Tintinalli's Emergency Medicine > Section 9 ...

  • Doc File 165.00KByte



Tintinalli's Emergency Medicine > Section 9: Gastrointestinal Emergencies > Chapter 72. Acute Abdominal Pain >

Scope and Definitions

Acute abdominal pain is commonly defined as pain of less than 1 week's duration.1 This chapter discusses nontraumatic acute abdominal pain in postpubescent males and females. Abdominal pain in women in the third trimester of pregnancy or the first month postpartum is discussed in Chaps. 102, 105.

Epidemiology

Data from the U.S. National Center for Health Statistics indicate that stomach and abdominal pain was the principal reason offered by patients for visiting EDs in 2000 (annual incidence approximately 63/1000 ED visits).2 Admission rates for abdominal pain vary markedly, ranging from 18 to 42 percent, with rates as high as 63 percent reported in patients over 65 years of age.

Pathophysiology

Abdominal pain is traditionally divided into three categories: visceral, parietal, and referred. In general, visceral (autonomic) and parietal (somatic) are considered the two basic causes of abdominal pain. Referred pain can be considered separately as a cortical misperception of either visceral or parietal afferent stimuli. Although each type of pain is thought to have a different neuropathophysiology, the categories are not entirely discrete. For example, visceral pain often blends with parietal pain as a pathologic process evolves. Still, these distinctions are clinically useful ways of thinking about abdominal pain.

Visceral Pain

Visceral abdominal pain is usually caused by stretching of fibers innervating the walls or capsules of hollow or solid organs, respectively. Less commonly, it is caused by early ischemia or inflammation. Severity ranges from a steady ache or vague discomfort to excruciating or colicky pain. Because the visceral afferents follow a segmental distribution, visceral pain can be localized by the sensory cortex to an approximate spinal cord level determined by the embryologic origin of the organ involved. For example, foregut organs (stomach, duodenum, biliary tract) produce pain in the epigastric region; midgut organs (most small bowel, appendix, cecum) cause periumbilical pain; and hindgut organs (most of colon, including sigmoid) as well as the intraperitoneal portions of the genitourinary tract cause pain initially in the suprapubic or hypogastric area.

Because intraperitoneal organs are bilaterally innervated, stimuli are sent to both sides of the spinal cord, causing intraperitoneal visceral pain to be felt in the midline, independent of its right- or left-sided anatomic origin. For example, stimuli from visceral fibers in the wall of the appendix enter the spinal cord at about T10. When obstruction causes appendiceal distention in early appendicitis, pain is initially perceived as midline periumbilical area, corresponding roughly to the location of the T10 cutaneous dermatome.

Parietal Pain

Parietal or somatic abdominal pain is caused by irritation of fibers that innervate the parietal peritoneum, usually the portion covering the anterior abdominal wall. Because parietal afferent signals are sent from a specific area of peritoneum, parietal pain—in contrast to visceral pain—can be localized to the dermatome superficial to the site of the painful stimulus. As the underlying disease process evolves, the symptoms of visceral pain give way to the signs of parietal pain, causing tenderness and guarding. As localized peritonitis develops further, rigidity and rebound appear.

Referred Pain

Referred pain is felt at a location distant from the diseased organ. Similar to visceral pain, and in contrast to parietal pain, referred pain produces symptoms, not signs. Unlike visceral pain, referred pain is usually ipsilateral to the involved organ and is felt in the midline only if the pathologic process is also located in the midline. This is because referred pain, in contrast to visceral pain, is not mediated by fibers providing bilateral innervation to the cord. Similar to visceral pain, referred pain patterns are based upon developmental embryology. For example, the ureter and the testes were once anatomically contiguous, and therefore share the same segmental innervation, supplying afferent fibers to the lower thoracic and upper lumbar segments of the spinal cord. Thus acute ureteral obstruction is often associated with ipsilateral testicular pain. Other sites of referred pain reflect similar dermatomal sharing, providing explanations for otherwise puzzling associations, e.g., supra- or subdiaphragmatic irritation and ipsilateral supraclavicular or shoulder pain; gynecologic pathology and back or proximal lower extremity pain; biliary tract disease and right infrascapular pain; and myocardial ischemia and midepigastric, neck, jaw, or upper extremity pain.

