Are These Patients Pregnant?
For each of the following cases, the clinician may need to determine the probability that the patient is pregnant.
A 36-year-old woman telephones her primary care physician, complaining of symptoms consistent with uncomplicated sinusitis. Before treating her with an antibiotic, you ask her about the possibility of pregnancy; she states her last menstrual period was 3 weeks ago and she is not pregnant.
A sexually active 16-year-old girl requests birth control pills and asks during the pelvic examination, when her mother has stepped out of the room, if you can tell whether she is pregnant. Her last menstrual period was 8 weeks ago, her home pregnancy test result was negative, and findings on her pelvic examination were normal.
A 41-year-old woman presents with breast tenderness, and her last menstrual period was 6 weeks ago. She wants to know whether she is “going through the change.”
Why Is This an Important Question to Answer With A Clinical Examination?
Frequent laboratory analyses are performed in the outpatient clinic and emergency department to rule in or to rule out the possibility of pregnancy. Generally accepted clinical indicators of pregnancy include amenorrhea, morning sickness, tender or tingling breasts, and, after 8 weeks’ gestational age (defined as weeks since the last menstrual period), an enlarged uterus with a soft cervix. Standard textbooks of obstetrics do not indicate the value (ie, sensitivity and specificity) of these symptoms and signs as predictors of the diagnosis of early pregnancy.
In the outpatient clinical setting, there are many reasons to determine whether the patient is pregnant, including avoiding nonurgent radiographs; avoiding teratogenic drugs, such as anticonvulsants; initiating early prenatal care; reassuring the patient; and explaining the multiple nonspecific complaints easily confused with the early symptoms of pregnancy.
We are reviewing a common problem facing the primary care physician: When treating or evaluating a woman of childbearing years, what is the value of historical or physical examination features in determining the probability of early pregnancy? We will focus on the patient's medical history and physical examination findings that help the clinician rule in or rule out early pregnancy. We intend to answer the following questions: (1) What is the value of history and symptoms in determining the probability of early pregnancy? (2) How accurate are home pregnancy tests (often part of the patient's medical history) for determining early pregnancy? (3) What is the value of physical examination findings in determining the probability of early pregnancy?
Anatomic and Physiologic Origins of the Signs and Symptoms of Pregnancy During the First Trimester
Pregnancy is suspected whenever a woman of childbearing years who has had regular menstrual cycles notices abrupt cessation of her menses. However, cessation of menses is a difficult symptom to evaluate in patients with previously irregular bleeding patterns. Occasionally, women have unexplained cyclic bleeding during pregnancy, especially in the first few months, and thus lack the symptom of amenorrhea. About 8% of pregnant women have a small amount of bleeding on or before the 40th day, which is thought to be related to implantation.1
The term morning sickness refers to the tendency of many women (approximately 50%) to develop nausea, often with vomiting, between 6 and 12 weeks’ gestational age.1 Usually the nausea is worse when the pregnant woman awakens in the morning, whereas it tends to diminish as the day progresses.
Shortly after missing her first period, the pregnant woman may notice a heavy sensation in her breasts, accompanied by tingling and soreness. These symptoms relate to hormone stimulation of the ducts and alveoli of the breast parenchyma, but may occur in identical form just before a menstrual period. As early as 6 weeks’ gestational age, there may be noticeable enlargement of the breasts, with engorgement of the superficial veins in the breasts.2 During the first trimester, the nipples darken and become more sensitive. The areolar areas darken and become puffy. These symptoms and signs are thought to be of more value in primigravida because in multigravida women, areolar and nipple changes often remain from previous pregnancies.3
A few weeks after implantation (6 weeks’ gestational age), distinct enlargement of the uterus may be felt on bimanual palpation. In early pregnancy, the uterus becomes softened and changes from a pear-shaped configuration to a globular contour.1 The congestive hyperemia of the pelvis in early pregnancy is manifested by a softening of the vagina and cervix, as well as a change in the color of these tissues. A significant increase in uterine artery pulsatile activity may occur as blood flow to the pregnant uterus increases.4 In early pregnancy, the enlarging uterus exerts pressure on the bladder. Some patients note an increase in urinary frequency and nocturia during the first trimester.
