As many as half of patients who are evaluated for abdominal pain do not receive a precise diagnosis. And for about half of those who are given a diagnosis, the diagnosis is wrong. In this article, I will use actual cases (not "textbook" examples) to illustrate an approach to abdominal pain that begins with a careful differential diagnosis. I also offer some general guidelines for evaluating patients.
Most cases of abdominal pain do not result from a life-threatening condition. When they do, they may not differ obviously from cases in which the cause of pain is benign.
As many as half of patients who are evaluated for abdominal pain do not receive a precise diagnosis. And for about half of those who are given a diagnosis, the diagnosis is wrong.1,2 Misdiagnosis can divert attention from a serious disorder or lead to ill effects of unnecessary treatment.
In this article, I will use actual cases (not "textbook" examples) to illustrate an approach to abdominal pain that begins with a careful differential diagnosis. I also offer some general guidelines for evaluating patients (Box I).
ELDERLY WOMAN WITH NAUSEA AND VOMITING
Initial evaluation. An 84-year-old woman presented with a 2-day history of nausea, vomiting, and constant lower abdominal pain; she had no anorexia, hematemesis, diarrhea, constipation, or urinary tract symptoms. She had no significant medical conditions or allergies and was taking no medications.
Heart rate was 97 beats per minute; blood pressure, 94/60 mm Hg; respiration rate, 18 breaths per minute; temperature, 37.6°C (99.7°F); and oxygen saturation, 95% on room air. Head and neck, cardiac, and pulmonary examination results were normal. The patient exhibited bilateral lower abdominal tenderness with rebound tenderness and guarding. The remainder of the examination revealed no abnormalities.
Differential diagnosis in elderly patients. Abdominal pain in the elderly is a medical Pandora's box. Of patients 65 years or older who present to the hospital with abdominal pain, about 7% die and 22% to 40% need surgery. Furthermore, the cause of the pain remains unknown almost twice as frequently in elderly as in younger patients.3
Always keep in mind the dangerous triad of vascular catastrophes that can cause abdominal pain, especially in older patients:
In this case, we can exclude myocardial infarction, since the patient had lower abdominal pain only, with rebound tenderness and guarding confined to that area. Generally, however, when a patient older than 50 years presents with diffuse abdominal or upper abdominal pain, consider an acute coronary syndrome. If another plausible explanation for the pain is not apparent, admission and cardiac workup are usually warranted.
We must still consider the other 2 components of the dangerous triad--AAA and mesenteric ischemia--as well as more common serious pathologies. A reasonable differential diagnosis for this patient includes:
The next steps. Blood test results were normal, including a white blood cell (WBC) count of 9000/µL. Urinalysis showed 16 WBCs per high-power field (hpf), 7 red blood cells (RBCs) per hpf, and a few epithelial cells. Findings on abdominal films were normal.
The abnormal urinalysis results and normal WBC count and temperature were consistent with a urinary tract infection. However, the abdominal tenderness required further evaluation. A helical CT scan suggested perforating appendicitis (Figure 1). Surgical exploration revealed acute suppurative appendicitis with rupture.
Had the results of abdominal CT been normal, it would have been necessary to evaluate this patient for mesenteric ischemia. Although routine multidetector CT (MDCT) may yield positive findings in about 80% of cases of mesenteric ischemia, it cannot reliably exclude the diagnosis.4 No combination of laboratory tests can rule out this deadly condition. Only laparotomy and conventional mesenteric angiography or, in some centers, a dedicated MDCT mesenteric angiogram can do so.5 One of these procedures should be performed when mesenteric ischemia is suspected after a nondiagnostic initial workup.
Appendicitis. About 7% of the population has had an episode of appendicitis.6 The cause is obstruction of the appendiceal lumen by fecal matter. As secreted fluids accumulate, distention of the appendix leads to ischemic necrosis and bacterial infection of the devitalized tissue. Appendicitis is generally well tolerated by young, otherwise healthy patients. For elderly or immunocompromised patients, however, the associated morbidity and mortality are much higher.
The diagnostic sensitivity and specificity of various clinical signs and symptoms of appendicitis have been well studied (Table 1).7 No single finding is reliable in isolation; diagnosis based on the history and physical examination remains an "art." The presentation of appendicitis is often atypical in elderly patients, who display fever, rebound tenderness, and guarding in fewer than 50% of cases. Furthermore, 20% to 30% of patients with appendicitis have symptoms or urinalysis findings that may suggest urinary tract infection.8 Presumably, inflammation of the appendix induces a sterile pyuria in the adjacent ureter, a condition that may also accompany diverticulitis.
