An Attack with Sarin Nerve Gas on the Tokyo Subway System and Its

Chapter 22

An Attack with Sarin Nerve Gas on the Tokyo Subway System and Its Effects on Victims

Downloaded by UNIV MASSACHUSETTS AMHERST on September 23, 2012 | http://pubs.acs.org Publication Date: December 20, 1999 | doi: 10.1021/bk-2000-0745.ch022

1,4

2

3

1

1

Y. Ogawa , Y. Yamamura , H. Ando , M . Kadokura , T. Agata , M . Fukumoto , T. Satake , K. Machida ,O. Sakai , Y. Miyata , H. Nonaka , K. Nakajima , S. Hamaya , S. Miyazaki , M . Ohida , T. Yoshioka , S. Takagi , and H. Shimizu 1

1

3

1

1

1

1

1

1

3

1

1

1

1

Jikei University School of Medicine, Minato-ku, Tokyo 105-8461, Japan Professor Emeritus, St. Marianna University School of Medicine, Miyamae-ku, Kawasaki 216-0015, Japan Criminal Investigation Laboratory, Metropolitan Police Department, Chiyoda-ku, Tokyo 100-8929, Japan 2

3

On the morning of March 20, 1995, the Tokyo subway system was attacked with nerve gas. Liquid, in plastic bags, left on the subway cars was analyzed and sarin, hexane, and N,N-diethylaniline were detected as the main components. The health effects of victims were studied through questionnaire and hospital records. Plasma Cholinesterase levels were used as exposure indicators. Muscarine-like symptoms appeared generally and could be used as early warning signs. Nicotine-like symptoms and effects on central nervous system appeared in more severely exposed cases suggesting that they can be used as severity indicators. Muscarine-like effects to the eye and respiratory system must be induced by the direct contact of sarin gas to mucous membranes. On the other hand, nicotine-like effects are caused by a systemic exposure to sarin.

On the morning of March 20, 1995, the Tokyo subway system was attacked with nerve gas. According to an announcement from the Tokyo Metropolitan Fire Department, twelve people were killed and 5498 were injured (1). Checking the hospital patients suffering from the attack, we found that shrinkage of the pupil was a common symptom, and that plasma Cholinesterase levels of many patients were under the normal range. These symptoms indicated that some kind of organophosphorus substance must be the culprit, and it strongly suggested the use of nerve gas. This incurrent address: National Institute of Industrial Health, 6-21-1 Nagao, Tama-ku, Kawasaki 214-8585, Japan. ©2000 American Chemical Society In Natural and Selected Synthetic Toxins; Tu, A., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1999.

333

334


cident reminded us of the Matsumoto's case, which was thought of as sarin attack executed by a religious cult Aum Shinrikyo in June 27,1994. The liquid left in the plastic bags on the subway cars was analyzed, and sarin was detected as a major component. There are a few reported incidents of accidental exposure to sarin (2-4) and a single report on experimental exposure (5), but there had been no study of mass exposure to sarin until the attack occurred at Matsumoto (6,7). Reports of severe cases (8-11), relation between pupil size and acetylcholinesterase (AChE) level (12,13), and a survey of subjective symptoms in patients (14,15) affected by the Tokyo attack have already been reported to journals, but there has not yet been a formal evaluation of the relation between the occurrence of symptoms and dose of exposure to sarin. In this study, health effect of victims at acute phase was studied and its relation to the exposure dose was analyzed. Analyzing Sarin Attackers left plastic bags filled with liquid in the subway cars. Volume of liquid in each bag was about 600 ml and was separated into two layers. The top layer was a transparent liquid of about 200 ml, and the bottom layer was a dark brown liquid measuring about 400 ml. A group of Metropolitan Police lead by one of our authors analyzed these materials using GC/MS analysis, NMR, and other methods. GC/MS Analyses. The total ion chromatogram of the material left in the subway cars was obtained using temperature programmed methods and capillary column chromatography (Figure 1). From retention time analysis and mass spectrum data of standard material of each substance we identified diisopropyl-fluoro-phosphonate (DFP) at a retention time of 6 minutes, diisopropyl-methyl-phosphonate (DIMP) at a retention time of 7.2 minutes, a small peak of triisopropyl-methyl-phosphonate (TIP) at a retention time of 8.2 minutes, and a huge peak of N,N-diethylaniline (DEA) at a retention time of 8.8 minutes. A peak at 4 minutes retention was suggested to be sarin from the peaks of its fragmented ions in EI-GC/MS (electrical ionization mode) spectrum,fromwhich we could detect demethylated substance of sarin at m/z 125 and methyl-fluorophosphonate dihalide at m/z 99, we could not detect the ion of sarin itself which must be at m/z 140. From CI-GC/MS (chemical ionization mode) we could detect molecular-related ion (M+l) of sarin at m/z 141. 3l

NMR Studies. Using P-NMR spectrum to detect phosphorus atom signals we received a spectrum as shown in Figure 2. Signals at -10.42 ppm and 29.32 ppm were identified from their standard materials as DFP and DIMP respectively. A signal at 29.62 ppm (d, J=1037 Hz) was within the range of the chemical shift of sarin (16), the

In Natural and Selected Synthetic Toxins; Tu, A., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1999.


