Article Text
PDF
Statistics from Altmetric.com
Q In patients with pulmonary embolism (PE), can those who are at low risk of short term mortality (and candidates for outpatient treatment) be identified from medical history and clinical findings?
Clinical impact ratings GP/FP/Primary care ★★★★★★☆ Internal medicine ★★★★★★☆ Emergency medicine ★★★★★★☆
METHODS
Design:
separate derivation, internal validation, and external validation cohorts.
Setting:
186 non-governmental hospitals in Pennsylvania, USA (derivation and internal validation cohorts) and 3 university hospitals in Switzerland and France (external validation cohort).
Patients:
15 531 inpatients ⩾18 years of age (45% ⩾70 y of age, 60% women) with a primary discharge diagnosis of acute PE were randomly divided into derivation (n = 10 354) and internal validation (n = 5177) cohorts. The external validation cohort consisted of 221 patients (49% ⩾70 y of age, 55% women) with confirmed PE who had presented to the emergency department and been enrolled in a diagnostic study.
Description of prediction guide:
the prediction rule was (i) age ⩾70 years; (ii) any of the comorbid conditions: cancer, heart failure, chronic lung disease, chronic renal disease, or cerebrovascular disease; or (iii) any of the clinical features: pulse ⩾110 beats/min, systolic blood pressure <100 mm Hg, altered mental status, or arterial oxygen saturation <90%. Patients with none of these factors were defined as low risk.
Outcomes:
all cause mortality at 30 days. Secondary outcomes were death within 7 days and the composite end point of non-fatal cardiogenic shock or cardiorespiratory arrest during the initial hospital stay.
MAIN RESULTS
The prediction rule identified as low risk 22% of the derivation cohort, 22% of the internal validation cohort, and 34% of the external validation cohort. The table shows the risks of outcomes in the predicted low and high risk groups of each cohort. The positive likelihood ratios were moderate (1.3 to 1.5), but the negative likelihood ratios of the prediction rule for 30 day mortality were strong at 0.06 in the derivation cohort, 0.15 in the internal validation cohort, and 0 in the external validation cohort.
Event rates (95% CI) by prediction rule risk classification in pulmonary embolism*
CONCLUSION
A simple prediction rule, based on history and clinical variables, accurately identified patients with pulmonary embolism who had a low risk of short term mortality and non-fatal adverse events.
Commentary
Risk stratification in pulmonary embolism is needed to select appropriate diagnostic and therapeutic strategies. The Aujesky rule is an accurate risk prediction tool and a promising one as a therapeutic decision aid. Because it is simpler and easier to use than existing rules using laboratory or imaging data, it will be useful in a wide range of settings. Its external validation adds strength to the guide’s ability to classify patient risk, but it should be noted that the external validation population showed lower mortality, even in the high risk group. However, the rule retained its ability to separate groups of different risk accurately, facilitating its extrapolation to other populations and contexts. It is interesting that the guide made an accurate prediction without using other variables that have been shown to be associated with adverse outcomes (eg, biochemical markers, echocardiographic findings, and history or presence of deep vein thrombosis).1
There is evidence that ambulatory treatment of selected patients with pulmonary embolism is safe.2 Goodman suggested that patients with small or peripheral pulmonary embolism without deep venous thrombosis and good cardiopulmonary reserve could be left untreated as they probably have a low risk of recurrence,3 but this has not yet been shown. The Aujesky rule may help clinicians select patients with pulmonary embolism who are at low risk and could be treated with simpler interventions outside the hospital. However, the ultimate proof of its clinical utility as a therapeutic decision aid will require a randomised controlled trial testing the hypothesis that ambulatory care is as efficacious as usual management in patients selected with this simple and well validated decision rule.
Footnotes
-
For correspondence: Dr D Aujesky, University of Lausanne, Lausanne, Switzerland. aujesky{at}swissonline.ch
-
Source of funding: US National Heart, Lung, and Blood Institute.