History and Physical Examination
With AUB, the diagnostic goal is exclusion of pregnancy or cancer and identification of the underlying pathology to allow optimal treatment. During initial evaluation of abnormal bleeding, a thorough menstrual history is collected. Topics typically include age at menarche, date of last menstrual period, birth control method, and the timing and amount of bleeding. Associated symptoms such as fever, fatigue, bulk symptoms, tissue passage, or pain can also direct evaluation. Importantly, medications are reviewed, as abnormal bleeding can accompany use of nonsteroidal antiinflammatory drugs (NSAIDs), anticoagulants, and agents associated with hyperprolactinemia (Table 12-2). Less robust evidence implicates herbal supplements such as ginseng, garlic, ginkgo, don quai, and St. John wort (Cordier, 2012).
Most gynecologic disorders do not consistently display a specific bleeding pattern, and patients may complain of HMB or intermenstrual bleeding or both. Thus, the pattern for a particular woman may be of limited value in diagnosing the underlying bleeding cause but can be used to assess improvement with treatment.
Of pain symptoms, dysmenorrhea often accompanies abnormal bleeding caused by structural abnormalities, infections, and pregnancy complications. This seems intuitive because of the role of prostaglandins in both HMB and dysmenorrhea. Painful intercourse and noncyclic pain are less frequent in women with AUB and usually suggest a structural or infectious source.
Following a historical inventory, physical examination attempts to identify findings that may suggest an etiology. Moreover, the site of uterine bleeding is confirmed, because vaginal, rectal, or urethral bleeding can present similarly. This is more difficult if there is no active bleeding, and urinalysis or stool guaiac evaluation may be helpful adjuncts.
To complement physical findings, blood tests, cervical cytology, sonography (with or without saline infusion), endometrial biopsy, and hysteroscopy are used primarily (Fig. 8-4). In many cases following history and physical examination, these tools may not be required or may be individually selected based on patient variables, suspected diagnosis, available resources, and/or provider training. Test suitability for a given patient is discussed next.
Diagnostic algorithm to identify endometrial pathology in patients with abnormal uterine bleeding.
aStudy obtained as indicated by patient history.
bPatients with chronic anovulation, obesity, ≥45 years of age, tamoxifen use, or other risks for endometrial cancer.
cBoth comparable in sensitivity and specificity. Either or both may be selected depending on patient characteristics and physician preference.
CBC = complete blood count; D&C = dilatation and curettage; GC/CT = Neisseria gonorrhoeae and Chlamydia trachomatis; EMB = endometrial biopsy; Hb = hemoglobin level; mgmt. = management; PRL = prolactin level; SIS = saline infusion sonography; TSH = thyroid-stimulating hormone level; TVS = transvaginal sonography; tx = treatment.
β-Human Chorionic Gonadotropin and Hematologic Testing
Miscarriage, ectopic pregnancies, and hydatidiform moles may cause life-threatening hemorrhage. Pregnancy complications are quickly excluded with determination of urine or serum β-human chorionic gonadotropin (hCG) levels. This is typically obtained on all reproductive-aged women with a uterus.
Additionally, in women with AUB, a complete blood count (CBC) will identify anemia and the degree of blood loss. With chronic loss, erythrocyte indices will reflect a microcytic, hypochromic anemia and show decreases in mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), and mean corpuscular hemoglobin concentration (MCHC). Moreover, in women with classic iron-deficiency anemia from chronic blood loss, an elevated platelet count may be seen. In those for whom the cause of anemia is unclear, those with profound anemia, or in those who fail to improve with oral iron therapy, iron studies are often indicated. Specifically, iron-deficiency anemia produces low serum ferritin and low serum iron levels but an elevated total iron-binding capacity. As discussed further on page 192, screening for disordered hemostasis is considered in women and adolescents with HMB and no other obvious cause.
“Wet Prep” Examination and Cervical Cultures
Cervicitis often causes intermenstrual or postcoital spotting. Accordingly, microscopic examination of a saline preparation of cervical secretions or “wet prep” can be informative. With mucopurulent discharge, sheets of neutrophils (>30 per high-power field) and red blood cells are typical. With trichomoniasis, motile trichomonads are also found. Cervicitis-related bleeding is frequently reproduced during sampling from an inflamed cervix with a friable epithelium.
