I Original Article Effects of Hydrogen Sulfide on Neurobehavioral Function Kaye H. Kilburn) MD Background: Nineteen hydrogen sulfide (H:!S)-exposed patients were compared with 202 unexposed subjects. 111j - 1 997-to-2001 cae-referent series was comp8red with 16 (1991 - 1996) case-referent patients. Methods: New patients were bystahders ofH 2 S exposure and [lone had been unconsci.olls, In contrast. 13 members of the prior group were exposed at work and 7 had been unconsclolls. The three groups were compared on the basis of 8 pbysiologic and 12 ps, cbolog; ical measurement·. Observed measurements were compared ilh pre- d.icted ones after adjllsti_og for age, sex educational at1ainment (years). and other signi ficant factors (obse rved/predicted X 100). Results: The IJew group, perfomled poorly compared with unex- posed COrltTOls and were similar to the first group On balance, reac- tion time, color discrimination. visual performance, hearing, Culrure- Fair digit symbol, vocabulary, verbal recall, peg placement, trail mak.ing A and B, and information. Conclusion: HzS impairments associated with H 2 S were similar in 19 worker (44% had been uDcenscioos) and in 16 who had not been unconscious. Key Words: balance impairITlellt, brain dysfullct.ion. bystander. oo.lor discrimjnation elTors, reactio,n time slowlng C -hristison 1 des-cribed deaths attributable to hydrogen sul- . fide (H 2 S) inhalation in 1845. He stated that survivors recovered completely a.nd this was believed for 150 years. But neuropsychological testing, a decade ago, showed per- sistent impairment in six "recovered" patients INho had been From the EnvirolU11enrai Scienc es Laboratory, Kc:ck . chool of Medic-inl!, University of Sourhern California, Los Angeles. CA. The author devel tlp e(1 [he-apparatus 1'0 measure balance. reaction lime. and and ha ol d tht:!>e device. for perfonnRnt: e testing. All participant. provided info nncd consent. and the study prot col WH:" appro ed b the Human rueij es Re &carch Committee of [he Keck $Ch001 of 1edicine. University of Soulbem Califomia. Patients paid for t heir eX t1minclti0Il s .- Reprint requests to Kaye H. Kilbu!1). MD. Environmental Sciences l'Ibora- tory Kcck chool of Medicim'; University of Soulhem California. 2025 Zonal Avenue, SC 201. L s Angele, A 90033. EmaiJ: k::ilburn us .edu Ac elwd March 26, 2003. Copyright 2003 by The Southern Medical Associ,:)(- ion 0038-4348/ 0) /9607 -063 <} ' SOli/hem Medicl1l .Irmmal • Volume 96, Number 7. July 2003 unconscious after exposure , to H 2 S (with one man. being de- mented and. bedfa t)2 3J1d in other groupS?·4 Unanswered questions include: 1) Do exposures wi.thout knockdown to uTlcOnSCioll ne s' have adverse oeQrobehavioral effects? 2) Doe impainrlent follow protracted exposure to H 2 S levels between those detected by the bl1man nose (approximately 30 parts per billion [ppb]) and 5 parts per million (ppm)? 3) Does eX'posure to a few breaths of 5 to 250 ppm of H 2 S cause impailmenf? Although the effects of brief human exposures to 2 to 10 ppm of H 2 S on pulmonary and cardiovascular fu.nctiol1 seem minimaL 5 ,6 they provided no data on neurobc- havioral function. Rats repeatedly exposed to 1-J 2 S at 125 ppm showed Im- paired learning and perfonna.nce speed in. a maze task. ' How- ever. a precise ratio with which to predict hwnan effects 011 the basis of the ratio of rat-to-human effects is lacking. In a rat model for H 2 S effects, altificial ventilation decreased the brain damage in rats that were administered sodiuin sulfide intTaperitoneally while anesthetized with balothane. s Because H 2 S concentrations less than 25 ppm generally only stimulate or do not affect human ventilntion,5 however, damage as a result of hypoxia from apnea i -. IlOt relevant. Impaired brain function in rats exposed to H 2 S 7 : R and in human testini- 4 . 9 recommended system' at-ic testing of additional subjects ex- posed to H 2 S in i_ncidents, "experiments of nature." The like- lihood of neurobehavioral ill e ffects from F-l 2 S, reviewed above, made the deliberate exposure of humans to this entity unthinkable and LLOethi aL ConcenlTations ofH:2S during human exposures vary mo- mentarily without opport'unities for measurements or even "grab" air s3mplil1g in lU1Schcduled slich ambient measurements ' usuaUy lag actual conditions. Although blood Key Points • Central nervo.us system effects of hydrQgen suUide can b l1lea'sured. • exposed at nome are impaired. • Dose measurements are needed but I ppm may be too much. • Brief expo ures 25 ppm e<!]'uate with months at I' ppm , 639
Transcript
I Original Article
Effects of Hydrogen Sulfide on Neurobehavioral Function Kaye H. Kilburn) MD
Background: Nineteen hydrogen sulfide (H:!S)-exposed patients were compared with 202 unexposed subjects. 111j - 1 997-to-2001
cae-referent series was comp8red with 16 previo~s (1991 - 1996)
case-referent patients.
Methods: New patients were bystahders ofH 2S exposure and [lone had been unconsci.olls, In contrast. 13 members of the prior group
were exposed at work and 7 had been unconsclolls. The three groups were compared on the basis of 8 pbysiologic and 12 ps, cbolog;ical measurement·. Observed measurements were compared ilh pred.icted ones after adjllsti_og for age, sex educational at1ainment (years). and other signi ficant factors (observed/predicted X 100).
