Table 1.

Manuscripts including the Poisoning Severity Score

Manuscript	Exposure and population	PSS breakdown
Abahussain and Ball [18]; Kuwait, 2010	Chemical or pharmaceutical poisoning in children <12	978: 578|364|34|0
	Rather than incorporating the PSS, it would be clearer to state that over 95% of patients developed minimal or no symptoms. Combined PSS of 2 and 3
Adams et al. [19]; England, 2010	Accidental pesticide exposures called in to the National Poisons Information service	1284:–|1138/1247|–|–
	Authors made multiple modifications to equate the PSS with another questionnaire. Given all the modifications, may have been easier to have everyone only use the questionnaire. Combined PSS of 2 and 3 and reported enquirers and SPI results, respectively
Akdur et al. [5]; Turkey, 2010	Patients with organophosphate poisoning presenting to the emergency department	54:0|32|6|13|3
	Modified the PSS. Authors claimed that both the PSS and GCS were effective tools as those patients with a PSS of 4 had significantly lower GCS scores (mean = 4.7; p < 0.5)
Bucaretchi et al. [20]; Brazil, 2012	Exposures to rodenticides with acetylcholinesterase inhibitors called into the poison center	76:4|9|27|33|3
	Modified PSS by including treatments when deriving the score
Caldas et al. [21]; Brazil, 2008	Pesticide exposures reported to the Center of Toxicology and Assistance of the Federal District	709:162|342|97|90|18
	It would be simpler and clearer to list the muscarinic symptoms or bleeding in patients exposed to organophosphates or coumarins, respectively, instead of using the PSS to demonstrate the severity of illness
Cevik et al. [22]; Turkey, 2006	Patients with carbon monoxide toxicity presenting to the University Medical Center	182:2|134|28|18|0
	No patients were given a PSS of 4, yet 6 died. Moderate correlation between CO level and PSS (r = 0.329, p < 0.001)
Charra et al. [23]; Morocco, 2013	Any intoxicated patient >14 years old admitted to the intensive care unit	214:–|–|–|–|–
	A PSS ≤ 2 was associated with a good prognosis (OR 0.11, CI 0.049–0.237. p < 0.001)
Churi et al. [7]; India, 2012	All patients exposed to organophosphates presenting to the emergency department	136:0|83|36|10|7
	The authors found a significant association (p < 0.001) between the PSS, GCS, APACHE II, and predicted mortality. They do not state which score was the best. Or the one they preferred using
Churi et al. [24]; India, 2012	All patients admitted following a poisoning or envenomation	212:0|165|42|1|4
	There was a moderate correlation between the GCS and PSS (r = 0.51, p < 0.001). There was also a significant association between clinical outcomes, the GCS, and the PSS
Davies, Eddleston, and Buckley [6]; Sri Lanka, 2008	All patients with an acute, solitary exposure to an organophosphate	1365:0|973|138|254|0
	Authors modified the PSS and 184 deaths were scored as a PSS of 1–3. A PSS of 2 at admission had a sensitivity of 0.78 and specificity of 0.79 for predicting death. Authors also concluded that the GCS was as accurate as the PSS in predicting mortality (AUC 0.84 vs. 0.81)
Deguigne et al. [25]; France, 2011	Metam sodium exposures reported to the Poisons and Toxicovigilance Centre	102:101|0|1
	As it was not clear how the authors scored the PSS, it would be clearer to state that 99% of patients had minimal or no symptoms. PSS of 0 and 1 and PSS of 3 and 4 were combined
Deguigne et al. [26]; France, 2013	Patients with acute-on-chronic lithium intoxication reported to the Toxic Exposure Surveillance System	59:29|30
	The PSS confirmed what is already intuitive, that a lower dose and a lower lithium concentration are associated with less severe symptomatology. PSS of 0 and 1 and PSS of 2–4 were combined
Eizadi-Mood et al. [27]; Iran, 2012	All poisoned patients presenting with a CPK > 250 IU/L	80:–|–|–|–|–
	The PSS was calculated but the results were not reported. No explanation was given for this so it is unclear what if anything the PSS contributed to the manuscript
El Salam et al. [11]; Egypt, 2011	Patient without chronic cardiac, pulmonary, or renal disease admitted within 24 h of a hydrocarbon ingestion	100:32|50|14|4|0
	They correlated the PSS with final outcome, ICU admission, and intubation. Authors recommended patients with PSS ≥ 2 are admitted to the ICU. They concluded that the PSS was useful in predicting outcomes.
