The Role of Echocardiographic Findings in Estimating Survival Probability of Intensive Care Unit Admitted Aluminum Phosphide Poisoned Patients

Abstract

Introduction

Cardiotoxicity represents the primary cause of death in acute aluminum phosphide (AlP) poisoning. Prompt supportive care can improve patient survival. This study assessed the role of echocardiography in estimating the survival probability of AlP-poisoned patients admitted to the intensive care unit.

Methods

A prospective cohort study of symptomatic acute AlP poisoned patients was conducted between September 2019 and December 2020. Patients were subjected to history taking, clinical examination, To be included, patient evaluation needed to include electrocardiographic (ECG) and echocardiographic studies. The statistical analysis assessed the association between patient survival and relevant factors. Survival analysis was performed using the Kaplan–Meier survival curve and Cox proportional hazard regression.

Results

A total of 90 patients met inclusion criteria. Electrocardiographic abnormalities were detected in 38.1% of survivors and 82.6% of non-survivors (p < 0.001). Survivors had a higher mean left ventricle ejection fraction (LVEF) (50.86 ± 6.30% vs. 26.52 ± 7.64%, respectively, p < 0.001) and a lower percentage of global LV hypokinesia (4.8% vs. 94.2%, p < 0.001). The mean survival time was higher among patients with LVEF ≥ 50% than those with LVEF = 41–49% and ≤ 40% (p = 0.014 and 0.001, respectively). The hazard of death was 4.42 and 5.40 times greater in patients with LVEF ≤ 40% or with global LV hypokinesia, respectively. Regression revealed that the global LV hypokinesia, ECG abnormalities, and decreased LVEF and oxygen saturation were significantly associated with the risk of death (hazard ratios: 4.382, 3.348, 0.957, and 0.971, respectively).

Conclusions

Echocardiography represents a valuable diagnostic tool to assess cardiac function in acute AlP poisoning. Both LVEF and global LV hypokinesia significantly impact the survival of AlP-poisoned patients. Echocardiography was superior to ECG changes in terms of accuracy for the prediction of mortality.

Background

Aluminum phosphide (AlP) is a metal phosphide that is widely used as a pesticide. Easy availability, effectiveness, and low price made AlP an ideal pesticide to protect stored grains in several agricultural countries. It is marketed in the form of grey or dark brown 3 g tablets (Phostoxin, Bhostoxin, Quickphos, Phosphume, and Phostek) or 0.6 g tablets (Quickphos, Alphos, and Cellphos) [1]. In Egypt, AlP is commonly used in grain storage industry and is known among the public as “the wheat pill” [2].

Unfortunately, the same factors which favored the use of AlP as a pesticide made it a popular means of suicide via ingestion of the tablets [3]. Suicidal acute AlP poisoning is a considerable problem in many developing countries, including Egypt [2]. The mortality rate of acute AlP poisoning is high, ranging from 40 to 80% [4].

When AlP comes in contact with water or gastric HCL, phosphine (PH3) gas is liberated and is rapidly absorbed. Several mechanisms of toxicity are proposed, with the primary one being mitochondrial dysfunction related to the inhibition of oxidative phosphorylation by PH3, thereby reducing the formation of ATP and producing reactive oxygen species (ROS) [5, 6]. In addition, PH3 inhibits the detoxifying enzymes catalase and peroxidase resulting in oxidative stress, lipid peroxidation, and DNA damage [7]. Mostly, all body organs are affected by PH3 toxicity, but the heart is one of the main target organs [8].

Several reports indicate that cardiotoxicity is the principal mechanism of death in AlP-poisoned patients [7, 9, 10]. Further, studies have shown that toxic cardiac affection due to different toxic agents can significantly predict mortality [11–13].

Previous studies have described cardiac toxicity in patients with acute AlP poisoning through the assessment of cardiac biomarkers [14–16], ECG [14, 17, 18], or echocardiography [17, 19–21]. However, the utility of echocardiography as a predictor of survival in AlP-poisoned patients has not been investigated.