Results: The IJew group, perfomled poorly compared with unexposed COrltTOls and were similar to the first group On balance, reaction time, color discrimination. visual performance, hearing, CulrureFair digit symbol, vocabulary, verbal recall, peg placement, trail mak.ing A and B, and information.
Conclusion: HzS impairments associated with H2 S were similar in 19 worker (44% had been uDcenscioos) and in 16 bystaJld~rs who had not been unconscious.
C-hristison 1 des-cribed deaths attributable to hydrogen sul.fide (H2S) inhalation in 1845. He stated that survivors
recovered completely a.nd this was believed for 150 years. But neuropsychological testing, a decade ago, showed persistent impairment in six "recovered" patients INho had been
From the EnvirolU11enrai Sciences Laboratory, Kc:ck . chool of Medic-inl!, University of Sourhern California, Los Angeles. CA.
The author devel tlpe(1 [he -apparatus 1'0 measure balance. reaction lime. and bli~.k and ha old tht:!>e device. for perfonnRnt:e testing. All participant. provided infonncd consent. and the study prot col WH:" appro ed b the Human rueijes Re&carch Committee of [he Keck $Ch001 of 1edicine. University of Soulbem Califomia. Patients paid for their eX t1minclti0Ils.-
Reprint requests to Kaye H. Kilbu!1). MD. Environmental Sciences l'Iboratory Kcck chool of Medicim'; University of Soulhem California. 2025 Zonal Avenue, SC 201. L s Angele, A 90033. EmaiJ: k::ilburn us .edu
Ac elwd March 26, 2003.
Copyright 2003 by The Southern Medical Associ,:)(-ion
0038-4348/0) /9607 -063 <}'
SOli/hem Medicl1l .Irmmal • Volume 96, Number 7. July 2003
unconscious after exposure ,to H2 S (with one man. being demented and. bedfa t)2 3J1d in other groupS?·4 Unanswered questions include: 1) Do exposures wi.thout knockdown to
uTlcOnSCioll nes' have adverse oeQrobehavioral effects? 2) Doe impainrlent follow protracted exposure to H2S levels between those detected by the bl1man nose (approximately 30 parts per billion [ppb]) and 5 parts per million (ppm)? 3) Does eX'posure to a few breaths of 5 to 250 ppm of H2S cause impailmenf? Although the effects of brief human exposures to 2 to 10 ppm of H2S on pulmonary and cardiovascular fu.nctiol1 seem minimaL5
,6 they provided no data on neurobchavioral function.
Rats repeatedly exposed to 1-J 2S at 125 ppm showed Impaired learning and perfonna.nce speed in. a maze task. ' However. a precise ratio with which to predict hwnan effects 011
the basis of the ratio of rat-to-human effects is lacking. In a rat model for H2S effects, altificial ventilation decreased the brain damage in rats that were administered sodiuin sulfide intTaperitoneally while anesthetized with balothane.s Because H2S concentrations less than 25 ppm generally only stimulate or do not affect human ventilntion,5 however, damage as a result of hypoxia from apnea i -. IlOt relevant. Impaired brain function in rats exposed to H2S7
:R and in human testini- 4
.9
recommended system'at-ic testing of additional subjects exposed to H2S in i_ncidents, "experiments of nature." The likelihood of neurobehavioral ill effects from F-l 2S, reviewed above, made the deliberate exposure of humans to this entity unthinkable and LLOethi aL
ConcenlTations ofH:2S during human exposures vary momentarily without opport'unities for measurements or even "grab" air s3mplil1g in lU1Schcduled incid~nts ; slich ambient measurements 'usuaUy lag actual conditions. Although blood
Key Points • Central nervo.us system effects of hydrQgen suUide
can b l1lea'sured. • Su~jects exposed at nome are impaired. • Dose measurements are needed but I ppm may be too
much. • Brief expo ures ~bove 25 ppm e<!]'uate with months at
I' ppm,
639
Kilburn • Effects of Hydmgen Sulfide on Neurobehavioral Function
sulfide is a biologic marker, it must be measured qujckly after adding zinc acetate to,the senlm to trap H2S as zinc sulfide. 10
Animal studies for effects of sulfides administer sodium hyposul fide solutions rather than inhaling H2S.ll The H2S ef· fect-to-do e range for human subjects are death at over 500 ppm, eye and respiratory trritation at 100 to 500 ppm, and v.ruiable discomfort of eyes and breathing between 5 and 50 pptu. Human subjects develop ol.tactory "fatigue~ ' so they have no perception of exposure after seconds or minutes.9
Nineteen new patients were enlisted in rhe 4 years after the report of 16 patients in 1997.9 Both exposed groups were compared with lli1expo~ed controls from whom prediction equatiolls were developed 12 and consistency and pattern of response were ' exa:mined.
Methods Nineteen patients exposed to H2S (10 aT work and 9 at
home) from nine states and Alberta were stl1clied (Table 1). The 9 women and 10 men had a mean age of 45.1 years and a mean educationai level of 12.7 years. Their exposl.lres to H1S varied.
H"2S Exposure
Four work exposures WeI' in oil and natural gas site~: five were enviroru:i1~ntal (thre.e from a natural gas storage ~ite); nvo were from hog manure lagoons; two were from buildings' 'sewers; onem a paper mill; two from chemioal explosions' two from workjng with gran.ite and foam glas insulation' and one who lived acros-s the road from a waste dump for gypSWll board. Because these were transient, mainly outdo.or exposures in unstabJe circumstances, there were no opportunities to measure concentrations. Durations of expo~ sure varied from 20 minute to 9 years with five duTIltions less than 24·hours. Subjects were studied 1.7 to 22 years after their acute symptoms.