Eyer et al. [28]; Germany, 2011	Patients admitted to the intensive care unit following a solitary quetiapine overdose with a PSS of 2 or 3	20:0|0|9|10|1
	Used the PSS to determine which patients to include. Only included patients with a PSS ≥ 2
Giannini et al. [29]; Italy, 2007	All patients with amatoxin poisoning	111:0|62|18|31|0
	The PSS was modified by stratifying the scores using transaminase concentrations. Two patients that died were scored a PSS of 3
Hrabetz et al. [12]; Germany, 2013	All patients admitted to the intensive care unit treated for cholinergic toxicity following an organophosphate exposure	33:–|–|–|–|–
	The authors did not include a breakdown of the PSS and rationale for this decision is not included. The authors included the PSS, GCS, APACHE II, and SOFA but did not compare the scores
Hunfeld et al. [13]; Netherlands, 2006	Patients reported to the pharmacy after an intentional quetiapine overdose	14:0|6|5|3|0
	The authors list specific symptoms as minor, moderate, or severe. What was included is not the complete list of symptoms in the PSS so the authors may have modified the PSS
Hung et al. [30]; Vietnam, 2008	All poisonings or toxic exposures admitted to the poison control center	1836:135|1214|298|188|1
	The authors misclassified the PSS as 21 patients died but only 1 was assigned a PSS of 4
Hung et al. [31]; Vietnam, 2009	All patients admitted to the intensive care unit following Bungarus multicinctus envenomations	60:0|3|0|50|4
	While it was reasonable to include a severity score, describing the frequency of interventions more transparently would better establish the severity of illness compared to using the PSS
Jimmink et al. [32]; Netherlands, 2008	All citalopram overdoses identified by the Central Hospital Pharmacy Laboratory	26:0|3|9|14|0
	The PSS was modified. Two fatalities were scored a PSS of 3
Jose et al. [33]; India, 2012	Children ≤18 years old admitted following a poisoning	121:24|67|17|10|3
	The PSS was used to triage patients. However, a shorter and less subjective list of signs or symptoms would more easily and rapidly triage patients
Karbakhsh and Zandi [34]; Iran, 2008	Patients >60 years old following an acute poisoning	299:16|76|156|50|0
	PSS was misapplied as there were 35 fatalities but no patients scored a PSS of 4
Kuzelová et al. [35]; Slovakia, 2009	Children ≤18 years of age with alcohol intoxication	537:10|256|250|21|0
	The problem of adolescent ethanol consumption was effectively communicated by demonstrating the increase in BAC and admission rates between the 2 time periods without including the PSS
Lauterbach et al. [36]; Germany, 2005	Patients with hydrogen phosphide exposures reported to the poison center	188:55|70|17|7|2
	The follow-up rate was low, 48.3%, which is a significant limitation in using the PSS
Malina et al. [37]; Croatia and Hungary, 2010	All admitted patients following Balkan adder envenomations	54:5|24|12|12|1
	The authors modified the PSS. Fatalities were not recorded during part of the study, although one was included
Palenzona et al. [14]; Switzerland, 2004	Patients >16 years old with solitary acute olanzapine overdoses reported to the Swiss Toxicological Information Centre	26:0|14|11|1|0
	The authors may have modified the PSS, although this is not clear. There was a significant association between increasing olanzapine dose and PSS (p = 0.025).