Early identification of patients with AlP-induced cardiac toxicity is vital as the observed cardiogenic shock in those patients is temporary and can be managed with adequate supportive care [22, 23]. Therefore, the present study aimed to assess the role of echocardiography in estimating the survival probability of intensive care unit (ICU) admitted AlP-poisoned patients.

Methods

Study Design, Settings, and Ethical Considerations

This prospective cohort study started from 1 September 2019 to the 31 December 2020 after the approval of the Research Ethics Committee, Faculty of Medicine, Tanta University (approval number: 33333/09/19). We assigned a code for each participant to maintain the confidentiality of the patients’ personal and clinical data. Written informed consent was signed from each patient (or his/her guardian) before recruitment and after an explanation of the aim of the study.

We recruited patients from the Tanta University Poison Control Centre (TUPCC), Emergency hospital, Tanta University. It is a tertiary-care center that provides emergency and supportive treatment for all acute poisoning patients, with the referral of the patients to the ICU for completing their treatment as conditions dictate. The ICU is a multidisciplinary unit at the Emergency hospital, Tanta University, that belongs to the Emergency Medicine and Traumatology Department.

Inclusion Criteria

All adult patients with the age of 18 years or older, of either sex, with symptomatic acute AlP poisoning and were admitted to the ICU.

Diagnosis of acute AlP poisoning depended on the history of a single exposure to AlP, the suggestive clinical manifestations, besides the documentation of the compound by requesting its container from the patient attendants. Further, we performed the silver nitrate test for detecting phosphine gas in the gastric aspirate [24].

Exclusion Criteria

We did not include patients who gave a history of acute AlP exposure, but no symptoms appeared during their in-hospital follow-up. A history of concomitant ingestion or exposure to other substances in addition to AlP and associated comorbidities including cardiovascular diseases, renal, or hepatic failure dictated exclusion from the study. Pregnant and lactating women and patients who received treatment for acute AlP poisoning in any hospital before admission at TUPCC were also excluded.

Data Collection

We consecutively enrolled all patients who fulfilled the eligibility criteria. We designed a specific data collection sheet to prospectively record the patients’ data. The collected data were obtained on admission to the ICU, and they included demographic data (age and sex), poisoning data (circumstances and route of poisoning, the alleged amount, and the estimated time interval between AlP exposure and the admission to the TUPCC), the first obtained values of the Glasgow coma scale (GCS), vital signs including systolic blood pressure (SBP), diastolic blood pressures (DBP), pulse rate, and respiratory rate. The GCS was categorized into normal (GCS = 15), reduced consciousness (GCS = 9–14), or coma (GCS = 8 or less). As well, the first obtained values of the arterial blood acid–base balance (pH: reference range 7.35–7.45), bicarbonate concentration (HCO3: reference range 22–26 mEq/L), partial pressure of carbon dioxide (P_aCO2: reference range 35–45 mmHg), and oxygen saturation (SaO2: reference range 95–100%) were recorded.

Upon admission in the ICU ward, we performed a standard 12-lead electrocardiogram (ECG) at a paper speed of 25 mm/s by using CardiMax FCP-7101 ECG machine (Fukuda Denshi Co., Ltd., Tokyo, Japan). The assessment was done by the consultant cardiologist from printed ECG. This included the rhythm (sinus or not); detection of extrasystole; atrial fibrillation; ventricular tachycardia or other types of dysrhythmias; measurement of the PR, QRS, and QT intervals; evaluation of the shape of waves; and detection of conduction abnormalities (such as heart block or bundle branch block). Any deviation from normal ECG was considered ECG abnormality. Additionally, an epicardial echocardiography was performed by using Affinity 50G Digital ultrasonic imaging system (Philips, Eindhoven, The Netherlands) to assess the cardiac function. The echocardiographic parameters were collected and evaluated by a consultant cardiologist in our team within 3 h of admission. In twenty out of the ninety included patients, another consultant cardiologist was available at the same time, and the echocardiographic parameters were reassessed. The main echocardiographic abnormalities were documented. The LVEF was categorized according to the guidelines of the European Society of Cardiology [25] into preserved (LVEF ≥ 50%), borderline (41–49%), and reduced (≤ 40%). Global left ventricular (LV) hypokinesia was defined as LVEF of less than 40% in all standard echocardiographic views with no visible segmental wall motion abnormalities.

