N-acetylcysteine Use in Treatment of Acute Aluminium Phosphide Poisoning: Systematic Review and Meta-Analysis

Background: Aluminum phosphide (AlP) is a popular used rodenticide. It inhibits oxidative phosphorylation, and causes depletion of glutathione, resulting in cellular wall dysfunction. N-acetylcysteine (NAC) is a glutathione precursor that would be effective in treatment of AlP poisoning. Aim of the work: Provide evidence based systematic review about role of NAC in treatment of AlP poisoning which may help in developing clear guidelines for treatment of such lethal poisoning. Methodology: We followed PRISMA guidelines during preparation of this study. PubMed, EKB, ScienceDirect and Cochrane CENTRAL were searched to identify the published literature from inception to June 2022. In addition, we searched for ongoing studies, reference lists for additional studies. We included randomized cotrolled trials (RCTs) and observational studies (OSs) published in English, those fulfilling inclusion criteria. Results : The study included four RCTs and two OSs with total 286 participants. The current study revealed that there was a significant reduction in mortality rate (OR 0.38, 95% CI [0,23 to 0.66]) and duration of hospital stay in survivors (SMD -1.73 days, 95% CI [-2.35, -01.11]) as well as a significant increase in survival time in non survivors in patients who received NAC, compared with those who did not receive NAC (SMD 0.87 day, 95% CI [0.37, 1.37]). There was no significant difference between NAC and control groups regarding the need for mechanical ventilation (OR 0.51, 95% CI [0,23 to 1.10]). Conclusion: N-acetylcysteine in treatment of acute AlP poisoning can reduce the mortality rate and duration of hospital stay in survivors and increase survival time


Introduction
odenticides are considered a global challenge to public health. Annually, 250,000 to 370,000 people die from deliberate ingestion of pesticides, which is responsible for about one-third of suicidal attempts worldwide (Manouchehri, et al., 2019).
Phosphides are normally found as powders or pellets, usually in the form of zinc or aluminium phosphide (Zn3P2 and AlP, respectively), Calcium and magnesium phosphides are also available (Altintop and Tatli, 2017).
Aluminum phosphide is a highly popular indoor and outdoor pesticide used in many developing countries to protect grain in stores and during transportation. Even 500 mg of this compound can be fatal for humans with mortality rates as high as 70-100% in various studies (Nourbakhsh, et al., 2019).
The toxicity of aluminium phosphides is due to production of deadly phosphine gas in contact with water or diluted acids. Phosphine gas is typically produced within 30 minutes of phosphide consumption (Yan et al., 2018). The main mechanisms of toxicity are electron transfer blockage and non-competitive inhibition of cytochrome oxidase c, which inhibits oxidative phosphorylation, and in turn, cellular respiration resulting in activation of peroxide radicals. In addition, phosphine can inhibit catalase and deplete glutathione, resulting in cellular wall dysfunction (Ari et al., 2022).
Metal phosphides can result in serious systemic poisoning; cardiovascular collapse and cardiogenic shock may occur due to their direct effects on myocytes, intravascular fluid leakage into the third space, severe metabolic acidosis, and poor tissue perfusion (Bansal et al., 2017).
N-acetyl cysteine (NAC) is a novel thiol compound, commonly used as a mucolytic agent, and a precursor of L-cysteine and reduced glutathione (GSH). In addition, NAC is a source of sulfhydryl groups in cells and free radical scavenger as it interacts with reactive oxygen species (ROS) such as OH and H2O2 (Colovic et al., 2018).
Although NAC is widely known as an antidote to acetaminophen overdose, it has multiple other uses supported by various levels of evidence. These diverse clinical applications are linked to its ability to support the body's antioxidant and nitric oxide systems during stress, infections, toxic assault, and inflammatory conditions (Tenório et al., 2021).
Although phosphide is well known as a lethal poison with neither an available effective antidote nor a specific treatment (Abdelhamid et al., 2023), in animal studies, NAC has been shown to have a protective role against phosphide-induced cardiovascular complications by protecting myocytes from the oxidative stress induced by phosphine, thus stabilizing blood pressure and pulse with dramatic improvement of outcome (Asghari et al., 2017). In addition, human studies revealed that NAC decreases mortality rates, length of hospitalization, and the frequency of intubation and mechanical ventilation after phosphide poisoning (ELabdeen et al., 2020). So, it is important to do systematic review of the existing studies about NAC usage in acute aluminium phosphide poisoning to assess its efficacy in treating such lethal condition.