How to Elicit These Symptoms and Signs
Although patients may give a simple description such as “I may be pregnant,” the examiner should seek a more complete medical history. Histories that indicate an increased likelihood of pregnancy include amenorrhea, morning sickness, breast symptoms (swelling, tingling, or tenderness), sexual activity, not using or inconsistent use of contraception, patient suspects she is pregnant, and a positive home pregnancy test result. Specific questions to ask include the following: (1) When was your last menstrual period, and was it normal? (2) Do you use any form of contraception? (3) Do you have any symptoms of pregnancy? (4) Is there a chance you are pregnant?
Frequently, the patient may report, “My home pregnancy test was positive, and I want to know whether I am pregnant.” Important questions regarding this type of history would be these: (1) How many days or weeks after your last menstrual period did you perform the test? (2) Did you feel comfortable performing the test? (3) Did the instructions seem complicated to you? (4) What kind of home pregnancy test did you use? (5) Did you repeat the test and get a similar result?
To diagnose pregnancy, the clinician might examine the patient's breasts, as well as the vaginal wall, cervix, and uterus, by bimanual examination. The breasts may become engorged and enlarged, with darkening of the areolar area. The venous pattern over the breasts becomes increasingly visible as pregnancy progresses.5
Vaginal examination can be performed to elicit the Chadwick sign associated with early pregnancy. As early as 8 to 12 weeks’ gestational age, the mucous membranes of the vulva, vagina, and cervix become congested and take on a bluish-violet hue (Chadwick sign).1 This hue is especially well defined in the anterior vaginal wall but is also present to some extent throughout the vagina and on the cervix. The Chadwick sign is rarely seen before 7 weeks’ gestational age.6
On bimanual examination, softening of the cervix (Goodell sign) may be detected by 8 weeks’ gestational age.7 The cervix of a nonpregnant woman is fibrous and normally feels like the tip of the nose. By contrast, the progressive edema that develops during pregnancy softens the consistency of the cervix tip to approximate that of the lips (Goodell sign).
Examination of the uterus on bimanual examination can be performed to detect changes in uterine consistency and size. A palpable softening of the lowermost portion of the corpus occurs at about 6 weeks’ gestational age (Hegar sign).7 To elicit this sign, when the uterus is anteverted, the examiner places two fingers in the anterior vaginal fornix (or the posterior fornix in the presence of a retroverted uterus) and then compresses behind the fundus at the lower uterine segment with the other hand, using suprapubic pressure (Figure 42-1). In this way, a distinct area of uterine softening is observed between 2 firmer structures: the fundus above and the cervix below.5 Occasionally, the softening at the isthmus is so marked that the cervix and the body of the uterus seem to be separate organs.3
Examination Eliciting the Hegar Sign
The Hegar sign is a softening of the lower uterine segment that can be appreciated during a bimanual examination.
Another early sign of pregnancy is the uterine artery pulsation that can be palpated on a bimanual examination.4 During a bimanual examination, the second and third digits of the examining hand can be placed in the lateral vaginal fornix, and the presence of uterine artery pulsations can often be palpated with minimal pressure on the parametrium.4
A few weeks after the embryo has become implanted, a distinct enlargement of the uterus may be felt on bimanual examination. The uterus remains confined in the pelvis until 12 weeks’ gestational age, when the fundus becomes palpable above the pubic symphysis (Figure 42-2).
Uterine Height at Different Gestational Weeks
The height of the fundus at comparable gestational dates varies greatly among patients. Those shown are the most common. A convenient rule of thumb is that, at 20 weeks of gestation, the fundus is usually at or slightly above the umbilicus.
The identification of the fetal heart rate distinct from the maternal heart rate establishes a diagnosis of pregnancy. Transvaginal ultrasonography can detect fetal heart activity as early as 5 weeks’ gestational age, and transabdominal ultrasonography can detect this activity as early as 6 weeks’ gestational age. Instruments that use the Doppler effect can detect fetal cardiac activity at 10 to 12 weeks’ gestational age. The fetal heart can usually be auscultated with a fetoscope by 20 weeks’ gestational age.
Reference Standard for Diagnosing Early Pregnancy
In this review, the detection of the β subunit of human chorionic gonadotropin (HCG) in urine or serum is the routine reference standard (or gold standard) for diagnosing early pregnancy. The diagnostic reliability of both the serum and urine HCG tests is comparable. The sensitivity and specificity for the diagnosis of pregnancy for both tests are between 97% and 100% when performed in the laboratory.8 In this review, we also report the results of studies conducted before the development of the HCG test. These earlier studies used delivery as the reference standard.