Follow-up data included time from hospital admission to death or discharge from the hospital (h). Finally, we recorded the patient outcome at the discharge time from the hospital as either complete uncomplicated recovery (group 1 or survivors) or in-hospital death (group 2 or non-survivors).

Treatment

The attending clinical toxicologist managed all acute AlP poisoned patients according to the local TUPCC protocol. Treatment included the emergency care of the airway, breathing, and circulation concomitant with proper decontamination measures done by gastric lavage using normal saline mixed with sodium bicarbonate solution (2 ampoules sodium bicarbonate 25% added to each 500 cc saline) for patients presenting within 2 h of tablets ingestion. Then, a single (50 gm) dose of activated charcoal was administered.

Hypotension and refractory shock were treated by intravenous fluids (guided by central venous pressure measurement) and vasopressors (norepinephrine) IV infusion. Additionally, metabolic acidosis was corrected by intravenous sodium bicarbonate. Magnesium sulfate was administered in the dose of 1 g IV infusion every 1 h for the first 3 h, followed by 1–1.5 g every 6 h for 24 h. The ICU admission was decided for patients who needed close vital monitoring and hemodynamic stabilization and/or mechanical ventilation.

Statistical Analysis

The sample size was calculated by G power 3.1.9.4 software program depending on the assumed differences in the mean LVEF between survivors and non-survivors. Based on the assumptions of a Cohen’s effect size of 0.8, 5% margin of error, power of 90%, non-survivors to survivor’s ratio of 3 to 1 (based on the assumed mortality rate of 75.0%), and a two-sided hypothesis.

Statistical analysis and presentation of data were conducted using SPSS (Statistical Package for the Social Sciences) version 22 computer program. Categorical data were represented as numbers and percentages. A Chi-square test was applied to investigate the association between the categorical variables. Alternatively, Fishers’ exact test was applied when the expected cell counts were less than 5. For continuous data, they were tested for normality by the Shapiro–Wilk test. Normally distributed data were expressed as mean ± standard deviation and were analyzed by the independent samples T-test. On the other hand, skewed data were summarized as the median and interquartile range (25th–75th percentiles), and the Mann–Whitney U test was used for comparison. Pearson’s or Spearman’s rank order correlations were performed. To assess the relationship between echocardiographic findings and survival probability, the Kaplan–Meier survival curves were compared using the log-rank test. Then, we performed a Cox proportional hazard regression model (backward stepwise) to determine the survival while controlling for possible confounding baseline characteristics in all patients, including the alleged amount (tablet), time from toxic ingestion to hospital admission (h), GCS, vital signs, and arterial blood gases results. We reported the hazard ratios (HR) for survival, and 95% confidence intervals (CIs). The sensitivity, specificity, and accuracy of predicting mortality using ECG and echocardiography were assessed. The sensitivity was calculated as the percentage of truly predicted non-survivors. The specificity was calculated as the percentage of truly predicted survivors. The accuracy represented the percentage of truly predicted non-survivors and survivors out of all studied patients. Additionally, Pearson’s correlation coefficient and Cohen’s kappa statistics were calculated for the LVEF and global LV hypokinesia to assess the interrater reliability. A p < 0.05 was considered statistically significant.

Results

Figure 1 illustrates the flow diagram of the enrollment of patients in the study. Patients who were admitted to our institution starting from the 1 September 2019 were screened until the end of December 2020. During this period, a total of 173 patients were assessed for potential eligibility, and out of them, 95 patients were eligible. Out of the 95 eligible patients, 5 were not recruited because echocardiography was not performed due to the unavailability of the cardiology consultant. Finally, 90 patients were included in the study and followed up until discharge from the hospital or death, with no loss to follow-up.

