Diagnostic accuracy of urinary amanitin in suspected mushroom poisoning: a pilot study.

Journal of Toxicology
CLINICAL TOXICOLOGY
Vol. 42, No. 6, pp. 901–912, 2004
Diagnostic Accuracy of Urinary Amanitin in Suspected Mushroom
Poisoning: A Pilot Study
Raffaella Butera, M.D.,* Carlo Locatelli, M.D., Teresa Coccini, Ph.D.,
and Luigi Manzo, M.D.
Pavia Poison Center and Toxicology Laboratory, IRCCS Maugeri Foundation
and University of Pavia, Pavia, Italy
ABSTRACT
Background. Amatoxin-containing species are responsible for the most severe cases
of mushroom poisoning, with high mortality rate. Therefore, this poisoning should
be ruled out in all patients presenting gastrointestinal symptoms after wild mush-
room ingestion. Objective. To determine sensitivity, specificity, positive predictive
value, negative predictive value, and diagnostic efficacy (DE) of urinary amanitin
analysis in cases of suspected mushroom poisoning. Methods. All cases of mush-
room ingestion referred to a Poison Center during a one-month period were
analyzed. Amanitin measurements were performed by ELISA method (functional
least detectable dose 1.5 ng/ml; cut-off value not clearly established). Gastrointes-
tinal symptoms latency and initial clinical assessment were considered alternative
diagnostic tools. Definitive diagnosis was used as the reference standard. Results.
Among 61 patients included in the study, amatoxin poisoning was diagnosed in
10 cases. Urine samples were collected 5.5 to 92 hours after mushroom ingestion.
Urinary amanitin DE was 91.8%, 93.4%, and 80.3%, based on the cut-off value
considered (1.5, 5.0, and 10.0 ng/ml, respectively). Symptoms latency longer than
6 hours and initial clinical assessment DE were 70.5% and 67.2%, respectively.
To identify amatoxin poisoning, initial clinical assessment resulted more sensi-
tive and urinary amanitin analysis more specific. Conclusions. Urinary amanitin
analysis is a valuable diagnostic tool and may significantly contribute to the man-
agement of suspected mushroom poisoning. At present, the best diagnostic accuracy
can be obtained taking advantage of both the high sensitivity and negative pre-
dictive value of the clinical assessment performed by an experienced toxicologist,
*Correspondence: Raffaella Butera, M.D., Pavia Poison Center and Toxicology Laboratory, IRCCS Maugeri Foundation and
University of Pavia, Via Ferrata 8, 27100 Pavia, Italy; Fax: +39-0382-24605; E-mail: rbutera@fsm.it.
901
DOI: 10.1081/CLT-200035472 0731-3810 (Print); 1097-9875 (Online)
Copyright D 2004 by Marcel Dekker, Inc. www.dekker.com

and the high specificity and positive predictive value that characterize urinary
amanitin analysis.
Key Words: Mushroom poisoning diagnosis; Amanita; Amatoxins; Amanitins
analysis; Amanitins diagnostic use; Sensitivity and specificity.
INTRODUCTION
Mushroom poisoning is a common medical
emergency in many European countries, where the
hunting and eating of wild mushrooms is a popular
activity. Amatoxin-containing species are responsible
for the most severe cases of poisoning, with a mortality
rate up to 20% (1–4). Therefore, amatoxin poisoning
should be ruled out in all patients presenting gastro-
intestinal symptoms after wild mushroom ingestion.
Amatoxin poisoning is believed to be clinically
diagnosed with a high degree of confidence by the
patient’s history and initial symptoms (5,6). Patients
usually present with a severe gastroenteritis (repeated
vomiting, profuse watery diarrhea) occurring not before
6 to 12 hours after mushroom ingestion. However,
confusion may result from the ingestion of amatoxin-
containing mushrooms together with other toxic species
producing early symptoms. In these cases, the long
latency of gastrointestinal symptoms can not be rec-
ognized. Another, similarly insidious scenario occurs
when the toxic etiology of gastrointestinal symptoms is
not suspected.
Definitive diagnosis of amatoxin poisoning can be
made sometimes late, when organ failure becomes
evident. Conversely, it is commonly accepted that mea-
sures to improve the survival rates can be taken if the
poisoning is diagnosed correctly and as early as pos-
sible (2,7). Although the consequences on outcome of
delayed therapy are unproven, theoretical arguments
and data obtained from 2108 hospitalized amatoxin
poisoned patients (1) indicate that therapeutic proto-
cols including specific pharmacological treatments are
more effective in reducing mortality than supportive
measures alone.
