The predictive value of PaO₂/FIO₂ and additional parameters for in-hospital mortality in patients with acute pulmonary embolism: an 8-year prospective observational single-center cohort study

Acute pulmonary embolism (PE) occurs frequently and may cause death or serious disability [1]. Most deaths in patients with shock occur within the first few hours after admission [2]; therefore, rapid stratification and appropriate treatment are critical to save patients’ lives. It has been a challenge to predict the outcome of patients with acute PE, and risk stratification of patients with acute PE may physicians identify patients who may benefit from additional surveillance or therapy [3, 4]. Unfortunately, none of the clinical prediction tools perform well when applied to all patients with acute PE. Of all the clinical prediction tools evaluated, the one proposed by the European Society of Cardiology (ESC) is the best at risk stratifying patients [5]. However, several studies have still shown that more than 50% of patients with acute PE are hemodynamically stable on admission but have a high risk of death according to clinical models [6‐8]. Additionally, another research study that assessed the ability of the 2014 ESC model to predict 30-day death after acute PE showed that stratification of patients at intermediate risk requires further improvement [9]. As more studies support the hypothesis of PE severity as a clinical continuum, it is important to find more scores, parameters, or biomarkers that would enable more accurate risk stratification in patients with acute PE [10].

Hypoxemia is common in acute PE and a very important mechanism in the pathogenesis of PE that leads to an adverse outcome. Although arterial blood gas analysis has been extensively evaluated in the clinical diagnostic algorithm of acute PE, it is not a criterion for evaluating the disease risk stratification. Studies have shown that risk stratification of patients with PE can be improved by integrating respiratory features into the 2014 ESC model [11, 12]. Among the parameters used to evaluate hypoxemia, arterial partial pressure of oxygen (PaO₂) was not suitable because most patients were supplied oxygen before blood gas analysis was performed. The PaO₂/fraction of inspired oxygen (FIO₂) ratio, which is measured by the arterial partial pressure of oxygen to fraction of inspired oxygen, has been used as criterion for acute respiratory distress syndrome (ARDS) categories [13]. It was postulated that this ratio could be used to predict outcome in ARDS and several diseases [14‐18], but no study has assessed the predictive value of PaO₂/FIO₂ in patients with acute PE.

In this study, we planned to assess the associations of PaO₂/FIO₂ with the risk of in-hospital mortality using a large registry of patients with PE. First, we aimed to assess the relationship between the PaO₂/FIO₂ ratio and in-hospital mortality, determine the optimal cutoff value of PaO₂/FIO₂, and determine if this value, quick and easy to obtain on admission, is a predictor of in-hospital mortality in this population. Second, we aimed to evaluate the potential additional determinants including laboratory parameters that may affect in-hospital mortality.

In this study, 408 patients were screened, 40 patients were excluded (25 had missing inpatient records, 2/25 were transferred to another hospital, and 15 lacked PaO₂/FIO₂ data), and 368 consecutive patients were included in the final analysis (Fig. 1). The median age of the cohort was 76 (IQR 65–83) years, and 179 patients (48.6%) were men. Complete baseline characteristics are presented in Table 1. Among the 368 patients diagnosed with acute PE by lung scintigraphy or CTPA during the study period, 317 (86.1%) had acute PE confirmed by contrast-enhanced CT, and 51 (13.9%) had acute PE confirmed by ventilation/perfusion lung scan. Three hundred twenty-nine patients (89.4%) underwent a sonographic examination of their leg veins. The groups were compared according to their demographic characteristic, medical history, and predisposing factor of PE. Those who survived to hospital discharge had a higher BMI (24.4 vs. 21.2 kg/m², P < 0.001) and higher incidence of cancer (41 vs. 11, P < 0.001) than those who did not. There were no significant differences in medical history between the two groups.

In our study cohort, PaO₂/FIO₂ appeared to be associated with higher in-hospital mortality in patients with acute PE, and the optimal cutoff value of PaO₂/FIO₂ for predicting mortality was 265.

