Abstract
Context: Antidote shortage is a global problem. In Thailand, the National Antidote Project (NAP) has operated since November 2010 to manage the national antidote stockpile, educate the healthcare providers on appropriate antidote use, and evaluate antidote usage.
Objective: To evaluate the effect of NAP implementation on mortality rate and antidote use in cyanide poisoning cases arising from ingestion of cyanide or cyanogenic glycoside.
Methods: This is a retrospective cohort of poisoning cases involving cyanide or cyanogenic glycoside ingestion reported to Ramathibodi Poison Center from 1 January 2007 to 31 December 2015. Mortality rate, antidote use, and appropriateness of antidote use (defined as correct indication, proper dosing regimen, and administration within 90 min) before and after NAP implementation were compared. Association between parameters and fatal outcomes was analyzed.
Results: A total of 343 cases involving cyanide or cyanogenic glycoside ingestion were reported to Ramathibodi Poison Center. There were 213 cases (62.1%) during NAP (Project group) and 130 cases (37.9%) pre-NAP implementation (Before group). Implementation of NAP led to increased antidote use (39.9% in Project group versus 24.6% in Before group) and a higher rate of appropriate antidote use (74.1% in Project group versus 50.0% in Before group). All 30 deaths were presented with initial severe symptoms. Cyanide chemical source and self-harm intent were associated with death (OR: 12.919, 95% CI: 4.863–39.761 and OR: 10.747, 95% CI: 3.884–28.514, respectively). No difference in overall mortality rate (13 [10.0%] deaths before versus 17 [8.0%] deaths after NAP) was found. In subgroup analysis of 80 cases with initial severe symptoms, NAP and appropriate antidote use reduced mortality (OR: 0.327, 95% CI: 0.106–0.997 and OR: 0.024, 95% CI: 0.004–0.122, respectively). In the multivariate analysis of the cases with initial severe symptoms, presence of the NAP and appropriate antidote use independently reduced the risk of death (OR: 0.122, 95% CI: 0.023–0.633 and OR: 0.034, 95% CI: 0.007–0.167, respectively), adjusted for intent of exposure, cyanide source, age, and sex.
Conclusions: After NAP implementation, both antidote use and appropriate antidote use increased. In cases presenting with severe symptoms, presence of the NAP and appropriate antidote use independently reduced the risk of mortality.
Introduction
Antidote shortage is a global problem in poison management that leads to delayed treatment, suboptimal alternative management options, increased complications, and increased mortality [Citation1–3]. Expert consensus guidelines for the stocking of antidotes for emergency care recommend immediate availability for certain antidotes such as digoxin-specific antibody fragment (DSFab) for digoxin poisoning, Cyanide Antidote Kit or hydroxocobalamin for cyanide poisoning, and methylene blue for methemoglobinemia [Citation4]. The implementation of an antidote logistics system may alleviate this shortage problem.
In Thailand prior to the establishment of the National Antidote Project (NAP) in 2010, sodium nitrite and sodium thiosulfate were not routinely stocked in hospitals, only a few of which had the capacity to prepare doses of sodium nitrite and sodium thiosulfate. Hence, if a patient was diagnosed with cyanide poisoning in a hospital near Ramathibodi Poison Center (RPC), the antidote was sent from the poison center. If the treatment hospital was far from Ramathibodi Poison Center, the poison center would also contact the hospital pharmacists to initiate antidote preparation. Each of these steps in the process led to delay in emergent administration of antidote.
Since November 2010, the NAP has operated to catalog the available antidote stockpile, manage distribution systems, educate the healthcare providers on proper antidote usage, and analyze antidote utilization across every region of Thailand. The project is a collaboration between the National Health Security Office (NHSO), Queen Saovabha Memorial Institute Thai Red Cross Society, The Thai Society of Clinical Toxicology, Government Pharmaceutical Organization (GPO), the Food and Drug Administration (FDA), and poison centers. The roles of each institution are summarized in the Supplemental Materials.