Clinical Features

Conceptual Framework

Classification

The classification scheme divides abdominal pain into two main categories: Intraabdominal (i.e., arising from within the abdominal cavity or retroperitoneum) and extraabdominal. Intraabdominal causes are divided by organ system into the "3-G's": GI (gastrointestinal), GU (genitourinary), and GYN (gynecologic), plus a fourth, less common but often catastrophic group of VASCULAR emergencies. Each of these four is further subdivided into specific diagnoses within that organ system. Pain of extraabdominal origin, which is substantially less common, is similarly divided into four broad etiologic categories of cardiopulmonary, abdominal wall, toxic-metabolic, and neurogenic. A systematic evaluation is necessary in the assessment of acute abdominal pain.

Finally, nonspecific abdominal pain (NSAP), which is the most common cause of abdominal pain among ED patients, is listed as a third category. Nonspecific abdominal pain stands alone since it is not known to what extent it may represent an underlying intra- vs. extraabdominal problem.

Abdominal Topography

By combining the four-quadrant approach traditionally used by U.S. physicians with selected aspects of a strategy widely employed throughout Europe and Asia, a simple model of abdominal topography can be developed. In addition to the four standard quadrants (RUQ, RLQ, LUQ, LLQ), this model includes four areas of the abdomen that are not discrete, but rather constitute combinations of all or part of two or more quadrants: (1) upper half of abdomen (UHA), which includes an area of pain as small as the mid-epigastrium, or as large as the RUQ + LUQ combined; (2) lower half of abdomen (LHA), which similarly includes an area of pain as small as the midhypogastrium or as large as the RLQ + LLQ combined; (3) central (CTL), which includes an area of pain composed of the centermost "quarters" of all four discrete quadrants, such that carving out these areas from each quadrant defines a periumbilical or central quadrant; and (4) generalized (GEN), which includes poorly localized pain encompassing much, perhaps most, of the abdomen, including at least some portion of all four discrete quadrants.

This topographic configuration incorporates both the early (visceral, poorly localized) and late (parietal, better localized) pain of an evolving intraabdominal pathological process, as well as the more generalized pain associated with toxic-metabolic derangements.

However, the association between the location of overlying pain or tenderness and underlying disease is so variable that about one case of abdominal pain in every three that comes to operation presents in a fashion that clinicians retrospectively regard as atypical. Failure to appreciate this may represent the largest single reason that error in the clinical diagnosis of abdominal pain is so common.

Historical Features

Historical data can be conveniently divided into attributes of pain, associated symptoms, and past history.

Pain Attributes

The principal characteristics of abdominal pain include location, quality, severity, onset, duration, aggravating and alleviating factors, and change in any of these features over time.

Associated Symptoms

These can be subdivided into one of the four main organ systems associated with intraabdominal pain.

Gastrointestinal Symptoms

Anorexia, nausea, and vomiting (unless bloody) are among the least helpful symptoms in altering the conditional probability that a patient does or does not have a GI cause of abdominal pain. For example, vomiting has been reported in over 40 percent of patients with salpingitis, and in over 60 percent of patients with renal colic. Lower GI symptoms such as nonbloody diarrhea or constipation are similarly too insensitive and nonspecific to significantly alter the probability of a GI cause of abdominal pain.

Genitourinary Symptoms

The hallmark of abdominal pain of GU origin is the concomitant development of some, often subtle, alteration in micturition, e.g., dysuria, frequency, urgency, hematuria, incomplete emptying, or incontinence (usually overflow). Non-GU pathology may develop in organs contiguous to the GU system, giving the appearance of an intrinsic GU problem. For example, an inflamed appendix lying across the bladder may cause urinary frequency.

