Comparison of the new-Poisoning Mortality Score and the Modified Early Warning Score for predicting in-hospital mortality in patients with acute poisoning

Figures & data

Table 1. Comparison of the case characteristics between the survivor group and the in-hospital-death group.

	Survivors (n = 16,206)	In-hospital deaths (n = 364)	P value
Demographics
Age (years) median (IQR)	41 (24–58)	72 (56-81)	<0.001
<40 (%)	7,705 (47.5)	42 (11.5)
40–59 (%)	4,877 (30.1)	68 (18.7)
60–69 (%)	1,641 (10.1)	54 (14.8)
70–79 (%)	1,150 (7.1)	90 (24.7)
≥80 (%)	833 (5.1)	110 (30.2)
Sex
Males (%)	6,518 (40.2)	251 (69.0)	<0.001
Female, (%)	9,688 (59.8)	113 (31.0)
Poisoning-related factors
Intent of poisoning			<0.001
Unintentional, (%)	4,596 (28.4)	49 (13.5)
Intentional (%)	11,560 (71.3)	309 (84.9)
Unknown (%)	50 (0.3)	6 (1.6)
Route of poisoning			0.417
Dermal, ocular, or contact (%)	66 (0.4)	1 (0.3)
Ingestion (%)	13,725 (84.7)	300 (82.4)
Inhalation (%)	2,415 (14.9)	63 (17.3)
Category of substances			<0.001
A (%)^a	8,269 (51.0)	38 (10.4)
B (%)^b	1,834 (11.3)	16 (4.4)
C (%)^c	674 (4.2)	3 (0.8)
D (%)^d	2,421 (14.9)	176 (48.4)
E (%)^e	147 (0.9)	30 (8.2)
F (%)^f	24 (0.1)	35 (9.6)
G (%)^g	2,314 (14.3)	63 (17.3)
H (%)^h	523 (3.2)	3 (0.8)
Initial vital signs in the emergency department
Systolic blood pressure (mmHg)			<0.001
≥100 (%)	13,965 (86.2)	223 (61.3)
70–99 (%)	1,453 (9.0)	49 (13.5)
≤69 (%)	788 (4.9)	92 (25.3)
Heart rate (beats/min)			<0.001
70–119 (%)	12,426 (76.7)	183 (50.3)
30–69 (%)	2,247 (13.9)	114 (31.3)
120–159 (%)	1,483 (9.2)	30 (10.2)
≥160 (%)	50 (0.3)	30 (8.2)
Respiratory rate (breaths/min)			<0.001
12–24 (%)	15,132 (93.4)	215 (59.1)
≤11 or ≥25 n (%)	1,074 (6.6)	149 (40.9)
Body temperature (°C)			<0.001
<39 (%)	16,180 (99.8)	347 (95.3)
≥39 (%)	26 (0.2)	17 (4.7)
Mental status (%)			<0.001
Alert	11,045 (68.2)	88 (24.2)
Verbal response	3,055 (18.9)	51 (14.0)
Pain response	1,897 (11.7)	95 (26.1)
Unresponsive	209 (1.3)	130 (35.7)
New-Poisoning Morality Score, median (IQR)	31 (23–40)	66 (57–77)	<0.001
Modified Early warning Score, median (IQR)	3 (2–4)	6 (4–10)	<0.001

IQR: interquartile range.

^a Hormones, hormone antagonists, contraceptives, diagnostic reagents, vitamins, dietary supplements, topical preparations, paracetamol (acetaminophen), antipsychotics, antidepressants, zolpidem, doxylamine, unspecified sedatives, antipsychotics, hypnotics, benzodiazepines.

^b Peptic and gastrointestinal drugs, antihistamines, cold and cough preparations, unspecified therapeutic drugs, anticonvulsants, cardiovascular drugs, unspecified analgesics, antibiotics, antifungals, opioids, stimulants, street drugs, asthma therapies, oral hypoglycemic drugs.

^c Alcohols (liquor, ethanol, methanol), heavy metals, hydrocarbons, chlorine bleach, sodium hypochlorite.

^d Unspecified artificial toxic substances, unspecified alkali, unspecified acids, unspecified corrosive agents, rodenticide, unspecified insecticides, pyrethroid, unspecified pesticides, unspecified herbicides, glyphosate.