The antidotes are stocked and distributed in every region of the country based on urgency of the poisoning. A web-based program is used for real-time monitoring of the quantity of each antidote in each stockpile hospital. The same program is also used to locate the nearest stockpile hospital with available antidote to the requesting hospital. In clinical use, all antidotes are provided to Thai residents and foreign workers free of charge through the Thai Universal Coverage scheme [Citation5,Citation6]. Antidotes managed by the project include methylene blue, sodium nitrite, sodium thiosulfate, dimercaprol, succimer, calcium disodium edetate, DSFab, glucagon, and botulinum antitoxin [Citation6,Citation7]. To educate healthcare providers on poisoning and antidote use, there were annual countrywide educational programs by the Thai Society of Clinical Toxicology and the poison centers.
Poison centers supervise antidote use, coordinate with stockpile sites, follow-up on cases, educate on poison management, and evaluate antidote utilization. If antidotes are administered before poison center consultation, the poison center is notified by the project to follow the cases to resolution.
Sodium nitrite and sodium thiosulfate are the most widely distributed antidote in the project. They are stocked in every provincial hospital, university hospital, poison center, and general hospital with distance greater than a 1-h drive from the nearest stockpile site. The maximum allowable time between initial contact stockpile management to antidote administration is 90 min. The time from stockpile contact to antidote administration of sodium nitrite and/or sodium thiosulfate, specifically, was recorded and used as a performance indicator for the stockpile and supply system. In these cases, poison centers are also responsible for confirmation of antidote transfer, administration, response after administration, determination if antidote should be repeated, progression, and case conclusion. Hydroxocobalamin is not available in Thailand because of its high cost, and is therefore not included in the project [Citation8].
This study evaluates the effect of the NAP implementation on (1) poisoning-related mortality rates and (2) antidote utilization. The study focuses solely on the management and outcomes of cyanide and cyanogenic glycoside poisoning.
Methods
Study design
This is a retrospective cohort study of cases involving cyanide or cyanogenic glycoside ingestion reported to the Ramathibodi Poison Center from 1 January 2007 to 31 December 2015 (46 months before establishment of the NAP through 62 months after). This study was approved by the Committee on Human Rights Related to Research Involving Human Subjects, Faculty of Medicine Ramathibodi Hospital, Mahidol University.
Ramathibodi Poison Center serves the entire Thai population of 69 million people and receives approximately 20,000 calls per year. It has operated 24 h a day, 7 days a week since 1996. Service is free of charge. All of the operational costs are derived mainly from budgets of the Faculty of Medicine, Ramathibodi Hospital, Mahidol University [Citation9]. Calls are received by specialists in poison information (SPIs), who are nurses or pharmacists whose additional training includes 30 h of didactics and 20 sessions of supervised practical instruction. For complicated cases, uncertain cases, or cases potentially involving the NAP, consultations with medical toxicologists are available. All records, including case diagnosis and severity, are reviewed and verified during daily staff meetings. Cases are followed until completion including this verification step prior to entry into the electronic database.
Data collection
Case records were electronically abstracted by a trained senior SPI with experience in extracting data from the poison center database. Case records were identified by active ingredient name, coded as “cyanide” or “cyanogenic glycoside”. All records were reviewed and verified by the conference of medical toxicologists and SPIs. Any differences of opinion were resolved by discussion.
Demographic data, symptoms, intent of exposure, source of exposure, treatment (including administration time and dose of antidote), initial severity, and outcome of the case were recorded.
Definitions
Cases reported before NAP implementation were defined as the “Before group” and cases reported during the project were defined as the “Project group”. The definition of initial severity and medical outcome in the records system was adopted from the Poison Severity Score developed by the International Programme on Chemical Safety, the Commission of the European Union, and the European Association of Poison Centres and Clinical Toxicologists (IPCS/EC/EAPCCT) [Citation10]. Initial severity was defined as the severity at the first presentation at healthcare facility. Medical outcome was determined by the most severe effect during the clinical course.
Antidote administration that met all criteria of (1) correct indication, (2) proper dosing regimen, and (3) timely administration were determined to be “appropriate”. Cases that fail to meet any one of the criteria were determined to be “inappropriate”.