The association between mucopurulent cervicitis and cervical infection with Chlamydia trachomatis and Neisseria gonorrhoeae is well established (Brunham, 1984). Thus, the Centers for Disease Control and Prevention (CDC) (2015) recommend testing for both when mucopurulent cervicitis is found. Moreover, even without frank discharge, these organisms can cause endometritis (Eckert, 2004). Thus, bleeding or spotting alone may merit screening for these two in at-risk populations listed in Table 1-1. Last, herpes simplex virus (HSV) may manifest as diffuse erosive and hemorrhagic ectocervical lesions (Paavonen, 1988). In patients with such findings who lack a known HSV history, directed culture or serologic testing can be obtained.
Cervical Cytology or Biopsy
Both cervical and endometrial cancers can bleed, and evidence for these tumors may be detected during diagnostic Pap smear evaluation. The most frequent abnormal cytologic results involve squamous cell pathology and may reflect cervicitis, intraepithelial neoplasia, or cancer. Less commonly, atypical glandular or endometrial cells are found. Thus, depending on the cytologic results, colposcopy, endocervical curettage, and/or endometrial biopsy may be indicated as discussed in Chapter 29. Moreover, at times, a visibly suspicious vaginal or cervical lesion may bleed and warrant direct biopsy with Tischler forceps.
In women with AUB, sampling and histologic evaluation of the endometrium may identify infection or neoplastic lesions such as endometrial hyperplasia or cancer.
AUB is noted in 80 to 90 percent of women with endometrial cancer. The incidence and risk of this cancer increases with age, and most affected women are postmenopausal (National Cancer Institute, 2014). Thus, in postmenopausal women, the need to exclude cancer intensifies, and endometrial biopsy is typically indicated. Of premenopausal women with endometrial neoplasia, most are obese or have chronic anovulation or both. Thus, women with AUB in these two groups also warrant exclusion of endometrial cancer. Specifically, the American College of Obstetricians and Gynecologists (2012) recommends endometrial assessment in any woman older than 45 years with AUB, and in those younger than 45 years with a history of unopposed estrogen exposure such as seen in obesity or polycystic ovarian syndrome (PCOS), failed medical management, and persistent AUB.
For years, dilatation and curettage (D & C) was used for endometrial sampling. However, because of associated surgical risks, expense, postoperative pain, and need for operative anesthesia, other suitable substitutes were evaluated. In addition, investigators have demonstrated incomplete sampling and missed pathology even with D & C (Grimes, 1982; Stock, 1975).
Initial office techniques used metal curettes. Endometrial samples that are removed with these curettes show significant positive correlation with histologic results obtained from hysterectomy specimens (Stovall, 1989). Thus, they are deemed adequate sampling methods. However, disadvantages include patient discomfort and rare procedural complications such as uterine perforation and infection.
To minimize these, flexible plastic samplers have been evaluated for endometrial biopsy. Advantageously, samples from these catheters have comparable histologic findings with tissues obtained by D & C, hysterectomy, or stiff metal curette (Stovall, 1991). Moreover, they afford greater patient comfort.
Prior to performing endometrial biopsy, pregnancy is excluded in women of reproductive age. With Pipelle insertion, patients frequently note cramping, which can be allayed by a preprocedural NSAID. For some, slow transcervical intrauterine instillation of 5 mL of 2-percent lidocaine using an 18-gauge angiocatheter can lower perceived pain scores (Kosus, 2014).
After patient education and consent, a speculum is placed, and the cervix is cleansed with an antibacterial solution, such as povidone-iodine solution. In many cases, a single-tooth tenaculum is needed to stabilize the cervix and permit passage of the Pipelle through the cervical os and into the endometrial cavity. When placing the tenaculum on the anterior cervical lip, closing the clamp slowly can decrease discomfort. Some evidence also supports topical anesthetic use. Examples are 10-percent lidocaine spray immediately prior or 5-percent lidocaine/prilocaine cream (EMLA cream) 10 minutes before tenaculum placement (Davies, 1997; Zullo, 1999).
With sampling, the Pipelle is directed toward the fundus until resistance is met (Fig. 8-5). Markings on the Pipelle allow measurement of uterine depth, and this value is recorded in the procedure note. The inner Pipelle stilette is then retracted to create suction within the cylinder. Several times, the Pipelle is withdrawn to the level of the internal cervical os and advanced back to the fundus. The device is gently turned during its advance and retraction to allow thorough sampling of all endometrial surfaces. Uncommonly, a vagal response can follow Pipelle insertion. In this instance, the procedure is terminated, and patient support is provided.