We searched the MEDLINE database for English-language articles concerning the diagnosis of pregnancy that were published between 1966 and 1996. The key words used were “pregnancy,” “diagnosis,” and “pregnancy tests.” Additional articles listed in the bibliographies of standard obstetric texts and references cited in articles included in our study were also included among the articles considered.
Articles were systematically reviewed by authors and given a grade of A, B, or C according to the study design and level of evidence (see Table 1-7 for a summary of Evidence Grades and levels).9 Articles were excluded if the results of the symptom or sign being investigated were not compared with the gold standard or the results could not be classified into a contingency table (attempts were made to reach authors of potential articles to obtain additional information needed to create contingency tables).
Through the MEDLINE, textbook reference, and bibliography searches, we initially identified 55 articles, 40 of which were rejected because the test was not compared with the gold standard (urine or serum HCG test) or a pregnancy outcome. The remaining 15 articles were then analyzed by us, and 6 more were excluded because the reported data were not sufficient to permit construction of contingency tables. Therefore, the results of 9 studies form the basis for this review.
We used data from contingency tables to calculate sensitivity and specificity. Likelihood ratios were also calculated to characterize the behavior of the diagnostic tests. The positive likelihood ratio (LR+) is defined as sensitivity/(1 – specificity) and expresses the change in odds favoring a disease, given a positive test result (LR+ values are ≥ 1), whereas the negative likelihood ratio (LR–) is defined as (1 – sensitivity)/specificity and expresses the change in odds favoring disease, given a negative test result (LR– values are 0 to 1).10 Data were sufficiently similar in design to assess for statistical similarity. The data were pooled when the Breslow-Day test for homogeneity was not significant (P > .05).11
Accuracy of History and Symptoms for Pregnancy Diagnosis
Several studies have been performed to evaluate the value of patient history in ruling in or ruling out early pregnancy compared with the gold standard HCG test (Tables 42-1, 42-2, 42-3, and 42-4). Among 208 consecutive patients for whom a qualitative serum HCG determination is ordered, emergency department physicians recorded the date of the patient's last menstrual period, whether her menstrual period was on time, if birth control had been used, and whether the patient suspected she was pregnant.12 The main indication for ordering a pregnancy test in this study was abdominal pain (138 patients). Sixty-eight women (33%) were pregnant. Three historical variables were statistically less likely to be associated with pregnancy: a last menstrual period that was on time, the patient thinking that she was not pregnant, and the patient stating that there was no chance that she could be pregnant (P < .001). Combinations of historical criteria were unsuccessful at ruling out pregnancy; there was still a 10% chance of pregnancy's being overlooked using any combination of these historical variables.
Table 42-1Does a Delayed Menstrual Period Predict Pregnancy?a |Favorite Table|Download (.pdf) Table 42-1 Does a Delayed Menstrual Period Predict Pregnancy?a
|Study ||Evidence Gradeb ||Characteristics ||Pregnant ||LR (95% CI) |
|Yes ||No |
|Robinson and Barber15 ||A ||Delayed menses ||618 ||248 ||1.6 (1.4-1.7) |
|Menses on time ||361 ||365 ||0.62 (0.56-0.69) |
|Ramoska et al12 ||A ||Delayed menses ||58 ||58 ||2.1 (1.6-2.6) |
|Menses on time ||10 ||82 ||0.25 (0.14-0.45) |
|Stengel et al13,c ||B ||Delayed menses ||3 ||43 ||1.0 (0.38-2.9) |
|Menses on time ||9 ||136 ||0.99 (0.70-1.4) |
|Zabin et al16 ||A ||Delayed menses ||703 ||1078 ||1.1 (1.0-2.9) |