^e Glacial acetic acid, organophosphates, carbamates.

^f Paraquat.

^g Carbon monoxide, unspecified gases.

^h Natural toxic substances.

Figure 1. Area under receiver operating characteristic curve of the new-Poisoning Mortality Score and the Modified Early Warning Score for predicting in-hospital mortality in acute poisoning (0.947 versus 0.800, respectively).

Table 2. The performance of the new-Poisoning Mortality Score and the modified early warning score for predicting in-hospital mortality in acute poisoning.

Statistics	New-Poisoning Mortality Score	Modified early warning score
Hosmer-Lemeshow goodness-of-fit test^a	P = 0.1425	P = 0.109
Area under receiver operating characteristic curve (95% CI)	0.947 (0.934–0.959)	0.800 (0.774–0.827)
Optimal cutoff value	52	5
Sensitivity (95% CI)	0.863 (0.823–0.894)	0.667 (0.618–0.714)
Specificity (95% CI)	0.912 (0.908–0.916)	0.817 (0.811–0.823)
Accuracy (95% CI)	0.911 (0.906–0.915)	0.814 (0.808–0.820)

CI: confidence interval.

^aHosmer and Lemeshow test is used to indicate a good fitting model when P is >0.05 [10].

Figure 2. Comparison of the calibration curves for the new-Poisoning Mortality Score and the Modified Early Warning Score for predicting in-hospital mortality in acute poisoning.

Table 3. Risk group stratification by the new-Poisoning Mortality Score for predicting in-hospital mortality in acute poisoning.

Risk group	New-poisoning mortality score	Predicted mortality^a	Observed mortality
Very low	0–27	0.09%	3/6,455 (0.05%)
Low	28–40	0.43%	20/5,968 (0.33%)
Intermediate	41–55	2.38%	57/3,023 (1.89%)
High	≥ 56	22.50%	284/1,124 (25.30%)

^a Predicted mortality rate = 1/(1 + e^−z), z = − 9.763 + 0.126 × new-Poisoning Mortality Score.

Table 4. Sub-analysis for the performance of the new-Poisoning Mortality Score for predicting in-hospital mortality in acute poisoning according to the substance category groups.

Category groups	Sensitivity	Specificity	Accuracy	Area under receiver operating characteristic curve
A	0.853	0.737	0.831	0.853
B	0.750	0.880	0.878	0.905
C	0.667	0.960	0.957	0.837
D	0.898	0.743	0.754	0.902
E	0.733	0.869	0.863	0.865
F	0.914	0.792	0.864	0.900
G	0.937	0.949	0.948	0.986
H	1.000	0.982	0.982	0.989
All groups	0.841	0.866	0.864	0.918

^aHormones, hormone antagonists, contraceptives, diagnostic reagents, vitamins, dietary supplements, topical preparations, paracetamol (acetaminophen), antipsychotics, antidepressants, zolpidem, doxylamine, unspecified sedatives, antipsychotics, hypnotics, benzodiazepines.

^bPeptic and gastrointestinal drugs, antihistamines, cold and cough preparations, unspecified therapeutic drugs, anticonvulsants, cardiovascular drugs, unspecified analgesics, antibiotics, antifungals, opioids, stimulants, street drugs, asthma therapies, oral hypoglycemic drugs.

^cAlcohols (liquor, ethanol, methanol), heavy metals, hydrocarbons, chlorine bleach, sodium hypochlorite.

^dUnspecified artificial toxic substances, unspecified alkali, unspecified acids, unspecified corrosive agents, rodenticide, unspecified insecticides, pyrethroid, unspecified pesticides, unspecified herbicides, glyphosate.

^eGlacial acetic acid, organophosphates, carbamates.

^fParaquat.

^gCarbon monoxide, unspecified gases.

^hNatural toxic substances.

Appendix Table 1. Multivariable logistic regression for the calculation of the new-Poisoning Mortality Score in the patients with acute poisoning.