Gynecologic Symptoms

Distinguishing GI from GYN causes of acute abdominal pain is one of the most challenging clinical dilemmas in emergency practice. A thorough gynecologic history is indicated, including menses, mode of contraception, fertility, sexual activity, sexually transmitted diseases, vaginal discharge, recent dyspareunia, and a past gynecologic history, to include pregnancies, deliveries, abortions, ectopics, cysts, fibroids, pelvic inflammatory disease, and laparoscopy.

Vascular Symptoms

History of MI, other ischemic heart disease or cardiomyopathy, atrial fibrillation, anticoagulation, congestive failure, peripheral vascular disease, or a family history of aortic aneurysm are all pertinent historical features in older patients.

Past Medical History

This includes a history of recent and current medications (including nonsteroidal anti-inflammatory drugs and antibiotics), past hospitalizations, in- or outpatient surgeries, diabetes, other chronic diseases (including HIV status and risk factors), and any history of recent trauma. A social history that includes habits (tobacco, alcohol, and other drug use), occupation, possible toxic exposures, and living circumstances (homeless, dwelling heated, running water, living alone, other family members ill with similar symptoms) provides important background and context in which to place the presenting complaint of acute abdominal pain.

Physical Examination

General

The patient's general appearance, including facial expression, diaphoresis, pallor, and degree of agitation provides information about the severity of pain. Although this is critically important in determining the need for analgesia, intensity of abdominal pain may bear no relationship to the severity of illness. For example, the pain of early mesenteric ischemia may be a vague discomfort, in contrast to the excruciating pain of ureteral colic. Nevertheless, uncomplicated kidney stones have no short-term mortality, while the majority of patients with ischemic small bowel go on to die.

Patients with colicky pain, which is characteristically visceral due to distention of a hollow organ, are often unable to lie still, while those with peritonitis prefer to remain immobile.

Vital Signs

A reliable means of obtaining a core temperature is important, although absence of fever, especially in the elderly, has no predictive value. Careful counting of rate and observation of depth of respirations for 15 s is often overlooked. However, it can provide crucial information about tachypnea or hyperpnea, which may be subtle. Pulse and blood pressure should include orthostatic changes if, after obtaining the history, there is any reason to suspect intravascular volume contraction. A pulse increase of thirty points lying to standing at 1 minute (or the development of symptoms of presyncope) has been shown to be highly specific for the loss of a liter of blood or its equivalent (roughly 3 L of NS). Changes in blood pressure have not been shown to be discriminatory, probably because they are late findings representing failure of a reflex tachycardia to maintain cardiac output. The tilt-test threshold of thirty points of pulse change may not be applicable to patients on medications such as beta-blockers, diabetics (who may have an autonomic neuropathy), and among the elderly, due to the effects of aging on the cardiac conducting system.

Abdomen

Inspection

The abdomen should be inspected for distention (with air or fluid), scars, and masses.

Auscultation

Contrary to conventional teaching, absent or diminished bowel sounds provide little clinically useful information. Patients with operative confirmation of peritonitis due to perforation of peptic ulcer have been noted to have normal or increased bowel sounds preoperatively. Hyperactive or obstructive bowel sounds, although of limited value, are somewhat more helpful for the diagnosis of small bowel obstruction (SBO). However, many with SBO can also have absent or diminished bowel sounds. It appears, therefore, that only hyperactive or obstructive bowel sounds have clinical utility, increasing the likelihood of SBO by about fivefold; however, normal or absent bowel sounds appear very nearly valueless, as evidenced by their occurrence with roughly the same frequency in both SBO and perforated peptic ulcer.