Commuruty reference subjects were picked at random from voter registration rolls of Wickenburg, AZ, -and were interviewed to exclude occupational exposure to neurotoxic 'chemicals and medical and neurolog"ic diseaSes. All s11bjects gave informed consent and the protQcol was approved by the Human Stit.dies Research Committee of the Reck Sohool of Medicine at the University of Southem California. Referent subjects were reimbursed for their time.
Completed queMionnaireswere checked by c:omputerguided reacting 89 that su.bjects rectified omissions. Questionnaires in.cluded the frequencies of 35 C0mmon health COD)
pJaintsl 3 rated from never to daily 011 an ll~point scale; the American Rhewnatism Association 11 lupus erythematosus questionS· IA a standard respiratory questionnaire; 15 histories of occ-1.lpational and other exposures to chemicals, pesticides and herbicides; tobacco, aLcohol, and dmg use (prescription and ilhcit); history of unconSclousTlCSS' anes thesia; and bead
640
trauma and new-ologie and medical histories. 13 The questionnaires and the neuTophysiol.ogical and neuropsychological test battery had evolved througb previous studies for formaLdehyde effects L6 -of firemen exposed to therrilolysis products of PCBs j 7 and people exposed to toluene-rich chemical waste l 3-and included several groups of unexposed subjects. 12
.13 Alcohol was measured in air expired after a. 20-
second breath-hold lIsing a fuel cell analyzer.
NeUl'opbysiological Tests
Simple reaction time (SRT) and visual two-choice reaction time (CRT) were measured from the appearance on the computer screen ofa J O~em block A to its , canceUation by tapping a keypad A for simple and A or S for 'choice with a computerized instrument. t8 The lowest m-edian score of the last seven in each of two trials of 20 was'· accepted flJ}" 'SRT and for CRT. Body 'balanc:e was measured with the 'subject standing erect with feet together. The position of the head was tTacked by two microphones from a 'soWld-geneTating Stylus on a headband, proc.essed in a computer, and ex:pressed us mean speed of sway in em/s. l9 The minimal sway. speed of three consecutive 20-sec.Qnd trials was the value used for sway \vith the eyes ope_ll and sway with the eyes clos{ld..
The blink reflex was measllred with surface electromyographic electrodes (EMG) from the latera] orbieulari,s ocuLi muscles bilaten,lUjO,21 after tapping the right and left supraorbital notches with a light hammer, which triggered a recording compLlter. Ten fi rings of R -I were averaged to fiI1d the mean response for each side and failures to respond were recorded. 2 1 Color confusion index was measured with the desatllrated Lanthony 15 hue test under constant i1Iumina~ tion2 2 and scored by tbe method of BO\V,ffian?3 Heari.ng was measured in the left and right ears with .standard audiQmeters (Model ML-AM; Microaudiometrics, S'outh Daytona, FL) at stepped frequencjes of 500~8,000 Hz. The sum of deficits for each ear was the hearing score.
Threshold. testing of visual 'fields used a 'computerized (JvJed Lab Teclinologies North Wales~ PA) automated peTim;.. eter that mapped the central 30 degrees 'of the right and left eyes individually. l'he performance value of each ey.e was the sum (in decibeJ.s) of the threshold values of 80 points 'Within the central 30 degrees.
Neuropsychological Tests
Immediate memory or recaU was mea~red with nvo stories from Wechsler's Memory Seale-:Revised.24 Culture Fair tested nonverbal nonarithmetic itltelligence based on the selection of designs for similarity clifferellce, completioJl, and pattern recognition and b:ansfer? 5.26 Culture Fair resembles Raven's progressive matrices.27 The 46-word vo'cabulary test was from the mult~dimensional aptitude battery?8 Digit symbol substitution from the Wechsler Adullt lnteUigence
nmirc -Iu Igc Cleantng t I c.ullin g.ran ilc al work
·armer/miJler UT H()g lag,oollJ'
rnlle 01 crator FL Papt>rl11 ill
Pip ~ i ll. ulator
PoUt:c
if d 'lIer
ncmpJoycd
IJi;ablcd. uncmpl ycd
Varmin t' gems
~chur
H tI C', if~
Repairman
PipefiH ~r
- It:: ITici 11
Mechanic
CA
L
Foam gin 0 11
ruea"ffi lin 070 ppm HJ
en mical cxpltd n
OK asmg on il \ ell
CA R fintlf}'
NM
ez;pl :i n
I1Im1 'ns tOl1'lgC
NM NaturaJ gas tOr'dge-
UT Boo Lugooru
CA R/,;\fincry
AL Oil \ ells
lau\ral gas torngc
TX Building scw,.;r
c Refinery
C. 'ewer gi
MI alUml ", \ . ell
FL Wa .re dl lmp
Scale~revi...,ed (WAf -R 2'/ le, t d attention and i.ntegraiive capac.ity. rn fonni1ti n pi lure compl li 11 , and ~ imil ari ties. al 0
f r m the ~ I\J ~R. tc,: ted IOR r'l - ( 'nn ( mbedded) mem ry. Time needed to pIa t: 25 peg~ in the Lafayette slotted pegboard wa mea ured a \.\'ere tim ornplc-re trail mak ing
Lo !\ of ill'mar . • , hOrtlles. 01 brtlllb. futigu
Hl!adill1he. oueh. di orient d ~
Headache. dizziness. breathing distre 's
2 .. 1.0
15. )
I~ .. 'i
Il.O
II.
11.0
10. -
7.
7.S
7.0
.1.
