Contrast Agent Pooling (C.A.P.) sign and imminent cardiac arrest: a retrospective study

Background The sign of contrast agent pooling (C.A.P.) in dependent part of the venous system were reported in some case reports, which happened in the patients before sudden cardiac arrest. Until now, there is no solid evidence enough to address the importance of the sign. This study aimed to assess the accuracy of the C.A.P. sign in predicting imminent cardiac arrest and the association of the C.A.P. sign with patient’s survival. Methods This is a retrospective cohort study. The study included all patients who visited the emergency department, who received contrast computed tomography (CT) scan and then experienced cardiac arrest at the emergency department (from January 1, 2016 to December 31, 2018). We evaluated the occurrence of the C.A.P. sign on the chest or abdominal CT scan, patients with ECMO were excluded. With positive C.A.P. sign, the primary outcome is whether in-hospital cardiac arrest happens within an hour; the accuracy of C.A.P. sign was calculated. The secondary outcome is survival to discharge. Results In the study, 128 patients were included. 8.6% (N = 11) patients had positive C.A.P. sign and 91.4% (N = 117) patients did not. The accuracy of C.A.P. sign in predicting cardiac arrest within 1 h was 85.94%. The C.A.P. sign had a positive association with IHCA within 1 h after the CT scan (adjusted odds ratio 7.35, 95% confidence interval [CI] 1.27 – 42.69). The relative risk (RR) of survival to discharge was 0.90 with positive C.A.P. sign (95% CI 0.85 – 0.96). Conclusions The C.A.P. sign can be considered as an alarm for imminent cardiac arrest and poor prognosis. The patients with positive C.A.P. sign were more likely to experience imminent cardiac arrest; in contrast, less likely to survive. Trial registration IRB No.108107-E. Supplementary Information The online version contains supplementary material available at 10.1186/s12873-022-00634-4.


Introduction
In-hospital cardiac arrest (IHCA) remains a largely unpredictable event, which can happen at any time, and has outcomes that are highly dependent on rapid diagnosis and treatment. Multiple studies suggest that healthcare providers often fail to detect changes or abnormalities in the vital signs of patients, hours before an IHCA event [1]. Currently, there are no evidencebased modalities or examinations for predicting cardiac arrest events. Computed tomography (CT) plays an important role in diagnosing diseases. Dependent venous pooling of CT contrast agent was first described in 2002 [2]. CT features of impending cardiac arrest are characterised by pooling of contrast agent in the dependent parts of the right side of the body, including the venous system, right renal vein, and right lobe of the liver (Fig. 1).
To the best of our knowledge, this is the first study to evaluate the accuracy of the C.A.P. sign in predicting imminent cardiac arrest.
We hypothesise that the risk of occurrence of cardiac arrest, within 1 h, increases from the detection of the C.A.P. sign on the CT scan. Hence, the C.A.P. sign can be considered a new early warning sign to provide in-time interventions for critically ill patients, thereby preventing cardiac arrest incidents.

Study design and setting
This retrospective cohort study was conducted in an emergency department (ED) of a medical centre between 1 January 2016 and 31 December 2018. The study has been approved by the Research Ethics Review Committee (IRB No.108107-E), which is organized under, and operates in accordance with, the Good Clinical Practice guidelines and government laws and regulations. Informed consent is waived as the study involved retrospective charts and images review. The study adheres to STARD statement whenever possible. The medical centre has 1,300 inpatient beds, has 120,000 yearly visits in the ED, and is the only medical centre in the region with a population of 1,205,570 people.

Selection of participants
The study included all patients who visited the ED during the study period, underwent intravenous contrast computed tomography (CT), and then experienced cardiac arrest in the ED. Patients who received extracorporeal membrane oxygenation (ECMO) support before CT scan were excluded from the study.
We collected patients' demographic characteristics (age and sex), underlying comorbidities (hypertension diabetes mellitus, chronic kidney disease, coronary artery disease, heart failure and malignant diseases), the latest vital signs before the CT scan (heart rate and systolic blood pressure), and clinical manifestations (use of inotropic agents and endotracheal intubation). Medians and ranges of age, heart rate, systolic blood pressure were obtained. Shock index was calculated by dividing heart rate by systolic blood pressure. Optiray ® 320 ioversol injection 68% was used as an intravenous contrast agent in our hospital. Its specific gravity is 1.371; the average specific gravity of adult blood is approximately 1.0506 at 37 °C [10]. The results of CT scans were evaluated by two senior emergency physicians (with Cohen's kappa coefficient 0.82) who were masked to C.A.P sign status and patients' outcomes.

Measurements
This study evaluates the occurrence of the C.A.P. sign on the chest or abdominal CT scan. It was defined as positive either when the contrast agent accumulated in the renal or hepatic vein or dependent part of liver, or when the contrast agent layering over vena cava was detected.