|Menses on time ||331 ||707 ||0.81 (0.68-0.76) |
Table 42-2Probability of Pregnancy if Patient Reports Symptoms of Pregnancya |Favorite Table|Download (.pdf) Table 42-2 Probability of Pregnancy if Patient Reports Symptoms of Pregnancya
|Study ||Evidence Gradeb || ||Pregnant ||LR (95% CI) |
|Yes ||No |
|Robinson and Barber15 ||A ||Morning sickness ||380 ||88 ||2.7 (2.2-3.3) |
|No morning sickness ||599 ||525 ||0.71 (0.67-0.76) |
|Bachman14 ||A ||Any pregnancy symptoms ||59 ||34 ||2.4 (1.7-3.4) |
| || ||No pregnancy symptoms ||59 ||131 ||0.63 (0.52-0.77) |
Table 42-3Probability of Pregnancy if Patient Reports Not Using Birth Control |Favorite Table|Download (.pdf) Table 42-3 Probability of Pregnancy if Patient Reports Not Using Birth Control
|Study ||Evidence Gradea || ||Pregnant ||LR (95% CI) |
|Yes ||No |
|Ramoska et al12 ||A ||No birth control ||61 ||96 ||1.3 (1.1-1.5) |
|Birth control ||7 ||44 ||0.33 (0.16-0.69) |
|Stengel et al13,b ||B ||No birth control ||9 ||88 ||1.5 (1.1-2.2) |
|Birth control ||3 ||91 ||0.49 (0.18-1.3) |
|Pooledc || ||No birth control ||70 ||184 ||1.5 (1.3-1.7) |
| || ||Birth control ||10 ||135 ||0.29 (0.16-0.53) |
Table 42-4Probability of Pregnancy if Patient Thinks There Is a Chance She Is Pregnant |Favorite Table|Download (.pdf) Table 42-4 Probability of Pregnancy if Patient Thinks There Is a Chance She Is Pregnant
|Study ||Evidence Gradea ||Patient Thinks She Is ||Pregnant ||LR (95% CI) |
|Yes ||No |
|Bachman14 ||A ||Pregnant ||109 ||95 ||1.6 (1.4-1.8) |
|Not pregnant ||9 ||70 ||0.18 (0.09-0.34) |
|Ramoska et al12 ||A ||Pregnant ||58 ||63 ||1.9 (1.5-2.3) |
|Not pregnant ||10 ||77 ||0.27 (0.15-0.48) |
|Stengel et al13b ||B ||Pregnant ||11 ||52 ||3.2 (2.4-4.2) |
|Not pregnant ||1 ||127 ||0.12 (0.02-0.77) |
|Zabin et al16 ||A ||Pregnant ||789 ||640 ||2.1 (2.0-2.3) |
|Not pregnant ||254 ||1148 ||0.38 (0.34-0.42) |
|Pooled resultsc || ||Pregnant ||967 ||850 ||2.1 (2.0-2.2) |
| || ||Not pregnant ||270 ||1422 ||0.35 (0.31-0.39) |
Women may not associate symptoms with early pregnancy. Investigators measured the effectiveness of a standardized patient history questionnaire in detecting unrecognized pregnancies.13 Consecutive fertile women (n = 191) presenting to the emergency department for any reason completed a menstrual and sexual history questionnaire and had a pregnancy test. This study reports a 6.3% prevalence of unrecognized pregnancy, defined as a “pregnancy not definitely known to exist” when the patient presented to the emergency department.13 Among those with abdominal pain or pelvic complaints (70 patients), the prevalence of unrecognized pregnancy was found to be 13%. Historical factors were analyzed for correlation with positive pregnancy test results. Two factors were found to be statistically significant correlates: the patient thought there was a chance she could be pregnant and an abnormal last menstrual period (P < .001). One factor, the delayed menstrual period, was not found to be significant (LR+, 1.0). Among the historical factors analyzed, “Is there any chance that you could be pregnant now?” was the most sensitive for pregnancy (92%), with a specificity of 71% (David Seaberg, MD, University of Pittsburgh, Pennsylvania, unpublished data, June 1995).
Unlike women who do not associate symptoms with early pregnancy, others self-diagnose pregnancy and request medical confirmation. Women (n = 283) with late menstrual periods who requested evaluation in a health center completed a structured contraception and sexual history questionnaire that included questions on whether the woman believed she was pregnant and whether subjective symptoms of pregnancy were present.14 The patient sealed her answers to the questionnaire in an envelope before the results of the pregnancy tests were available. One hundred eighteen women (42%) were pregnant. Women were better at ruling out pregnancy (sensitivity, 92%) than ruling in pregnancy (specificity, 42%).
