Is it just me or have we lost our collective minds?
A recent trial published in JAMA, by Freund et al (1), illustrates just how far our diagnostic psychosis has progressed.
The authors conducted a crossover cluster–randomized non-inferiority trial examining a strategy utilizing the pulmonary embolism rule out criteria (PERC) rule to rule out pulmonary embolism (PE). The authors enrolled patients deemed to be low risk for a pulmonary embolism.
Each center was randomized to start with a 6-month control period (usual care) or 6-month intervention period (PERC based strategy), followed by 6-months of the alternative strategy. Patients were included with new onset or worsening of shortness of breath, chest pain and a low clinical probability of PE, estimated by the treating physician to be below 15%.
In the intervention group, the diagnostic workup included an initial calculation of the PERC score. In patients with a PERC score of zero, the diagnosis of PE was considered ruled out and no further testing was required. If the PERC score was positive, the usual diagnostic strategy was applied. In the control group, the diagnostic workup for PE included the use of D-dimer testing, followed by a CTPA, if the D-dimer was positive. The diagnosis of PE was considered to be excluded if one of these two tests was negative. Any CTPA with emboli was considered positive, including isolated subsegmental PE. If the CTPA was judged inconclusive, the patients would undergo further testing (pulmonary ventilation-perfusion scan or lower-extremity doppler).
1749 patients were enrolled, 902 patients in the control group and 847 in the intervention group. Overall the authors found no difference in their primary outcome, the rate of thromboembolic events at 3-months, in patients who underwent a PERC-based diagnostic strategy (3%) and those who received usual care (3%). All cause mortality was also similar (0.3% vs 0.2%, respectively). Additionally, patients in the PERC-based strategy underwent less CTPAs (13% vs 23%), had shorter ED length of stays, and were admitted to the hospital less frequently.
And while it is easy to get caught up in the methodological success of this study, it is important to put these results in their proper context. We have discussed in a previous post our continued quixotic need to diagnose clinically unimportant pulmonary embolisms in the young and healthy who will likely fare well independent of our diagnostic endeavors, while simultaneously ignoring the sick vulnerable patients in which the diagnosis of PE is far more likely and clinically relevant. The study by Freund et al illustrates this point. For a patient to meet the PERC criteria and not require any further diagnostic testing, he or she must have an oxygen saturation of 94% or more, a pulse rate of less than 100/min, can be no older than 49, demonstrate no signs of unilateral leg swelling, hemoptysis, or have a history of recent trauma or surgery, prior PE or deep venous thrombosis (DVT), or exogenous estrogen use. Essentially the PERC rule is designed to rule out PEs in a subset of patients without the disease in question. It fails to assist in the diagnostic workup for patients truly at risk for clinically relevant disease. Such was the case in the Freund et al cohort. The mean age of this cohort was only 44 years old. Only 3% of the population had chronic respiratory pathology, only 2% had heart failure. Approximately 16% were tachycardic on presentation, only 4.5% tachypneic, and only approximately 5% had an Spo2 < 95%. A history of estrogen use, clinical signs of DVT, a past history of PE or DVT, recent surgery or trauma, hemoptysis, or malignancy was present in a small minority of this cohort. Only 2% of the entire cohort was diagnosed with a pulmonary embolism on their initial visit, with an additional 3% experiencing VTE during follow-up.
Given the rarity of disease in this cohort, the success of the PERC rule should come as no surprise. A simple coin toss would have reduced the use of CTPA as reliably, and likely would have missed just as many pulmonary emboli. In fact, one could argue that this group was already at such a low risk of disease, that they were already at the level of the test threshold, and no further testing should have been offered independent of the individual patients PERC status.
More interesting than the diagnostic prowess of the PERC rule, Freund et al clearly illustrates the dangers of overdiagnosis in our overeager search for pulmonary emboli. In theory, because of its randomized design, the rate of PEs in each group should be similar. But what the authors found was for patients in the traditional diagnostic pathway group, the rate of PE was higher than in the PERC-based strategy (2.7 vs 1.5). And while more PEs were diagnosed on initial presentation, 3 month outcomes were identical in both groups, meaning the PEs that went undiagnosed on initial presentation in the PERC-based strategy were likely clinical meaningless. Simply put, the more often we go looking for a PE, the more frequently we will be rewarded for our efforts. And yet these diagnostic success stories do not translate into improvement in clinically meaningful outcomes. Rather we have identified a radiological disease that has no clinical consequences. One wonders what the rate of disease would have been if there had been a third arm studied in this trial. One in which the diagnostic strategy was simply to do nothing.
I am sure many will view these results as a validation of the PERC rule’s clinical utility. But in reality the Freund et al trial demonstrates just how far we have fallen. We now require a decision aid to determine when we should not pursue the diagnosis of PE in a cohort of patients who clearly have no risk for PE in the first place. This trial illustrates the ability of the Emergency Physician to risk stratify these patients using nothing more than their unstructured clinical judgment. In fact, for patients deemed low risk by the treating physician, appropriate for enrollment into this trial, the risk of PE was so low, no further diagnostic workup was required.
There have been a multitude of diagnostic strategies proposed for the work up of pulmonary embolism, each one claiming to decrease the rate of invasive diagnostic procedures. But how could they not when compared to the absurd diagnostic straw-man of a D-dimer and CT pulmonary angiogram on all comers. From this perspective any such strategy will decrease needless testing. But each pathway represents a slippery slope, easily accessible for the treating physician. These algorithms provide us a veil of certainty to offload the cognitive stress imposed by a patient’s diagnostic dilemma. These diagnostic pathways fail to address what is clearly illustrated by the results presented by Freund et al, the most important decision is the clinician’s ability to say no. In a patient who is already at low enough risk and requires no further work up, any decision aid, no matter its diagnostic performance, will only increase the rate of overdiagnosis and overtreatment.
- Freund Y, Cachanado M, Aubry A, Orsini C, Raynal P, Féral-Pierssens A, Charpentier S, Dumas F, Baarir N, Truchot J, Desmettre T, Tazarourte K, Beaune S, Leleu A, Khellaf M, Wargon M, Bloom B, Rousseau A, Simon T, Riou B, Effect of the Pulmonary Embolism Rule-Out Criteria on Subsequent Thromboembolic Events Among Low-Risk Emergency Department PatientsThe PROPER Randomized Clinical Trial. 2018;319(6):559–566. doi:10.1001/jama.2017.21904