“Correct indication” was defined as those antidotes used in cases with symptomatic cyanide poisoning, excluding cases with minor symptoms such as only nausea, vomiting, and/or dizziness.
“Timely administration” was defined as the time interval from a clinician or poison center initially making contact with an antidote stockpile site to the time of initiating antidote infusion was less than or equal to 90 min.
“Proper dosing regimen” was defined as administration of sodium nitrite and sodium thiosulfate, or sodium thiosulfate alone, at the following dose: sodium nitrite 300 mg in adults or 6 mg/kg in children intravenously, and sodium thiosulfate 12.5 g in adults or 250 mg/kg in children intravenously [Citation11,Citation12]. If hemoglobin concentration was available before sodium nitrite administration, the sodium nitrite dose would be adjusted accordingly [Citation13]. If the patients did not improve within 30 min after completion of antidote administration, a further half-dose could be administered. Administration of sodium nitrite alone or any other dosing regimen was deemed to be incorrect.
Statistical analysis
The data are presented as descriptive statistics. Comparisons of case characteristics before versus after the establishment of the NAP and surviving cases versus deaths were performed using a Chi-square test or Fisher’s exact test. Age of patients before and after the NAP was compared using the rank-sum test. Clinical effects, initial severity, treatment, antidote use, appropriateness of antidote use, and medical outcome were compared in both groups (Before and Project) using logistic regression analysis. A multicollinearity between the NAP, appropriate antidote use, cyanide source (chemical or cyanogenic glycoside), intent of exposure (intentional or unintentional), age, and sex was explored and indicated the absence of collinearity among these factors. In addition, there were different characteristics between the Before and Project groups, therefore these factors were also adjusted for age, gender, source exposure, and intent of exposure. Because the antidotes might not influence the outcomes of asymptomatic or minor symptom cases and all deaths in this cohort presented in initial severe symptoms group, subgroup analysis of only cases with initial severe symptoms was performed to determine the effect of the NAP, antidote use, and appropriate antidote use on mortality. Univariate and multivariate logistic regression analyses were performed to determine the independent effect of the NAP and the appropriate antidote use on mortality rate.
A p value of less than .05 was deemed to be statistically significant. All analyses were performed using Stata version 12 (College Station, TX).
Results
During the study period, 350 cyanide or cyanogenic glycoside exposures were reported to Ramathibodi Poison Center. Seven were excluded: three dermal exposures with mild irritation and four house-fire exposures with minor symptoms. A total of 343 symptomatic, single-substance ingestion cases were included for analysis, all of which had complete follow-up. Most were diagnosed by history of exposure without confirmatory testing. Twelve cases (3.5%) had confirmed diagnosis with blood cyanide concentration. Of the 12 cases, four cases also had cyanide concentration performed on gastric lavage content (two cases) or substances found on the scene (two cases).
There were 130 cases (37.9%) in the Before group and 213 cases (62.1%) in the Project group (). Median age was five years (interquartile range [IQR]: 3–29 years). Fifty-two percent of patients were male. Common sources of cyanide were cyanogenic glycosides (n = 245, 71.4%) including 243 from cassava, one from bitter almond, and one from bamboo shoot. There were 98 (28.6%) cyanide chemical exposures, 70 of which were exposed to gold or silver polishing solutions. Most exposures were unintentional (316, 92.1%). There were more cases with self-harm intent in the Project group (22, 10.3%) than in the Before group (five, 3.9%) (p value .038). Of the 27 self-harm cases, the majority involved ingestion of cyanide chemical (26; 96.3% of 27 cases), followed by cyanogenic glycoside (one, 3.7%). Most of the cases were symptomatic at presentation (338, 98.5%). Eighty patients (23.3%) had initial severe symptoms. There was no difference in initial severity between the two groups ().
Table 1. Characteristics of cyanide ingestion cases reported to the Ramathibodi Poison Center.
Table 2. Initial severity of cyanide ingestion cases reported to the Ramathibodi Poison Center.
The number of reported cases increased in 2008, stabilized until 2014, then increased again in 2015 (). Cases were reported in every season, with a peak during late winter in December and January (). Cases were reported from every region of the country, mostly from north-eastern and eastern regions ().