. 1
.0
H )
.0
~ . 1
and hese te Is rom rJle Hal tead-Reitan b, ner)r ) I m a ti red d Xl lity, coord ination. and decision making. Peril)h ral sen ~t ion perception was m asured wi th finge rtip number w- iting rrors. ubjecl m ad' \ ere apprai d b re p Ol to 65 (eml d . ri bing emorional tam for th
641
Kilburh • Effects of Hydrogen Sulfide on Neurobehavioral Function'
Table 2. Hydrogen sulfide-exposed subjects (1.9~ compared with 202 referent subjects as percentage of predicted, ,means, and standard deviations (SD), P valu~s by analysis of variance
19 Exposed Referent mean ± SD mean ± SO P value Holm's P
Age (yr) 45. t .± 11.6 46.6 ± 20.6 .757
12.7 ± 2.2 12.9:!: 2.3 .735
108.1 ::!:: 7.8 99.9 :t 3.7 .0001*
106.4 ± 5.8 100.0 ± 2.5 .0001'"
Educati0Dal ie'Ve.l (yr)
Simple reaction tim~ (ms)
Ch~)ice reacti0fl time (ms
Balance- ~way speed (em! ) Eyes open
Eye closed
Right
208.1 :t 166.4 100.2 :': 20.0 .0001":
.0017*
.0016*
.0015*
.0014:t
.0065*
.0012*
.0310'"
243.1 ± 141.9 103.1 ± 26.8 .0001'*
Blink reflex latency R I (ms)
Hearing losses
Color score
Visual pe'rfonnance
Grip ~t..rengt:h
Culture Fair A
Digit symbol substitution
Vocabulary
Verbal recall
Pegb-oard
Trails· A
TraHs B
Finger writil1~
lnfonnabon
PictuTe completion
Simjlarities
POMS score Frequency of symptoms mean
* = Sf(Jti.s~ica/~" sign(jicallf.
Left
Right
Left
Right
Ldt
RighI
Left
Right
Left
Immediate
Delayed
Right
Left
113.3 :!:: 12.4
lJ-1.4 ± 14.5
118.3 :!:: 46.6
109.7 ± 31.6
64.8 ±. 57.7
53.7 ± 40.2
'81.5 :!z2 1.4
78.0 ± 29.4
89.5 :!: 15.4
93.8 ± 16.7
89,7 ± 25.3
90.0 ± 12.1
77.3 :t 27.1
78.3 ± 29.2
62.3 ± 36.9
81.0 ± 19.3
I09~0 ± 12.6
105.9 ± 9.3
103.4 :!: 8.2
103.3 ± 9,.2
71.5 ± 29.4
75.8 ± 32.6
79.1 :... 36.1
77.1 .± 36.3
5.8 .:!: 1.6
w.eek usi.ng the profile of mood states (POMS)?2 Recall of the Rey 15 forms tested whether recaIJ was appropriate or suggested malingering.33
Respiratory flows and vital capacities were measured from a fun inspiration willie subjects stoqd ·and blew forcefully into ·a volume displacement spirom€ter (Ohio 822; Sensm M-edicis, Anaheim.~ CA) while using a nose clip. Thi's maneHver w.as repeated until two forced expirations agreed within 5%.34 Records of volume-and flows w.ere traced with a digitizer and were measured by a computer. Pred.iction equations adjusted f0r height age, sex, and smoking status.35
Statistical Analysis
Scores and computed data were entered into an ffiMc mpatible microcomputer. Descriptive. and analytical com-
642
99.4 ± 14.6
6.4 ::t: 13.2
101.5 ± 24.6
99.3 :!:: 21 ,.8
102.6 :t 51.l
102.6 ± 51.1
100.0 :t 22.8
101.1 ± 21.7
99.3 :t 17.5
99.t:t17.5
101.2 ± 20.0
101.5 ± 9.2
99.1 ± 3.0.8
99.8 :t 31.1
99.9 :!:: 41.3
101.8 ± 25.7
100.3 ± 8.3
100.4 :!:; 7.5
100.'0' ± 7.5
100.0 ± 7.8
101.5 ± 39.4
99.3 :!:: 32.2
98-, I :t 41.2
21.0 ± 31.6
2.6' :t 1.2
,0005*
.0001"'
.0310"
.0930
.0026'"
.000'*
.0019*
.0002*
.0200*
.2030
.0200~
.0001 ¥
.0030·
.0040*
.0002*
.0007'"
.0010*
.0030*
.0850
. 1125
.0014'"
.0027*
.0530
.00OI'~
.0001'"
.0268:j1
.0010*
.0012'·
.0120·
.0200"
.0200·
.0009*
.0090*
.0:080*
:0014*
.003S"'
.0008'·
.006*
.0056 '~
.0080*
putations adjusted for differences in age, education., sex, height, and weight using stepwise linear regre-ssion modeling that used Stata statistical software (Stata Corp., CoUege StatiOD, IX). These prediction equations were based on measurements of the functions of 202 subjects. Each was symmet,ricaJly dlstributed.l 2 or was transformed mathematically for symmetry. The observed measurements and scores for each patient were compared with i_ndividual predioted values and were expressed as percent -predicted. Then, tl1e exposed group ' s percent predicted values were compared with the control group 's by analysis of variance (ANOVA). (Other factors such as family inconle" hours of general anesthesia. POMS score, .and depressi'on score were testedf0f influence in equations but were excluded because their coefficients wer;e not significant.) S~atistical significance was deftned as P <
Table 3. Comparison of 16 (1997) patients and 19 (2001) patients as mean values of percentage predicted
1997 2001 Referent (16) (19)
Age (yr) 46.6 44.7 45.1
Education level (yr) 12.7 10.0 ,12.9
Balance sway _ peed Eyes open 100.0 1-9.0* 208.0*
Eyes closed ,103.0 146.0* 243.0~
Simple reaction Lime 100.0 151.0;) 108.0·
,Choice ["e.acrion time 100.0 1.30.0* L06.0'"
Blink reflC"x Imeney R-I Right 99.0 87.0· 113.0*
Left 96.0 95.0 [11.0*
Visua'l perfonnance Right 100.0 72.0~ 81 .5*
Len 101.0 55.0* nto'" Color score Ri.gbl 103.0 75.0* 65.0*
Len 103:0 64.0* 54.0"',
Hearing Right 100.0 160.0* 118.0'"
Left 100.0 174.0* 110.0
Gripsu-ength RighI 99.0 94.0 90.0 >1>
Len 99.0 82.0') 94.0
Culture Fair 101.0 'H5.0* 90.0*
DiSH' symbol 104.0 77.0"" 90.0*
Vocabulary 99.0 56.0* 7-7.0*
Verbal recall ImmcdjalC 99. 69 . 0~' 78.0*
Delayed 99.9 60.0" 62.0"
Pegboard 102.0 87.5* 1.0*
Trail making A 100.0 178 .0* tOt.O'"
Trail mak-ing B 100.0 140.0'" 106.0"" Fingcn'ip number RighI 100.0 102.0 103.0
writing error.;
Lft 100.0 104.0 103.0
Information 100.0 .69.0~ 72.0"
Pkture. completion 98.0 "'4.0 76.0*
'imilarities 96.0 R4.0* 79.0
POMS score 21.0 83.2 77.1
SYluptom frequency _.6 4.0 - .8
*== Sta'fi~ .. [ically -ignijicq/IJ.