Although universally obtained when appendicitis is suspected, laboratory studies are often unhelpful and can be misleading. The WBC count has an overall sensitivity of about 80% and a much lower specificity. It is more likely to be normal early in the course of the disease; in immunocompromised persons; or at the extremes of age, when the presentation of appendicitis tends to be more obscure. The C-reactive protein assay does little to enhance the accuracy of the WBC count. I do not recommend routine use of the assay for the diagnosis of appendicitis.9
Can imaging studies improve the diagnosis of appendicitis? Ultrasonography results may be helpful if positive, but negative results do not exclude appendicitis. Its sensitivity was summarized in a recent systematic review and ranges from 38% to 82%.10 Ultrasonography may be a preferred first step in gravid patients or in children who lack the visceral fat needed to identify inflammation adjacent to the appendix on CT.11 Ultrasound scanning may also detect other diseases that are important in the differential diagnosis of appendicitis, especially pelvic disorders in female patients. However, it is highly operator- and patient-dependent, which limits its usefulness in most centers.
Helical CT scanning is the best imaging test for most patients. In expert hands, helical CT is highly sensitive (96% to 100%). The sensitivity of 100% obtained by Rao and coworkers12-14 required the use of colonic contrast material. If only intravenous or oral contrast material is used, the sensitivity may be somewhat lower, although it probably still ranges from 90% to 95%.15,16
Helical CT is also useful in diagnosing other abdominal abnormalities that are included in the differential diagnosis, such as AAA, diverticulitis, nephrolithiasis, cholelithiasis, colon cancer, and gynecologic disorders. Unlike ultrasonography, helical CT can rule out appendicitis by showing a nondistended appendix that fills with contrast material.
When the clinical evidence of appendicitis is obvious, the diagnosis can be made on the basis of the history and physical examination alone and the patient can proceed to surgery. When the diagnosis is in question, helical CT is the imaging test of choice in most cases.
LESSONS FROM THIS CASE
YOUNG WOMAN WITH SEVERE ABDOMINAL PAIN
Initial evaluation. A 29-year-old woman presented with severe right lower quadrant pain that had begun abruptly 3 hours earlier. The pain awoke her from sleep and was accompanied by vomiting and chills. The patient did not have fever, vaginal discharge or bleeding, or dysuria. One day earlier, she had been evaluated at another hospital for a similar episode and trimethoprim-sulfamethoxazole was prescribed for a presumptive urinary tract infection.
She had no significant medical disorders and took no other medications. Her last normal menses had occurred 6 days earlier.
Heart rate was 105 beats per minute; blood pressure, 120/81 mm Hg; respiration rate, 28 breaths per minute; temperature, 37°C (98.6°F); and oxygen saturation, 99% on room air. Head and neck, cardiac, pulmonary, skin, and extremity examination results were normal. The patient's abdomen was diffusely tender, with rebound tenderness and guarding in the right lower quadrant. Pelvic examination showed a closed cervical os with bilateral adnexal and uterine tenderness but no cervical motion tenderness or palpable masses. Rectal examination revealed tenderness on palpation that was greater on the right side than on the left and heme-negative brown stool.
Differential diagnosis in women of childbearing age. Ectopic pregnancy remains the most important life-threatening cause of abdominal pain for women who are in the first trimester of pregnancy. When a woman of childbearing age presents with abdominal pain, ectopic pregnancy must be ruled out by negative pregnancy test results or further evaluation is warranted. If the patient is pregnant, the only physical findings that can exclude this diagnosis are fetal heart tones and passage of the conceptus.17
For nonpregnant patients, the evaluation of acute abdomen focuses on 2 main areas, GI and gynecologic. Although a common cause of lower abdominal pain, urinary tract disease (infection or stones) rarely produces tenderness as pronounced as that found in this patient's examination.
The nature of this patient's pain--its abrupt and episodic occurrence, with simultaneous onset of vomiting--made appendicitis relatively unlikely, but this common disorder often presents atypically and may be as difficult to diagnose in fertile women as it is in elderly persons. Appendicitis is initially misdiagnosed in about one third of young women who ultimately prove to have the condition.6
A striking but inconclusive finding in this patient's physical examination was right-sided rectal tenderness. Although medical dogma dictates a rectal examination in all cases of suspected appendicitis, Dixon and associates18 found the rectal examination unhelpful in establishing that diagnosis. Patients with abdominal pain should be tested for fecal occult blood to exclude GI bleeding. Generalized or lateralizing rectal tenderness, however, seems to have no diagnostic or prognostic significance.
