Family Practice Advance Access originally published online on June 17, 2005
Family Practice 2005 22(5):474-477; doi:10.1093/fampra/cmi039
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Results of computed tomography in family practitioners' patients with non-acute abdominal pain
a Department of Gastroenterology, Kaiser Permanente Medical Care Plan, San Diego and b Department of Research and Evaluation, Southern California Permanente Medical Group, Pasadena, California
Correspondence to George F Longstreth, Department of Gastroenterology, Kaiser Medical Center, 4647 Zion Avenue, San Diego, California 92120, USA; Email: George.F.Longstreth{at}kp.org
Received 20 May 2005; Accepted 1 April 2005.
Master SS, Longstreth GF and Liu AL. Results of computed tomography in family practitioners' patients with non-acute abdominal pain. Family Practice 2005; 22: 474–477.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Objectives. The utility of abdominal computerized tomography (CT) for evaluating abdominal pain in non-emergency, primary care outpatients is unknown. Family practice patients commonly report abdominal pain.
Methods. We reviewed the records of health maintenance organization primary care outpatients <2 years after they had undergone CT for abdominal pain and assessed demographic variables, clinical and laboratory data, CT findings and final diagnoses.
Results. We studied 137 patients: age 58.1 ± 16.1 years (mean ± SD), 80 (58.4%) females. Fifty (36.5%) patients had 
1 warning clinical or laboratory feature. Positive (etiologic) and negative CT reports and unrelated and multiple CT findings occurred in 16 (11.7%), 104 (75.9%), 16 (11.7%) and 1 (0.7%) patients, respectively. Positive findings occurred in 16 (32.0%) patients with 1 warning feature and 1 (1.2%) patient (including the 1 patient with multiple findings) with no warning feature (P < 0.0001). One (6.3%) unrelated finding led to treatment, ovariectomy for a benign tumor. Fifty-four (39.4%) patients had a final diagnosis.
Conclusions. A majority of outpatients who had CT for abdominal pain received no diagnosis, and CT was rarely diagnostic for patients lacking a warning feature. Positive and unrelated CT findings were equally prevalent, and the latter were not beneficial.
Keywords. Abdominal pain, computed tomography.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Abdominal pain is a common outpatient symptom that has organic and functional etiologies.1–3 Computed tomography (CT) is useful for evaluating patients with an acute abdomen 4–5 and accurately identifies pathology in abdominal and pelvic organs, but its utility as an imaging procedure for diagnosing the etiology of abdominal pain in non-emergency, primary care outpatients has not been studied. Since CT images multiple organs, it can reveal abnormalities in anatomic regions that are unrelated to symptoms, which could lead to further evaluation that may not benefit the patient.6 Knowledge of patient features that predict a positive CT and awareness of the benefits and disadvantages following CT reports of unrelated findings could help family practitioners. This information could also help gastroenterologists who see referred patients with abdominal pain.
We reviewed the records of outpatients whose abdominal pain was evaluated by their primary care physicians with CT. We categorized the types of CT findings and final diagnoses and assessed the diagnostic rate of CT in relation to warning clinical and laboratory features.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
We conducted the study in the Kaiser Permanente Medical Care Plan, a group-model health maintenance organization (HMO) in which prepayment covers most expense, and all members have a primary physician and a unified medical record. The family practitioners have unrestricted access to laboratory and radiological tests. They have lists of patients for whom they are responsible to provide care. The institutional review board approved the study. We identified 383 consecutive patients from radiology department records who had undergone abdominal or abdominal/pelvic CT, and one investigator (SSM) reviewed their medical records at least 2 years later. Inclusion criteria were age at least 18 years and CT ordered by a primary care physician in an outpatient, non-emergency setting to evaluate abdominal pain. We aimed to study only patients who had undergone CT for abdominal pain that was new or had not been evaluated with gastroenterologist consultation or recent CT. The reviewer excluded patients who had undergone CT for indications other than undiagnosed abdominal pain (usually for evaluation of known malignancy) (163 patients); patients with a prior radiological, sonographic or endoscopic abnormality (71 patients); or those who had undergone specialty consultation for abdominal pain within 1 year (8 patients) or had undergone CT for the same symptom within 1 year (4 patients). Therefore, 137 patients were studied. Diagnostic investigation prior to CT varied widely and played no role in patient selection unless in contributed to the exclusion criteria. They underwent routine CT, using both oral and intravenous contrast in most cases, without other special techniques.
We classified the CT findings into 4 categories: 1) a positive finding identified the etiology of the pain; 2) a negative finding revealed no abnormality or an insignificant finding, i.e. one that led to no specialty consultation or additional imaging and was not established as the etiology of the pain during follow-up; 3) an unrelated finding, i.e. an abnormality for which specialty consultation or additional imaging was requested by the primary physician but was not established as the etiology of the pain during follow-up; and 4) multiple finding categories. Using clinical judgment, we agreed on warning symptoms and signs for recording: unintentional weight loss >10%, rectal bleeding, vomiting, dysphagia, gross hematuria, fever, palpable abdominal mass, anemia (females, hemoglobin <12.0 g/dl; males, <13 g/dl),7 leukocyte count >11 000/mm–3), and liver test abnormality (total bilirubin >3 mg/dl; alkaline phosphatase, aspartate aminotransferase or alanine aminotransferase > twice normal values).