Steps of endometrial biopsy. A. During biopsy, the Pipelle is inserted through the cervical os and directed to the uterine fundus. B. The stilette of the Pipelle is retracted to create suction within the cylinder. C. Several times, the Pipelle is withdrawn to the level of the internal cervical os and advanced back to the fundus. The Pipelle is gently turned during its advance and retraction to allow thorough sampling of all endometrial surfaces.
Despite its advantages, there are limitations to endometrial sampling with the Pipelle device. First, a tissue sample that is inadequate for histologic evaluation, such as from endometrial atrophy, or an inability to pass the catheter into the endometrial cavity is encountered in up to 28 percent of biopsy attempts (Smith-Bindman, 1998). Cervical stenosis and large submucous leiomyomas are classic obstructions. An incomplete evaluation often necessitates further investigation with D & C, transvaginal sonography with or without saline infusion, or diagnostic hysteroscopy (Emanuel, 1995). Second, endometrial biopsy has a cancer-detection failure rate of 0.9 percent. Thus, a positive histologic result is accurate to diagnose cancer, but a negative result does not definitively exclude it. Therefore, if an endometrial biopsy with normal tissue is obtained, but abnormal bleeding continues despite conservative treatment or if the suspicion of endometrial cancer is high, then further diagnostic efforts are warranted. Finally, endometrial sampling is associated with a greater percentage of false-negative results with focal pathology such as endometrial polyps. In a study of 639 women evaluated by diagnostic office hysteroscopy and endometrial biopsy, Svirsky and colleagues (2008) found that the sensitivity of endometrial sampling for detection of endometrial polyps and submucosal fibroids was only 8.4 and 1.4 percent, respectively. Because of these limitations with endometrial sampling, investigators have evaluated sonography, hysteroscopy, or both to replace or complement endometrial sampling.
With improved resolution, this technology is chosen by many instead of endometrial biopsy as a first-line tool to assess AUB. Advantageously, it allows assessment of both the myometrium and the endometrium. Thus, if AUB stems from myometrial pathology such as leiomyomas, sonography offers anatomic information that is not afforded by hysteroscopy or endometrial biopsy. In addition, transvaginal sonography (TVS) compared with these other two typically offers greater patient comfort and suitable detection of postmenopausal endometrial hyperplasia and cancer (Karlsson, 1995; Van den Bosch, 2008). That said, no tool, including TVS, is recommended for routine endometrial cancer screening in asymptomatic women (Breijer, 2012).
When the endometrium is imaged in a sagittal view, opposed endometrial surfaces appear as a hyperechoic endometrial stripe down the center of the uterine body (Fig. 8-6 and Fig. 2-16). In postmenopausal women, this endometrial thickness has been correlated with endometrial cancer risk. Although endometrial thickness varies among patients, ranges have been established. Granberg and coworkers (1991) found thickness measurements of 3.4 ± 1.2 mm in postmenopausal women with atrophic endometrium, 9.7 ± 2.5 mm in those with endometrial hyperplasia, and 18.2 ± 6.2 mm in those with endometrial cancer. Subsequent investigations have similarly focused on endometrial thickness as it relates to hyperplasia and cancer risks in postmenopausal women. For endometrial cancer, negative predictive values >99 percent have been reported using a measurement of ≤4 mm (Karlsson, 1995; Tsuda, 1997). Use of hormone replacement therapy (HRT) does not appear to affect the threshold used (Smith-Bindman, 1998). In those using cyclic HRT, completing TVS on day 4 or 5 following cycle bleeding is recommended (Goldstein, 2001). For postmenopausal women, an endometrial thickness >4 mm typically requires additional evaluation with saline infusion sonography (SIS), hysteroscopy, or endometrial biopsy.
The sonographic endometrial stripe in a sagittal plane represents the thickness created by the apposed anterior and posterior endometrium. In premenopausal women, stripe thickness will vary during the menstrual cycle as the endometrium gradually thickens and then is sloughed.
Consensus, however, is lacking regarding the asymptomatic postmenopausal women in whom a thick endometrium is found. The American College of Obstetricians and Gynecologists (2013d) notes that this finding need not routinely prompt evaluation but that further testing is directed by coexistent patient risks. Focal lesions are common in this subgroup and thus may favor SIS or hysteroscopy if additional evaluation is indicated (Schmidt, 2009).