Palpation

The vast majority of clinical information obtained from examination of the abdomen is acquired through gentle palpation, using the middle three fingers, and beginning at a distance from the area of maximum pain. Voluntary guarding (contraction of the abdominal musculature in anticipation of or in response to palpation) can be diminished by asking patients to flex their knees. Those who remain guarded following this maneuver will often relax if the clinician's hand is placed over the patient's, and the patient is then asked to use their own hand to palpate their abdomen. In contrast to the symptom of pain, tenderness is a sign in which pain is produced by palpation. Optimally, the patient's tenderness will be confined to one of the four discrete quadrants. However, this is often not the case, and one finds more diffuse tenderness encompassing one or more of the four combined areas noted above. Peritoneal irritation is suggested by rigidity (involuntary guarding or reflex spasm of abdominal muscles), as is pain referred to the point of maximum tenderness when palpating an adjacent quadrant.

Rebound tenderness, often regarded as the clinical criterion standard for peritonitis, has several important limitations. In patients with peritonitis, the combination of rigidity, referred tenderness, and especially pain with coughing usually provides sufficient diagnostic confirmation that little additional information is gained by eliciting the unnecessary pain of rebound. False positives occur in about one patient in four without peritonitis, perhaps due to a nonspecific startle response. Based on this, one might reasonably question whether rebound has sufficient predictive value to justify the discomfort it causes patients.

Enlargement of the liver or spleen, and other masses, including a distended bladder, should be sought. One should also examine for hernias in both men and women, particularly those that are tender, suggesting incarceration or strangulation.

In women, the pelvic examination—like the pregnancy test—may provide the clinician with relevant information that would not have been expected on the basis of the history. For this reason, it is wise to perform a pelvic examination in the evaluation of abdominal pain, particularly in women of reproductive age.

Although the rectal examination is widely regarded as an essential component in the assessment of abdominal pain, particularly in suspected appendicitis, there is little evidence that rectal tenderness in patients with RLQ pain provides useful incremental information beyond what has already been obtained by other components of the physical examination. Grossly melanotic, maroon, or bloody stool indicates GI bleeding. The test for occult blood, although routinely done, loses sensitivity if not performed serially over several days. Conversely, repeated rectal examinations performed over several hours by multiple examiners tends to reduce the specificity of the test for occult blood, presumably due to local trauma. Among patients with a final diagnosis on follow-up of NSAP, 10 percent had a positive stool test for occult blood.

Basic Laboratory and Radiographic Tests

The complete blood count and plain abdominal film are among the most overutilized tests in emergency practice. Neither test offers sufficiently powerful likelihood ratios (see below) to revise disease probability. One approach to the use of both these tests is to take note only of high threshold abnormalities, e.g., a very elevated WBC (>20,000/mm3), but to resist the temptation to draw any reassurance from a "normal" WBC or a "nonspecific bowel gas pattern."

Complete Blood Count

The limited clinical utility of the CBC can be demonstrated most readily by examining its performance characteristics in the three most common causes of abdominal pain: Appendicitis, biliary tract disease (principally cholecystitis), and NSAP. Based upon a metaanalysis of three studies containing a total of over 1800 patients, a WBC exceeding the threshold value of 10,000/L only doubled the odds of appendicitis, while a WBC below this cutoff point reduced the odds by only about half. As noted below in the discussion of likelihood ratios (LRs), an LR (+) = 2 and an LR (–) = 0.4 are of marginal clinical value.

For acute cholecystitis, the LRs of the WBC count are virtually identical to those seen in appendicitis, and of equally limited clinical utility.

In one large, well-conducted series of patients with NSAP, 28 percent (95% CI; 22 to 34%) of patients were reported to have WBC counts >10,500/L. In the development of a decision rule for identification of NSAP, investigators did not find the CBC to be of value in distinguishing patients with NSAP from other, more serious diagnoses. Because of the design of studies on NSAP, it is not possible to calculate a specificity or likelihood ratios for the performance of the WBC count in this setting. However, using 28 percent as the approximate sensitivity of the test, it is possible to estimate that, in order for leukocytosis to be of any value in NSAP (defined as producing LRs that deviate significantly from 1), the WBC count would have to demonstrate substantially better specificity than was seen in either appendicitis or cholecystitis.