0.05. AbnomlaJities tor each patient were cOtmted (Table I) after assigning most bilateral tests a value of 0.5 per side (for example, bearing). Visual perf o lTn an ce was scored I per side and balance was assigned 2 for the eyes open test ;:md 2 for the eyes closed test. Two exposure variables, durat.ion and latency from exposure to testing ·and profi.1e of mode states score and depression score, were tested for influence on total abnormalities and sped-fic measl1rements ie, balanc\! wjth eyes closed using regression analysis.
Results
New Subjects
The 19 exp0sed subjects were statistica.lIy significclJ)tly different from Ule unexposed (control) subjec t':! for simple
SOllthern Melfical Journal • Volume 96. Number 7. July 2003
Original Article
and two-choice visual react'ion times (Table 2). Referent subjects had values near 1"00% of predicted (96.4 -103. n for aI/ test.s (Tables 2 ·and 3). Balance was affected as sway speed was increased with eyes open and witb eyes closed . Blink reflex W(lS slowed. Color error scores, a cone function, were abnonnally elevated, and visual perronnance by visual fields <l rod function, wa decreased. Grjp slTength and bearing were abnormal on the right.
Cognitive perf0n11anCe was decreased for Culture Fair, digit symbol substitution. tuld vocabulary. Immediate verbal recall for two stories was reduced and became more abnormal after 30 minutes (delayed). Peg plac'ement and trail making A and B were abnormally slow. In contrast, fingertip number writing errors were not increased. The [1J,nd of information and picture completion (recognizing missing items) scores· of t.he l0l1g-term memory tests were diminished but sjlud~ritie" (clas . ..ifying two items, ie, dog and lion are anima.ls) were not. After adjusti.ng P values for the simultaneous inference, {) aU differences between the groups remained signJficant (Table 2).
Comparison of Groups
Test abnormalities in the 19 oew subjects matched the 16 reported io 19979 with minor vruiation (Table 3 . Only hearing and grip ~lTength varied.; and the stabstical significance·of two long-teml memory tests, picture completio,t1 and simi1ar~ ities tests, reversed. Perhaps having been wlconsciolls . lowed imple reacrion tjme and decreased hearing in the members of
the 1997 grouP. but the smalJ differences did not suggest that unconsciousness was important .
The mean profile of mood states score was e levated almost fourfold (Table 2). Tension, depression, anger, fatigue, and confusion were all elevated significantly; and vigor was below unexposed control values (not detailed). However neither total POMS score nor depression score .influenced total abnormalitie or specific m:easurements. The mean frequency of 35 symptoms was 5.8. more than double the mean of 2.6 in unexJJosed (control) subjects (Table 2).
Respi.ratory symptoms were significantly more prevalent in exposed subjects. particularly shortness of breath and wheezing (Table 4). Their mean pulmonary funotion values were slightly higher than unexposed subjects and midtlow (FEV 25- 75 ) and the ratio of I-second forced expiratory volUlrie to vital capacity (FEY /FVC were 'significantly better.
Neither duration of exposute nor the time between exposure and testing latency had significant coefficients for total abnormalities or specific measurements by [egression anaJy is.
Discussion The replication of ob e.rvations in a second case-referent
seJies of patients exposeo to H2S added e:vid nee for neorobehavioral abnomlalities after H_S exposure and confirmed
643
Kilburn • Effects of Hydrogen Sulfide on Neurobehavioral Function
Table 4. Respiratory symptoms e~posed and control compared by ANOVA with P values
Exposed (19) mean ± SD Unexpos.ed (202) mean ± SD p
Pblegm
hort of breath
Wheezing
Shon of breath with wheezing
Rc.st
Walking
Stair '
36.8
63.2
84.2
100.0
57.9
6},2
10.3
5.2
8.6
32.S
IO.}
13 .8
.002*
.OOOl *
.0001 *
.0001*
.0001 *
.0001*
Pulmonary function for 19 hydrogen-sulfide exposed compared with 202 unexposed subjects
Exposed mean ± SD Unexposed mean ± S.D p
five rEv.1
r EF,s. 7S
FEF75-8.ti
FE JIFVC
,., sTafistica lly significant.
*. j? .\·'(Jtis'ical~r sigl1ificant~\· betfer than COli troIs.
