In this post, I link to and excerpt from emDocs’ Speculations on the Speculum: Is a pelvic exam ever needed in the ED?
APR 8TH, 2019 KIMBERLY CHRISTOPHE from emDocs.
Authors: Kimberly Christophe, MD and Mark Silverberg, MD (Kings County Hospital/SUNY Downstate Emergency Medicine) // Edited by: Alex Koyfman, MD (@EMHighAK) and Brit Long, MD (@long_brit)
All that follows is from the above resource.
A 19-year-old female presents to your busy emergency department with a complaint of malodorous vaginal discharge for 2 days. She has recently started dating a new partner and is concerned about having contracted a sexually transmitted infection. She is afebrile and hemodynamically stable with a negative pregnancy test. What is your next step in the workup and management of this patient?
Introduction
Classical medical teaching deems the pelvic exam a necessary component of the emergent evaluation for female patients presenting with a genitourinary or abdominal complaint. We are taught that when considering intra-abdominal disease, a speculum and bimanual exam should be performed to rule out pelvic pathology as the cause for these patients’ symptoms. But the utility of the pelvic exam as a diagnostic tool has been called into question for being both inaccurate and poorly reproducible.1-3 In addition, the time and privacy required for its appropriate utilization can add considerable inconvenience to the workflow of an already overextended emergency practitioner. Given the expanded availability of ultrasound and PCR technology, it is not clear that the pelvic exam is always needed. Indeed, while some medical textbooks suggest that the pelvic exam may occasionally be useful as one of many diagnostic tools available to the emergency practitioner, there are many cases in which it cannot be relied upon and will likely provide no actionable diagnostic information.4
The Pelvic Exam as a Diagnostic Test
The ED pelvic exam typically consists of a speculum examination during which the cervix and vaginal walls are visualized, as well as a bimanual examination, at which time a provider evaluates for tenderness, masses, uterine morphology, and the state of the cervical os. Unfortunately, abnormal findings can be both difficult to discern and highly subjective. This is typified in an observational analysis by Close et al. wherein patients receiving a pelvic exam during usual care were examined by two emergency physicians of similar experience. There was weak inter-examiner reliability with respect to abnormal findings, with only 33% agreement when an examiner noted uterine tenderness, 23% agreement when a pelvic mass was palpated, and 17% agreement when cervical motion tenderness was documented.1 Lest one infer that better training of emergency physicians to perform pelvic exams might improve accuracy, it does not appear that gynecologists fare significantly better. Padilla’s small cohort study tasked gynecologists to perform bimanual pelvic examinations while patients were under general anesthesia and compared their results with intra-operative findings. Although residents and attendings performed better than medical students, exams nonetheless demonstrated less than 50% sensitivity for adnexal masses, regardless of the gynecologist’s level of training.2 In a similar observational series by Ueland et al., not even half of the ovaries evaluated in their study were palpable on pelvic exam under anesthesia (again by gynecologists), with fewer than 10% of ovaries discernible in women weighing 200 lbs or more; this is in contrast to 85% of ovaries observable with ultrasound, including the ability to detect 73% of ovaries in patients over 200 lbs with ultrasound. Of those ovaries that were palpable on exam, only 40% of the size estimates came close to approximating the true dimensions as determined by post-oophorectomy pathology, compared to 70% of size estimates by ultrasound.3 Given its inconsistent results and low sensitivity, it is not clear that the pelvic exam is high yield for detecting gynecologic emergencies when more advanced diagnostic methods are available.
-Pelvic exam findings are highly subjective.
-Accuracy of pelvic exam results may vary depending on patient body habitus.
Non-pregnant patients
When evaluating non-pregnant female patients with abdominal or genitourinary complaints, a provider must consider pathology likely to cause significant morbidity—infection (upper or lower genital tract), ovarian torsion, trauma, or retained foreign body. While the latter two conditions oblige a provider to perform a pelvic exam for both diagnostic and therapeutic benefit, an unambiguous history from a patient may be sufficient to exclude these etiologies. Regarding the former two conditions, however, the role of the pelvic exam is somewhat less obvious.