All of the above refers only to individual WBC counts. There is some evidence that serial counts may assist in the identification of appendicitis. However, in this setting, it would seem wiser to obtain a CT rather than risk a perforation or other complication while obtaining serial WBCs and waiting for development of leukocytosis.

Plain Abdominal Radiograph

The plain abdominal radiograph (PAR) is often ordered as an "abdominal series," the meaning of which is variously defined. In some institutions, this includes an upright abdomen, in others an upright chest; in still others, only a single supine film is obtained. The utility of the erect abdominal film, when added to the combination of the supine abdominal and erect chest film, is generally low and does not impact management. Abdominal films in suspected appendicitis, NSAP, or urinary tract infection are also unlikely to be helpful, and can be misleading.

An additional limitation of the plain abdominal radiograph is poor interrater reliability for commonly used radiographic signs.

Restriction of PARs to patients with suspected obstruction or perforation would reduce utilization by over 80 percent with no adverse impact on management. Ultrasound may be a more sensitive test for detection of free air than the combination of upright chest and left lateral decubitus plain films (93 vs. 79 percent), which is one of the principal uses for plain radiography in abdominal pain.3 Ultrasound can be extremely helpful, particularly as a rapid bedside screening test, but it is highly operator-dependent and limited by overlying gas and obesity. Computed tomography (CT) is markedly superior for identifying virtually any abnormality that could be seen on plain films, particularly SBO and renal colic (Tables 72-1 and 72-2). Bedside sonography, combined with computed tomography would seem to be the key to obviating the need for continued use of the PAR in the future.

Table 72-1 Diagnostic Tests for Small Bowel Obstruction

Target Diagnosis Test Sensitivity (Range) Specificity (Range) LR (+) LR (–)

Small bowel obstruction (SBO) Plain abdominal films 63% (44–71%) 54% (38–65%) 1 0.7

SBO high-grade CT with IV +/– PO contrast 90% (81–97%) 96% (85–98%) 22 0.1

SBO low- & high-grade CT with IV +/– PO contrast 64% (55–85%) 79% (68–88%) 3 0.5

SBO with ischemia CT with IV +/– PO contrast 83% (32–100%) 88% (61–100%) 7 0.2

Table 72-2 Diagnostic Tests for Renal Colic

Test Sensitivity [95% CI] (Range) Specificity [95% CI] (Range) LR (+) LR (–)

Microscopic Urinalysis 84% [81–87%]* 48% [43–53%] 2 0.3

Plain abdominal film 58% (39–68%) 74% [47–88%] 2 0.6

Unenhanced helical CT (criterion standard) — — — —

Intravenous pyelogram (IVP) 78% [67–88%] 95% [91–99%] 16 0.2

Ultrasonography (without Doppler) 74% (19–100%) 95% (90–100%) 15 0.3

Doppler ultrasound (resistive index) 90% [79–97%] 100% [94–100%] 30 0.1

* Brackets indicate 95% CI; parentheses indicate range.

Diagnosis and Testing

Diagnosis is now more closely linked to appropriate disposition and treatment than was the case when the only interventions in abdominal pain were laparotomy or observation with medical management.

Accurate diagnosis is extremely difficult using only clinical information and basic laboratory tests. When initial and final diagnoses are compared, diagnostic accuracy falls somewhere between 50 and 65 percent overall. Diagnostic error in adults with abdominal pain increases in proportion to age, ranging from a low of 20 percent if only young adults are considered, to a high of 70 percent in the very elderly.

Although some improvement in clinical diagnostic accuracy occurs with experience, most is due to diagnostic imaging.