Outcomes
The primary outcome was the accuracy of the C.A.P. sign in predicting imminent cardiac arrest (denoted as cardiac arrest occurring within 1 h after the CT scan). The secondary outcome was whether a patient survived up to his or her discharge from the hospital (denoted as survival to discharge) with the presence of C.A.P. sign.

Statistical analysis
Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated by standard formulas for a binominal proportion. The corresponding 95% confidence intervals (CIs) were calculated by the Wilson interval method. The accuracy of the test was determined. The distributions of the selected covariates were compared by the presence of C.A.P. sign. Categorical variables were reported as percentages and analysed with Pearson's chi-squared tests. A multivariable regression model was established to evaluate C.A.P sign as a predictive factor for primary outcome. Additional variables that showed univariable association with the primary outcome was included in the model if evidence of a significance had been confirmed (p < 0.05). A two-tailed p-value < 0.05 was considered to be statistically significant. All statistical analyses were conducted using R software, version 3.5.3.

Characteristics of the study subjects
In the study period, 726 patients experienced IHCA during their stay in the ED of FEMH. Among them, 138 patients underwent CT before cardiac arrest. Ten patients who received ECMO therapy before the CT scan were excluded, and therefore, 128 patients were included in the study.
Demographic characteristics of the study population and outcomes are summarised in Table 1. Most patients were male (60.2%) and were older than 65 years (57%). The profiles of underlying diseases were as follows: history of heart failure (8.6%, N = 11), history of coronary artery disease (11.7%, N = 15), history of chronic kidney disease (8.6%, N = 11), history of diabetes mellitus (22.7%, N = 29), history of hypertension (44.5%, N = 57), and history of malignant diseases (17.2%, N = 22). As for the underlying diseases, there was no significant difference between 2 groups. Before cardiac arrest, heart rates higher than 100 bpm and systolic blood pressure lower than 90 mmHg were detected in 43.8% (N = 56) and 39.1% (N = 50) of the included patients, respectively. A total of 14.1% (N = 18) of the included patients survived to discharge. No patient with a positive C.A.P. sign was diagnosed with diabetes mellitus previously. The heart rates of patients in the group with a negative C.A.P. sign were higher than those of patients in the group with a positive C.A.P. sign (46.2% vs. 18.2%); although, the difference was not statistically significant. Shock index was higher in patients with positive C.A.P sign (1.05 vs 0.85), with most of patients in the group had shock index higher or equal to 0.9 (85.7% vs 43.2%, p = 0.03). Regarding to clinical outcomes, patients with positive C.A.P sign had lower survival rate (0% vs 15.4%, p = 0.16), and are more likely to experience cardiac arrest within 1 h after the CT scan (63.6% vs 12%, p < 0.0001). Table 2 shows the odd ratios and their 95% CIs for selected variables. Among those, cardiac arrest within 1 h after CT scans was significantly associated with positive C.A.P sign (OR 12.88, 95% CI 3.34 to 46.634). Shock index ≧ 0.9 was slightly associated with positive C.A.P sign (OR 7.90, 95% CI 2.92 to 68.22).
The C.A.P sign was positively associated with episode of cardiac arrest within 1 h after CT scans (aOR 7.35, 95% CI 1.27 to 42.69) and negatively associated with survival to discharge (RR 0.9, 95% CI 0.85 to 0.96). Likewise, shock index ≧ 0.9 could be identified as a predictor factor for both clinical outcomes (aOR 1.70, 95% CI 0.47 to 6.07; RR 0.33, 95% CI 0.10 to 1.12), though it was not statistically significant. The results are presented in Table 3.
The clinical and imaging findings, as well as the outcome of the patients with positive C.A.P. sign were summarized in Table 4. Among 11 patients, there were 10 patients (90.9%) with contrast agent pooling at IVC, 6 patients (54.5%) with contrast agent pooling at hepatic veins, 2 patients (18.2%) at renal veins and 1 patient (9.1%) at SVC. No patient in this group was discharged alive.
Changes in IHCA incidence percentage with respect to time are presented in Fig. 3. About 45.5% of the patients with a positive C.A.P. sign collapsed within 30 min after the CT scan. A total of 7 out of the 11 patients (63.7%) with a positive C.A.P. sign collapsed within 1 h after the CT scan. In contrast, only 4.3% of the patients with a negative C.A.P. sign were reported as having IHCA within 30 min after the CT scan, and the incidence percentages increased with time.

