Clinical background
In acute pulmonary embolism the central early question is not only whether the patient should receive anticoagulation, but where it should be given and how closely the patient should be monitored. Patients in manifest shock are by definition high risk and are managed acutely with reperfusion treatment. Among normotensive patients, however, there is a heterogeneous group in which the majority do well on anticoagulation alone, while a minority suffer haemodynamic decompensation or pulmonary embolism-related death within 30 days. It is this minority that must be identified in order to justify hospital care with monitoring, not infrequently on a high-dependency unit.
The Bova score was developed precisely to single out this intermediate-high-risk group among normotensive patients. It rests on four variables available clinically at presentation, each of which has been shown to be independently associated with early complications: a mildly reduced systolic blood pressure, tachycardia, a raised troponin and signs of right ventricular dysfunction on imaging [1]. The score is therefore a complement to, not a replacement for, the clinical low-risk assessment with, for example, sPESI or the Hestia criteria.
Calculating the Bova score
The score is the sum of four variables:
where is an indicator function taking the value 1 if the condition is met and 0 otherwise. The score ranges from 0 to 7 and is divided into three stages: stage I (0 to 2 points), stage II (3 to 4 points) and stage III (more than 4 points).
The derivation cohort consisted of 2,874 normotensive patients with acute symptomatic pulmonary embolism, assembled from individual patient data from six prospective studies [1]. The primary outcome was a composite of pulmonary embolism-related death, haemodynamic collapse or recurrent pulmonary embolism within 30 days, which occurred in 198 patients (6.9%). The four variables were identified by multivariable logistic regression with adjusted odds ratios of 2.45 for a systolic blood pressure of 90 to 100 mmHg, 1.87 for a heart rate of at least 110 beats/min, 2.49 for a raised troponin and 2.28 for right ventricular dysfunction. The odds ratio for complications per point was 1.55 (95% CI 1.43 to 1.68) [1].
Interpretation in practice
Stage-specific complication rates differ between the derivation cohort and later validation studies. The largest external validation, Fernández 2015, reported higher rates in stages II and III than the derivation [2]. The calculator displays the figures from the Fernández cohort, since they come from an independent population and therefore better reflect what can be expected outside the derivation setting.
| Stage | Score | 30-day complications | PE-related mortality | Clinical action |
|---|---|---|---|---|
| I | 0–2 | 4.4% | 3.1% | Standard anticoagulation; hospital care if otherwise indicated, but no need for intensified monitoring on the basis of the score alone |
| II | 3–4 | 18% | 6.8% | Hospital care with regular monitoring of vital signs; consider a high-dependency unit if further risk factors are present |
| III | >4 | 42% | 10.5% | Close monitoring, preferably on an intensive care or high-dependency unit; be prepared for reperfusion treatment if the patient deteriorates haemodynamically |
It is important to note that even stage I carries a complication rate of 4.4% in the validation cohort. The Bova score is therefore not a tool for identifying patients who can safely be sent home, but for grading risk among patients who already have a confirmed pulmonary embolism and who are not in shock. Decisions on early discharge should be made with sPESI or the Hestia criteria, not with the Bova score.
Validation and performance
Fernández et al. validated the Bova score retrospectively in 1,083 normotensive patients diagnosed in the emergency department of a Spanish university hospital [2]. The distribution was 80% in stage I, 15% in stage II and 5% in stage III. Complication rates rose with increasing stage (4.4%, 18% and 42% respectively), and interobserver reliability was high (ICC 0.93, κ 0.80). Pulmonary embolism-related mortality was 3.1%, 6.8% and 10.5% respectively. Compared with the derivation cohort, complication rates were higher in stages II and III, which may reflect differences in the patient population or in the definition of the outcome.
A meta-analysis of nine studies with a total of 8,342 patients reported pooled complication rates of 3.8%, 10.8% and 19.9% for stages I, II and III [3]. The pooled AUC for the composite outcome was 0.73 (±0.09). Pooled sensitivity for stage III was low, 0.25 (95% CI 0.22 to 0.29), while specificity was high, 0.93 (95% CI 0.92 to 0.93). This means that the score is better at ruling out high risk than at identifying it: a stage III result is fairly specific, but the majority of patients who suffer complications do not fall into stage III.