Performance Characteristics of Diagnostic Tests

Tables 72-1, 72-2, 72-4, 72-5, 72-6, 72-7, 72-8, 72-9, 72-10, and 72-11 provide a summary of the performance of diagnostic tests used in the ED work-up of acute abdominal pain. These test properties are displayed as sensitivity, specificity, and likelihood ratios. When derived from a metaanalysis of several studies, sensitivity and specificity, bounded by 95 percent confidence intervals (CIs), are calculated using the Summary Receiver Operating Characteristics (SROC) methodology, which adjusts for interstudy variation in diagnostic threshold.4 Under conditions where merged studies are too clinically or statistically heterogenous for valid metaanalysis, aggregate sensitivity and specificity are calculated as weighted means, bounded by ranges.

Table 72-4 Diagnostic Tests for Appendicitis

Test Sensitivity [95% CI] (Range) Specificity [95% CI] (Range) LR (+) LR (–)

Plain abdominal radiograph 48% [41–54%] 58% [54–62%] 1 0.9

Abdominopelvic ultrasound (real-time, graded compression, gray-scale) 55% [48–62%] 95% [93–97%] 11 0.5

Abdominopelvic ultrasound (color Doppler added to gray-scale) 84% [77–91%] 96% [88–100%] 21 0.2

Abdominopelvic unenhanced helical CT (no PO, IV, or colonic contrast) 88% [82–94%] 97% [94–99%] 29 0.1

Abdominopelvic helical CT (double [PO + IV] contrast; no colonic contrast) 91% [81–98%] 95% [90–98%] 18 0.1

Focused appendiceal (RLQ) unenhanced helical CT (no PO, IV, or colonic contrast) 87% [78–93%] 97% [92–99%] 29 0.1

Focused appendiceal (RLQ) helical CT (PO contrast only; no IV or colonic contrast) 76% [62–87%] 95% [90–98%] 15 0.3

Focused appendiceal (RLQ) helical CT (PO + colonic contrast; no IV contrast) 100% [94–100%] 95% [84–99%] 21 0.03

Focused appendiceal (RLQ) helical CT colonic contrast only; no PO or IV contrast) 98% [90–100%] 98% [89–100%] 49 0.02

MRI (gadolinium-enhanced) 97% [85–100%] 92% [75–99%] 12 0.03

Table 72-5 Diagnostic Tests for Biliary Tract Disease

Target Diagnosis Test Sensitivity [95% CI] (Range) Specificity [95% CI] (Range) LR (+) LR (–)

Cholelithiasis Plain abdominal radiograph 64% [59–68%] 68% [52–83%] 2 0.5

Cholelithiasis Ultrasound (US) 91% [84–97%] 97% [95–99%] 30 0.1

Cholelithiasis CT 85% (77–96%) 97% (86–99%) 28 0.2

Acute cholecystitis US 86% (65–97%) 97% (87–100%) 29 0.1

Acute cholecystitis Color velocity imaging & power Doppler US 93% (77–100%) 97% (88–100%) 31 0.1

Acute cholecystitis Radionuclide scanning 95% [91–98%] 90% [86–94%] 10 .05

Common duct obstruction US 90% (38–95%) 92% (48–97%) 11 0.1

Common duct obstruction CT 83% (51–90%) 87% (44–94%) 6 0.2

Common duct obstruction Radionuclide scanning 93% (81–99%) 92% (84–100%) 12 0.1

Common duct obstruction MR cholangiography 95% (85–96%) 97% (85–99%) 32 0.05

Common duct stone US 85% (19–76%) 89% (52–98%) 8 0.2

Common duct stone CT 71% (29–82%) 86% (55–92%) 5 0.3

Common duct stone MR cholangiography 95% (86–100%) 96% (87–100%) 24 0.05

Table 72-6 Diagnostic Tests for Acute Pancreatitis

Target Diagnosis Test Sensitivity (Range) Specificity (Range) LR (+) LR (–)

Inflammation Serum amylase 82% (72–93%) 85% (78–94%) 5 0.2

Inflammation Serum lipase >2x normal 90% (79–99%) 92% (85–98%) 11 0.1

Pancreatic necrosis CT with PO & bolus IV contrast 92% (75–100%) 95% (92–100%) 18 0.1