Discussion
The results of this study reveal that the presence of the C.A.P. sign on the CT scan strongly associated with the likelihood of imminent cardiac arrest. The patients with positive C.A.P. sign were more likely to experience cardiac arrest within 1 h after the CT scan. In our study, all the patients with positive C.A.P. sign failed to survive; our result also showed that the C.A.P. sign negatively associated with survival to discharge slightly. C.A.P. sign occurs owing to the effect of gravity. The density of contrast agent is much higher than that of blood; thus, during normal physiological blood flow, specific gravity has little effect on the contrast agent dynamics. In patients with cardiogenic shock or impending circulatory collapse, both the arterial and venous blood flow dramatically decreases, and the contrast agent tends to accumulate in the dependent parts of the venous system [11]. Additionally, a lack of sufficient blood pressure during shock also inhibits penetration into the organ parenchyma and contributes to the pooling of contrast agent in major vessels [8]. About 52% of the cases demonstrate this effect of gravity after cardiac arrest [12]. In a retrospective study that reviewed 127 patients who underwent contrast-enhanced CT of the chest or abdomen and echocardiography, high injection rate (> 3 ml/ sec), tricuspid regurgitation, pulmonary hypertension, and right ventricular systolic dysfunction were found to be an independent predictor of retrograde contrast agent pooling over inferior vena cava or hepatic vein (p< 0.001, < 0.01, = 0.05, and < 0.005, respectively) [13]. This study further demonstrated that the phenomenon of contrast agent pooling is caused by weak venous blood flow and may be influenced by contrast agent injection rate.
We found 25 articles, reporting 59 cases with the sign of dependent contrast agent pooling . Age ranges from 6 to 87-year-old. 36 of 59 are male, 21 are female and 2 patients without mentioning of gender. The CT scans with contrast of chest and abdomen were arranged for the survey of trauma (18 cases), aortic dissection, pulmonary embolism, sepsis and abdominal mass, as shown in the supplementary table [Additional file 1].
The site of contrast agent pooling is mostly found over inferior vena cava (IVC). Contrast agent pooling at IVC was described in 55 cases (93.2%). In 4 cases without contrast agent pooling at IVC, 3 cases are traumatic    There was one case (1.7%) who did not experience cardiogenic shock or cardiac arrest despite of positive contrast agent pooling sign; however, extreme low ejection fraction was noted at follow up echocardiography [20].
Four pediatric cases were found in literature review, three of them were collapsed soon after CT scan [17,29]. One case with history of chronic constrictive pericarditis developed neither cardiac arrest nor cardiogenic shock was proven with severe right ventricular failure at follow up echocardiography.
Summarized from literature reviews and our cases, the feature of C.A.P. sign includes contrast agent pooling in IVC, right hepatic vein, left hepatic vein, dependent liver parenchyma, right renal parenchyma, lumbar vein, splenic vein, superior mesenteric vein, superior vena cava, right atrium, right ventricle, coronary sinus, great cardiac vein, azygos vein and hemiazygos vein. The most common site of contrast agent pooling is IVC (93.2% from the reported cases, 90.9% from our case series).
In our study, 63.7% of the patients with a positive C.A.P. sign experienced cardiac arrest within 1 h, and all patients with a positive C.A.P. sign experienced cardiac arrest within 8 h. Our result coincides with those of the previous studies, indicating that C.A.P. sign may be used as a predictor for imminent cardiac arrest.
Contrast-enhanced computed tomography may not be a diagnostic tool for impending cardiac arrest and severe cardiogenic shock. Nevertheless, if the patients underwent CT to survey critical illnesses, the physician has an opportunity to inspect for the C.A.P. sign. Recognition of the sign can effortlessly alert primary care physicians, and rapidly indicate the possibility of extremely low cardiac output or impending circulatory failure. Hence, C.A.P. sign should grab physician's attention to activate timely evaluation and earlier intervention to save the patient's life. In some cases, earlier termination of CT may be necessary to provide immediate cardiopulmonary resuscitation and prevent end-organ ischemia or death. In clinical scenario, if a patient with positive C.A.P. sign is clinically unstable, we propose earlier intubation and inotropic agent usage to stabilize the patient, and rapidly search and treat potential reversible causes.
This study has several limitations. First, the presentation of dependent venous pooling is not well defined. In some cases, contrast agent pooling was noted to be confined to the right hepatic lobe, right renal vein, inferior or superior vena cava, and even subclavian vein. This condition may imply specific haemodynamic disturbances in these cases.
Second, there are no statistical data on the incidence rate of C.A.P. sign in the normal population or the population with low cardiac output. Further studies are needed to investigate the incidence rate of C.A.P. sign in the normal population and in populations with extremely Third, selection bias. Our study focuses on the CT findings of the patients experienced cardiac arrest in emergency department of single medical centre. The statistic values of C.A.P. sign may not apply to all population. However, as the occurrence of C.A.P. sign in normal population is extremely rare, we believe the accuracy of C.A.P. sign in predicting cardiac arrest would be much higher if the study could include more patients.

Conclusions
From our findings, the C.A.P. sign could be addressed as an imaging feature of circulatory failure, it could be used to predict imminent cardiac arrest and should be considered a warning sign for clinical physicians to provide intime interventions for critically ill patients.