Although there was a significant difference between the survivors and non-survivors in only the pulse rate, there were no significant differences in other signs and symptoms, such as dyspnea, fever, cough, expectoration, etc. (Table 2). There was significant elevation of the WBC count in the non-survivors, which is in accordance with results of another study [20], which showed that the WBC count (OR, 1.9; 95% CI 1.2–3.5) predicted short-term (30-day) mortality following PE. There was a significant elevation of the D-dimer level in the non-survivors in this study Klok [21]. showed that high D-dimer levels were also correlated with centrally located pulmonary emboli and 15-day mortality, and there are also several studies evaluated the correlation between D-dimer levels and the burden of PE [22, 23]. We also found a significant difference between survivors and the non-survivors on the LDH level. Until now, there have been few studies about this biomarker. The cTNI was significantly increased in the non-survivors group in this study, Barrios [7] showed that elevated troponin levels were associated with a high all-cause mortality (OR, 4.3; 95% CI,2.1–8.5%), however, he also pointed out that troponin by itself did not appear to clinically significantly change the pretest to posttest probability of death, and the usefulness of basing therapeutic decision making solely on troponin levels does not appear warranted. So we combined all the above parameters with other prognostic instruments for PE.After adjusting for body mass index, history of cancer, PaO2/FiO2 ratio < 265, pulse rate, cTNI level, LDH level, WBC count, and D-dimer level, we found that the risk stratification, PaO2/FiO2 ratio < 265, and history of cancer continued to be associated with an increased risk of in-hospital mortality of acute PE (Table 4).

Acute PE impairs the efficient transfer of oxygen across the lung and leads to hypoxia, which is a very important mechanism in the pathogenesis of PE. Hypoxia is responsible for physiological consequences including but not limited to tachycardia, dyspnea, peripheral vasodilatation, and increased cardiac output. Hypoxia-mediated vasoconstriction is one of the causes of acute pulmonary hypertension, which is an important mechanism of acute right heart failure in PE [24]. These factors also lead to long-term outcomes like pulmonary hypertension or right ventricular failure [24]. However, hypoxia is not a criterion for evaluating disease risk stratification yet. Studies have shown that risk stratification of patients with PE can be improved by integrating respiratory features into the 2014 ESC model [11, 12]. Among the parameters used to describe hypoxia, PaO₂ is the most common parameter, but it was not suitable for most patients with PE who were supplied oxygen on admission. One study on other parameters provided evidence that oxygen saturation or the respiratory rate could be added to the 2014 ESC strategy for risk stratification in order to further identify hemodynamically stable patients with PE at increased risk for death who are potentially candidates for more aggressive treatment [11]. Another study showed that both sPESI and pulse oximetry measurements are moderately accurate identifiers of low-risk patients with PE [25]. Although the PaO₂/FIO₂ ratio is widely used in the clinical setting for evaluating hypoxemia, there is some doubt about it [26, 27]. The PaO₂/FiO₂ ratio was used as a predictor of outcome in patients with ARDS, congenital diaphragmatic hernia, and cardiac surgery [14‐18]. Studies also showed that the PaO₂/FIO₂ is not an independent predictor of mortality in patients with ARDS in multivariate analyses that controlled for other measures of severity of illness [28]. Because of the mechanism of right-to-left shunt, the variation of the PaO₂/FIO₂ ratio with an increasing FIO₂ is complex [29]. It have been demonstrated that the PaO₂/FIO₂ ratio and its variation with changes in FIO₂ depends on many clinical variables, and it may not be the only parameter for determining the state of arterial hypoxemia. In our cohort, PaO₂/FIO₂ ratio < 265 was related to the in-hospital mortality of patients with acute PE. Low values of the PaO₂/FIO₂ ratio may due to both the pathological conditions of the respiratory disease and alterations in the hemodynamic status of acute PE.

In our study, PaO₂/FIO₂ ratios were lower in patients at risk, so a simple determination of the PaO₂/FIO₂ ratio at <265 may provide important information about poor in-hospital prognosis. It is supposed that the hemodynamically stable patients with PaO₂/FIO₂ ratio < 265 may need more surveillance. According to the 2014 ESC model, high-risk PE is characterized by overt hemodynamic instability and the need for immediate advanced therapy, including consideration of fibrinolysis [30], and further study is warranted to determine whether the patients with a PaO₂/FIO₂ ratio at <265 are potentially candidates for more aggressive treatment.

In summary, the PaO₂/FIO₂ ratio may be useful for identifying in-hospital mortality of patients with acute PE on admission. PaO₂/FIO₂ ratios are lower in patients at risk, and the value of 265 is the cutoff point of predicting in-hospital mortality. A simple determination of the PaO₂/FIO₂ ratio at <265 may provide important information about patients’ in-hospital prognosis, and these patients may need more surveillance. Integrating respiratory features into the 2014 ESC risk stratification model of PE is also important, and future studies or clinical trials will be required to clarify the clinical utility of the PaO₂/FIO₂ ratio in a larger sample of patients.