The determination of amatoxins in biological fluids
may be helpful to achieve early specific diagnosis, before
the typical syndrome of amatoxin poisoning would be
clear enough to confirm or modify the initial clinical
suspect (8). Over the last 25 years radioimmunoassays
(RIA) (9–11) and chromatographic (8,12–14) methods
have been developed for the detection of amanitin in
urine of suspected poisoned patients. Recently, an
enzyme-linked immunosorbent assay (ELISA) has been
introduced as a substitute method instead of RIA (15,16).
Although urinary amanitin analysis has long been
available, there is still a lack of knowledge about its
diagnostic value in the clinical setting. Up to now ama-
nitin analysis was mostly used to confirm a diagnosis
already established by the patient’s history and clinical
course (17–21). A Medline search from 1966 to 2003
failed to identify studies other than a case report (22)
addressed to evaluate the usefulness of this assay for
the early diagnosis of amatoxin poisoning; however,
linguistic barriers precluded to evaluate some possibly
pertinent articles (23–25).
According to the above mentioned considerations,
a pilot study was undertaken in order to assess the
diagnostic accuracy of urinary amanitin analysis in
patients with suspected mushroom poisoning.
OBJECTIVE
The objective of the study was to determine
sensitivity (SENS), specificity (SPEC), positive predic-
tive value (PPV), negative predictive value (NPV) and
diagnostic efficacy (DE) of urinary amanitin analysis in
cases of suspected mushroom poisoning.
Moreover, the study compared the diagnostic
accuracy (SENS, SPEC, PPV, NPV and DE) of urinary
amanitin analysis with that of alternative diagnostic
tools that can be used for the diagnosis of amatoxin
poisoning, that is 1) the long latency of gastrointestinal
symptoms and 2) the initial clinical assessment made
by Poison Center (PC) physicians.
METHODS
Study Design and Setting
A pilot study was performed in a PC involved in
the management of approximately 10,000 cases of
suspected poisoning annually, 86% of which referred
from health care facilities, mostly Emergency Depart-
ments (EDs).
Recruitment centers had a homogeneous geograph-
ic distribution, were located at acceptable distance
from PC (because of transport of biological samples),
902 Butera et al.

and had common procedures for the management of
poisoning cases.
Conduct and reporting of the study were assessed
by using the STARD checklist (26,27).
Study Population
All consecutive cases of mushroom ingestion
referred to the PC in the period August 1st-31, 2002
because of poisoning concerns were considered eligible
in the study.
Patients were included in the study if they had
gastrointestinal symptoms (nausea, vomiting, abdomi-
nal pain, and/or diarrhea) and were evaluated in an ED.
Asymptomatic subjects who ate the same mushrooms
were also included, if they too were evaluated in
the ED, in view of the possibility that symptomatic
patient(s) could be the sentinel event and asymptomatic
subjects would develop symptoms later on.
Exposures clearly judged non-toxic by the PC phy-
sician (e.g. complaints after edible mushrooms inges-
tion) and symptoms unrelated to mushroom toxicity
(e.g. symptoms after rotten toadstool ingestion) were
excluded. Patients for whom the result of urinary ama-
nitin analysis was already available at the time of PC
call were also excluded.
No further patients’ selection was made. Patients
or ED physicians who declined to perform urinary
amanitin analysis were considered a drop-out and were
excluded from the analysis.
Study Protocol
All patients in the study underwent an evaluation
that included history and physical examination. Initial
clinical assessment (see below, under Measurements)
was made and treatment was started accordingly.
Patients with suspected amatoxin poisoning were
treated with gastrointestinal decontamination, repeated
doses activated charcoal, intravenous N-acetylcysteine
and forced diuresis (3); treatment was withhold as soon
as amatoxin poisoning was excluded, and continued in
patients with confirmed poisoning or with a still
pending diagnosis, on the basis of the criteria used to
establish the definitive diagnosis (see below, under
Measurements). The other patients, with a negative
initial clinical assessment for amatoxin poisoning, were
treated according to the suspected disease.
Upon presentation, serum transaminases levels
were obtained, and a urine sample for amanitin
analysis was collected. Patients were evaluated by PC
physicians at least once a day until hospital discharge.
A liver enzymes control on day 3 after mushroom
ingestion was proposed to all patients.
Data were collected by PC physicians on call. In
addition to standard PC medical record, a template for
mushroom ingestions was used, complete with all
relevant information that otherwise may have been
difficult to retrieve in detail from PC charts. Recorded
demographic data were patient age and gender. His-
torical items included the mushroom species supposed
to be ingested, the way the dish was prepared, and
mycological identification when available. Presenting
gastrointestinal symptoms were carefully assessed and
characterized by latency, severity and duration. Clinical
status of subjects who ate the same mushrooms (if
any) was also noted. Initial clinical assessment, uri-
nary amanitin levels, definitive diagnosis, and out-
come were recorded as soon as they were determined.
