Review: Ultrasound is best for detecting cholelithiasis and radionuclide scanning is best for detecting acute cholecystitis
ACP J Club. 1995 May-June;122:76. doi:10.7326/ACPJC-1995-122-3-076
Shea JA, Berlin JA, Escarce JJ, et al. Revised estimates of diagnostic test sensitivity and specificity in suspected biliary tract disease. Arch Intern Med. 1994 Nov 28;154:2573-81.
To estimate, by meta-analysis, the sensitivity and specificity of common tests used to diagnose gallstones and acute cholecystitis.
English-language studies were identified using MEDLINE (1966 to September 1992) with the Medical Subject Headings cholelithiasis, cholecystitis, cholecystography, ultrasonography, ultrasonics, tomography, nuclear magnetic resonance, and radionuclide imaging.
Studies were selected using a clearly defined 4-stage process. Inclusion criteria were the following: The data were original and data on gallstones and acute cholecystitis could be differentiated from other diagnoses; ≥ 20 patients were included and < 10% were lost to follow-up; the criteria for positive and negative test results were sufficiently described; the diagnosis was confirmed by an acceptable diagnostic standard; and the sensitivity and specificity of the test could be calculated.
Data on the number of patients, the definition of positive and negative test results, the diagnostic standard, and the sensitivity and specificity of the test were extracted. Cluster-sampling methods were used to obtain estimates of sensitivity and specificity across studies. Adjustments were made to estimates that were biased because the diagnostic standard was applied preferentially to patients with positive test results.
30 of the 1614 studies identified met the selection criteria and were included in the analysis. Ultrasound had the best unadjusted sensitivity (97%, 95% CI 95% to 99%) and specificity (95%, CI 88% to 100%) for evaluating patients with suspected gallstones. The adjusted values were 84% (CI 76% to 92%) and 99% (CI 97% to 100%), respectively (Table). The adjusted and unadjusted results for oral cholecystogram were lower (sensitivity 60% and 90%, specificity 97% and 95% respectively), as were the unadjusted results for computed tomography (sensitivity 79%, specificity 99%) (Table). Radionuclide scanning had the best sensitivity and specificity for evaluating patients with suspected acute cholecystitis (Table). Unadjusted sensitivity and specificity of ultrasound for evaluating patients with suspected acute cholecystitis were lower: 94% (CI 92% to 96%) and 78% (CI 61% to 96%). The adjusted values were 88% (CI 74% to 100%) and 80% (CI 62% to 98%), respectively (Table).
Ultrasound is superior to oral cholecystogram and computed tomography for diagnosing cholelithiasis, and radionuclide scanning is the most accurate test for diagnosing acute cholecystitis.
Source of funding: Agency for Health Care Policy and Research.
For article reprint: Not available.
Table. Test characteristics for diagnosing gallstones and acute cholecystitis*
|Outcomes||Tests||Sensitivity (95% CI)||Specificity (CI)||+LR||-LR|
|Gallstones||Ultrasound||84% (76 to 92)†||99% (97 to 100)†||84||0.16|
|Oral cholecystogram||60% (41 to 79)†||97% (94 to 100)†||20||0.41|
|Computed tomography||79% (Not reported)||99% (Not reported)||79||0.21|
|Acute cholecystitis||Radionuclide scanning||97% (96 to 98)||90% (86 to 95)||9.7||0.03|
|Ultrasound||88% (74 to 100)†||80% (62% to 98%)†||4.4||0.15|
|Oral cholecystography||63% (Not reported)||100% (Not reported)||∞||0.37|
*LRs defined in Glossary and calculated from data in article.
†Sensitivity and specificty were adjusted for estimates when the diagnostic standard was applied preferentially to patients with positive test results.
The first conclusion of the meta-analysis by Shea and colleagues is that ultrasonography is superior to oral cholecystography for the diagnosis of gallstones. This, of course, is old news. Ultrasonography supplanted cholecystography in most centers almost 15 years ago, partly because of the large fraction of indeterminate cholecystograms (estimated to be 42% in this study). This substitution of technologies is so complete that my medical students have never heard of cholecystography. This meta-analysis does indicate, however, that ultrasonography is not as accurate in the diagnosis of gallstones as is commonly stated in textbooks. The best estimates by Shea and colleagues are between 88% and 90% for sensitivity and between 97% and 98% for specificity. For a patient with an estimated 50% probability of developing gallstones before testing, the positive predictive value is 97% and the negative predictive value is only 90%.
The choice of which technologies to use to diagnose acute cholecystitis is more complex. This meta-analysis concludes that radionuclide scanning is superior to ultrasonography. Although this is probably true, other considerations are also important. Ultrasonography is faster and less expensive than radionuclide scanning, more readily available in smaller hospitals, and easier to obtain at night and on weekends. Ultrasonography can also image other abdominal organs that can produce symptoms similar to those of gallbladder disease, and it does so without exposing the patient to radiation.
These advantages may outweigh the relatively modest differences in test accuracy. For example, for a patient with an estimated 50% probability of developing acute cholecystitis before testing, the positive predictive value of ultrasonography is 81% compared with 91% for scanning; the negative predictive values are 90% and 97%, respectively. Given its practical advantages, ultrasonography could be considered the test of first choice, with radionuclide scanning reserved for instances in which uncertainty persists.
Andrew K. Diehl, MD, MSc
The University of Texas Health Science Center at San AntonioSan Antonio, Texas, USA