Stein PD. NEJM 1 June 2006; 354(22): 2317-27.
RESULTS
Randomization: 7284 screened, 3262 eligible, 1090 enrolled. 238 did not receive reference diagnosis, 824 underwent subsequent analysis (11% of those screened, 25% of those eligible).
Baseline: mean age 51.7y, majority had low or moderate probability of pulmonary embolism (Wells score).
Reference Diagnosis: 192/824 diagnosed with PE (23%), of the 632 ruled out, 592 had interpretable CTA, 590 did not receive anticoagulatants, 2/590 had initially unrecognized PE.
Wells Score | Table 4 | Table 5
Results of CTA:
Sensitivity 83% (76-92%)
Specificity 96% (93-97%)
Likelihood ratio for a positive test was 19.6 (13.3 - 29.0)
Likelihood ratio for a negative test was 0.18 (0.13 - 0.24)
PPV 86% (79-90%)
NPV 95% (92-96%)
PPV for PE in main or lobar artery 97%
PPV for PE in segmental vessel 68%
PPV for PE in subsegmental vessel 25%
Results of CTA-CTV:
Sensitivity 90% (84-93%)
Specificity 95% (92-96%)
Likelihood ratio for a positive test was 16.5 (11.6 - 23.5)
Likelihood ratio for a negative test was 0.11 (0.07 - 0.16)
PPV 85% (78-94%)
NPV 97% (94-97%)
Results using the Wells score:
High Clinical Probabiltiy: PPV 96%, NPV 60%
with CTV, PPV 96%, NPV 82%
Discussion: In high clinical probability patients, a positive CTA provides sufficient evidence to start treatment. Venous phase imaging increased the sensitivity by 7%. In high clinical probability patients who have a negative CT, the false negative rate was 40% which poses a dilemna regarding treatment.
Recent IC Attempts