98.0 :t 14.6
93 .7 :t 14.2
105.9 2: 29.4
86 .0 :t' 35.1
77.7 ± 5.6
our e'arlier sene .Y These included. both physiologic and psychological impairrrient. Long periods between exposure and testing suggest' t.hat these effects are permanent. The absence ofuJ1consciotlsness after H2S exposure to the new group compared with 44% who had been unconsciousness in the 1997 group suggests that this is not a determinant of protracted impainnent from H2S and fails to support the argument that hypoxiaJ 7 is neces&ary for neurobehaviora.l damage. The twofold elevated symptom frequencies and eleva.ted scores for depressi'on. tension and confusion. shown on POMS suggest that adverse effects on the limbic brain sites of emotion and memory coincided with physiologic impairment. In addition, the frequency of 35 symptoms was elevated in all eight cat· egories: irritation, indlgestion balance mood. sleep, memory, limbic, and respiratory. HzS appears to damage three domains: the physiologic, the psychological, and the moods or limbic. Tn control subjects unexposed to neurotoxic chemicals, the tbtee domains were independent, but damage from H2S caused them to increase togctber. 13
Exten ive brain damage is deduced 'fi'om t.he impaimlent of vision (second cranial nerve and oGcipital c"Ortex); blink (cranjal nerves V and VB)' hearing and balance ,(Cranial Nerve VI'll, cerebellum) proprioceptive and motor effeetor tracks).; reaction time (vi ual perception and eye-hand coordination, parietal lobe) ' and as ooiative and memory areas of the temporal occipital and parietal frontal lobes. Deere·ased memory and upset moods indicate impaired temporaJ lobe and limbic system function .
The five patients with minutes to hours of ex.posure ,averaged 11 abn"Om1alities compared with 8.6 tn others . This suggested that patients who were brietly exposed received 'higher doses ofH ·S and thus~ bad greater effects than those
644
10l.6 :t 15.1
93 .6 ± 15 .8
88.1 :!: 35.0
78.1 ± ~2.7
72.S .± 9.5
.320
.9 3
.033 ",ot.R
.521
.029 "'.* R
who were exposed for longer periods of time. Neither duration of exposure nor latency to measurement was a factor in total abnormalities or speci fie measurements that cannot 'be considered surrogates for exposure to H1 S.
Limitations
These observations were made as retrospective analyse.s of symptomatic. people (this led to an i.nability to control all confounders). Collected p'eople environmentally exposed to H2S, are less homogenous than occupationally exposed groups, Of the workers, only two were in oil refmeries. A dose-response analysis was not possible because H:2S concentrations were rarely measured. Workplaces should be monitored; however, only 6 of these 1.9 patients were exposed at work and only 3 were in petroleum recovery and reilning where air anaJysis is done. Other reduced sulfur gases such as carbon oxide sulfide, mercapt'lns, and thiophene should be measured simultaneously because they contribute to toxicityY' 10
The possibility thaf test results were altered consciously by patients to increase evidence of adverse effect (because they were contemplating lawsu.its) seems unlikely because of the cO'nsistency and appropriateness of test results. t11e absence 0 f effect of mood state scores on specific meaSLlrementS and on total abnortnalities, and the similarity to the earlier series (Table 3),9 The 14 physiologic measurements, simple reactioD time through grip stTength (Tables 2 and 3 resjst conscious interference that produces easily recognized inconsistencies in multiple trials and between tests in the same domain. This second group of exposed subjeGts essentially dupticaled the results of the earlier seties.9
(~ 2003 SOJJ.(hel'fl ·Mediaa! AssociaJion
Mechanism
It is Dot clear bow the inhibitillg effects of H2S on 1nJ.
tochondrial oxidation equate with enhanced metabolism of H
2S at this loCUS'O,II . ..>8 .39 but failure of neuronal respiration
cou ld account for the apparently enhanced damage from doses of HiS above 50 ppm acting briefly (as after a few breaths in human subjects).
As sulfide binds iron in cytochrome enzymes in mitochondria, hair ceUs in the semicircular canals and cochlea cease functioniog - decreasing the hearing and balanoe function and probably damaging the ba.lance-correcting pathways from the cerebellum and through vision. 38
.39 Impaired bal
ance leads to falls and is a seriou impainnent. The metabolic demand of the retina 1S high, especiaUy the cones of the macuJa and the rods furthest from the optic disk. 8 Cognitive functions, thns intelligence: are reduced and attributed to overall brain slowing as ceUs ·are killed or damaged by H2S.
If catecholamine and 5-hydroxytryptamine levels in the brainstem are enhanced in huma.n sLlbjects (as was shown after sul'fi.de administration in rats), respiration would be stimulated. 11 A possible mechanism was proposed when sulfide reversibly aboli. hed Na + currents in a model ystem of neuroblastoma cells.4o Such effects were pre iously found by others in synaptosomes.41 Learning and memory ill the rat, as measured in maze 'running, were adversely affected by repea.ted exposures to H2 S at doses of 125 ppm for five 8-hour days. n.us impairment lasted for weeks after cessation ofH2 S exposu.re.! matching tl1c permanence we observed in human subjects.
lrritation of the a.irways has been altribu ted4 ~ 6 to the hydratloD of H2S, fomling Ilydmsu.lfurolls acid. The absence of ai:rway obstruction suggests that airways are more resilient than the brain to H2S effects. Effects of rl}S may extend to the heart as coronalY disease mortality was more elevated in Finnish pulp mill workers exposed to H2S (standard mortality ratio [SMR] 1.50) than in those exposed to sulfur dioxide .(S02) (S.MR 1.23) a.fter adjusting for smoki.ng and common risk factors .42
Brain exci.tation folJowed by torpor and collapse was found in workers exposed to carbon disulfide (CS2 ) and re:. sembled the effects fmm H2S.43
.44 Also, epileptiform seizures
and psychosi.s ha e been described resu.lting from H2S:n and CS2 exposure.45 Because the effects of H2S and CS2 exposure (U'e si.milar central nervous system effects should be monitored in subjects after HzS exposures.7.