Additionally, the time of mushroom ingestion, time of
symptoms appearance, time of medical evaluation, and
time of urine sampling were abstracted from the
medical records.
For the study purposes, data were reviewed for
consistency before statistical analysis by one of the
authors. Changes were allowed if they did not alter
patients’ clinical assessment, and were discussed in
advance with the PC physician involved in the man-
agement of the case.
Measurements
The latency of gastrointestinal symptoms was
calculated considering the time elapsed between
mushroom ingestion and the first episode of vomiting
or diarrhea. According to reference textbooks and
current clinical practice, long latency syndromes were
defined by symptoms ensued more than 6 hours after
mushroom ingestion.
Initial clinical assessment was made at the time of
the first clinical evaluation by the PC physician on call,
and was based on patient’s history, initial symptoms
and clinical status in the patient and, when available, in
the other subjects who shared the meal. When
available, the result of mycological analysis was also
included. Since no formal, validated diagnostic algo-
rithm to suspect amatoxin poisoning exists, the initial
clinical assessment relied on the overall judgment of
the physician on call and was taken ‘‘as it is.’’ Initial
clinical assessment was considered positive or negative
for amatoxin poisoning.
Urines were sampled not earlier than 4 hours after
mushroom ingestion. Samples were stored at +2 to
+8�C if not immediately processed. Urinary amanitin
Diagnostic Accuracy of Urinary Amanitin 903

analyses were carried out as soon as biological samples
were received by the laboratory, and were performed
by experienced laboratory technicians unaware of the
patients’ clinical status.
Urinary amanitin measurements were performed
with Bu¨hlmann Amanitin ELISA Kit (Bu¨hlmann
Laboratories, Allschwil, Switzerland). The method has
a stated functional least detectable dose of 1.5 ng/ml
(28), but the cut-off value is not clearly established.
Therefore, in the study three cut-off values to define
positive results were evaluated: 1) 1.5 ng/ml, the
functional least detectable dose of the method (28),
2) 5.0 ng/ml, an arbitrary cut-off, and 3) 10.0 ng/ml,
the cut-off previously used for the RIA method (29),
initially suggested also for the ELISA kit (15).
Definitive diagnosis was made at the time of hos-
pital discharge; it was discussed and agreed between
the PC physician and the hospital physician in charge
of the patient. Correspondence between study definitive
diagnosis and ICD-9-CM (International Classification
of Diseases—Ninth Revision—Clinical Modification)
discharge diagnosis was not considered, because of
the lack of detailed items to code mushroom-related
diseases. Definitive diagnosis was established on his-
tory, symptoms at presentation, overall clinical course
and laboratory results. Diagnostic criteria for ama-
toxin poisoning were 1) the presence of gastrointes-
tinal symptoms after wild mushrooms ingestion and
2) two among the following criteria: symptoms ensuing
after more than 6 hours after mushroom ingestion,
clearly positive urinary amanitin levels (�10.0 ng/ml),
increase in serum transaminases greater than 3 times
the normal values. In collective poisoning cases, the
diagnosis made in subjects who shared the same mush-
rooms was also considered. In the study, definitive
diagnosis was considered positive or negative for
amatoxin poisoning.
Data Analysis
Descriptive statistics, Fisher exact test and Mann-
Whitney test were used to evaluate differences in
demographic and clinical characteristics between drop-
out cases and enrolled patients, as well as between
amatoxin poisoned patients and subjects with other
definitive diagnosis. A value of p<0.05 was consid-
ered significant.
SENS, SPEC, PPV, NPV and DE (and exact 95%
confidence intervals for binomial proportions, CIs)
were calculated for urinary amanitin analysis, symp-
toms latency, and initial clinical assessment, consider-
ing the definitive diagnosis as the reference standard.
DE was defined by the number of true results out of
the total number of assessments performed. Correction
for the prevalence of the disease was not applied (30),
since amatoxin poisoning prevalence is not realistically
known and may have relevant annual and geographical
variations (31–33).
Receiver operator characteristic (ROC) curves
were constructed for urinary amanitin, and the area
under the curve was calculated. The agreement
between urinary amanitin analysis and initial clinical
assessment was analyzed by the McNemar’s test. A
value of p<0.05 was considered significant.
Data analysis was performed on the entire
population, and in the subgroup of patients presenting
in the ED within 36 hours after mushroom ingestion.
Statistical analyses were carried out using Analyse-it
for Microsoft Excel software version 1.67 (Analyse-it
Software Ltd, Leeds, UK).
RESULTS
Patients Selection
In the study period, 126 eligible cases were iden-
tified. Excluded patients were 4 cases from geograph-
ical locations not covered in the study, 15 cases with
symptoms after ingestion of edible mushroom, and
8 cases with symptoms clearly unrelated to mushroom
toxicity. In 38 cases that met the inclusion criteria but
did not meet any exclusion criteria, the patient or the
ED physician declined to perform urinary amanitin
analysis (drop-out cases). Definitively, 61 patients were
enrolled in the study.