Upper and/or lower genital tract infections (i.e. pelvic inflammatory disease and cervicitis) are frequently diagnosed by emergency providers in patients presenting with symptoms of vaginal discharge, dysuria, abdominal pain, pelvic pain, or vaginal bleeding. Pelvic exams are often performed reflexively on these patients with at least some thought that direct visualization and palpation might help identify the likely pathogen and determine if empiric treatment is warranted. However, in Farrukh et al.’s study, the ED pelvic exam was not shown to be particularly useful in identifying which patients actually had infection. In this observational study, emergency practitioners predicted the likelihood of positive STI test results before and after performing a pelvic exam on adolescent patients; the results were then correlated with NAAT testing for gonorrhea and chlamydia, as well as wet mount for trichomonas, which were utilized as the gold standards for this study. The results demonstrated no identifiable pattern with which providers altered their predictions after performing the pelvic exam, resulting in sensitivity for the combined history and pelvic exam of only 48% with a specificity of 61%—values statistically indistinguishable to that of the history alone.7 Nonetheless, CDC guidelines advocate a very low threshold for empiric treatment of cervicitis or PID, the bases for which rest almost entirely on pelvic exam findings. The diagnosis of cervicitis includes visualizing purulent endocervical discharge or friable endocervical tissue, while pelvic inflammatory disease is diagnosed with cervical motion, uterine, or adnexal tenderness in the setting of unexplained pelvic or abdominal pain.8,9 Returning to Farrukh’s investigation, if providers were to have treated empirically based on their suspicion for infection after performing the pelvic exam, it would have likely resulted in 8% of patients spared unnecessary antibiotic treatment while another 8% received unnecessary antibiotics; another 4% of patients would have received appropriate antibiotic treatment, while 5% would have had treatment withheld despite having a sexually transmitted infection.7 That is to say that on a population scale, the pelvic exam was essentially useless in determining which patients had genital tract infections and did not lead to improved patient care.
Pelvic exams are also not necessary for accurate STI testing, provided that NAAT is performed as CDC guidelines do not distinguish among endocervical, vulvovaginal, or urine samples for diagnosing chlamydia, gonorrhea, and trichomonas.8-10 Urine samples and vulvovaginal swabs can be obtained privately by patients, and may be preferable for many women.11 Not only are self-collected vulvovaginal specimens simpler to obtain, but they may provide increased sensitivity over provider-collected endocervical swabs, as revealed through linked studies by Schoeman and Stewart. These paired studies evaluated the test characteristics of self-collected vulvovaginal specimens for diagnosis of chlamydia and gonorrhea, respectively. In each of the studies, patients received NAAT-based STI testing of both a physician-collected endocervical swab and a self-collected vulvovaginal swab. Results revealed overall sensitivities of 96% (endocervical) versus 99% (vulvovaginal) for gonorrhea and 88% (endocervical) versus 97% (vulvovaginal) for chlamydia, though a statistically significant difference was observed only in testing for chlamydia. That said, among symptomatic patients—a subset perhaps more representative of emergency department patients—while sensitivity for chlamydia remained 88% and 97% for endocervical and vulvovaginal specimens, sensitivity for gonorrhea was 100% for either method of collection.12,13 Comparatively, culture detected only 84% of gonorrhea infections in symptomatic patients (judged against a gold standard of at least one positive provider-collected or patient-collected NAAT test), making it a poor substitute for amplification assays.13
-Absence of adnexal mass and/or lack of severe pelvic tenderness does not rule out ovarian torsion.
-Pelvic exam should not be relied on to rule out sexually transmitted infections.
-Patient-collected vaginal swabs or urine specimens are acceptable and accurate for STI testing.
Early pregnancy vaginal bleeding or abdominal pain
In addition to the same diagnoses attributable to non-pregnant women, the differential diagnosis for vaginal bleeding and/or abdominal pain in early pregnancy includes (among other less common or non-gynecologic conditions) ectopic pregnancy, retained products of conception, spontaneous abortion, and normal first trimester bleeding. High quality evidence for the usefulness of emergent pelvic exam in these patients is lacking as it primarily comes from small observational studies and studies that omit patients without confirmed intrauterine pregnancy. The available data indicate that in general, when emergency ultrasound is readily available, the pelvic examination adds marginal information to direct immediate management.14-16 This principle should be tempered by provider gestalt and cannot be blindly applied to patients who are hemodynamically unstable or where there is concern for trauma or foreign body.
