How Decision Rules Rule Out Pulmonary Embolism Without a Scan

Can a doctor safely rule out a lung clot without a CT scan?#

Yes, and the reason is a single number: a validated decision rule can drive the chance of a missed pulmonary embolism below roughly one to two percent, which is the point at which imaging tends to cause more harm than it prevents. A clinician grades a short list of findings, sometimes adds a blood test, and if the result falls inside a defined low-risk band, guidelines treat the diagnosis as excluded. Wells, PERC, age-adjusted D-dimer, and YEARS are the tools that reach that band, and each one earned its place by being tested on real patients who were then followed to see whether a clot surfaced later.

Key points#

• A rule is "safe" when the three-month rate of a later clot in the untested group stays under about one to two percent.
• That ceiling comes from the historical miss rate after a normal pulmonary angiogram, so it is a benchmark, not a guess.
• PERC decides whether to test at all; D-dimer strategies decide what a test result means; Wells and YEARS set the pretest probability that frames both.
• Age-adjusted and YEARS thresholds exist because a fixed D-dimer cutoff turns positive too easily in older and sicker patients.
• The point is not zero missed clots. It is fewer scans, a higher share of scans that find something, and a miss rate held below the accepted line.

Why not simply scan everyone#

CT pulmonary angiography is fast, widely available, and accurate, so the tempting answer is to image anyone with breathlessness or chest pain. The catch is that pulmonary embolism hides behind symptoms that overlap dozens of harmless conditions, so most people who might have a clot do not. Scanning all of them delivers radiation, contrast reactions, kidney strain, and a stream of incidental findings that spark their own workups. It also turns up tiny clots that may never have caused trouble, which then get treated with blood thinners that carry bleeding risk. When the great majority of scans are negative, the yield is low and the collateral damage is real.

Decision rules exist to lift that yield. The aim is not to catch every conceivable clot at any cost. It is to find the patients whose risk is genuinely low enough that expected harm from imaging outweighs expected benefit, using logic you can write down, teach, and audit.

The benchmark that makes "safe" mean something#

Before you look at any individual rule, it helps to fix what safe means. Researchers apply a rule, withhold imaging from the group it labels low-risk, and then wait. Three months later they count how many of those untested people came back with a confirmed clot. That count, expressed as a percentage, is the failure rate, and a rule is considered acceptable when it stays under roughly one to two percent. The figure is not arbitrary. It mirrors the residual miss rate that remained even after a normal pulmonary angiogram, the older reference test. In other words, no strategy reaches zero, so the honest question is whether a rule matches the safety of the best tool clinicians already trusted. Everything below is an appraisal against that line.

Setting the pretest probability: Wells and YEARS#

The first job is to grade how likely a clot is before any test. Two tools dominate.

The Wells score converts clinical judgment into points. It rewards findings such as signs of a leg clot, a heart rate over 100, recent surgery or immobilization, and a prior venous thromboembolism, plus one telling item: whether the clinician thinks pulmonary embolism is the single most likely explanation. That last line writes intuition directly into the arithmetic. Patients who land below the threshold can be cleared with a normal D-dimer. The strength of Wells is that it makes reasoning explicit. Its weakness is downstream: D-dimer rises with age, pregnancy, infection, cancer, and surgery, so a fixed cutoff turns positive in exactly the older patients who are hardest to clear, sending them to the scanner anyway.

YEARS trims the list to three questions, signs of a leg clot, blood in the sputum, and whether pulmonary embolism is the most likely diagnosis, then does something clever with the threshold. Think of it less as a fixed gate and more as a sliding one. With none of the three items present, the D-dimer cutoff rises to 1000; with any present, it holds at 500. In the prospective Dutch cohort published in The Lancet in 2017, this excluded pulmonary embolism without imaging in nearly half of patients, up from about a third under the older Wells-plus-fixed-D-dimer approach, while the failure rate among those managed without a scan stayed near half a percent.

Deciding whether to test at all: PERC#

PERC works one layer earlier than the others. Instead of interpreting a D-dimer, it asks whether ordering one is even worth it. The rule is eight yes-or-no findings: age under 50, pulse under 100, oxygen saturation above 94 percent, no one-sided leg swelling, no blood in the sputum, no recent trauma or surgery, no prior clot, and no estrogen use. When a clinician has already judged the risk low and all eight are satisfied, PERC says stop, because a D-dimer here would generate more false alarms than the clots it would catch.

The best evidence for PERC is the PROPER randomized trial published in JAMA in 2018. Emergency departments were assigned to either a PERC-based strategy or usual care, and the three-month rate of a later clot was 0.1 percent with PERC against 0 percent with conventional testing, a gap that met the trial's noninferiority margin. That is the kind of result that licenses withholding a test: not a perfect record, but a miss rate low enough that the imaging avoided is a fair trade.

Fixing the threshold, not the test: age-adjusted D-dimer#

The age-adjusted D-dimer targets the exact weakness in the Wells pathway. Rather than a flat cutoff of 500, it raises the ceiling for anyone over 50 to their age multiplied by ten, so a 75-year-old may register a D-dimer up to 750 before imaging is triggered. The ADJUST-PE study in JAMA in 2014 validated this across thousands of patients. Among those sent home without a scan because they fell below the age-adjusted line, the three-month rate of venous thromboembolism stayed well under one percent, while the adjustment let far more older patients skip imaging than the fixed number allowed. It is a small edit to a threshold with an outsized effect on how many scans happen.

Reading the failure-rate evidence honestly#

The question that matters is not which rule is most elegant but whether skipping the scan is safe, and safety has that precise meaning: the three-month clot rate in the untested group. That is the number to ask for whenever someone tells you a pathway is safe. A pragmatic Australian study in Emergency Medicine Australasia in 2019 wove Wells, PERC, and age-adjusted D-dimer into a single flowchart and measured real practice. Imaging fell by roughly 40 percent, the share of scans that actually found a clot rose, and no missed clots were traced to the pathway. That is the signature these rules are engineered to produce: fewer scans, a higher hit rate, and a failure rate held under the accepted ceiling.

None of this replaces judgment. These rules are validated in adults with suspected pulmonary embolism. They are not built for pregnancy without a dedicated pathway, not for people already on blood thinners, and not as a reason to ignore a patient who simply does not fit the low-risk picture. Ask which population a rule was validated in before you carry it anywhere else. A decision rule is a floor for safe reasoning, not a ceiling on it.