Aim of the Work
Provide evidence based systematic review about role of NAC in treatment of phosphide poisoning which may help in developing clear guidelines for treatment of such lethal poison.

 Study design:
This is a systematic review and meta-analysis study. We followed PRISMA statement guidelines during preparation of this systematic review and metaanalysis.  Criteria for considering studies for this review: A. Only articles fulfilling the inclusion criteria were included for further steps of data collection, analysis, and reporting. We recorded the selection process in detail to complete a PRISMA flow diagram. b) Data extraction and management: Data were extracted independently by the first and the fifth authors and any discrepancies were resolved by five-member discussion and consultation with the original study. For missing information, we contacted the trial's authors for incomplete data. We extracted the following study characteristics and outcome data from the included studies: Methods; study design, study setting, date and duration of study, participants; mean age, age range, gender, severity of the condition, inclusion and exclusion criteria, intervention; intervention, comparison, and any co interventions, outcomes; specified and collected outcomes, time points reported, notes; comments on quality of studies, notable conflicts of interest of trial authors, funding of trial. III. Assessment of risk of bias in included studies: Risk of bias was assessed by the first author then revised by the other four authors according to recommendations of the Cochrane Handbook for Systematic Reviews of interventions (Higgins et al., 2019).
A. Assessment of risk of bias in randomized controlled trials: We used COCHRANE ROB tool for randomized clinical trials studies. For each domain, we judged the risk of bias as low, high, or unclear if there was insufficient information to assess risk of bias. We resolved any disagreement with five-member discussion. The following definitions were used in the assessment of risk of bias in RCTs: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. If the trial had been assessed at low risk of bias in all the above domains, we judged it as having low risk of bias. If the trial had been assessed at unclear or high risk of bias in one or more of the above domains, we judged it as having high risk of bias. B. Assessment of risk of bias in observational studies: We also used Newcastle Ottawa Scale (NOS) to assess quality and risk of bias of observational studies. Newcastle Ottawa Scale is a 9-star scale for observational studies assessing the quality of selection (maximum 4 stars), comparability (maximum 2 stars), and outcome in cohort studies or exposure in case control studies (maximum 3 stars). We judged the study as good quality, fair quality or poor quality as follows: Good quality: 3 or 4 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain; Fair quality: 2 stars in selection domain AND 1or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain; Poor quality: 0 or 1 star in selection domain OR 0 stars in comparability domain OR 0 or 1 star in outcome/exposure domain. IV. Measures of outcomes: For evaluation of the dichotomous outcomes (mortality rate and need for mechanical ventilation), we recorded the total number of people with one or more events within each study and we presented comparisons between groups as odds ratio [with corresponding 95% confidence intervals (CIs)] instead of risk ratio to resolve heterogeneity that appeared when using risk ratio in meta-analysis. For continuous outcomes (duration of hospital stay in survivors and survival time in non survivors), we recorded the mean, standard deviation, and total number of people in both groups of each study, and we presented comparison between groups as standard mean difference [with corresponding 95% confidence intervals (CIs)]. V. Dealing with missing data: We contacted trial's authors for clarification about missing data in identified publication reports and then incorporated data when provided by the authors. We do all analyses according to the intention-totreat principle by including all participants who were randomized in the statistical analysis and analyzing them according to the group they were originally assigned, irrespective of compliance or follow up (McCoy, 2017).