335

S a r i n DEA

DIMP

DFP

T IP • ï « • 4.00

1

»ï • 5.00

TfTi 6.00

ï

i^T"i~T 7.00

JU ï ï

ï

ï

8.00*

ï

; ï

9.00

ï

I

• I l '

10.00

I

I

1

I

'

11.00

Figure 1. Total ion chromatogram of the substance left in the subway cars.


'

'

I »

min.


336

S a r i n

c52 2 . 2 1 /

\

; = 9 6 2Hz

(MP F )

D

F

P

D IMP JL ppra

60

50

40

30

20

10

0

-10

Figure 2. ^P-fHjNMR spectrum of the substance left in the subway cars.


337

signal was at 28.95 ppm (d, J =1036 Hz) (17); we concluded that this signal was from the phosphorus atom of sarin. These substances identified by NMR studies were comparable with those identified by the GC/MS analyses. Unfortunately we could not identify the substancefromwhich we received the phosphorus signal at 22.21 ppm (d, J=96.2 Hz).


PF

Analysis of Sodium Ethoxide Reaction Products. Sodium ethoxide was added to the material left in the subway cars and mixed. The water soluble portion and ether soluble portion were separated and both portions were treated to induce tertbutyldimethylsiryl (TBDMS) derivatives (18) and analyzed using EI-GC/MS. We could not detect the peak of sarin but could detect the peak of methyl-ethylphosphonate isopropyl ester in the mass spectrum analysis of ether soluble portion. In the spectrum of the water soluble portion we detected TBDMS derivatives of methylphosphonate monoethyl ester, methylphosphonate (MPA), and isopropylmethyl-phosphonate (IMPA). The reactions which occurred between the materials and sodium ethoxide can be explained by the mechanism shown in Figure 3. So called Williamson's synthesis shows that sodium ethoxide reacts with halogens but not with hydroxy residue (19). Accordingly, sarin was converted to methyl-ethyl-phosphonate isopropyl ester, but MPA and IMPA were not affected and remained intact. The important aspect of this reaction was that methylphosphonate monoethyl ester was produced, which means that a content of the material in the bags was methylphosphonate-fluoride (MPF). Considering this result and the structure of MPF, 3l

the signal in the P-NMR spectrum which we could not identify must have been from MPF. Concentration of the Contents of the Material.

The existence of sarin in the

materials left on the subway cars was proved using GC/MS analyses,

31

P-NMR

studies and other methods. P-NMR spectra were measured in *H decoupling mode. Sarin and its related chemicals do not have hydrogen atom bound to phosphorous atom directly, and furthermore, the phosphorous atoms were coordinated by four atoms other than hydrogen. This means by stereochemistry that there are no influential hydrogen atoms near the phosphorus atoms. These facts suggests that there is little change in P-NMR signals by nuclear Overhauser effect (NOE). Using trimethylphosphonate as an internal standard for the phosphorous contents of the liquid left in the subway car, we analyzed it with P-NMR. Combining these results and the results obtainedfromthe -H-NMR spectrum, and gas chromatography, the concentrations (w/w) of phosphorous compounds sarin, MPF, DIMP, and DFP were found to be 35%, 10%, 1%, and 0.1% (trace amount) respectively, and those of 31

3l

31


338 hexane and DEA, which were used as solvents, or a reaction promoter, were 16% and 37% respectively.