In a post hoc analysis of 868 normotensive patients in the Italian IPER registry, the Bova score performed less well than both the ESC 2019 algorithm and the modified FAST score [4]. The AUC for an in-hospital adverse outcome was 0.64 (95% CI 0.55 to 0.73) for Bova compared with 0.69 (95% CI 0.58 to 0.79) for ESC 2019 and 0.67 (95% CI 0.59 to 0.76) for the modified FAST score. The Bova score failed to discriminate between intermediate-low and intermediate-high risk in this cohort: the odds ratio for an adverse outcome in the intermediate-high compared with the low-risk group was 1.6 (95% CI 0.7 to 3.7, p = 0.266) and therefore not statistically significant [4].
In the largest external validation to date, 5,036 haemodynamically stable patients in the prospective multicentre COPE study, the c-statistic for the Bova score was 0.741 for pulmonary embolism-related death [5]. This was comparable to ESC 2014 (0.712), PEITHO (0.738) and NEWS2 (0.742), but lower than TELOS (0.777). The positive predictive value for all models was below 10%, with the exception of TELOS and NEWS2, while the negative predictive value exceeded 98% for Bova [5].
A retrospective study of 531 patients with acute pulmonary embolism confirmed the pattern: the Bova score had high specificity (90.4%) for 30-day mortality and a high negative predictive value (90.4%), but low sensitivity and a low positive predictive value [6].
Limitations
The Bova score applies only to normotensive patients with acute pulmonary embolism. Patients with a systolic blood pressure below 90 mmHg or signs of shock are by definition high risk and should not be stratified with the Bova score, but managed according to a high-risk protocol with immediate reperfusion treatment.
The score requires both troponin and imaging (echocardiography or CT) to be calculated in full. If either investigation is unavailable at presentation, stratification is delayed, and treating missing values as normal, as has been done in some validation studies [4], may lead to underestimation of risk.
The low sensitivity is the most important clinical limitation. Several studies show that the Bova score misses a substantial proportion of patients who later suffer complications [3, 4]. This applies particularly to patients who fall into stage II, where the complication rate in the validation cohort was 18% [2], a level that hardly justifies regarding an intermediate classification as reassuring. In the IPER cohort the score was unable to discriminate at all between intermediate-low and intermediate-high risk [4].
The Bova score has not proved superior to the ESC risk stratification algorithm. In the comparative studies published, the ESC 2019 algorithm generally performs better or equivalently [4, 5]. The Bova score may, however, be simpler to calculate in the acute situation since it does not require sPESI to be computed as a separate step, and it gives a continuous score rather than a binary classification.
References
- Bova C, Sanchez O, Prandoni P, et al. Identification of intermediate-risk patients with acute symptomatic pulmonary embolism. Eur Respir J 2014;44(3):694-703. PMID: 24696111
- Fernández C, Bova C, Sanchez O, et al. Validation of a Model for Identification of Patients at Intermediate to High Risk for Complications Associated With Acute Symptomatic Pulmonary Embolism. Chest 2015;148(1):211-218. PMID: 25633724
- Chen X, Shao X, Zhang Y, et al. Assessment of the Bova score for risk stratification of acute normotensive pulmonary embolism: A systematic review and meta-analysis. Thromb Res 2020;193:99-106. PMID: 32534329
- Hobohm L, Becattini C, Konstantinides SV, et al. Validation of a fast prognostic score for risk stratification of normotensive patients with acute pulmonary embolism. Clin Res Cardiol 2020;109(8):1008-1017. PMID: 32025793
- Becattini C, Vedovati MC, Colombo S, et al. Identification of hemodynamically stable patients with acute pulmonary embolism at high risk for death: external validation of different models. J Thromb Haemost 2024;22(9):2502-2513. PMID: 38810699
- Surov A, Thormann M, Bär C, et al. Validation of clinical-radiological scores for prognosis of mortality in acute pulmonary embolism. Respir Res 2023;24(1):195. PMID: 37543614