Researchers have also attempted to create endometrial thickness guidelines for premenopausal women. Merz and colleagues (1996) found that the normal endometrial thickness in premenopausal women did not exceed 4 mm on day 4 of the menstrual cycle, nor did it measure more than 8 mm by day 8. However, endometrial thicknesses can vary considerably among premenopausal women, and evidence-based abnormal thresholds that have been proposed range from ≥4 mm to >16 mm (Breitkopf, 2004; Goldstein, 1997; Shi, 2008). Thus, a consensus for endometrial thickness guidelines has not been established for this group. At our institution, no additional evaluation is recommended for a normal-appearing endometrium measuring ≤10 mm in a premenopausal female experiencing AUB if she has no other risk factors to prompt further testing. Risk factors for endometrial carcinoma include extended AUB, chronic anovulation, diabetes mellitus, obesity, and tamoxifen use.
Qualities other than endometrial thickness are also considered because textural changes may indicate pathology. Punctate cystic areas within the endometrium may indicate a polyp. Conversely, hypoechoic masses that distort the endometrium and originate from the inner layer of myometrium most likely are submucous leiomyomas. Although there are no specific sonographic findings that are characteristic of endometrial cancer, some findings have been linked with greater frequency (Fig. 33-3). For example, intermingled hypo- and hyperechoic areas within the endometrium may indicate malignancy. Endometrial cavity fluid collections and an irregular endometrial-myometrial junction have also been implicated. Thus, with these findings, even with a normal endometrial stripe width in postmenopausal patients, endometrial biopsy or hysteroscopy with biopsy is considered to exclude malignancy (Sheikh, 2000).
Although these criteria can safely reduce endometrial biopsy rates for many patients, others consider false-negative rates as too high with this strategy for evaluation of postmenopausal women (Timmermans, 2010). Some advocate hysteroscopy with direct biopsy or D & C to evaluate postmenopausal bleeding (Litta, 2005; Tabor, 2002). In other patient populations, the 4-mm guideline may also be inappropriate. For example, van Doorn and coworkers (2004) reported decreased diagnostic accuracy in diabetic or obese women, and they recommend consideration of endometrial sampling.
A major limitation of TVS is its higher false-negative rate for diagnosing focal intrauterine pathology. This results in part from the physical inability of TVS to clearly assess the endometrium when there is concurrent uterine pathology such as leiomyomas or polyps. In these cases, SIS or hysteroscopy may be informative.
Saline Infusion Sonography
This simple, minimally invasive, and effective sonographic procedure can be used to evaluate the myometrium, endometrium, and endometrial cavity (Chap. 2). Also known as sonohysterography or hysterosonography, SIS allows identification of common masses associated with AUB such as endometrial polyps, submucous leiomyomas, and intracavitary blood clots. These masses frequently create nondescript distortion or thickening of the endometrial lining when imaged with TVS. Thus, compared with TVS, SIS typically permits superior detection of intracavitary masses and differentiation of lesions as being endometrial, submucous, or intramural (Fig. 8-7). In addition, Moschos and colleagues (2009) describe a method of endometrial biopsy during SIS using a sonography-guided Pipelle. Although not yet widely used, this technique enables directed histologic sampling of endometrial pathology and has proved superior to blind endometrial biopsy in providing a diagnosis for AUB in peri- and postmenopausal women.
Transvaginal sonography of the uterus in the sagittal plane. A. The endometrium is thickened in this postmenopausal patient. B. Saline infusion sonography reveals a posterior endometrial mass and further delineates its size and qualities. (Used with permission from Dr. Elysia Moschos.)
SIS has also been compared with hysteroscopy to detect uterine cavitary focal lesions. De Kroon and coworkers (2003) performed a metaanalysis of 24 studies and reported SIS to equal the diagnostic accuracy of hysteroscopy. Importantly, neither hysteroscopy nor SIS can reliably discriminate between benign and malignant focal lesions. Thus, because of the malignant potential of many focal lesions, biopsy or excision of most structural lesions, when identified, is recommended for those with risk factors. For this, operative hysteroscopy is typically used.
SIS has other limitations. First, it is cycle dependent and best performed in the proliferative phase to minimize false-negative and false-positive results. For example, focal lesions may be concealed in a thick, secretory endometrium. Also, the amount of endometrial tissue that can develop during the normal secretory phase can be mistaken for a small polyp or focal hyperplasia. Second, SIS usually has more patient discomfort than TVS, and approximately 5 percent of examinations cannot be completed because of cervical stenosis or patient discomfort. As expected, stenosis is more prevalent in postmenopausal women, and the incompletion rate mirrors that of diagnostic hysteroscopy.