Health Effects The main substances used in the attack were DEA, sarin, and hexane. Clinically manifested symptoms suggested the effects of organophosphorus anticholinesterase compounds. The toxic effects of DEA are similar to those of aniline but less severe. The typical sign of exposure is cyanosis due to methemoglobinemia (20). The acute toxic effects of hexane are euphoria, dizziness, and numbness of limbs, but in those toxic cases the concentration of inhaled hexane is very high and its odor would be irritating (21); this was not the case in the subway attack. According to the collected facts on the attack, the main agent responsible for the toxic symptoms in the victims was decided to be sarin. We collaborated with doctors from 7 hospitals to conduct a survey about the health effects of sarin on people suffering from the attack. This survey was intended to study their subjective symptoms as they occurred at acute phase, and its time course plus the relations between the incidence of those symptoms and the dose of the exposure. Exposure Indicator. We needed a dose indicator of the exposure to sarin. Organophosphorus agents bind to AChE and inactivate them, which is an irreversible reaction with nerve gas. These reactions produce; 1) stimulation of muscarinic receptor responses at autonomic effector organs; 2) stimulation, followed by depression or paralysis, of nicotinic receptor at all autonomic ganglia and skeletal muscle; and 3) stimulation, with occasional subsequent depression, of cholinergic receptor sites in the central nervous system (CNS) (22). Due to this relation, reduced amount of AChE activity correlates with the intensity of manifested effects in the acute phase (5,23). We used the activity of butyrylcholinesterase or benzoylcholinesterase in plasma, also known as plasma Cholinesterase (ChE), as a dose indicator of exposure either than the activity of AChE (24). This enzyme is not a specific esterase for acetylcholine but it can be inactivated by organophosphorus agents (22). Accordingly, the change of this dose level in plasma does not directly correlate with the toxic symptoms by organophosphorus agent, but rather it works as a dose indicator of the exposure; further more, the activity of this enzyme in plasma has been commonly measured in clinical situations when organophosphorus pesticide poisoning was suspected (25). Methods of ChE activity measurement were different among hospitals. To construct a common indicator we used the ratio of ChE level to lower limit value of


339 normal range in percent (%ChE) instead of ChE level itself which was defined as follows. %ChE =

100x(ChE level) " (fower limit value of normal range of ChE level)


%ChE was compared with the reduction rate of ChE level (rChE) , which is defined as follows. {(level after two years) - (level at the day of the attack)} (level after two years) rChE correlates well with the dose of exposure to organophosphorus agents and is considered as a reliable indicator (23,26). Figure 4 shows the relation between %ChE and rChE among 32 victims. It shows a good linear relationship between the two variables especially under 100%. Subjects. As summarized in Table I, 1089 people suffered from the attack and had medical treatment at 7 hospitals. All of these people were listed and at the end of May 1995 questionnaires were mailed to them. Among those people 681 subjects responded to the questionnaire and agreed to participate in our survey. The response rate was 62.5%. Participants consisted of 369 men and 312 women, with an average age of 40.2 (14-72) and 28.2 (16-70) respectively. Parenthesis shows the range. We checked the hospital records further and selected those people who had their ChE level measured on March 20 and before treatment, because PAM (pyridine-2aldoxime methyl chloride; pralidoxime) recovers ChE level drastically within a few hours (27). Finally 454 subjects remained, 259 men and 195 women, with an average age of 39.8 (16-70) and 28.2 (16-70), respectively. Figure 5 shows the time course of the cases treated by PAM. The low ChE level cases, treated with-in a certain hour following the exposure (our cases were all within

Table I. Study Subjects Male

Female

Respondent Subjects with ChE

Total 1089

Mailed 369

312

(40.2)

(28.2)

259

195

681

(62.5%)

454

(41.6%)

(28.2) (39.8) ( ) under the number: Average age ( ) outer right column: Rate to mailed number in percentage


340

O CH -P-F

+

3

6cH(CH ) 3

NaOC H 2

^

s

3

2

s

OCH(CH ) 3

2

2

W a t e r - i n s o l u b l e

S a r i n

9 Downloaded by UNIV MASSACHUSETTS AMHERST on September 23, 2012 | http://pubs.acs.org Publication Date: December 20, 1999 | doi: 10.1021/bk-2000-0745.ch022

CH -P-OC H

+

CH,-p-F

NaOC H 2

CH -P-OC H 3

s

2

s

OH

OH

W a t e r - s o l u b l e

MPF

0 CH3-Ç-OH 6CH(CH ) 3

2

+

IMPA

NaOC H 2

U n r e a c t i v e

5

0 CHj-P-OH OH MPA Figure 3. Results of sodium alkoxide reactions with the substance left in the subway cars.

1.0 0.8 H

0.6

W U

0.4-1

„ o %o 00 o o O o

0.2 H

0

#

0

0.0

0

100

200

%ChE Figure 4. Relation between %ChE and rChE among 32 victims.



Figure 5. The time course of ChE level of cases treated by PAM. (Left) the cases of very low ChE level on arrival. (Middle) the cases that were in normal range of ChE level on arrival. (Right) the cases that were not treated by PAM. (A) On arrival, (B) after administration, (C) next morning.


342 4 hours), quickly recovered to the normal range. However, recovery of untreated cases and that of normal level cases were not prominent.


