PERC Rule Calculator

The PERC rule exists to answer a narrow question: in a patient whose pre-test probability of pulmonary embolism is already low, is any testing at all worth doing? The logic rests on the testing threshold concept, the probability below which the harms of investigating exceed the benefit. For PE that threshold is conventionally placed near 1.8 percent, because D-dimer assays have false positive rates around 15 to 25 percent in symptomatic emergency populations, and each false positive leads to CT pulmonary angiography with roughly 3 to 7 millisieverts of radiation, contrast nephropathy risk, and frequent incidental findings. PERC is deliberately built from eight easily obtained variables, all of which are risk factors or physiologic markers with independent association to PE: age at least 50, heart rate at least 100 beats per minute, oxygen saturation below 95 percent on room air, unilateral leg swelling, hemoptysis, surgery or trauma within four weeks, prior DVT or PE, and exogenous estrogen use. Because it is designed for rule-out, PERC is all-or-nothing: a single positive item voids it, and the number of positives carries no gradation. Reported sensitivity is roughly 97 percent with a negative predictive value above 99 percent when applied correctly, but that performance collapses if the rule is used outside its intended population. It must not be applied when clinical gestalt or a Wells score suggests moderate or high probability, in hemodynamically unstable patients, in those already anticoagulated, or in patients with known thrombophilia. Pregnancy is a particular pitfall, since estrogen effects and leg swelling are physiologic and baseline VTE risk is several-fold higher.

The rule was derived by Jeffrey Kline, then at Carolinas Medical Center in Charlotte, North Carolina, with colleagues who were studying why so many emergency department PE workups ended negative. Their derivation used a large prospective emergency department dataset of several thousand patients and applied logistic regression and recursive analysis to find the smallest set of bedside variables whose complete absence pushed post-test probability below the testing threshold. The resulting eight-item rule was published in the Journal of Thrombosis and Haemostasis in 2004. Kline and colleagues then carried out a prospective multicenter validation across more than a dozen United States emergency departments, published in the same journal in 2008, in which PERC-negative low-risk patients had a PE rate near 1 percent at 45-day follow-up. Early European validation attempts were less favorable, partly because baseline VTE prevalence and D-dimer practice differ across health systems, and this fed a decade of debate about generalizability. The question was largely settled by the PROPER trial, a cluster-randomized noninferiority study conducted in French emergency departments by Yonathan Freund and colleagues and published in JAMA in 2018, which found that a PERC-based strategy was noninferior to conventional workup for excluding PE in low-risk patients while reducing imaging and length of stay. PERC now appears in American College of Emergency Physicians clinical policy and in most modern PE diagnostic algorithms, typically as a step applied after clinical gestalt or Wells has established low probability.