that none of the prior studies included systematic examination of platelet function in their study population [5–7]. Interestingly, Edlund et al.  performed bleeding times and found that 5/30 had prolonged bleeding times without evidence of VWD, suggesting that these women may have had platelet defects. Based on these results one might surmise that platelet dysfunction may have been at least as prevalent as VWD in this population. In addition, our study population was composed of approximately 20% African-American women who have recently been reported to have a lower prevalence of VWD [7,10]. The present data suggests that, compared with white women, black women are more likely to have platelet dysfunction associated with menorrhagia.
Bleeding time, a measure of the platelet–microvascular interaction , has been described as a screening test for platelet function for nearly a century , provided it is performed in a standardized manner. A substantial portion (43%) of the women with abnormal platelet aggregation studies in our study had a prolonged bleeding time, but significant numbers of women (57%) with abnormal platelet aggregation studies would have been missed if the bleeding time was used to screen for further tests of platelet function. Our data does not support earlier literature suggesting a relationship between prolonged bleeding time and anemia [15,16].
The most commonly found platelet function defect was a reduced aggregation response to ristocetin in 30% women. Most of the women with decreased ristocetin induced platelet aggregation had normal VWF:RCo and VWF:Ag, and normal VWF:CB suggesting the possibility of differences in the glycoprotein IB–IX–V receptor complex or its interaction with VWF. This defect was particularly common in black women with menorrhagia, occurring in 56% of black women tested. We also found low ristocetin induced aggregation in 20% of black controls, but this was significantly less prevalent than in black women with menorrhagia. Decreased ristocetin-induced platelet aggregation in black controls compared with white controls has been previously reported [9,10,17,18], but the mechanism of this defect is not known. One author  has postulated a plasma inhibitor to ristocetin aggregation based on an inhibition of normal ristocetin aggregation by their plasma.
We found 16 women with impaired platelet aggregation with epinephrine, including 11 who had either subnormal collagen and/or ristocetin induced aggregation. The five women with isolated impaired epinephrine aggregation also had subnormal ATP release and 3/5 had a prolonged bleeding time in addition. Thus abnormal epinephrine aggregation, in addition to being significantly more prevalent in women with menorrhagia, was not found as an isolated defect in this population but rather was uniformly found to be associated with other aggregation or release defects. Epinephrine aggregates human platelets and also potentiates the aggregation of platelets with other agonists . Familial decreases in epinephrine responsiveness have been reported by several investigators in apparently normal individuals [20,21] as well as in individuals with mild bleeding diathesis [22,23]. Decreased epinephrine responsiveness has been associated with decreases in α2-adrenergic receptors [21–23] and to defects in activation of the fibrinogen receptor [24,25]. The prevalence of epinephrine aggregation abnormalities in individuals with mild bleeding diathesis has not been previously reported in the literature. Similar to our findings with ristocetin induced aggregation, black women with menorrhagia had a higher prevalence of defective epinephrine aggregation than did white women with menorrhagia.
The diagnosis of VWD is complex and, in the absence of genetic analysis, considerable variability in prevalence exists between studies. Such variability can be explained in part based on whether blood type and/or race specific ranges were used, the racial composition of the cohort, whether borderline concentrations were included, and if menstrual cycle specific testing was utilized. In our study, the prevalence of VWD varied from 6.7% to 13% depending on whether race and blood type-specific ranges were used. Using race and blood type specific ranges, Dilley et al. and Miller et al.[7,10] reported a VWD prevalence of 6.6%, whereas Edlund et al. and Kadir et al. using VWF levels of < 55% and 50%, respectively, reported prevalences of 20% and 13%. Differences in the racial composition of the cohorts studied further confound comparison of the prevalence rates between the studies. When white cases were analyzed separately, Dilley et al.  found a prevalence of VWD of 15% compared with an overall prevalence of 6.6%.
A number of authors [26–30] have suggested that fluctuations in VWF levels occur with the menstrual cycle. Although data on the optimal days for testing vary [26–28], it has been suggested that restricting sampling to specific days of the cycle reduces interindividual variation [6,24]. Cycle specific blood draws were obtained in the present study and by Edlund et al. but not in other reported studies examining the prevalence of bleeding disorders in women with menorrhagia [5,7,10]. The effects of menstrual cycle, if any, on platelet function have not been reported to date.
Our findings extend the results of earlier studies and emphasize the importance of comprehensive hemostatic testing, including platelet function testing, in the evaluation of women with unexplained menorrhagia. Studies elucidating specific platelet defects have largely been limited to individual family studies. Our results demonstrate that mild platelet defects are common in women with unexplained menorrhagia. Since approximately 5% of women are estimated to have menorrhagia  and half of them have no organic explanation , an extrapolation of our data to the general menstruating female population suggests a high prevalence of platelet dysfunction in the general population. Further elucidating their specific platelet abnormalities and developing targeted medical treatment options should have a profound impact on the health and quality of life of large numbers of women.
The authors gratefully acknowledge the technical assistance of Chona Baquiran, Elizabeth Haff, Jaya Lakra, Jean Platt, Maria Propst, Peggy Rawlins and Fran Siciliano. Meena Sehadan's assistance with data analysis is also gratefully acknowledged.
Supported by Association of Teachers Preventive Medicine Grant #TS-284.