Drop-out cases were significantly different from
enrolled patients, based on the latency of gastrointes-
tinal symptoms (long latency syndromes were observed
in 10.5% of cases vs. 42.6% in enrolled patients,
p<0.05). Study diagnostic criteria were not applied to
drop-out cases, because of the lack of amanitin levels;
overall clinical course was highly suggestive for
amatoxin poisoning in one patient.
Definitive Diagnosis
Among 61 patients included in the study, defini-
tive diagnosis of amatoxin-containing mushroom poi-
soning was made in 10 cases (16.4%). Acute hepatitis
with ALT values greater than 2000 UI/l was observed
in 4 cases. Mild to moderate increases in serum trans-
aminases were observed in 4 additional patients. No
deaths were recorded.
In the remaining 51 patients, definitive diagnosis
was mushroom poisoning due to gastrointestinal
904 Butera et al.

irritants (28 patients), mushroom-induced gastroenteri-
tis (15 patients), Amanita pantherina poisoning (1 pa-
tient), and non-toxic gastroenteritis (5 cases); two
asymptomatic subjects who presented in the ED after
eating the same mushrooms than symptomatic patients
(subsequently diagnosed with non-toxic gastroen-
teritis) were discharged after a short observation
with no illness. None of the patients who attended
the follow-up visit on day 3 (n=22) showed liver en-
zymes abnormalities attributable to an overlooked
amatoxin poisoning.
Clinical characteristics of patients with definitive
diagnosis of amatoxin poisoning, compared to patients
with other diagnosis, are shown in Table 1.
Performance of Diagnostic Tools
Twenty-six patients presented in the ED because
of gastrointestinal symptoms with latency longer than
6 hours. Among them, definitive diagnosis of amatoxin
poisoning was made in 9 cases. One additional case of
amatoxin poisoning was observed in the group of
patients with a shorter latency.
Initial clinical assessment made by PC physicians
was positive for amatoxin poisoning in 30 patients. The
suspected diagnosis was confirmed in 10 of them. No
cases of amatoxin poisoning were observed among
patients (n=31) in whom the initial clinical assessment
was negative for amatoxin poisoning.
Urinary amanitin levels were evaluated on samples
collected 5.5 to 92 hours after mushroom ingestion; in
8 patients urine samples were collected more than
36 hours after mushroom ingestion because of late ED
presentation. The result of urinary amanitin analysis
was available to be included in the patients’ clinical
management within 12 hours after sampling in 58.3%
of cases, and between 12 to 24 hours in 37.5%. In two
cases (4.2%) the analysis was performed 2 days after
sampling because of shipment delay. Urinary amanitin
levels, according to the definitive diagnosis, are
Table 1. Demographic and clinical characteristics of the study patients according to definitive diagnosis.
All patients (n=61)
Definitive diagnosis
Amatoxin
poisoning (n=10)
Other
diagnosis (n=51)
Age (years, mean±SD) 45.6±19.8 54.9±15.9 43.8±20.1 n.s.
Gender (% of males) 65.6 66.7 60.0 n.s.
Time from mushroom ingestion to
. Onset of symptoms (hours, mean±SD) 6.4±4.2* 10.0±3.2 5.6±4.1* p<0.05
. Clinical evaluation (hours, mean±SD) 17.7±18.4 40.2±28.7 13.2±11.7 p<0.05
. Urine sampling (hours, mean±SD) 20.7±20.2 41.0±28.8 16.5±15.1 p<0.05
Urine sampling within 36 hours (%) 86.9 60.0 92.1 p<0.05
Mycological identification (%) 18.0 0.0 21.6 n.s.
Acute hepatitis, liver toxicity (n. patients) 8** 8 0**
*7 Patients who presented in the ED with neither vomiting nor diarrhea were not included in the analysis.
**29 Patients missed follow-up visit on day 3 for serum transaminases control.
Table 2. Urinary amanitin levels according to definitive diagnosis.
Urinary amanitin levels (ng/ml)
<1.5 1.5–4.9 5.0–9.9 �10.0
All patients included in the study (n=61) 45 10 1 5
Patients with amatoxin poisoning (n=10) 3 1 1 5
Patients with other diagnosis (n=51) 42 9 0 0
Patients evaluated within 36 hours (n=53) 41 6 1 5
Patients with amatoxin poisoning (n=6) 0 0 1 5
Patients with other diagnosis (n=47) 41 6 0 0
Data in the table refer to the number of patients.
Diagnostic Accuracy of Urinary Amanitin 905