Author: Allison Lee, PGY-2
Background:
- Risk scoring patients with acute PE can identify:
- patients low-risk for adverse outcomes (and are candidates for outpatient treatment) VS
- patients high-risk (and require inpatient/critical care treatment)
- Two of the tools most widely used are the
- Simplified Pulmonary Embolism Severity Index (sPESI) and
- Hestia criteria.
- Multiple studies have validated the sPESI and the Hestia criteria, however, studies have not compared the performance of these two criteria.
- N.B: The expected mortality rate for low-risk sPESI patients is 1% and that for low-risk Hestia patients is 4% within 30 days after the diagnosis of acute PE.
Article Reviewed:
Clinical Question:
- How do the sPESI and the Hestia Criteria compare for predicting short-term mortality in “low-risk” patients with PE?
- How does interobserver reliability compare for the sPESI and the Hestia Criteria for classifying low- vs. high-risk patients with acute PE?
Population:
- Patients with acute symptomatic PE confirmed by CT, V/Q, or LE ultrasound positive for a proximal DVT in patients with an inconclusive V/Q scan
Outcomes:
- Primary:
- 30-day all-cause mortality after diagnosis of PE
- Secondary:
- Interobserver reliability of low- vs high-risk score for sPESI and Hestia criteria
- Intervention:
- All patients were scored with both sPESI and Hestia Criteria by 2 clinicians
- Design:
- Single-center, prospective, cohort
- Inclusion Criteria:
- Objectively confirmed PE and able to give consent
Primary Results:
Critical Findings:
Strengths:
- Included a wide range of patients with a diverse comorbidity profile
Limitations:
- Study cohort had low event rates
- Small cohort (n=488), single center study in Spain
- The two clinicians scored the patients have expertise in PE diagnosis, and Hestia criteria has some subjective items — may not apply to the general population who are using the scoring systems
- Study did not address nonfatal events including early recurrent VTE and major bleeding event rates
Discussion:
- In comparison to the Hestia criteria, the sPESI had better interobserver reliability, greater objectivity, and a lower burden of administration (fewer # of terms)
- sPESI classified similar proportion of patients as low risk (28%) compared to Hestia criteria (27%)
- Patients classified as low risk by sPESI and Hestia criteria had a similar rate of all-cause mortality (sPESI 1/135 or 0.7%, Hestia 3/132 and 2.3%).
- 2 patients who died in the Hestia low risk group had cancer, which automatically placed them in the sPESI high risk group
- Patients classified as high risk by sPESI (30/353, or 8.5%) and Hestia (28/356, 7.9%) had similar 30 day all-cause mortality
- The odds ratio for 30-day mortality for high-risk compared to low-risk groups was 12.5 for sPESI and 3.7 for Hestia.
- Kaplan-Meier analysis showed that the survival difference between low- versus high-risk groups was statistically significant for both sPESI and Hestia criteria
- In total, 1.4% of patients were classified differently using the sPESI and Hestia criteria
- Interobserver reliability of sPESI was very good (K=0.97) with disagreement about 1.2 of patients
- Interobserver reliability of Hestia was good (K=0.80) with disagreement in 7.6% of patients.
- sPESI has 6 clearly defined terms that are objective. Hestia has 11 terms and many are subjective and interpretation may vary depending on clinician expertise.
- Authors’ Conclusions:
- Both the sPESI and the Hestia criteria performed well- similar classification and prognostic accuracy for 30-day mortality after PE.
- In this cohort (n=488), patients defined as low-risk by sPESI had 0.7% rate of mortality and by Hestia, 2.3% mortality.
- Patients defined as high-risk by sPESI had 8.5% mortality and by Hestia, 7.9% mortality.
- sPESI may have more reproducible results among clinicians that vary in expertise in comparison to Hestia criteria
- Clinical Bottom Line:
- sPESI and Hestia are both clinically valid scoring systems and have similar prognostic accuracy.
- However, the Hestia criteria has more subjective items than sPESI and may not apply well to all clinicians that vary in expertise.
- sPESI was shown in this study to have better interobserver reliability