Although accurate for identifying focal lesions, SIS may not add to the value of TVS for evaluation of diffuse lesions such as hyperplasia and cancer. Therefore, in postmenopausal women with AUB, and in whom the exclusion of cancer is more relevant than evaluating focal intracavitary lesions, SIS alone as an initial diagnostic tool may not have advantages over TVS.
Additional Sonographic Techniques
In selected instances, other imaging modalities can provide information beyond that obtained from TVS and SIS. Of these, color and pulsed Doppler, by demonstrating vascularity, may better highlight suspected focal abnormalities (Bennett, 2011). Similarly, 3-dimensional (3-D) sonography and 3-D SIS are most helpful to clarify focal lesions (Benacerraf, 2008; Makris, 2007). With power Doppler, finding multiple irregularly branching vessels may suggest malignancy (Opolskiene, 2007). 3-D power Doppler has been employed to differentiate malignant and benign endometrium, but its value is still undefined (Alcazar, 2009; Opolskiene, 2010). Last, although preferred to computed tomography (CT), magnetic resonance (MR) imaging is rarely needed for AUB evaluation but can display endometrium in cases in which sonographic views are obstructed.
With this procedure, an endoscope, usually 3 to 5 mm in diameter, is inserted into the endometrial cavity as explained in detail in Section 44-12. The uterine cavity is then distended with saline or another medium for visualization. In addition to inspection, biopsy of the endometrium allows histologic diagnosis of abnormal areas and has been shown to be a safe and accurate means of identifying pathology. Also, focal lesions can be diagnosed and completely removed in the same session. In fact, many studies examining the accuracy of TVS or SIS for intracavitary pathology evaluation use hysteroscopy as the “gold standard” for comparison.
The main advantage of hysteroscopy is detection of intracavitary lesions such as leiomyomas and polyps that might be missed using TVS or endometrial sampling (Tahir, 1999). It also permits simultaneous removal of many lesions once identified. Thus, some advocate hysteroscopy as the primary tool for AUB diagnosis. However, the invasiveness and cost of hysteroscopy is balanced against improved diagnostic efficiency. Moreover, although accurate for identifying endometrial cancer, hysteroscopy is less accurate for endometrial hyperplasia. Accordingly, some recommend endometrial biopsy or endometrial curettage in conjunction with hysteroscopy (Ben-Yehuda, 1998; Clark, 2002).
Hysteroscopy has other limitations. Cervical stenosis will sometimes block successful introduction of the endoscope, and heavy bleeding may obscure and hinder an adequate examination. Hysteroscopy is more expensive and technically challenging than TVS or SIS. Costs can be lower with office hysteroscopy rather than that in an operative suite. However, patient discomfort may limit complete examination during some office procedures. Use of a smaller diameter or flexible hysteroscope may diminish this procedural pain (Cicinelli, 2003). In either arena, associated infection and uterine perforation have been reported, but their incidences are low (Bradley, 2002; Vercellini, 1997). Last, peritoneal seeding with malignant cells may take place during hysteroscopy via retrograde flow through the fallopian tubes in some women subsequently diagnosed with endometrial cancer (Bradley, 2004; Zerbe, 2000). Despite the risk of peritoneal contamination by cancer cells with hysteroscopy, patient prognosis overall does not appear to be worsened (Cicinelli, 2010; Polyzos, 2010). The American College of Obstetricians and Gynecologists (2011) considers hysteroscopy acceptable for AUB evaluation in those without advanced-stage uterine or cervical cancer.
Summary of Diagnostic Procedures
There is no one clear sequence to the use of endometrial biopsy, TVS, SIS, and hysteroscopy when evaluating AUB. None of these will distinguish all anatomic lesions with high sensitivity and specificity. That said, TVS for several reasons is a logical first step. It is well tolerated, is cost-effective, and requires relatively minimal technical skill. Additionally, it can reliably determine stripe thickness and whether a lesion is myometrial or endometrial. Once potential anatomic lesions have been identified, subsequent evaluation requires individualization. If endometrial hyperplasia or cancer is suspected, then endometrial biopsy may offer advantages. Alternatively, possible focal lesions may be best investigated with either hysteroscopy or SIS. Ultimately, the diagnostic goal is to identify and treat pathology and specifically to exclude endometrial carcinoma. Thus, selection of appropriate tests depends on their accuracy in characterizing the most likely anatomic lesions.