Sabzghabae et al. [38]; Iran, 2011	Non-diabetics >18 years of age following a suicide attempt with an agent not associated with hyperglycemia or hypoglycemia and that did not receive dextrose or glucocorticoids	345:169|78|58|38|2
	As opposed to correlating blood glucose concentrations with the PSS, it would be clearer to associate the glucose concentrations with pertinent outcomes such as respiratory failure or death
Sam et al. [8]; India, 2009	All patients presenting to the emergency department following organophosphate or carbamate exposures	71:1|3|20|37|10
	The authors modified the PSS. They did conclude that the GCS, APACHE II, PMR, and PSS were all useful scoring systems
Schaper et al. [15]; Germany and France, 2004	All patients with rattlesnake envenomations reported to poison centers in France or Germany	21:0|8|5|8|0
	Extreme pain, severe thrombocytopenia, and compartment syndrome are all criteria for a PSS of 3, although the clinical significance of an isolated lab abnormality is very different than the other 2 findings.
Schaper et al. [39]; Germany and France, 2009	Patients reported to any of 4 different poison centers following envenomations from exotic pets	404:0|320|55|29|0
	While the inclusion of a severity score was reasonable, stating the number of patients that developed certain symptoms or required specific interventions would be a more useful descriptor than the PSS
Seok et al. [40]; South Korea, 2012	Patients >18 years of age following an intentional ingestion of a chloracetanilide herbicide	35:17|10|5|2|1
	It would be easier to state that few patients had significant symptoms and that the elderly were more likely to be symptomatic than it was to correlate findings with PSS
Turedi et al. [41]; Turkey, 2011	All patients with CO > 2% admitted to the emergency department without ischemic disease, hepatic failure, or heart failure	37:0|22|10|1|0
	The authors found a significant correlation between carbon monoxide levels and the PSS at admission, but there was no correlation between the PSS and ischemia-modified albumin
Von Mach et al. [42]; Germany, 2004	All patients reported to the poison center following an insulin overdose	160:27|59|40|34|0
	Follow-up was only obtained in 46.9% of patients. Four patients that died were not scored a PSS of 4
Von Mach et al. [43]; Germany, 2006	All patients following an overdose of insulin, sulfonylureas, or biguanides reported to a regional poison center	626:27|59|40|34|0
	Follow-up was poor for all 3 types of exposures
Walter and Persson [44]; Sweden, 1999	Patients >10 years of age with toxic exposures following home or leisure activities	1033:124|806|72|0|0
	An author was involved with the development of the PSS. Ninety percent of cases were scored a PSS of 0 or 1. The authors could just as easily have communicated the mildness of exposure by noting only 28% of cases required any medical care
Williams et al. [45]; England, 2012	Inquiries made to the UK National Poisons Information Service following household product exposures	5939:4117|1638|75|9|0
	Follow-up calls were obtained 4 h after original inquiry and were only successful in 55% of patients. As such, scoring may not accurately reflect the patient’s most severe symptomatology. PSS misapplied in 2 patients that died
Williams et al. [46]; England, 2012	Inquires made to the UK National Poisons Information Service following exposures to liquid detergent capsules	647:382|243|10|2|0
	The authors correctly used the PSS
Wong et al. [47]; Australia, 2010	All patients reported in the VEMD or VPIC following intentional overdose of an antidepressant	1833:1715|118|0
	The authors combined PSS of 0 and 1 and 2 and 3. It would be more useful to know how many patients developed certain adverse effects such as cardiotoxicity, seizures, or respiratory depression following a TCA overdose than knowing their PSS
Wong et al. [48]; Australia, 2012	Patients reported to the VPIC following hydrofluoric acid exposures	75:10|49|16|0|0
	All patients called into VPIC early had a PSS of 0 while most called in later had a PSS of 1 or 2. This study illustrates the problems with attempting to apply the PSS as a prognostic score based on initial symptoms rather than the most severe symptomatology

Key: Total patients: PSS 0|PSS 1|PSS2|PSS 3|PSS 4. It is noted in the table if the authors combined any categories. Due to missing or incomplete information, the amount of patients scored with the PSS does not always equal the total number of patients included in the study. Not all studies included a breakdown by PSS

BAC blood alcohol concentration, CPK creatine phosphokinase, TCA tricyclic antidepressant, GCS Glasgow Coma Scale, APACHE Acute Physiology and Chronic Health Evaluation, SOFA sequential organ failure assessment score, PMR predicted mortality rate, VEMD Victorian Emergency Minimum Dataset, VPIC Victorian Poisons Information Centre