Fig. 1.

Flow diagram of the study participants.

This study included 90 ICU admitted acute AlP poisoned patients. Survivors (group 1) were 21/90 (23.3%), while non-survivors (group 2) were 69/90 (76.7%). The highest percent were females (55/90, 61.1%), and their ages ranged from 18 to 53 years, with a median age of 20.0 (IQR = 18.0–26.0), with no significant differences between both groups. All patients alleged suicidal ingestion of AlP tablets. The median of the alleged ingested amount was significantly lower in survivors than non-survivors (0.5 vs 1.0 tablet, p = 0.018). The survivors exhibited a significantly longer median time passed from toxic ingestion to hospital admission (3.5 vs 2.0 h, p = 0.013). All the patients reported vomiting at least once after the ingestion. The means of the SBP (106.7 ± 16.5 vs. 71.7 ± 16.2) and the DBP (69.0 ± 10.0 vs. 43.3 ± 10.8) were significantly higher in survivors in comparison to non-survivors (p < 0.001), whereas the pulse and the respiratory rates showed a significantly lower means in survivors than non-survivors (p = 0.001). Additionally, there were significant differences between groups 1 and 2 regarding the results of arterial blood gases including pH, PaCO2, HCO3, and SaO2, with significantly higher mean values in survivors (p < 0.05). Eight out of 21 (38.1%) patients in group 1, compared to 57/69 (82.6) patients in group 2, developed ECG abnormalities, with a statistically significant difference (p < 0.001). The main recorded ECG abnormalities included sinus tachycardia (76%), atrial fibrillation (20%), ST segment depression (30%), ST segment elevation (10%), T wave inversion (23%), ventricular tachycardia (8%), and bundle branch block (2%).

Concerning the echocardiographic assessment, the interrater reliability was 0.91 and 0.90 for LVEF and global LV hypokinesia, respectively. The main abnormal findings in the studied patients included reduced LVEF and global LV hypokinesia. The mean LVEF was 50.86 ± 6.30% in group 1 compared to 26.52 ± 7.64% in group 2 (p < 0.001). All patients in group 2 had an LVEF below 50% and the majority had an LVEF ≤ 40% (66/69, p < 0.001). Global LV hypokinesia was detected in 1/21 (4.8%) patient in group 1 and most patients (65/69, 94.2%) in group 2. Further, the median time from hospital admission to death or discharge was 48 h in group 1 compared to 12 h in group 2, with a significant difference (p < 0.001) as demonstrated in Table 1.

Table 1.

Comparison between survivors and non-survivors regarding demographic, toxic, clinical, and echocardiographic parameters. Clinical parameters were assessed on admission and echocardiography was performed within 3 h of admission.