The investigator recorded initial demographic characteristics, clinical and laboratory data, and CT findings. She also recorded each patient's final diagnosis made by primary care or specialist physicians, if one was made. We categorized diagnoses as organic diseases (those with a structural or infectious etiology) and functional (other diagnoses). A second investigator (GFL) reviewed the records of patients with positive, unrelated and multiple findings. The two investigators independently classified these CT findings with 100% concordance.
We summarized continuous data as mean ± standard deviation (SD) and compared categorical variables with the chi square test or Fisher's exact test.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Patient features
The patients' mean age was 58.1 ± 16.1 years, and 80 (58.4%) were females. The pain duration was 8.8 ± 16.9 months (n = 123; unrecorded in 14) and
1 month in 116 (94.3%) patients. The distribution of abdominal pain sites was: lower, 61 (44.5%); upper, 55 (39.4%); upper and lower, 14 (10.2%); middle, 4 (2.9%); and unspecified, 4 (2.9%). Fifty (36.5%) patients had at least 1 warning feature.
CT findings and diagnoses
CT findings were positive, negative, unrelated and multiple in 16 (11.7%), 104 (75.9%), 16 (11.7%) and 1 (0.7%) patients, respectively. Positive findings were liver metastases (5, including 1 each with a pancreatic or cecal mass), ventral hernia (3), cholelithiasis (2), abdominal abscess (1), adrenal mass (1), abdominal wall hematoma (1), cholecystitis (1), ascites (1), and sigmoid colon thickening (1). Unrelated findings comprised organomegaly, masses or other lesions in gynecologic organs (6), liver (3), kidney (2), pancreas (2), spleen (1), and multiple organs (2). The radiologists reported that some of the unrelated findings were minor, but they often advised another imaging study to further delineate the abnormality. The single multiple category finding comprised an inguinal hernia (positive) and cecal thickening (unrelated). For patients with and without (breve)1 warning feature, positive CT findings occurred in 16 (32.0%) and 1 (1.2%) patients (including the 1 patient with multiple findings), respectively (P < 0.0001).
Fifteen of 16 (93.8%) of patients with unrelated findings had specialty consultation and/or additional imaging, as advised on the CT reports and/or selected by a primary care practitioner. The remaining patient had a small liver lesion on CT, and sonography was requested by a physician but not performed. All but 1 (6.3%) unrelated finding were concluded to be insignificant or were not confirmed after specialty consultation and/or additional imaging, and they warranted no therapy. The exception occurred in a woman who had resection of a benign ovarian tumor, which did not diminish her abdominal pain.
Follow-up to the last medical record entry was 29.3 ± 11.7 months. A final diagnosis was made in 54 (39.4%) patients (Table 1). Organic and functional diagnoses were made in 34 (63.0%) and 20 (37.0%) of diagnosed patients, 24.8% and 14.6% of all patients, respectively. An organic diagnosis was made in 21 of 50 (42.0%) of patients with 1 warning feature and in 13 of 87 (14.9%) of patients with no warning feature (P = 0.0008).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
In this study of 137 non-emergency, primary care outpatients at least 2 years after undergoing CT for abdominal pain, a majority of patients received no final diagnosis. The procedure was rarely diagnostic in the absence of a warning clinical or laboratory feature. A minority of diagnostic findings was accompanied by an equal proportion of findings that were unrelated to the pain, which led to specialty consultation and/or additional imaging in most cases; however, further evaluation of these patients resulted in specific therapy for only 1 patient that did not relieve her pain. We believe our findings are applicable to other primary care settings.
The diagnosis of an organic etiology in only one-fourth of patients resembles the 21% organic diagnosis rate reported in other medical outpatients who were evaluated for abdominal pain or dyspepsia.2 About one-half of the organic diagnoses are either usually detectable by physical examination or less expensive laboratory or imaging tests and/or cause no CT abnormality. The proportion of patients with a functional diagnosis was probably under-estimated, as we recorded only definite diagnoses, and some undiagnosed patients likely had functional disorders. The two most common functional diagnoses were abdominal wall pain, which is identified by physical examination but often overlooked,8 and irritable bowel syndrome, which usually can be diagnosed by typical symptoms and limited diagnostic testing.9 We could not distinguish physician and patient motives for ordering CT, but general medicine outpatients often expect diagnostic testing, and patients who expect a test are more likely to receive testing.10,11 Therefore, patient expectations could have played a role.