Figure 1. Cyanide and cyanogenic glycoside ingestions reported to Ramathibodi Poison Center by year.
Figure 2. Cyanide and cyanogenic glycoside ingestions reported to Ramathibodi Poison Center by month of the year.
Figure 3. Cyanide and cyanogenic glycoside ingestions reported to Ramathibodi Poison Center by region.
The common clinical effects were vomiting (218, 63.6%), central nervous system (CNS) depression (187, 54.5%), nausea (141, 41.1%), and acidosis (129, 37.6%) (). There were also 94 patients (27.4%) with respiratory failure, 68 (19.8%) with hypotension, and 35 (10.2%) with cardiac arrest. There were more reported cases of seizure, single-episode seizure, tachycardia, and CNS depression in the Project group (adjusted odd ratio [OR]: 2.859, 2.438, 4.217, and 1.785; 95% confidence interval [CI]: 1.423–5.742, 1.122–5.301, 2.244–7.924, and 1.104–2.886; adjusted p value .003, .024, <.001, and .018, respectively). There was less reported nausea in the Project group (adjusted OR: 0.380, 95% CI: 0.232–0.622, adjusted p value <.001). There were 20 cases (5.8%) of hyperglycemia (defined as capillary blood glucose >200 mg/dL; the median was 323 mg/dL with a range of 225–606 mg/dL). Only one case involved type 2 diabetes mellitus. Of the hyperglycemia cases, only two were treated with insulin. Two cases reported capillary blood glucose <50 mg/dL and required glucose infusion.
Table 3. Common clinical effects of cyanide ingestion cases reported to the Ramathibodi Poison Center.
The common treatments were intravenous fluid (263, 76.7%), oxygen supplementation (143, 41.7%), cyanide antidote (117, 34.1%), intubation and ventilatory support (111, 32.4%), and sodium bicarbonate (97, 28.3%) (). There were higher rates of overall cyanide antidote administration and sodium thiosulfate only administration in the Project group (adjusted OR: 2.382 and 2.242; 95% CI: 1.422–3.991 and 1.337–3.757; adjusted p value .001 and .002, respectively). Of the 117 cases using cyanide antidote, 79 (67.5% of 117) were determined to be appropriate (). There were more correct indications, and timely and overall appropriate cyanide antidote administration, in the Project group (adjusted OR not calculable since all use of the project group were correctly indicated, 3.605 and 3.600; 95% CI not calculable, 1.347–9.647 and 1.395–9.289; adjusted p value not calculable, .011 and .008, respectively).
Table 4. Common treatments of cyanide ingestion cases reported to the Ramathibodi Poison Center.
Table 5. Appropriateness of antidote use in cyanide poisoning from ingestion cases reported to the Ramathibodi Poison Center.
There were 30 deaths (mortality rate 8.8%), including 17 (mortality rate 8.0%) in the Project group and 13 (mortality rate 10.0%) in the Before group (). There was a higher proportion of cases with severe outcome in the Project group (adjusted OR: 3.436, 95% CI: 1.576–7.490, p value .002), but no difference in mortality between the two groups.
Table 6. Medical outcome of cyanide ingestion cases reported to the Ramathibodi Poison Center.
Regarding the 30 deaths, the median age was 23 years (IQR: 3–31 years) and 53.3% were male. The majority was cyanide chemical exposures (24, 80.0%). Eleven deaths (36.7%) were associated with self-harm intent. All of the deaths presented initially with severe symptoms (). Characteristics related to death were self-harm intent (OR: 10.746, 95% CI: 3.884–28.514, p value <.001), cyanide chemical source (solution or other chemical form, OR: 12.919, 95% CI: 4.862–39.761, p value <.001), overall antidote use (OR: 9.462, 95% CI: 3.590–29.032, p value <.001), and initial severe symptoms (OR and 95% CI were not calculable since non-severe initial presentation cases survived, p value <.001). Appropriate antidote use reduced the risk of death (OR: 0.032, 95% CI: 0.006–0.129, p value <.001).