		Group 1 Survivors (n = 21, 23.3%)	Group 2 Non-survivors (n = 69, 76.7%)	Total (n = 90)	P value
Gender; female	n (%)	12 (57.1)	43 (62.3)	55 (61.1)	0.670
Age (years)	Median (IQR)	24.0 (18.0–30.0)	19.0 (18.0–23.0)	20.0 (18.0–26.0)	0.083
Alleged amount (tablet)	Median (IQR)	0.50 (0.50–1.00)	1.0 (1.0–1.0)	1.0 (0.50–1.0)	0.018*
Time from toxic ingestion to hospital admission (h)	Median (IQR)	3.5 (1.5–5.0)	2.0 (1.0–3.0)	2.0 (1.0–4.0)	0.013*
GCS	15 n (%)	7 (33.3)	0 (0.0)	7 (7.8)	< 0.001*
	9–14 n (%)	14 (66.7)	69 (100)	83 (92.2)
Systolic blood pressure	Mean ± SD	106.7 ± 16.5	71.7 ± 16.2	79.8 ± 22.0	< 0.001*
Diastolic blood pressure	Mean ± SD	69.0 ± 10.0	43.3 ± 10.8	49.3 ± 15.2	< 0.001*
Pulse rate	Mean ± SD	90.8 ± 19.1	107.7 ± 20.6	103.7 ± 21.4	0.001*
Respiratory rate	Mean ± SD	22.3 ± 5.1	27.6 ± 6.8	26.3 ± 6.8	0.001*
PH	Mean ± SD	7.41 ± .05	7.27 ± .09	7.30 ± .10	< 0.001*
PaCO2	Mean ± SD	32.2 ± 8.6	26.0 ± 9.0	27.5 ± 9.2	0.006*
HCO3	Mean ± SD	19.6 ± 4.4	12.6 ± 4.5	14.2 ± 5.3	< 0.001*
SaO2	Mean ± SD	90.9 ± 6.5	75.2 ± 16.4	78.9 ± 16.1	< 0.001*
ECG abnormalities	n (%)	8 (38.1)	57 (82.6)	65 (72.2)	< 0.001*
LVEF (%)	Mean ± SD	50.86 ± 6.30	26.52 ± 7.64	32.20 ± 12.68	< 0.001*
≥ 50% (preserved)	n (%)	14 (66.7%)	0 (0.0%)	14 (15.6%)	< 0.001*
41–49% (borderline)	n (%)	5 (23.8%)	3 (4.3%)	8 (8.9%)
≤ 40% (reduced)	n (%)	2 (9.5%)	66 (95.7%)	68 (75.6%)
Global LV hypokinesia	n (%)	1 (4.8)	65 (94.2)	66 (73.3)	< 0.001*
Time from hospital admission to death or discharge (h)	Median (IQR)	48.0 (48.0–70.0)	12.0 (8.0–18.0)	13.5 (9.0–48.0)	< 0.001*

^*Significant at p < 0.05; GCS Glasgow coma scale, PaCO2 partial pressure of carbon dioxide, HCO3 bicarbonate, SaO2 oxygen saturation, LVEF left ventricle ejection fraction, LV left ventricle

Table 2 shows that LVEF had a significant positive, strong correlation with SBP (r = 0.866, p < 0.001) and DBP (r = 0.853, p < 0.001) and a significant moderate positive correlation with pH (r = 0.521, p < 0.001), HCO3 (r = 0.554, p < 0.001), and SaO2 (r = 0.356, p < 0.001). The LVEF correlated negatively and weakly with the amount of AlP (r = -0.284, p = 0.007).

Table 2.

Correlation between LVEF and each of age, ingested amount, blood pressure, serum bicarbonate, and oxygen saturation.

		Age	Amount	Systolic Bp	Diastolic Bp	pH	HCO3	SaO2
LVEF (%)	r	0.084a	-0.284b	0.866a	0.853a	0.521a	0.554a	0.356b
	p	0.433	0.007*	< 0.001*	< 0.001*	< 0.001*	< 0.001*	0.001*

LVEF left ventricle ejection fraction, r correlation coefficient, ^acoefficient of Pearson’s correlation, ^bcoefficient of Spearman’s rank-order correlation, *significant at p < 0.05

Kaplan–Meier survival analysis of the studied patients according to the LVEF and global LV hypokinesia is shown in Table 3 and Figs. 2 and 3. The mean survival time was significantly higher among patients who had an LVEF ≥ 50% (72 h) than those with an LVEF 41–49% (52 h, p = 0.014) and ≤ 40% (18.10 h, p = 0.001), and a significant difference was detected also between the latter two categories (p = 0.001). None of the patients with LVEF ≥ 50% died, while the risk of death was higher with LVEF ≤ 40% compared to LVEF of 41–49% (HR, 4.42; 95%CI, 2.11 to 9.28). The global LV hypokinesia was a significant risk factor for death; the hazard of death was 5.40 times greater in patients who showed global LV hypokinesia (HR = 5.40; 95%CI, 3.21–9.09). Figures 2 and 3 also show the overall survival percentage at intervals of 12 h (12, 24, 36, 48, etc.).