	B	Points = B/0.124	Odd ratio (95% confidence interval)	P Value
Demographics
Age (years)
<40	Reference	0	1	<0.001
40–59	0.815	7	2.26 (1.64–3.11)	<0.001
60–69	1.435	12	4.20 (2.97–5.93)	<0.001
70–74	2.003	16	7.41 (5.20–10.57)	<0.001
75–79	1.955	16	7.07 (4.93–10.14)	<0.001
≥80	2.395	19	10.97 (7.74–15.55)	<0.001
Sex
Female	Reference	0	1	<0.001
Male	0.436	4	1.55 (1.30-1.84)	<0.001
Poisoning-related factors
Intent of poisoning
Unintentional	Reference	0	1	<0.001
Intentional	1.039	8	2.83 (2.22–3.61)	<0.001
Unknown	1.073	9	2.92 (1.80–4.74)	<0.001
Route of poisoning
Dermal, ocular, or contact	Reference	0	1	0.274
Oral ingestion	1.006	8	2.73 (0.65–11.46)	0.169
Inhalation	0.592	5	1.81 (0.33–9.84)	0.493
Category of substances
A^a	Reference	0	1	<0.001
B^b	1.373	11	3.95 (2.48–6.27)	<0.001
C^c	1.817	15	6.15 (3.16–11.98)	<0.001
D^d	2.654	21	14.21 (10.13–19.93)	<0.001
E^e	3.36	27	28.80 (19.15–43.30)	<0.001
F^f	5.866	47	352.78 (241.57–515.19)	<0.001
G^g	1.801	15	6.05 (2.20–16.68)	<0.001
H^h	1.492	12	4.44 (1.99–9.93)	<0.001
Initial vital sings at emergency department
Systolic blood pressure (mmHg)
≥100	Reference	0	1	<0.001
70–99	0.734	6	2.08 (1.65–2.63)	<0.001
≤69	1.903	15	6.70 (4.56–9.85)	<0.001
Heart rate (beats/min)
70–119	Reference	0	1	0.001
30–69	0.124	1	1.13 (0.88–1.45)	0.323
120–159	0.458	4	1.58 (1.20–2.09)	0.001
≥160	0.984	8	2.68 (1.29–5.53)	0.008
Respiratory rate (breaths/min)
12-24	Reference	0	1	<0.001
≤11 or ≥25	0.663	5	1.94 (1.53–2.46)	<0.001
Body temperature (°C)
<39	Reference	0	1	0.001
≥39	0.684	6	1.98 (0.64–6.12)	0.235
Mental status
Alert	Reference	0	1	<0.001
Verbal response	0.61	5	1.84 (1.47–2.30)	<0.001
Pain response	1.017	8	2.77 (2.21–3.45)	<0.001
Unresponsive	2.033	16	7.64 (5.62–10.39)	<0.001

Base constant B was selected as the smallest regression coefficient in the model, which was 0.124. The new-Poison Mortality Score is the sum of the point of each variable. The possible range of new-Poison Mortality Score was 0 to 137 points.

^aHormones, hormone antagonists, contraceptives, diagnostic reagents, vitamins, dietary supplements, topical preparations, paracetamol (acetaminophen), antipsychotics, antidepressants, zolpidem, doxylamine, unspecified sedatives, antipsychotics, hypnotics, benzodiazepines.

^bPeptic and gastrointestinal drugs, antihistamines, cold and cough preparations, unspecified therapeutic drugs, anticonvulsants, cardiovascular drugs, unspecified analgesics, antibiotics, antifungals, opioids, stimulants, street drugs, asthma therapies, oral hypoglycemic drugs.

^cAlcohols (liquor, ethanol, methanol), heavy metals, hydrocarbons, chlorine bleach, sodium hypochlorite.

^dUnspecified artificial toxic substances, unspecified alkali, unspecified acids, unspecified corrosive agents, rodenticide, unspecified insecticides, pyrethroid, unspecified pesticides, unspecified herbicides, glyphosate.

^eGlacial acetic acid, organophosphates, carbamates.

^fParaquat.

^gCarbon monoxide, unspecified gases.

^hNatural toxic substances.

Hosmer DW, Lemesbow S. Goodness of fit tests for the multiple logistic regression model. Comm Stats. 1980;9(10):1043–1069. doi: 10.1080/03610928008827941.

 Web of Science ®Google Scholar