Table 7. Characteristics of deaths and surviving cases related to cyanide ingestion reported to the Ramathibodi Poison Center.
In the subgroup analysis of 80 cases with initial severe symptoms, the NAP and appropriate antidote use correlated with lower mortality (OR: 0.327 and 0.024; 95% CI: 0.106–0.997 and 0.004–0.122; p value .026 and <.001, respectively) (). Although overall antidote use correlated with death in the overall study population, it did not affect mortality in this subgroup.
Table 8. Univariate subgroup analysis of the 80 initial severe cases determining association between presence of the National Antidote Project (NAP), overall antidote use, appropriate antidote use, and death.
After confirmation of the absence in significant collinearity between each factor, a logistic regression analysis of the 80 initial severe cases to determine the effect of the NAP, appropriate antidote, intent of exposure, cyanide source, age, and sex on mortality was performed. The presence of the NAP and appropriate antidote use independently reduced the risk of death (OR: 0.122 and 0.034; 95% CI: 0.023–0.633 and 0.007–0.167; p value .012 and less than .001, respectively) ().
Table 9. Multivariate subgroup analysis of the 80 initial severe cases determining association between presence of the National Antidote Project (NAP), appropriate antidote use, age, sex, intent of exposure, cyanide source, and death.
Discussion
This study analyzed a cohort of cases of cyanide poisoning from either cyanide or cyanogenic glycoside ingestion before and during the NAP. Initial severe symptoms and self-harm intent increased the risk of death in cyanide poisoning cases. The presence of the NAP and appropriate antidote use were independently related to avoidance of fatal outcome in cases presenting initially with severe poisoning.
Cyanide antidote may have had little impact on mild poisoning cases. Patients with only nausea or vomiting might have recovered without antidote use. The cases that were more severe would benefit more from timely antidote administration. This may explain why the NAP failed to affect the overall outcome but correlated with lower mortality risk in the subgroup analysis of cases with initial severe symptoms ( and ). The same reason might explain why overall antidote use increased risk of mortality in the overall population (OR: 9.462) but not in the subgroup analysis.
During the NAP, overall antidote use, use with correct indication, timely administration, and appropriate antidote use were increased in comparison with the pre-NAP era. This resulted from centralized supply management, increased accessibility of cyanide antidote nationwide, and education on cyanide poisoning and antidote use. Appropriate antidote use, but not overall antidote use, decreased the risk of death ( and ). This finding suggests that the cyanide antidotes must be used appropriately, defined by the study as correct indication, proper dosing regimen, and timely administration in order to yield therapeutic benefit in patients with severe cyanide poisoning.
In addition to the supply management and distribution of antidotes, the NAP provided education on poison management as well as diagnostic and therapeutic recommendations to clinicians seeking assistance. Early education, diagnostic recommendations, early management guidance, and close monitoring also might have improved outcomes in the studied cases. As a whole, the presence of the NAP provided a reduction in mortality risk independent of appropriate antidote use () in cases of severe cyanide poisoning at presentation. Further studies focusing on poison education, time from presentation to diagnosis, and details of supportive treatment may elucidate the relative contributions of these aspects toward outcomes of cyanide poisoning.
Regionality and timing of cases in this study might be related to the large rate of exposure to cassava (243, 70.8%). The greatest number of cases was reported from north-eastern and eastern regions, both major cassava-producing areas of Thailand [Citation14]. The exposures peaked in December and January, in the middle of the cassava-harvesting period of November to March [Citation14]. Incidence of cases over time showed an increase in 2008 that plateaued until 2014 and then increased again in 2015. The initial increase in 2008 might have correlated with the promotion of cassava for bio-fuel production in 2008 [Citation15,Citation16] and the influence of local news outlets that reported a death from cassava ingestion in December 2007 [Citation17]. The increased incidence of cases in 2015 may be from greater vigilance and heightened awareness by health care providers after additional educational programs through the implementation of the NAP.