Table 3.

Kaplan–Meier survival analysis according to the left ventricle ejection fraction and the global hypokinesia.

Echocardiographic parameters		Mean survival time in hours (95%CI)	Hazard ratio (95%CI)	Log-rank test
LVEF	≥ 50%	72.00 (72.00–72.00)	-	< 0.001*a EF ≥ 50 vs. 41–49% = 0.014* EF ≥ 50 vs. ≤ 40% = 0.001* EF 41–49 vs. ≤ 40% = 0.001*
	41 to 49%	52.00 (30.40–73.60)	0.23 (0.11 to 0.47)
	≤ 40%	18.10 (14.01–22.19)	4.42 (2.11 to 9.28)
Global LV hypokinesia	Yes	17.60 (13.65–21.56)	5.40 (3.21–9.09)	< 0.001*
	No	64.84 (54.85–74.84)	0.184 (0.109–0.311)

^*Significant at p < 0.05; CI confidence interval, LVEF left ventricle ejection fraction, LV left ventricle, vs versus; ^a denotes the overall comparison of the survival curves among the three categories of ejection fraction before performing pairwise comparisons

Fig. 2.

Kaplan–Meier survival analysis curves according to the left ventricle ejection fraction.

Fig. 3.

Kaplan–Meier survival analysis curves according to the global LV hypokinesia.

Backward stepwise multivariable Cox regression for the hazard of death revealed a model that consisted of LVEF, global LV hypokinesia, ECG abnormalities, and SaO2 as significant risk factors contributing to death. The presence of global LV hypokinesia (HR = 4.382, 95%CI 0.078–17.806, p = 0.039) and ECG abnormalities (HR = 3.348, 95%CI 1.500–7.475, p = 0.003) was significantly associated with an increased risk of death, while the increase in LVEF (HR = 0.957, 95%CI 0.922–0.993, p = 0.020) and SaO2 (HR 0.971, 95%CI 0.954–0.989, p = 0.002) was associated with a reduced risk of death as shown in Table 4.

Table 4.

Backward stepwise multivariable Cox regression model for the hazard of death.

Model	Beta coefficient	p Value	Hazard ratio	95%CI
LVEF	-0.044	0.020*	0.957	0.922–0.993
Global LV hypokinesia	1.477	0.039*	4.382	1.078–17.806
ECG abnormalities	1.208	0.003*	3.348	1.500–7.475
SaO2	-0.029	0.002*	0.971	0.954–0.989

^*Significant at p < 0.05; CI confidence interval, LVEF left ventricle ejection fraction, LV left ventricle

The sensitivity, specificity, and accuracy of electrocardiographic and echocardiographic parameters as predictors of mortality are shown in Table 5. The highest accuracy was obtained by global LV hypokinesia (94.4%), followed by LVEF < 50% (92.2%), and lastly using ECG (77.8%).

Table 5.

Sensitivity, specificity, and accuracy of electrocardiographic and echocardiographic parameters as predictors of mortality.

	TP (n)	FP (n)	TN (n)	FN (n)	Sensitivity (%)	Specificity (%)	Accuracy (%)
ECG abnormalities	57	8	13	12	82.6	61.9	77.8
LVEF < 50	69	7	14	0	100.0	66.7	92.2
Global LV hypokinesia	65	1	20	4	94.2	95.2	94.4

FN false negative, FP false positive, n number of patients, TN true negative, TP true positive, LVEF left ventricle ejection fraction, LV left ventricle

Discussion

Several studies have documented the effect of AlP on the heart and proposed that cardiotoxicity is the primary cause of death in cases of acute AlP poisoning. Prompt supportive care for the patients who exhibit cardiac effects could improve their survival. Consequently, there is a need to identify those patients early. Several cardiac biomarkers along with ECG have been evaluated in such patients. Nevertheless, only a few studies have assessed echocardiography; the majority of which were case reports. We hypothesized that echocardiography could detect more accurately early cardiac dysfunction in such patients. Therefore, the present study aimed to assess the role of echocardiography in estimating the survival probability of the ICU-admitted AlP-poisoned patients.