VI. Assessment of statistical heterogeneity:
Heterogeneity which is a significant variation in the effect size of the included studies was assessed by the following tests: Cochrane Q chi square test: P-value < 0.1 is a statistically significant test, donated heterogeneity among the studies, and I-squared (I 2 ) index which is calculated as follows: as Q is cochrane Q chi square and df is degree of freedom. The I-squared is interpreted as follows: 0% to 40%: might not be important. 30% to 60%: may represent moderate heterogeneity. 50% to 90%: may represent substantial heterogeneity. 75% to 100%: considerable heterogeneity ( (Egger et al., 1997).  Summary of finding and assessment of the certainty of the evidence: We assessed confidence in the evidence from included RCTs using GRADE criteria (GRADEpro GDT) an online guideline development tool (http://gradepro.org), and we constructed 'Summary of findings' table that included our review outcomes and comparisons. We assessed five factors referring to limitations in the study design and implementation of included studies that suggest a high likelihood of bias: Study risk of bias, indirectness of evidence (population, intervention, control, outcome), unexplained heterogeneity or inconsistency of results, imprecision of results (wide confidence intervals), and high probability of publication bias (Schünemann et al., 2020).
The certainty of evidence is defined as the following: High certainty, we are very confident that the true effect lies close to that of the estimate of effect. Moderate certainty, we are moderately confident of the effect estimate. Low certainty, our confidence in the effect estimate is limited. Very low certainty, we have very little confidence in the effect estimate.

Results
We included six studies in our review; four randomized controlled trials (Tehrani et al., 2013;Bhalla et al., 2017;El-ebiary and Abufad, 2017;Emam et al., 2020) and two observational studies; a cohort study (Agrawal et al., 2014), and a case-control study (Taghaddosonijad et al., 2016) with total 286 participants of whom 145 received NAC. Four of the included studies (Agrawal et al., 2014;Taghaddosonijad et al., 2016;Bhalla et al., 2017;Emam et al., 2020) used NAC with a dose of 300 mg/kg intravenous over about 20 to 21 hour (continous infusion or divided as150 mg/kg over one hour then 50 mg/kg over four hours then 100 mg/kg over 16 hours) and the two remainder studies (Tehrani et al., 2013;Elebiary and Abufad, 2017) used intravenous NAC with a dose of 1.33gm/kg over 72 hour divided as (140 mg/kg as a loading dose then 70 mg/kg every four hours for 17 doses). The included studies were published between 2013 and 2020. Two studies were conducted in India, two in Iran, and two studies were carried out in Egypt. Baseline characteristics of the populations of the included studies are shown in (Table 1) and the summary of their designs and their main results are shown in (Table 2).
Three studies were excluded with reasons as follows; Bhat and Kenchetty (2015) was about rodenticides in general and AlP was not specified; Abdel-hady et al., (2019) had no details about the participants who received NAC, no additional data were received when contacted trials' authors; Tawfik (2020) was an abstract, so we contacted the author who send us the original study which was about metal phosphide poisoning (aluminium phosphide and zinc phosphide) and there were no isolated data about AlP alone. In addition, three ongoing studies (Irct20200724048192N1, NCT04509258, and NCT05370729) were excluded.

I. Results of the search
The results of our searches are detailed in a PRISMA diagram (figure 1). Our electronic searches retrieved 966 records. Searching of other resources produced five additional references. After removing 66 duplicate references by endnote reference manager, we evaluated a total of 905 records, of which we excluded 893 based on the title and the abstract using Rayyan online site. The remaining 12 records were checked as full texts; three studies might be eligible as ongoing; further information is in the ongoing studies. We excluded three studies with reasons.

II. Risk of bias of included studies
Risk of bias within studies was assessed by Risk of bias tool for RCTs using (Revman 5.4) and by Newcastle Ottawa Scale (NOS) for non-randomized studies. According to our protocol, when a single domain was assessed at high or unclear risk, the trial was classified as being at high risk. As demonstrated in the risk of bias assessment (figures 2&3), we classified the four RCTs to be at overall high risk of bias. Authors' judgement with justification are shown in Supplementary File N.1. Two observational studies were assessed by NOS; one of them was of good quality (Agrawal et al., 2014) while the other study was of poor quality (Taghaddosinejad et al., 2016) (Table 4).
We also did subgroup analysis according to type of studies (RCTs or OSs). The RCTs subgroup included four studies and revealed a statistically significant difference favoring NAC group (subtotal effect size 0.34, 95% CI [0.17 to 0.68], P = 0.002). In contrast, OSs subgroup included two studies and revealed a statistically non-significant difference between both groups (subtotal OR = 0.46, 95% CI [0.20 to 1.08], P = 0.08). Pooled studies in each subgroup were homogenous. Intergroup difference was not significant (Chi-square P = 0.58, I 2 = 0%) ( Figure 5).