44,45
The physiologic effects of H2S found in human slIbjects suggest additional animal experiments on mechanisms. Further human experimental studies5.6 appear unethica.l in light of those findings. M~anwhjle, physicians should adv]se their patients that the 'rotten egg odor promi es haml and that avoidance ofH2S exposure is essential to preserve their brain function. Where human expo me from leaks. flres and explosions are Likely. monitoring H2S and reduced sulfur gases should be
Southern Medical .Journal • Volume 96, Number 7. July 200)
Original Article
combined with an as e sment of workers) nClirobehavioral functions. Measurements of H2S concentrations in Ol,lnUre
lagoons, landfi.1ls. sewers, and oth r nonperrolewn hum.an exposures woul.d help develop dose-response relationships.
References 1. hristi son R. A Treatise Oil Poi OilS in RelllliolJ to Medical '/urispru
dence. Physiology. tlnd Ihe P1'llctice afPhysic. Edinburgh. Black, 1845. ed 4.
2. Tvedl B kyberg K, Aaserud 0, l-Iobbesland A, Marhiesen T. Brain damage cau ed by. bydrogen 'ulfide: A follow-up snldy of i patients. Alii J /lId Mer! 1991;20:91 - 101.
3. Waseh 1-IT-I. EstTin Wl Yip P, Bowler R one JE. Prolongation of the ]>-300 latency associated \\~tJl hydrogen sluJlde expo·surc. AI' II Nellral 19 9:46:902- 904 .
4. Kilbul1l KH , Warshaw RH. Hydrogen sulfide and reduccd- ulfur S'\ses adverse.ly affect neurophysiological function . To",icolll'ld Health 1995; 11:1.85-197.
5. 811ambhani Y, Singh M. Physiolo2"ical effeCTS of hydrog~n sulti de. jn· halation during exercise in healrhy !nell . J Apf. I Phy:sio/ 1991 ;7 1.: 1872-1877.
6. Bhal'nbhani Y. Bumham R. Snydmiller G, Maclean r. Effects of IO-ppm hydrogen sulfide inhalation in exercising men and women: ardiovascu13r. metabolic and biochemical responses . .l OCCIlP Ell vi}'@ Med I 997~ 39:122-129.
7. Partlo LA, Saillsbl.\.f RS, Roth SH. t eels of repcated hydrc)l1eo sulph iue (1-12 ) expo ure on Icarning and memory in the adult rnl. Neuroloxicology 2001 ;22: 177- J 89.
8, Baldelli RJ Green FH, uer RN. ulfide 'tOxic ity: Mechanillal ~ef\tilatiol1 and hYf1oteosion detenuin(; urvivaJ rate and 'bra,in n roS'is. J AppZ. Plly.'iol 19 3;75: 1348- 1353.
9. Kilburn KH. Exposme to reduced sulfur gases impairs neurobehavioral function. 501lth Med .J 1997 ;90:997-1006.
10. \,: arellycia MW. Go d\\ lh LR. Benishin eG. Rei.ffenslein RJ. Francom D 1, Taylor ro. et 31. Acute hydrogen sulfide poi oning: Demonstration of electi e uptake of sulfide by the brainstem by measurement of brain tilfide levels. Biochem PhaJ'mllco/ 19 9;3 :973- 9 !.
II. W arCI1) cia MW, milh KA , Bla hko ' S. Kombian SB. Reiffen ·tein R.T. Monoamine o~jdase' inhibilion a a . eque l of hydrogen sulfide intoxicali n: lncrea es in bre in care holamine and 5-hydroxynyptominc Ie -cis. Arch Toxico/1 989;63: 131 - 136.
12. Kilburn KH, Til rnl n JC, Hanscom B. Population-based prediction equations for neurobcha\ ioral tests. Arch Environ HeaLTh 1998;53:257-263.
13 . K.iJblUl1 KH. Warshaw RH . eUTOroxic effects fTom residcbtial expo!'ure to chemicals from all O'il reprocess iilg facility and "upe.rfund sileo Nellmloxico/ Temfo/ 1995; 17: 89 - I 02.
14. Levin RE. Weinstein A. Peterson M. Testa MA. Rothi'ield NF. A comparL On of the sensitivity of the 197 1 and 1982 ' American Rheumatism Associatipll criteria l'br rhe cia sifieal10n of • $lcmic lupl.1 crythemato· sus. Arthrifis Rh 11111 1 '984;27 :5"'O -53~.
IS. Ferri BG. Epjdemjology Standardizanon Project (American Thoracic Society . Am R l' Respir Dis 1978:0 ': 1-120.
16. Kilburn KH. Warshaw R. Tbornlon.rC. Fom)(lldchyde impai.ts memory. eqn ilibrium, and dexterity in h.i stoh)g, tec lllli t:ian : Effec t which persi,. t for day ' after exposure. Arch Ellvil'O,II Health 1987 ~42:117-120.
17. Kilburn KH. Warshaw RH. Shield.s MG . Neurobchavioral dysfunction in ti l:emeu exposed fo polychlorinated biphenyls (P 8s): Po ible improvement a.lkr detoxifIcation. Arch Enviro1/ Nealtll [9 9;44:345- 50.
18. Miller JA. ohen GS. Warshaw R. Thomton JC, Kilburn KH. CllOice ( RT) and simple reaction limes (SRT) compared in laboratory technician .: Fal.tors inl1u{'.nclug re.action limes and fl p~ wctive model. Alii .J {lid t (?d I 89~ 1 :687- 197.
645
Kilburn • Effects, of Hydrogen Sulfide on Neurobehavi'oral Function
19. Kill,uro rG-I. War.shaw RH. Hanscom B. Balance hleasured by ~ead (and truck) tracking and a force plutfonn in che.mically (PCB and TeE) exposed and referent subjects. Oceup EnvirOIl Med 1994; -1: 381 - 385 .