Figure 6 shows the histogram of the %ChE of our study group. Proportion of the people whose %ChE were under 100% was 28.2%. The 454 subjects were classified into two groups. High %ChE group (H), %ChE of whose members were over or equal 100%, and low %ChE group (L), %ChE of whose members were below 100%. Questionnaire. The questionnaire we used can be divided into two parts. The first part was asking about the exposure situations the participants experienced: times when they were aware of the involvement, where (in the subway car, on the platform, or at other places) they were involved, and when and how they accessed doctors for medical care. The second part asked about the symptoms which the participants experienced. The symptoms constituted three groups: 1) muscarine-like effects which show over secretion of glands, bronchoconstriction, contraction of sphincter muscle of iris and ciliary muscle, increased motility of gastrointestinal tract, and urinary bladder dysfunction: sneeze, rhinorrhea, cough, nasal speech, sore throat, dyspnea, dimness of vision, constricted visual field, eye pain, increased lacrimation, blurred vision, nausea, vomiting, headache, increased sweating, increased salivation, anorexia, abdominal pain, diarrhea or tenesmus, frequent and involuntary micturition; 2) nicotine-like effects which shows impaired muscle movements: dyspnea, double vision, slurred speech, disturbed mouth movement, muscular twitching or cramps, gait disturbance, difficulty in standing, dysphagia, generalized weakness, muscle pain; and 3) effects on central nervous system (CNS): dizziness, headache, slurred speech, gait disturbance, difficulty in standing, generalized weakness, anorexia, insomnia. Other selected symptoms included eye irritation, dys-osmia, numbness of extremities, nasal bleeding and subfever. The total symptoms we asked about were 37. Among them, 23 questions were from the questionnaire used at the survey on the incident of Matsumoto (7), the other questions were added by us referring to the study of Grob (2). The occurrence of these subjective symptoms were asked in four categories; symptoms which occurred during the whole course, symptoms which appeared earliest, symptoms which disappeared earliest, and symptoms which still remain. Statistical Analysis. Comparison of group means was first tested by ANOVA and then comparison between two groups was tested by Bonferroni correction. For comparisons of prevalence of symptoms among subjects belong to group L and group H, statistical significance was determined by Fisher's exact p-value. SAS statistical package was used for calculation and analysis. Results.

The distribution of the places where subjects were involved were shown



343

6 «M

O

%ChE

Figure 6. Histogram of %ChEfromour study group.


344


on Table IL Among 681 people 31.7% were involved in the subway cars and 43.3% were involved on the subway platform. Five hundred and ninety four people visited doctors to get medical care on the same day. Median visiting time of 594 people was 11:00 (8:00 - 22:00), which is approximately 3 hours after the incident. The methods used to reach clinics or hospitals were: by themselves 63.6%, by ambulance 14.3%, and with some one's help 13.6% (Table III). The proportion of cases belonging to group L was unevenly distributed among the groups of four places, the proportion in the subway car was the highest. Mean %ChE of the group involved in the subway car was significantly low compared to the group involved on the platform. Mean symptom count, which was defined as a number of positive symptoms during the whole course, of the group involved in the subway car was significantly high compared with those of the group involved on the platform, the group involved in the station and the group involved at other places (Table IV). The frequency of symptoms which occurred during the whole course in percentage were; dimness of vision 64.8, headache 52.3, dyspnea 45.5, constricted visual field 42.3, rhinorrhea 36.4, cough 35.4, eye irritation 33.9, blurred vision 30.5, nausea 27.3, and eye pain 25.6. Frequency of symptoms which appeared most early in percentage were; dimness of vision 35.8, dyspnea 28.9, cough 23.5, headache 17.5, constricted visual field 14.8, rhinorrhea 13.4, eye irritation 12.0, blurred vision 9.5, sore throat 8.8, and nausea 8.1. Frequency of symptoms which disappeared most

Table II. The distribution of places where subjects were involved (Results from 681 subjects) Places

%

In the subway car

31.7

On the platform

43.3

In the station

15.2

Other places

9.7

Table III. Ways to reach medical facilities (Resultsfrom681 subjects) Ways

%_

By ambulance

14.3

With someone's help

13.6

By himself (had symptoms)

53.8

By himself (had no symptoms)

9.8

Others

8.2


345

Table IV. Prevalence of Group L Subjects, Mean %ChE, and Mean Symptom Count (SC) by Places (Resultsfrom454 subjects)


n

SC

%ChE

% of group L"

31

(A) In the subway car

152

36.1

122.9

(55.2)

7.93

(5.05)

(B) On the platform

195 64

26.1

140.8

(56.0)

6.34

(4.16)

20.3

142.5

(55.2)

42

21.4

145.8

(51.4)

5.56 5.38

(3.25) (4.03)

(C) In the station (D) Other places Missing

1

1

Number in ( ): SD a

: p

An Attack with Sarin Nerve Gas on the Tokyo Subway System and Its

Recommend Documents