Perform a digital rectal examination in patients with suspected rectal carcinoma, fecal impaction, prostate disease, or cauda equina compression. In other patients with abdominal pain, stool occult blood testing may supplant the need for digital rectal examination.
For this patient, the differential diagnosis includes:
The next steps. Results of a urine pregnancy test were negative, and a urine dipstick test revealed no abnormalities. The patient's hemoglobin level, measured at the bedside, was 14 g/dL.
Surgical consultation was obtained 30 minutes after the patient's initial evaluation. Because administration of opiates might obscure the examination (Box II), the consultant requested that none be given to the patient beforehand. After the examination, the consultant recommended abdominal radiography and gynecologic consultation. The patient was then given fentanyl.
Urinalysis showed 3 WBCs per hpf and 2 RBCs per hpf. WBC count was 15,200/µL; all other blood test results were normal, including a hematocrit of 42.8%. Findings of abdominal radiography were normal, with no free air.
Pelvic ultrasonography showed a 9 3 10-cm cystic mass with multiple septations anterior to the uterus (Figure 2). The left ovary was normal, but the right ovary was not visualized.
The gynecology service recommended laparoscopy for a presumptive tubo-ovarian abscess. During surgery, a 10 3 10-cm hemorrhagic right ovary and 4-cm right fallopian tube were found to be twisted 2 full revolutions from their natural position. The adnexa appeared dusky, and the patient underwent right salpingo-oophorectomy.
Ovarian torsion. This condition, in which rotation of the adnexa around their vascular pedicle causes ischemic necrosis of the ovary and fallopian tube, can lead to infertility if not treated promptly. Suspicion of ovarian torsion mandates the same urgent diagnostic workup and treatment as its equivalent in men, testicular torsion.
Ovarian torsion most often occurs in the presence of ovarian cysts (mean diameter of 5 to 6 cm), but the ovaries are normal in up to 20% of cases.19 Some 15% of cases occur in infants and children because of the relative mobility of their adnexa. About 20% of cases occur during pregnancy, as a result of increased vascularity of the adnexa and their displacement by the growing uterus. Although two thirds of torsions are found on the right side, this may indicate "workup bias"; that is, patients with right-sided pain may be more likely to undergo further diagnostic evaluation.
The hallmark of ovarian torsion is acute, severe pain in the right or left lower abdomen. Nausea and vomiting, if present, start concurrently with the onset of pain. This can help distinguish torsion from appendicitis, in which the GI symptoms usually develop later. As this case demonstrates, many patients with ovarian torsion have a history of similar episodes that resolved spontaneously.20 Physical examination shows marked adnexal tenderness but less cervical motion tenderness than is seen with pelvic inflammatory disease. An adnexal mass is palpable in fewer than 50% of cases.21
The diagnosis remains challenging; consider ovarian torsion whenever it is consistent with the patient's presentation. Laboratory tests are not helpful in establishing the diagnosis. Fever and leukocytosis are rare; they occurred in fewer than 20% of cases in one study.22
Radiologic studies may be helpful but cannot exclude the diagnosis. Standard pelvic ultrasonography can reveal a large ovarian cyst but does not reliably detect torsion. Doppler ultrasonography confirms torsion if it shows absence of blood flow to the ovary, but normal flow does not exclude the diagnosis and may be found in 50% of patients with surgically proven ovarian torsion.19
A few investigators have described increased detection rates by using specialized color flow Doppler to depict arterial and venous flow to the ovary independently. However, these brief reports lack clear methods and investigator blinding; hence, this technique must be studied more scientifically before it can be used to rule out the diagnosis of ovarian torsion.23,24
The treatment of ovarian torsion entails emergent laparoscopy. Traditionally, salpingo-oophorectomy was performed without untwisting because of the theoretic risk of pulmonary embolism from a thrombosed vein and the inability to determine viability of the ovary. Studies performed during the past 10 years confirm the safety of laparoscopic detorsion; fertility was preserved in most cases in which blood flow was confirmed before surgery by color flow Doppler imaging.22,24-26 Contralateral oophoropexy is recommended in cases of torsion of normal adnexa when the torsed ovary is resected.25
LESSONS FROM THIS CASE
OLDER MAN WITH GROSS HEMATURIA
Initial evaluation. An obese 69-year-old man presented to the hospital with left lower quadrant pain of sudden onset and gross hematuria. The patient's pain was moderate, but he displayed significant diaphoresis.