In another study,15 general practitioners performed a study to determine the value of pregnancy symptoms (presence or absence of amenorrhea and morning sickness) in determining the probability of pregnancy. Information was collected prospectively about women who consulted their general practitioner for a diagnosis of pregnancy; the gold standard was a positive pregnancy test result. General practitioners throughout Scotland (n = 155) participated in the study, which was restricted to women between the ages of 16 and 45 years. Of the 1592 women enrolled, 979 (62%) were pregnant. The symptom of amenorrhea was 63% sensitive and 60% specific for pregnancy. Morning sickness as a symptom of pregnancy had a sensitivity of 39% and a specificity of 86%. This study did not ask the participants whether they thought they were pregnant.
In 1996, Zabin et al16 performed a similar study in a population of adolescents (younger than 17 years) to determine historical predictors of pregnancy. They performed a cross-sectional study of 2926 adolescents who presented to 52 clinics in the United States and requested a pregnancy test. The girls were asked to complete an anonymous questionnaire (98% response rate) while they waited for the results of their pregnancy test. Thirty-six percent of adolescents in this study were pregnant. A late menstrual period was the most frequent reason (63%) for the visit (for pregnancy: sensitivity, 68%; specificity, 40%).
Although a delayed menstrual period yields statistically significant results for predicting pregnancy, with an LR+ of 1.1 to 2.1 (Table 42-1), the results are inconsistent and, therefore, not a reliable symptom of pregnancy. Typical early symptoms of pregnancy provide more consistent results across studies and serve to increase slightly the likelihood of pregnancy (LR+, 2.4) (Table 42-2). Unfortunately, the absence of early symptoms of pregnancy, such as morning sickness, does not rule out pregnancy (LR–, 0.71). Likewise, the patient's use of birth control decreases the likelihood of pregnancy (LR–, 0.29), but not enough to efficiently rule it out (Table 42-3). Even the patient's suspicion of pregnancy statistically alters the likelihood of pregnancy, but not enough to be reliable (Table 42-4).
Accuracy of Home Pregnancy Tests
It has been reported that one-third of women who think they may be pregnant have used a home pregnancy test.17 A recent study of teenagers requesting pregnancy tests in health departments revealed that 28% of adolescents had used an in-home pregnancy test before their visit.16 In-home pregnancy test kits became available in 1976 and used the hemagglutination-inhibition method of detecting HCG. Currently, most test kits use monoclonal HCG antibodies, which can produce test results that can be read as a color change. The accuracy of these tests is claimed to be 97% to 99% by the manufacturers.18 Studies have shown that accuracy depends on several factors, such as whether the woman read the instructions carefully and the number of days beyond the missed menstrual period.19
In 1986, Doshi20 published a study measuring the accuracy of 3 in-home tests for early pregnancy. The author studied 109 women of childbearing age whose menses were late by at least 6 days, but not more than 20 days. Volunteers for the study were obtained from 3 sites; the majority were white and educated. Participants brought to the study site their first morning urine, which was then divided in half. One portion of the sample was returned to the participant to use in performing a pregnancy test at home. Using 1 of 3 study kits (Answer [Carter Products; Carter-Wallace, Inc, New York, New York]; Daisy 2 [Boehringer-Mannheim Corp, Ingelheim, Germany]; and e.p.t. [Warner-Lambert Co, Morris Plains, New Jersey]), the participants were instructed to follow the package directions in performing the test, call the site with results, and complete and return the data collection survey to the investigator. The investigator performed an identical test using the other portion of the urine sample. Despite manufacturer claims of 97% overall accuracy for the test kits used, the investigator found an accuracy of 77%. The participants had a sensitivity of 80% and specificity of 68% for detecting early pregnancy with the home pregnancy tests (LR+, 2.5; LR–, 0.29), with similar diagnostic efficiency observed for all 3 kits. These results concerned Doshi20 because of missed opportunities for early prenatal care and the postponement of discontinuing teratogenic substances.
In 1993, investigators from France published an extensive analysis of the reliability and feasibility of home pregnancy tests.21 They looked at 27 different test kits (manufacturers were not identified) and selected 11 kits for the study, which were found to have a 100% sensitivity and specificity under ideal laboratory conditions. Laywomen volunteers (aged 14-49 years; n = 638) were asked to test a home-use test kit for pregnancy using a coded urine specimen. They also were asked to complete a questionnaire after they performed the test. The results of the diagnostic study showed that 5 of the 11 kits had 100% specificity; the others had specificity values between 77% and 94%. Two kits had a high diagnostic sensitivity (>90%), and 2 kits were found to have a low diagnostic sensitivity (<10%). Whereas 90% of the participants claimed that the test was easy to perform, of the 478 positive (result positive for pregnancy) urine samples distributed, 230 were falsely interpreted as negative (sensitivity, 48%). The authors concluded that the main reason for the poor performance was difficulty in interpreting the instructions rather than the socioeconomic situation of the participants.