This study included 90 patients with acute AlP poisoning. The survivors constituted 23.3%, while non-survivors were 76.7%. The reported mortality rate of acute AlP poisoning varies widely in the literature, ranging from 40% up to 100% [3, 14, 18, 26–28].

Several factors can influence the likelihood of mortality following exposure to AlP, including the dose, freshness of tablets, vomiting, comorbidities, and delay before receiving medical care [15]. In the present study, the univariate statistical analysis revealed that the median alleged ingested dose was significantly lower in survivors than non-survivors (0.5 versus 1.0 tablet, p = 0.018). Moreover, the survivors spent a significantly longer median time from the toxic ingestion until hospital admission (3.5 versus 2.0 h, p = 0.013). This difference in delay time could be explained by the less severe symptoms that initially appeared in survivors due to the lower ingested dose, whereas the non-survivors had more severe toxic manifestations that urged them to seek medical care earlier. This explanation is supported by the differences in clinical manifestations on admission between the two groups, as the survivors showed significantly higher mean GCS, systolic and diastolic blood pressure, pH, bicarbonate level, and oxygen saturation.

The metabolic acidosis detected in this study, which was more profound in the non-survivors, is regarded as the hallmark of AlP toxicity, and numerous studies have reported its association with mortality [17, 29, 30]. Metabolic acidosis increases the risk of mortality due to its depressive effect on the contractility of the myocardium resulting in cardiogenic shock [31].

Cardiotoxic manifestations develop in 60 to 100% of AlP-poisoned patients in the form of hypotension, shock, arrhythmias, myocarditis, and pericarditis [32]. The mechanism of cardiotoxicity in such patients includes the direct toxic action of PH3 and the effect of metabolic acidosis on the myocardium. The PH3 inhibits the cytochrome oxidase and disrupts oxidative phosphorylation, resulting in cellular damage and inducing cell death [8].

Cardiovascular affection was manifested in the current study more profoundly in the non-survivors, as observed from blood pressure monitoring, ECG changes, and echocardiographic assessment. The survivors exhibited significantly higher means of the SBP and the DBP compared to those of the non-survivors, which is in line with the results reported by previous studies [17, 29].

The current study showed also that the survivors had a significantly lower rate of ECG abnormalities compared to the non-survivors. The development of ECG changes in AlP poisoning is attributed to the focal necrosis of myocardial cells and the changes in membrane action potentials, leading to the appearance of ischemic like changes in the ST segment and the T wave [4, 18, 33, 34]. Previous studies have reported that ECG changes were significantly associated with mortality in AlP poisoned patients [14, 17, 18, 35, 36].

Concerning the echocardiographic findings, the survivor group in the present study had a significantly higher mean LVEF than that of the non-survivor group (50.86 ± 6.30% vs. 26.52 ± 7.64%, p < 0.001). In addition, global LV hypokinesia was observed in only one patient (4.8%) in the survivor group compared to 94.2% of non-survivors. Some case reports [19, 22, 23, 34, 37] and two cohort studies [15, 17] have shown similar findings in echocardiography. They have also described that the changes in LVEF and wall motility are reversible and improve over few days.