The mortality rates from both cyanogenic glycosides (2.4%, six deaths out of 245 exposures) and cyanide chemicals (24.5%, 24 deaths out of 98 exposures) in Thailand exceeds the mortality reported in the 2014 Annual Report of the American Association of Poison Control Centers’ National Poison Data System (NPDS; 0.2% [two deaths out of 914 calls] of cyanogenic glycoside exposures, and 8.4% [nine deaths out of 107 calls] of cyanide chemical exposures) [Citation18]. The difference in mortality rates might be related to the inclusion of only ingestion cases and the higher rate of intentional exposure in our study (7.9% in this study versus 3.8% in the NPDS) [Citation18].
Interestingly, there were 20 cases (5.8%) with capillary blood glucose concentration higher than 200 mg/dL, two of which received insulin. Hyperglycemia has been reported with cyanide poisoning, and some previously reported cases demonstrated spontaneous resolution without insulin administration [Citation19,Citation20]. The mechanism of hyperglycemia in cyanide poisoning might be related to increased glycogenolysis and insulin resistance [Citation21,Citation22].
Limitations
The population in this study originated from a single poison center database, which combines both voluntarily reported data and NAP notifications. Cases undetected by these two methods are not reported to the poison center database, so the overall number of exposures might be underreported. The diagnoses of cyanide poisoning from the two reporting sources were based largely on reported histories rather than laboratory-confirmed exposures. However, cases with poison center consultation and/or with requests for antidotes might represent more clinically severe poisonings, while rates of clinical effects and fatality may not be generalizable to studies with other designs or populations.
The definition of timely antidote administration was arbitrarily derived from expected lag time between initial stockpile contact and initiation of antidote administration. This was set as a performance indicator for the stockpile and supply system. In severe cyanide poisoning, the antidote should be administered as soon as possible without a verified “gold standard” period of maximum antidote efficacy. As timely antidote administration in this study was not based on the time of exposure, cases with late presentation or delayed diagnosis might not have benefited from antidote administration within 90 min after stockpile contact. This possibly explains occurrence of deaths despite appropriate antidote administration according to the study definition.
The NAP was established within the setting of the Thai healthcare system. Characteristics of the system included small, widely distributed hospitals in each region, general practitioners as primary first responders, and universal coverage for medical care expense. The healthcare setting varies greatly by country. Hence, solutions proposed by the NAP may not be implementable in other systems or may require significant customizations.
Conclusions
The National Antidote Project was associated with more antidote administrations and more appropriate antidote utilization. The NAP and appropriate antidote use independently reduced the risk of death in cases of cyanide and cyanogenic glycoside exposure.
ICTX_Srisuma_et_al_Supplemental_Content.docx
Download MS Word (19.3 KB)Acknowledgements
We would like to thank the collaborators of National Antidote Project as followings: National Health Security Office (NHSO), Queen Saovabha Memorial Institute Thai Red Cross Society, The Thai Society of Clinical Toxicology, Government Pharmaceutical Organization (GPO), Food and Drug Administration (FDA), and staff members of the poison centers for carrying out the National Antidote Project.
We would like to thank Dazhe Cao, MD, Assistant Professor of Emergency Medicine at the University of Texas Southwestern Medical Center, and Christopher Hoyte, MD, Assistant Professor of Emergency Medicine at the Rocky Mountain Poison and Drug Center for suggestions for the manuscript and language support.
We would like to thank Pongsakorn Atiksawedparit, MD, emergency physician and lecturer at Department of Emergency Medicine, Faculty of Medicine Ramathibodi Hospital, Mahidol University for suggestions for the manuscript and statistical support.
Disclosure statement
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.
Additional information
Funding
References
- Mazer-Amirshahi M, Hawley KL, Zocchi M, et al. Drug shortages: implications for medical toxicology. Clin Toxicol. 2015;53:519–524.
- Thanacoody RH, Aldridge G, Laing W, et al. National audit of antidote stocking in acute hospitals in the UK. Emerg Med J. 2013;30:393–396.
- Wium CA, Hoffman BA. Antidotes and their availability in South Africa. Clin Toxicol. 2009;47:77–80.
- Dart RC, Goldfrank LR, Erstad BL, et al. Expert consensus guidelines for stocking of antidotes in hospitals that provide emergency care. Ann Emerg Med. 2017 [2017 July 14]. DOI:10.1016/j.annemergmed.2017.05.021.