Ectopic pregnancy, the life-threatening condition in which implantation occurs external to the uterine endometrium, occurs in an estimated 2-3% of pregnancies in the United States and must remain high on the differential when a woman presents with early pregnancy complaints.17,18 Unfortunately, as demonstrated by Crochet et al., there is no single physical exam finding that can reliably predict or rule out ectopic pregnancy.19 While the presence of cervical motion tenderness or an adnexal mass are among the most informative positive exam findings, they exhibit sensitivities of only 45% and 10%, with positive likelihood ratios of just 4.9 and 2.4, respectively. Likewise, the negative likelihood ratio of neither cervical motion tenderness (LR- 0.62), adnexal mass (LR- 0.94), adnexal tenderness (LR- 0.57), nor peritoneal findings (LR- 0.8) is adequate to rule out ectopic pregnancy on its own.19 This is supported by the work of Johnstone et al. in a small randomized controlled study wherein ED patients with vaginal bleeding in early pregnancy were randomized with respect to whether or not a pelvic exam was performed. Clinical diagnoses after history and physical (with or without pelvic exam) were compared to final diagnoses after ultrasound and quantitative hcg testing. Although the study was not adequately powered, it is notable that there was no difference in accuracy of the clinical diagnoses, regardless of whether a pelvic exam was performed—the groups were equally inaccurate with only 57% of final diagnoses matching provisional clinical diagnoses made without ultrasound.14 These findings are in agreement with a more recent randomized control trial by Linden et al., a study which unfortunately also fell short of its enrollment goal. In this trial, Linden randomized patients with abdominal pain or vaginal bleeding in the setting of ultrasound-confirmed early pregnancy to receive a pelvic exam or not as part of their ED evaluation. While it is may not be appropriate to draw practice-changing conclusions from an underpowered study, it is noteworthy that the data did not show significant difference in 30 day morbidity as a result of these patients forgoing ED pelvic examinations.15,20 If future research upholds the results of Linden’s study, it will likely be due to the fact that ultrasound provides greater information in both quantity and quality as compared to a pelvic exam performed by the most experienced practitioners. When transvaginal ultrasound shows an adnexal mass and no intrauterine pregnancy, the overall positive likelihood ratio for ectopic pregnancy is more than 100, with a negative likelihood ratio of 0.12.19 In patients with ultrasound-proven intrauterine pregnancy, the pelvic exam rarely has an impact on management or disposition of emergency department patients with early pregnancy, even in instances where there are unexpected findings on exam.21,22
Further undermining the purpose of the pelvic exam when ultrasound is available, ACEP guidelines suggest a pelvic ultrasound be performed in any pregnant patient presenting to an ED with vaginal bleeding or abdominal pain.23 Though not necessarily feasible in every hospital, bedside ultrasound by a proficient emergency physician is an accepted and accurate means for ruling out ectopic pregnancy, with a sensitivity of 99%.16
-Pelvic exam cannot reliably rule out ectopic pregnancy.
-When ultrasound is performed in patients with early pregnancy complaints, the pelvic exam is unlikely to change patient disposition.
When should you perform a pelvic exam?
Despite the pelvic exam’s lack of sensitivity and specificity for most pelvic pathology, there are emergencies where it should be performed. As with trauma to any other body part, pelvic trauma must be examined closely to identify the need for specific imaging and/or intervention.24 The unstable patient with vaginal bleeding also warrants a pelvic examination in order to identify a source and quantify blood loss. Additionally, depending on the level of expertise available in a given setting and an ED provider’s level of training, he or she might use this opportunity to provide definitive care to the actively hemorrhaging patient via manual aspiration of uterine contents (i.e. retained products of conception).25 Patients presenting with acute urinary retention with a history suggesting organ prolapse also warrant a pelvic exam in order to identify the prolapsed organ, which may be easily reduced.26
-Always perform a pelvic exam in cases of: trauma, hemodynamic instability, suspected organ prolapse causing acute urinary retention
Conclusions
After collecting your thoughts, you determine that regardless of what you may find on a pelvic exam, the patient is at a high enough risk that you are going to treat her empirically for gonorrhea, chlamydia, and trichomonas. You discuss this with the patient, and she strongly prefers not to have a pelvic exam. You provide her with instructions for self-swabbing and send the specimen to the lab. You treat her appropriately and discharge her home. A few days later you follow up on the results of her testing and see that it is positive for chlamydia. You give her a call, and she is already improving. She thanks you for your conscientiousness and promises to follow up with her doctor.
Key Points
-Pelvic exam findings are highly subjective.
-Accuracy of pelvic exam results may vary depending on patient body habitus.
-Absence of adnexal mass and/or lack of severe pelvic tenderness does not rule out ovarian torsion.
-Pelvic exam should not be relied on to rule out sexually transmitted infections.
–Patient-collected vaginal swabs or urine specimens are acceptable and accurate for STI testing.
-Pelvic exam cannot reliably rule out ectopic pregnancy.
–When ultrasound is performed in patients with early pregnancy complaints, the pelvic exam is unlikely to change patient disposition.
-Engage patients in shared decision making when you are questioning the utility of a pelvic exam.