IV. Certainty of evidence
The certainty of evidence was mentioned in the methodology and summarized in the summary of findings for the four RCTs using GRADE criteria (GRADEpro GDT) online Guidelines Development Tool (Table 4).

(4 RCTs) Moderate
Comments: Our confidence in this result is moderate, downgraded one level for serious risk of bias (single RCT study at high risk of bias in allocation concealment and the three studies at high risk of bias in blindness of participants and personnel).
N-acetylcysteine is likely to reduce mortality in population with acute aluminum phosphide poisoning.

Comments:
This outcome was not clearly reported in the included studies Duration of hospital stay in survivors -SMD 1.73 SD fewer (2.35 fewer to 1.11 fewer) -61 (2 RCT) Moderate Comments: Our confidence in this result is moderate, downgraded one level for serious risk of bias (the included study was with high risk of bias at both allocation concealment and blindness of participants and personnel). N-acetylcysteine is likely to reduce duration of hospital stay in acute aluminum phosphide survivors.

Number of participants (studies)
Certainty of the evidence (GRADE)

Risk with [control]
Risk with [nacetylcysteine] Duration of hospital stay in non survivors -SMD 0.27 SD fewer (0.3 fewer to 0.84 more) -63 (2 RCT) Very Low c Comments: Our confidence in this result is very low, downgraded one level for serious risk of bias (single study had unclear risk if bias at allocation concealment and selective reporting and high risk of bias at blindness of participants and personnel), downgraded one level for serious imprecision (wide confidence intervals crossing the line of no effect), and downgraded one level due to serious inconsistency (heterogeneity between studies p <0.00001). N-acetylcysteine may have no effect on survival time in patients expired after acute aluminium phosphide poisoning. Mechanical ventilation 48 per 100 32 per 100 (18 to 51) OR 0.51 (0.23 to 1.10)

Comments:
Our confidence in this result is moderate, downgraded one level for serious risk of bias (one study had high risk of bias at allocation concealment and two studies at high risk of bias in blindness of participants and personnel). Nacetylcysteine may have no effect on need for mechanical ventilation in population with acute aluminum phosphide poisoning.     Figure 8: Forest plot showing the difference between NAC and control groups as regards the need for mechanical ventilation with subgroup analysis according to NAC regimen.