Blood pressure was 161/95 mm Hg; heart rate, 61 beats per minute; respiration rate, 17 breaths per minute; and temperature, 36.4°C (97.5°F). His abdomen was soft, with normoactive bowel sounds and moderate left lower quadrant tenderness. He displayed no costovertebral angle tenderness.
A chest film revealed cardiomegaly and mild congestive heart failure. Urinalysis showed 3 WBCs per hpf and and 10 RBCs per hpf.
Differential diagnosis of severe pain with slight tenderness. The dangerous triad of vascular emergencies (AAA, mesenteric ischemia, and myocardial infarction) can cause abdominal pain of a severity that is out of proportion to the tenderness observed on physical examination. So, too, can the other conditions included in this differential diagnosis:
The next steps. The ECG showed sinus rhythm and no signs of myocardial ischemia. Because the differential diagnosis includes nephrolithiasis, many authorities recommend obtaining a screening radiograph of the kidneys, ureter, and bladder (KUB) to look for stones in this setting. This patient's KUB is shown in Figure 3.
An abdominal series should consist of 3 views, including flat, upright, and upright chest films. These radiographs are helpful only for diagnosing obstruction, perforation, foreign body, and occasionally, volvulus.27 Plain radiography is not a good screening tool for nephrolithiasis because the sensitivity is only 45% to 58%.28 Furthermore, clinicians must carefully interpret films that will have a delayed reading by a radiologist.
This patient had an AAA, which can have a subtle appearance on an abdominal x-ray film. In patients with AAA, abdominal radiographs may show a calcified outline of the aorta and lateral displacement of the calcified wall, which suggests the presence of the aneurysm. A sign of leakage or rupture of an AAA is the obliteration of the ipsilateral iliopsoas margin. In a patient with appropriate symptoms, the sensitivity of abdominal radiography for AAA is only 60% to 70%. Therefore, plain films should not be the initial diagnostic study when AAA is suspected. Instead, ultrasonography or CT should be used (Table 2).
Had the treating physicians in this case ordered an immediate bedside ultrasonogram, it might have looked like the image in Figure 4, which shows a large AAA.
Abdominal aortic aneurysm. This is a localized dilatation of all layers of the aortic wall. Although the cause of AAA is unknown, the condition is associated with long-standing hypertension and atherosclerosis. AAAs are present in about 6% of men older than 55 years and 14% of those older than 65 years. They are 5 times more common in men than in women.29
Most AAAs are asymptomatic unless they rupture or leak. With rupture, patients may present with abdominal or back pain and hypotension. The physical examination may be misleading because patients often lack significant tenderness and a pulsatile mass.
Radiologic studies establish the diagnosis in most cases. Plain lateral lumbar spine radiographs reveal evidence suggestive of AAA in 60% of cases. However, they should never be used to rule out AAA and are, therefore, of little use in the evaluation of a patient in whom AAA is suspected.
CT is the diagnostic test of choice for patients in stable condition and has excellent ability to detect other abdominal abnormalities. For patients whose condition is unstable, emergent bedside ultrasonography is preferable; however, when the clinical evidence of AAA is compelling, seek immediate laparotomy. Angiography is less sensitive and more invasive than other imaging tests. Historically, its main use has been the evaluation of patients before elective AAA repair; it has no role in the emergent evaluation of a patient with a suspected leaking AAA.
Treatment of a leaking AAA involves immediate surgical consultation and fluid and blood transfusions to maintain adequate organ perfusion. In prior decades, the mortality associated with emergent repair of a ruptured AAA was 50%, but a recent study suggests that it may now be as low as 41%.30 The mortality is higher when hemorrhage occurs in the abdominal cavity as opposed to the retroperitoneal space.31 One study found that emergency surgery in patients older than 80 years prolonged life, on average, by only 1 week.32
Newer treatments involve endovascular grafts placed via the femoral artery under fluoroscopic guidance. Currently, most centers use these grafts for elective AAA repair in patients who are poor candidates for surgery, but a few institutions have the capability to place them emergently.33
Hospital course. The treating physician received the radiologist's report identifying the AAA 5 hours after the patient arrived, and immediate surgical consultation was obtained. As the patient was prepared for surgery, he complained of increased abdominal and back pain. He was given analgesics but continued to have pain and became hemodynamically unstable. All resuscitative efforts failed, including blood transfusion and emergency thoracotomy, and the patient died less than 2 hours after the diagnosis was made.
LESSONS FROM THIS CASE