Accuracy of the Physical Examination
Only a few studies have analyzed at the accuracy of the physical examination for pregnancy. Unfortunately, no studies have examined interobserver or intraobserver reliability. In 1887, Chadwick6 published a study of 337 women evaluated weekly (until delivery for those women who were pregnant) to assess the presence of the Chadwick sign. He described the coloration of the vaginal wall as no color or doubtful color, suggestive color, characteristic color, and general deep color. He classified any vaginal wall with characteristic or general deep color to be “diagnostic.” With his criteria, the sensitivity of this physical sign is 51% and the specificity is 98%. No validation studies could be found.
Robinson and Barber15 performed a study in 1977 to determine the value and reliability of the physical examination for pregnancy compared with a pregnancy test. They examined the vagina for signs of pregnancy, palpated the fundus, and assessed breast changes on physical examination. The most common feature observed was breast signs (42%), with a sensitivity of 56% and a specificity of 79%. Thirteen percent of women were observed to have “signs” on vaginal examination (signs not specified, but presumably some combination of the Goodell, Hegar, and Chadwick signs) consistent with pregnancy, with a sensitivity of 18% and a specificity of 94%. Last, 6% of women were observed to have a palpable fundus at presentation for a pregnancy test (sensitivity, 9%; specificity, 97%).
Recently, a study was performed to determine whether palpable uterine artery pulsation is a reliable clinical indicator of early pregnancy.4 The authors conducted the study in 2 phases. During the first phase, one of the authors examined 299 women who were less than 6 weeks from their last menstrual period for palpable uterine artery pulsation; this examination was conducted after a medical history had been obtained, and thus the examiner was not blind to the clinical situation. During the second phase, one of the authors examined 155 women for palpable uterine artery pulsation but performed only the bimanual examination and was blind to all other historical and physical examination data. With data from the second phase only, palpation of uterine artery pulsations may be a valuable tool in diagnosing early pregnancy (sensitivity, 76%; specificity, 93%). According to the results of this study, physicians were encouraged to add uterine artery pulsation to their clinical examination in diagnosing early pregnancy.
Despite descriptive articles dating back to the 1880s, no studies could be identified that measured the value of the Goodell or Hegar signs. In 1908, McDonald7 reported the prevalence of early pregnancy findings in 100 women known to be pregnant. He followed up women with weekly pelvic examinations during their first trimester. In this descriptive study, pregnant women were found to have the following: Hegar sign, 94%; Goodell sign, 66%; and Chadwick sign, 61%. This study is included for historical interest. Knowing the pregnancy status of patients creates expectation bias that probably overstates the value and prevalence of these signs.
As summarized in Table 42-5, several physical findings significantly increase the likelihood of pregnancy. The most useful findings on physical examination for making the diagnosis of early pregnancy appear to be Chadwick sign (LR+, 29) and palpable uterine artery pulsation (LR+, 11), although validation studies are needed because these 2 studies had comparatively lower methodologic quality scores. Unfortunately, if any of these signs are absent, this does not rule out pregnancy.
Table 42-5Probability of Pregnancy if Physician Examination Findings Present |Favorite Table|Download (.pdf) Table 42-5 Probability of Pregnancy if Physician Examination Findings Present
|Study ||Evidence Gradea ||Characteristic ||Pregnant ||LR (95% CI) |
|Yes ||No |
|Chadwick6 ||C || |
|Present ||144 ||1 ||29 (4.1-200) |
|Absent ||137 ||55 ||0.50 (0.44-0.56) |
|Robinson and Barber15 ||A ||Breast signs |
|Presentb ||549 ||127 ||2.7 (2.3-3.2) |
|Absent ||430 ||486 ||0.55 (0.50-0.60) |
|Robinson and Barber15 ||A ||Vaginal examination signs |
|Presentc ||172 ||34 ||3.2 (2.2-4.5) |
|Absent ||807 ||579 ||0.87 (0.84-0.90) |
|Robinson and Barber15 ||A ||Palpable fundus |
|Present ||84 ||19 ||2.8 (1.7-4.5) |
|Absent ||895 ||594 ||0.94 (0.92-0.97) |
|Meeks et al4 ||B ||Uterine artery pulsation |
|Present ||19 ||9 ||11 (5.6-21) |
|Absent ||6 ||121 ||0.26 (0.13-0.52) |
Clearly, to establish a diagnosis of early pregnancy, a clinician should order a urine or serum HCG test. However, there may be circumstances in which it would be useful for patients or physicians to know the value of pregnancy symptoms, home pregnancy test results, and physical examination findings for the diagnosis of pregnancy.