Drainable collections Transabdominal ultrasound (US) 54% (23–83%) 88% (47–100%) 4 0.5

Drainable collections CT with PO & bolus IV contrast 90% (72–100%) 48% (32–85%) 2 0.2

Drainable collections MRI (unenhanced) 92% (66–100%) 88% (79–100%) 8 0.1

Acute hemorrhagic pancreatitis Unenhanced CT (criterion standard) — — — —

Biliary pancreatitis Serum alanine aminotransferase (ALT) >3x normal 54% (38–73%) 92% (77–96%) 7 0.5

Common bile duct obstruction US 90% (38–95%) 92% (48–97%) 11 0.1

Common bile duct obstruction CT 83% (51–90%) 87% (44–94%) 6 0.2

Common bile duct obstruction Radionuclide scanning 93% (81–99%) 92% (84–100%) 12 0.1

Common bile duct obstruction MR cholangiography 95% (85–96%) 97% (85–99%) 32 0.05

Table 72-7 Diagnostic Tests for Acute Diverticulitis

Target Diagnosis Test Sensitivity (Range) Specificity (Range) LR (+) LR (–)

Inflammation or abscess Ultrasonography (high resolution, graded compression) 83% (77–91%) 95% (86–99%) 17 0.2

Inflammation or abscess Helical CT with colonic contrast only (no IV or PO contrast) 98% (88–99%) 99% (96–100%) 98 0.02

Table 72-8 Diagnostic Tests for Acute Pelvic Inflammatory Disease

Target Diagnosis Test Sensitivity [95% CI] (Range) Specificity [95% CI] (Range) LR (+) LR (–)

Salpingitis (macroscopic laparoscopy) Erythrocyte sedimentation rate >15 mm per h 78% (45–81%) 44% (25–57%) 1 0.5

Salpingitis (macroscopic laparoscopy) C-reactive protein 70% (54–93%) 59% (48–90%) 2 0.5

Salpingitis (macroscopic laparoscopy) Endometrial biopsy 80% (70–89%) 76% (67–89%) 3 0.3

Salpingitis (macroscopic laparoscopy) Gonococcus or Chlamydia cultured from upper genital tract 65% [41–85%] 100% [75–100%] 5 0.4

Salpingitis (macroscopic laparoscopy) Transvaginal power Doppler 100% [83–100%] 80% (56–94%) 5 0.1

Endometritis (endometrial biopsy) Conventional transvaginal sonography 85% [54–98%] 100% [91–100%] 18 0.2

Salpingitis (fimbrial minibiopsy) Laparoscopy (macroscopic) 50% [29–71%] 80% [66–90%] 2 0.6

Endometritis (endometrial biopsy) Laparoscopy (macroscopic) 93% [68–100%] 67% [41–87%] 3 0.1

Salpingitis/endometritis (fimbrial minibiopsy or endometrial biopsy) Laparoscopy (macroscopic) 48% [30–67%] 79% [66–88%] 2 0.7

Chlamydia cultured from upper genital tract Laparoscopy (macroscopic) 53% [28–77%] 67% [22–96%] 2 0.7

Table 72-9 Diagnostic Tests for Ectopic Pregnancy

Target Diagnosis Test Specificity [95% CI] (Range) Specificity [95% CI] (Range) LR (+) LR (–)

Pregnancy Serum hCG [10 mIU/mL = (+)] 99% [92–100%] 98% [94–100%] 50 .01

Pregnancy Serum hCG [25 mIU/mL = (+)] 98% [91–100%] 99% [94–100%] 98 .02

Pregnancy Urine hCG [>20 mIU/mL = (+)] 98% [96–100%] 98% [96–99%] 49 .02

Pregnancy Urine hCG [>50 mIU/mL = (+]] 95% [90–98%] 99% [97–99%] 95 .05

IUP TVS on all patients w/ (+) hCG 94% [90–97%] 93% [88–97%] 13 .06

IUP TVS on patients w/ hCG ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download