The current study went further into assessing the impact of these echocardiographic changes on the survival of patients. Univariate analysis using the Kaplan–Meier survival curve demonstrated that both LVEF and hypokinesia significantly impacted the survival of patients. The mean survival time was significantly longer among patients who had LVEF ≥ 50% than those with an LVEF of 41–49% or ≤ 40%. The estimated relative risk of death was 4.42 and 5.40 times significantly greater in patients who had an LVEF ≤ 40% and global LV hypokinesia, respectively. The value of LVEF as a validated measure of systolic function has been demonstrated by Klaeboe and Edvardsen [38]. They have also confirmed a unique position of LVEF as a criterion in different therapeutic trials in cardiology. Further, comparable studies on carbon monoxide poisoning reported associations between various echocardiographic findings and the severity and outcomes of these patients [39, 40]. Another prospective study included that all hospitalized acute poisoned patients showed a significant reduction in the LVEF in many types of poisonings including pesticide exposures with a significant association with in-hospital mortality [13].

Backward stepwise multivariable Cox regression for the hazard of death revealed a model consisted of LVEF, global LV hypokinesia, ECG abnormalities, and SaO2 as significant factors contributing to death. Global LV hypokinesia (HR = 4.382, 95%CI: .078–17.806) and ECG abnormalities (HR = 3.348, 95%CI: 1.500–7.475) were significantly associated with a risk of death. The increase in LVEF or SaO2 was significantly associated with a reduced risk of death (hazard ratios: 4.382, 3.348, 0.957, and 0.971, respectively). This finding is in line with Lionte and Bologa [41] who concluded that ECG and echocardiographic assessment of acutely poisoned patients on admission to a medical ward or the ICU could accurately predict the worse outcomes and mortality.

The use of echocardiography in the assessment of AlP-poisoned patients provides several benefits to the management of patients. The availability of this diagnostic modality as a bedside test in the emergency department enables the evaluation of patients on admission and serially if the patients’ condition mandates further monitoring. It provides comprehensive information about the anatomy and function of the heart. The present study demonstrated that LVEF and global LV hypokinesia had higher sensitivity and specificity in predicting mortality than ECG changes. Echocardiography can effectively guide the treatment of cardiogenic shock [42], as the injudicious administration of intravenous fluids to treat shock in AlP-poisoned patients would place volume overload on the affected heart resulting in decompensated shock [43].

The present study was subject to some limitations. The diagnosis of AlP poisoning was based on history combined with clinical manifestations, examination of the container of poison, and silver nitrate test. The sensitivity of silver nitrate test ranges from 50 to 100% [44], and confirmation by gas chromatography was not available in our center. Likewise, the exclusion of concomitant ingestion of other poisons was based on history and clinical evaluation as a comprehensive toxicological evaluation was not available. Moreover, the serial assessment of echocardiographic parameters was not available in all patients. In addition, cardiac biomarkers were not measured. We compared the time to death in the non-survivors and time to hospital discharge in the survivors; however, other factors may impact the length of hospital stay in survivors other than the AlP-induced cardiotoxicity until discharge from the hospital. It was not feasible to investigate the clinical significance of different ECG abnormalities due to the occurrence of life-threatening arrhythmias took place in a small number of patients and were accompanied by variable combinations of other arrhythmias. It is worth to mention that fluid administration was guided by the CVP as echocardiography is not performed routinely in our ICU and was done for research purposes in this study.

In conclusion, echocardiography represents a valuable diagnostic tool to assess cardiac function in patients with acute AlP poisoning. Both LVEF and global LV hypokinesia significantly impact the survival of AlP-poisoned patients. The accuracy of echocardiographic parameters was superior to that of ECG changes for the prediction of mortality. The conduction of further studies to assess echocardiographic-guided management of AlP poisoning and its impact on patient outcomes is recommended, preferably with serial measurement of cardiac biomarkers and echocardiographic parameters that evaluate both systolic and diastolic functions of the right and left sides of the heart.

Author Contribution

FME and ASAFH conceived the study, contributed to the idea and the design of the research, and drafted the manuscript. FME also performed the statistical analysis of the data as well as edited and revised the manuscript in its final form. MAS shared the study design, collection of the data, AND interpretation of the results. OMS was responsible for performing and interpreting the echocardiography findings. All authors contributed substantially to the revision of the final manuscript.

Sources of Funding

None

Declarations

Conflict of Interest

^{The authors declare no competing interests.}