We can predict the likelihood of pregnancy for the patients in the clinical scenarios. For case 1, the woman with sinusitis has a prior probability of pregnancy of about 5%. Because she reports that her menses was on time (LR–, 0.62) and states that she is not pregnant (LR–, 0.35), the calculated probability of pregnancy might be from 1.7% to 3.1% for this patient. We would not order a pregnancy test for case 1. For case 2, the sexually active teenager, we can also calculate a probability that she might be pregnant. Zabin et al16 reported a pregnancy rate of 36% among teenagers presenting for a pregnancy test in their study. If we assume her prior probability of pregnancy is 36% and know her menses is late (LR+, 1.1), her home pregnancy test result was negative (LR–, 0.29), and her pelvic examination findings were normal (LR–, 0.87), her probability of pregnancy ranges from 10% to 41%, and we would recommend ordering a pregnancy test for this case. For case 3, the 41-year-old woman with a late menses and breast tenderness, the prior probability of pregnancy might be low (approximately 2%) because of decreased fecundity secondary to her age. If we consider her late menses (LR+, 1.6) and her breast tenderness (LR+, 2.4), her probability of pregnancy has increased approximately 2-fold to a range of 3.1% to 4.9%, and we would order a pregnancy test.
Patients may call their clinician asking for advice regarding a late period or symptoms of pregnancy. They may want to know whether they should perform a home pregnancy test, or they may request assistance in interpreting the test results. Evidence suggests that some historical features, when absent, are fair but not reliable for ruling out pregnancy. When diagnosing pregnancy, the patient or clinician should not rely on symptoms and signs of pregnancy or a home pregnancy test; a laboratory test should be requested.
Author Affiliations at the Time of the Original Publication
Departments of Internal Medicine, Obstetrics and Gynecology, Duke University Medical Center, and Comprehensive Health Center and the Center for Health Services Research in Primary Care, Veterans Affairs Medical Center, Durham, North Carolina (Dr Bastian); and the Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, North Carolina (Dr Piscitelli).
We thank David Seaberg, MD, University of Pittsburgh, Pennsylvania, for providing the unpublished data necessary to calculate sensitivity and specificity for the study by Stengel et al.13
WN. Danforth's Obstetrics and Gynecology. 7th ed. Philadelphia, PA: JB Lippincott; 1994.
BM. Pregnancy diagnosis. BMJ.
NF. Williams Obstetrics. 19th ed. East Norwalk, CT: Appleton & Lange; 1993.
GW. Palpable uterine artery pulsation as a clinical indicator of early pregnancy. J Reprod Med.
JE. A Guide to Physical Examination and History Taking. 6th ed. Philadelphia, PA: JB Lippincott; 1995.
JR. Value of the bluish coloration of the vaginal entrance as a sign
of pregnancy. Trans Am Gynecol Soc (1886).
E. The diagnosis of early pregnancy, with report of one hundred cases and special reference to the sign
of flexibility of the isthmus of the uterus. Am J Obstet Dis Women Child.
DB. Likelihood ratios with confidence: sample size estimation for diagnostic test studies. J Clin Epidemiol.
GG. Categorical Data Analysis Using the SAS System. Cary, NC: SAS Institute Inc; 1995.
GA. Myth or fact: can women self-diagnose pregnancy? J Med Soc N J.
JH. Early diagnosis of pregnancy in general practice. J R Coll Gen Pract.
RA. How frequently are home pregnancy tests used? results from the 1988 National Maternal and Infant Health Survey
SM. The pharmacist and home-use pregnancy tests. Am Pharm. 1992;32(1):57–60.
BC. Evaluation of home pregnancy test kits. Biomed Instrum Technol.
ML. Accuracy of consumer-performed in-home tests for early pregnancy detection. Am J Public Health.
et al.. Reliability
and feasibility of pregnancy home-use tests: laboratory validation and diagnostic evaluation by 638 volunteers. Clin Chem.