If the discovery of a CT finding that is unrelated to a patient's pain resulted in effective therapy, this occurrence would be important. However, our findings suggest that such an outcome would be infrequent. In contrast, 10% of patients undergoing screening computed tomographic colonography had highly important extracolonic findings.12 However, comparison of that study with our results is complicated by its higher proportion of males (unrelated findings in our patients were most often reported in gynecological organs), differing CT techniques, unreported clinical and laboratory data, and possible referral bias. In our patients, unrelated findings were generally followed by fruitless specialist consultations and imaging procedures.6
Our retrospective study could not measure the reassurance value that patients might have received from negative CT reports or the benefit to physicians of these reports in confirming management or preventing referral. This potential benefit could have reduced physician visits and other imaging procedures. Although we identified all primary care patients who underwent CT during the case finding period, we do not know what proportion of all primary care patients presenting with abdominal pain had CT.
To our knowledge, this is the first study of the utility of CT for diagnosing the cause of abdominal pain in non-emergency, primary care outpatients who have not already had an abnormal imaging procedure result or undergone specialty consultation. The most important results are the infrequency of diagnostic CT findings from patients lacking clinical and routine laboratory warning features and the common reporting of abnormalities that were subsequently found to be absent or insignificant after further evaluation. These preliminary findings could encourage physicians to rely more on the history and physical examination, which is mainly of value in ruling out disease, and routine diagnostic testing in managing patients with abdominal pain.6,13 A prospective, controlled study of CT in the management of abdominal pain, including assessment of the most useful clinical and laboratory predictors of a diagnostic CT, would be worthwhile. Further research could lead to guidelines for imaging as have been created for the evaluation of headache14 and low back pain.15 Our results suggest caution in expanding open CT access to family practitioners in areas where it is not routinely available, such as Britain.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: this research was funded by a grant from the Kaiser Foundation Health Plan and Hospitals.
Ethical approval: the institutional review board approved the study.
Conflicts of interest: none.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 Haubrich WS. Abdominal pain. In Haubrich WS, Schaffner F, Berk JE. Bockus Gastroenterology. 5th edn. Philadelphia: WB Saunders Co; 1995, 11–29.
2 Kroenke K, Mangelsdorff AD. Common symptoms in ambulatory care: incidence, evaluation, therapy, and outcome. Am J Med 1989; 86: 262–266.[CrossRef][Web of Science][Medline]
3 Kroenke K, Mangelsdorff AD. The prevalence of symptoms in medical outpatients and the adequacy of therapy. Arch Int Med 1990; 150: 1685–1689.
4 Ng CS, Watson CJE, Palmer CR et al. Evaluation of early abdominopelvic computed tomography in patients with acute abdominal pain of unknown cause: prospective randomized study. Br Med J 2002; 325: 1387–1389.
5 Rosen MP, Sands DZ, Longmaid III HE et al. Impact of abdominal CT on the management of patients presenting to the emergency department with acute abdominal pain. Am J Roentgenol 2000; 174: 1391–1396.
6 Mold JW, Stein HF. The cascade effect in the clinical care of patients. N Engl J Med 1986; 314: 512–514.[Web of Science][Medline]
7 DeMaeyer E, Adiela-Yagman M. The prevalence of anaemia in the world. World Health Stat Q 1985; 38: 302–316.[Medline]
8 Costanza CD, Longstreth GF, Liu AL. Chronic abdominal wall pain: clinical features, health care costs, and long-term follow-up. Clin Gastroenterol Hepatol 2004; 2: 395–399.[CrossRef][Medline]
9 Cash BD, Schoenfeld P, Chey WD. The utility of diagnostic tests in irritable bowel syndrome patients: systematic review. Am J Gastroenterol 2002; 97: 2812–2819.[CrossRef][Web of Science][Medline]
10 Kravitz RL. Measuring patients' expectations and requests. Ann Int Med 2001; 134: 881–888.
11 Jackson JL, Kroenke K. The effect of unmet expectations among adults presenting with physical symptoms. Ann Int Med 2001; 134: 889–897.
12 Gluecker TM, Johnson CD, Wilson LA, MacCarty RL, Welch TJ, Vanness DJ et al. Extracolonic findings at CT colonography: evaluation of prevalence and costs in a screening population. Gastroenterology 2003; 124: 911–916.[CrossRef][Web of Science][Medline]
13 Martina B, Bucheli B, Stotz M, Battegay E, Gyr N. First clinical judgement by primary care physicians distinguishes well between nonorganic and organic causes of abdominal or chest pain. J Gen Intern Med 1997; 12: 459–465.[CrossRef][Web of Science][Medline]
14 Silberstein SD, Rosenberg J. Multispecialty consensus on diagnosis and treatment of headache. Neurology 2000; 54: 1553. Full text available at www.neurology.org.
15 Jarvik JG, Deyo RA. Diagnostic evaluation of low back pain with emphasis on imaging. Ann Int Med 2002; 137: 586–597.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. F Longstreth Avoiding unnecessary surgery in irritable bowel syndrome Gut, May 1, 2007; 56(5): 608 - 610. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||