- Tangcharoensathien V, Pitayarangsarit S, Patcharanarumol W, et al. Promoting universal financial protection: how the Thai universal coverage scheme was designed to ensure equity. Health Res Policy Sys. 2013;11:25.
- Srithamrongsawat S, Jongudomsuk P, Suntorntham S. NHSO annual report fiscal year 2014. Bangkok (Thailand): National Health Security Office (NHSO); 2015. p. 7.
- Krairojananan W, Management for orphan drug and drug with limited access. In: Sriapha C, editor. Antidote 5. Bangkok (Thailand): The Thai Society of Clinical Toxicology; 2015. p. 1–8.
- Emergency Medical Products Inc. Cyanokit. 2017; [cited 2017 May 10]. Available from: http://www.buyemp.com/product/cyanokit-cyanide-antidote-kit.
- Wananukul W, Sriapha C, Tongpoo A, et al. Human poisoning in Thailand: The Ramathibodi Poison Center’s experience (2001–2004). Clin Toxicol. 2007;45:582–588.
- Persson HE, Sjoberg GK, Haines JA, et al. Poisoning severity score. Grading of acute poisoning. J Toxicol Clin Toxicol. 1998;36:205–213.
- Sodium Nitrite – Medical Countermeasures Database – CHEMM: US Department of Health & Human Services, Office of the Assistant Secretary for Preparedness and Response, National Library of Medicine. 2017 [updated 2017 Apr 29; cited 2017 May 20]. Available from: https://chemm.nlm.nih.gov/countermeasure_sodium-nitrite.htm.
- Sodium Thiosulfate – Medical Countermeasures Database – CHEMM: US Department of Health & Human Services, Office of the Assistant Secretary for Preparedness and Response, National Library of Medicine. 2017 [updated 2017 Apr 29; cited 2017 May 20]. Available from: https://chemm.nlm.nih.gov/countermeasure_sodium-thiosulfate.htm.
- Berlin CM Jr. The treatment of cyanide poisoning in children. Pediatrics. 1970;46:793–796.
- Ratanawaraha C, Senanarong N, Suriyapan P, Status of cassava in Thailand: implications for future research and development [updated 2006 Jun 12; cited 2016 Dec 10]. Available from: http://www.fao.org/docrep/009/y1177e/Y1177E04.htm.
- Russell T. Frymier. Bioethanol production in Thailand: a teaching case study comparing cassava and sugar cane molasses 2012 [updated 2012 Mar 19; cited 2016 Dec 10]. Available from: http://www.jsedimensions.org/wordpress/content/bioethanol-production-in-thailand-a-teaching-case-study-comparing-cassava-and-sugar-cane-molasses_2012_03/.
- Bloyd C, Foster N. An update on ethanol production and utilization in Thailand – 2014. VA: Energy USDo; 2014.
- Khaosodnews. A 3 year-girl died from eating raw cassava Bangkok: Center of Excellence on Hazardous Substance Management Chulalongkorn University; 2007 [updated 2007 Dec 26]. Available from: http://www.chemtrack.org/News-Detail.asp?TID=7&ID=36.
- Mowry JB, Spyker DA, Brooks DE, et al. 2014 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS): 32nd annual report. Clin Toxicol. 2015;53:962–1147.
- Akyildiz BN, Kurtoglu S, Kondolot M, et al. Cyanide poisoning caused by ingestion of apricot seeds. Ann Trop Paediatr. 2010;30:39–43.
- Garlich FM, Alsop JA, Anderson DL, et al. Poisoning and suicide by cyanide jewelry cleaner in the US Hmong community: a case series. Clin Toxicol. 2012;50:136–140.
- Vandebroeck A, Uyttenhove K, Bollen M, et al. The hepatic glycogenolysis induced by reversible ischaemia or KCN is exclusively catalysed by phosphorylase a. Biochem J. 1988;256:685–688.
- Singh BM, Coles N, Lewis P, et al. The metabolic effects of fatal cyanide poisoning. Postgrad Med J. 1989;65:923–925.