Discussion
A. Summary of main results This systematic review of N-acetylcysteine usage in acute aluminium phosphide poisoning included six studies: four RCTs and two OSs with total 286 participants.
The overall meta-analysis found that Nacetylcysteine could reduce the mortality rate which was reported in the six included studies and hospital stay duration in survivors which was reported in two of the included studies with a statistically significant difference.
Although quantitative analysis of duration of hospital stay in non survivors (survival time) which was reported in three studies showed significant prolongation, we could not consider meta-analysis because of significant unresolved heterogeneity between studies (Chi-square P =< 0.00001, I 2 = 95%). The cause of heterogeneity may be due to the difference between the included participants in each study as Bhalla et al. (2017) included only patients with severe toxicity manifested by hypotension and shock which would affect the survival time. Another reason may be due to difference in type of studies, two were RCTs and one was a cohort observational study.
Meta-analysis for the difference between NAC and control groups as regards need for mechanical ventilation revealed that NAC did not affect this outcome. Pooled studies were with moderate heterogeneity (Chi-square P = 0.18, I 2 = 43%). This heterogeneity was resolved by subgroup analysis according to NAC regimen.
Our subgroup analyses according to NAC regimens showed that 300 mg/kg IV NAC over 21 h (300 mg/kg over 20 h or 150 mg/kg over one hour then 50 mg/kg over four hours then 100 mg/kg over 16 hours) significantly decreased both mortality rates and duration of hospital stay in survivors. On the other hand, it significantly increased the survival time in non survivors, but there were significant heterogeneity (Chi-square P =< 0.00001, I 2 = 95%).
While the other regimen of 1.33 g/kg IV NAC over 72 h regimen (140 mg/kg as a loading dose then 70 mg/kg every 4 h up to 17 doses) significantly reduced each of the following: the mortality rate, duration of hospital stay in survivors, and the need for mechanical ventilation There was no difference between the two NAC regimens as regards reduction of both the mortality rate and duration of hospital stay in survivors.
Our subgroup analysis according to type of studies revealed that RCTs subgroup showed that NAC usage resulted in a significant reduction in mortality rate and hospital stay duration in survivors but did not affect both survival time and the need for mechanical ventilation.
While OSs subgroup showed that NAC usage resulted in a significant increase in survival time in non survivors, it did not affect the mortality rate. B. Quality of the evidence The certainty of evidence (quality of evidence) of the four included RCTs was summarized in the summary of evidence and as follows: Regarding primary outcomes, the certainty of evidence for the mortality rate was moderate, it was downgraded one level due to serious risk of bias of the included studies. While certainty of evidence for morbidity rate was not reported.
Regarding secondary outcomes, the certainty of evidence for duration of hospital stay in survivors was moderate, it was downgraded one level due to serious risk of bias of the included study. The certainty of evidence for duration of hospital stay in non survivors was very low, it was downgraded one level due to serious risk of bias, one level due to serious imprecision and one level for serious inconsistency. Finally, the certainty of evidence for mechanical ventilation was moderate, it was downgraded one level due to serious risk of bias. C. Potential biases in the review process We performed this review according to a predefined protocol, following guidance from the Cochrane Handbook for Systematic Reviews of Interventions, which we completed and published prior to beginning of the review process. We used a comprehensive search strategy to minimize possible publication bias. It is unlikely that this strategy missed any published studies or large unpublished studies. We could not formally evaluate publication bias due to the small number of trials identified.
We included both randomized clinical trials and observational studies to identify as large as possible data published on our topic, and this is one of the limitations in our study. Two studies (40%) are observational studies, one of them of poor quality and observation time was 24 hours from admission (Taghaddosonijad et al., 2016).
To overcome this limitation, subgroup analyses of RCTs alone were done and we created summary of findings (certainty of evidence) for only the four included RCTs.
Another limitation was different NAC regimens and doses. To overcome this issue, subgroup analyses according to NAC regimen were done.
The last detected limitation was significant, unresolved heterogeneity between studies reported duration of hospital stay in non survivors (survival time). D. Agreement and disagreement with other studies or reviews This study is the first systematic review and meta-analysis done in this topic.
A randomised controlled trial published after June 2022 (our search limit) was done by Ashraf and his colleagues to determine the effect of NAC on mortality rate in AlP acutely intoxicated patients. It was conducted in Lahore, Pakistan with 96 participants; 48 in each group (NAC and control). The study revealed a significant reduction in mortality rate favoring NAC group (p =0.024) which agreed with our study (P =0.0005) (Ashraf et al., 2022).
Recent systematic review and meta-analysis was done by Rashid and his colleagues about NAC use in rodenticide poisoning including yellow phosphorous, zinc phoshsphide, aluminium phosphide and others (Rashid et al., 2022).
Mortality in Rashid et al., 2022 study showed that meta-analysis of RCTs (OR: 0.25; 95% CI: 0.11-0.59; n = 2) and retrospective studies (OR: 0.34; 95% CI: 0.15-0.78; n = 3) showed a significant reduction in mortality, whereas pooled analysis of prospective studies recorded a non-significant effect. And thus, agreed with our study which showed significant reduction in that outcome in meta-analysis of RCTs but non-significant reduction in meta-analysis of OSs.
Unlike our study Rashid et al., 2022 showed a significant reduction of intubation or ventilation (OR: 0.25; 95% CI: 0.11-0.60; 2 RCTs) and a nonsignificant reduction in duration of hospital stay (P = 0.41) between patients who received NAC and who were not treated with NAC. This study is not similar to ours, as rodenticides are many types of different mechanisms of toxicity, but we discussed it because aluminium phosphide is one of the rodenticides.