20. hahani aT, Young R.R. Human orbicuJaris oculi reflexes. Nellro/ug,J 1972;22: 1"49- 154.
21 . KllbLUlI KH, fliomlon lC, Hao!)com B. A field method for blink reHex I.arency R-I (Bll R-l) lmd prediction IJquatiofl. for adults and children. E1ecrromyogr /ill Nellrophy. ioi 1 C)98:38:25-3I .
22. Lanthon~1 P. T'J1e desaturaled pan I D-15. Doc Oph/IJalllloll-978;46: J 5-189.
2~. , Bow.man KJ. A method for quantitative scoring of the Famswo!1h Panel D-1 5 . ...I.e/a· OpJulwlil/o/ (Copen") 1982:60:907- · 1·6.
24. Wechsler DA. Wechsler Memory Scale (WMS-Revi ed). San Almmio. TX. The Psycbological Corp. '1";)87; rev ed.
,25 . . attc.ll RB. Classk-al and standard score IQ !;tandardization'oftJlc IPAT: ClJlLUre~t"rC'e intellige.nce scale 2. J Con '·ull P.\~)!(" hol 1951:15 :154-1"9.
26. anen RB, Pdngokl SN. San~:son SB. A cul t'u.re free jntelli.gence. test: Pan 11. F.vallla ~ion of 'culluml influences on test perfoonance. J Edlt(: P::.ycliol 1941·32: 1-100.
27. [~\ven le. C urt Jfi Raven J. Ravt!1I '.\. Sumdl.lrd Progressive Nf(lfrices a:nd} ocabull1r~ Scale.,>,. Ox fon'!. Oxford Psychologj Is Press., L988, t::c1 2.
28. Jackson D . I\-fldlidlmc'ftsionol Aptitude 8qttelJ'. Port· Huron, M] , Sigma ·'ssessment Systems. Tnc. 19-R5.
29. Wec~s ler V A. ~tlec"${er Arill.ll inle(lige1lce Scale Mtlllllal- Revised. New Yol"k, The PsycholegiC3J Corp., 198 1.
30. Reitan RJvt. A research pf0gram on tbe p ychological effects of br'aiu Je:iions iil Innmm beings. Inl Rel Res lv/en/ Retard 1%6'1: 153- 2 J 8.
3:1. Reitan RM. Validity ,of the [ra il-Illakin o ,test a. an indicator of organiC brain damage. Pel."cept AlaI Skills 19!18'8 :27 1- 276.
2. Profile 0 . Mood States. San Diego, Educ.arionul and Industria l Testing Service, 1971J 1981.
:\3. Rey A . L :examell Cliniqlle ell PsycJwlogie . Paris. Presses Univcrsilaires de Fmnce, 1964.
34. American Thoracic Soc.iety. Srandardizat,ion f Spiron'letry. 1994 Update. A'ITJ J Respir Cdl Care Med 1995; 152:J.1 07-11 36.
3S, I\;fil1er A, Thornton Je, Warshaw R. Bernstein.r Se'Tikoff IJ Teirste·in AS. Mean and instantaneous expiratory ·flows FV .and FEV I : Prec\.iclion equations from a ,Probability ample of Mjcbig~. a large industrial sralc. 8,,/1 EliI' PhySivfJLJl~lol R' 'l~ir t 986;12:5 9- 597.
36. Wright ST>. AdjllSI~d P valu.e,s f, r . imulral'leou5 inference. Biometrics 1992"4-8 :1005-1013.
37. Snydcr JW, Safir EF, SLU11merville GP. Middleberg RA. Occupational fatali ty and persistent ne,·urological sequ.elae after mass exposUIe to .hydrogen sulfide. Am J Emerg Med 1995:13: 199- 203.,
38. Wallace DC. Mitochondrial DNA in .;\glng and disease. Sti Am 1997; 277:40- 47.
39. Gt)flnan AM, Cecca'tclH , Orrenius S. Role ofmitochondr.ia in neuronal <lpoptosis. Dev Neurosci 2000;22:348- 358.
40. Waronycia MW, Steele JA, Karpinski E, Reitfe!15!-ein RJ. Hydrogen slI.I:tIde in combination with laurinE-' or cys£eic acid reyerSibly abolishes sodium ClLrrents in neuJobLIlSloma reUs. N Cliroroxi".nlogy 1989: LO: 191-199,.
41. Rafalowska U. Zitlirtg A, avolainen H. MetaboliC changes in rbt brain ynaplosomes after exposure to ulfide in vivo. Toxicol Lett 1986·34:
193-200.
42. Jappinen P Tola S. ardiovnscular ruortalJty among pulp mill wo;rkel'S. Br J [lid J\t/eel 1990:47:259- 262.
43. Ii,L1er JR., Schi.lling RS, Morris .TN. Occupaliof,lal to"Xic fact r in il\Ort:a liry from coronary beart ,ctisease. Br Med .I1.968:4:4,0'7- tl.ll.
44. Gordy ST, Trumper M. Carbon disulfide poisonlng: With a report Of SIX
cases. JAMA 19J8: 110:1543'-1549.
45. Sweetnam PM. Taylor SW. Elwood PC. Exposun: to Carb0Jl dll>ulph.ide and ischael11ic heart di '~ai)~ LI1 a viscose rayon factory:. BI' J 1na lvJed 1987;44:220 22'7.
I Blessed is he who expects no gratitude, For he shall not be disappointed.
- w.e. Bennett
646 ~ 2003 Southern Medical ociat,ion
Copyright of Southern Medical Journal is the property of Lippincott Williams & Wilkins and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
