Clinical Probability and D-Dimer Testing.

In order to address the well-known inadequacies in diagnostic imaging for pulmonary embolism (PE), health care providers can utilize pretest probability estimates in conjunction with the likelihood ratios of diagnostic tests. This approach helps define more accurately the posttest probability of a disease with a given imaging result. In the case of PE, we have the benefit of proven clinical prediction rules to help assess the pretest probability. Furthermore, the D-dimer test, which measures the degradation product of a cross-linked fibrin blood clot, can also be employed in diagnostic algorithms. It is, however, important to keep in mind that although all D-dimer tests lie on the same receiver operating characteristic curve, some have higher sensitivity than others. Thus, when patients are clinically classified as low-probability, moderate-sensitivity D-dimer tests can be performed next, and a negative result will suffice to rule out PE. Because the negative likelihood ratio with these tests is approx 0.20, the patient's pretest probability of PE must be less than 10% to rule out PE with a negative D-dimer. If, on the other hand, a high-sensitivity D-dimer test is used in patients who have a low or moderate probability of PE, are PE-unlikely by the Wells model, or low- to moderate-probability by the Wicki clinical model, the physician can avoid the need for diagnostic imaging when the D-dimer is negative. In this latter case, the likelihood ratio of 0.06 with the high-sensitivity D-dimer implies that, if patients have a pretest probability of no more than 22%, a negative D-dimer will negate the need for diagnostic imaging. Importantly, combining D-dimer and pretest probability not only allows selection of patients appropriate for diagnostic testing, but also helps interpret the imaging test result as potentially false-negative or false-positive, and thus allows for standardization of the diagnostic workup of PE. [ABSTRACT FROM AUTHOR]