AN OUTBREAK OF TYPE C BOTULISM IN HERRING GULLS (LARUS

ARGENTATUS) IN SOUTHEASTERN SWEDEN

A. Neimanis,1,5 D. Gavier-Widen,2 F. Leighton,1,3 T. Bollinger,1,3 T. Rocke,4 and T. Morner2

1 Canadian Cooperative Wildlife Health Centre, Department of Veterinary Pathology, Western College of VeterinaryMedicine, 52 Campus Dr., University of Saskatchewan, Saskatoon, SK, Canada S7N 5B42 Department of Wildlife, Fish and Environment, National Veterinary Institute, SE-751 89 Uppsala, Sweden3 Department of Veterinary Pathology, Western College of Veterinary Medicine, 52 Campus Dr., University ofSaskatchewan, Saskatoon, SK, Canada S7N 5B44 United States Geological Survey, National Wildlife Health Center, 6006 Schroeder Rd., Madison, Wisconsin 53711–6223, USA5 Corresponding author (email: aleksija.neimanis@usask.ca)

ABSTRACT: From 2000 to 2004, over 10,000 seabirds, primarily Herring Gulls (Larus argentatus),died from an undetermined cause in the Blekinge archipelago in southeastern Sweden. In June2004, 24 affected Herring Gulls were examined clinically, killed humanely, and 23 were examinedby necropsy. Seven and 10 unaffected Herring Gulls collected from a local landfill site and fromIceland, respectively, served as controls. All affected birds showed similar neurologic signs, rangingfrom mild incoordination and weakness to severe flaccid paralysis of legs and wings, but generallywere alert and responsive. All affected gulls were in normal nutritional condition, but weredehydrated and had empty stomachs. No gross or microscopic lesions, and no bacterial or viralpathogens were identified. Type C botulinum toxin was detected in the sera of 11 of 16 (69%)affected gulls by mouse inoculation. Type C botulism was the proximate cause of disease in 2004.Sera from 31% of birds tested from outbreaks in 2000 to 2003 also had detectable type Cbotulinum toxin by mouse inoculation. No large-scale botulism outbreak has been documentedpreviously in this area. The source of toxin, initiating conditions, and thus, the ultimate cause of thisoutbreak are not known. This epidemic might signal environmental change in the Baltic Sea.

Key words: Avian, Baltic sea, bird, botulism, Herring Gull, Larus argentatus, paralysis,Sweden.

INTRODUCTION

In the summer of 2000, unusually largenumbers of seabirds were found sick ordead in the Blekinge archipelago in theBaltic Sea off of the southeast coast ofSweden (Morner et al., 2005) (Fig. 1).Mortality events recurred annually from2001 to 2004 to varying degrees, and alsooccurred in other locations in the BalticSea and inland waters. Over 10,000 sea-birds were affected during this period inBlekinge (Morner et al., 2005). Sick birdsdisplayed neurologic signs of weaknessand paralysis. Although a number ofdifferent species were involved in theoutbreaks, most of the affected birds wereHerring Gulls (Larus argentatus) (Morneret al., 2005).

Carcasses from 10 different species ofaffected birds were collected from 2000 to2003 and examined at the Department ofWildlife, Fish and Environment, National

Veterinary Institute, Uppsala, Sweden, butno definitive cause of the outbreak wasdetermined (Morner et al., 2005). In 2004,a detailed, qualitative investigation thattargeted live, clinically abnormal HerringGulls was conducted. The objective was todetermine the cause of morbidity andmortality in these gulls. Here we describethe clinical signs, pathologic findings, anddiagnostic test results from these affectedbirds, and compare findings with local andoff-site control samples. We also includea retrospective analysis of samples avail-able from previous outbreaks.

MATERIALS AND METHODS

Study site

Clinically abnormal Herring Gulls werecollected from the Blekinge archipelago,Sweden at locations ranging from 55u589N,14u269E to 56u089N, 15u159E (Fig. 1). Thearchipelago consists of numerous, small rockyislands that are rarely more than 800 m across

Journal of Wildlife Diseases, 43(3), 2007, pp. 327–336# Wildlife Disease Association 2007

327

in the largest dimension, and is located in theBaltic Sea where surface salinity averages 7%to 7.5% (Kullenberg, 1981). Most islands arecovered in low-lying grasses and bushes, withoccasional deciduous trees. Many of theseislands are protected from human activity fornesting seabirds. Herring Gulls and otherseabirds, including other gull species, ducks,geese, terns, and shorebirds breed in thearchipelago every summer and many of thesebirds overwinter here (Mullarney et al., 1999).

Gull collections

A summary of specimens, sample sizes,examinations, and tests is presented in Ta-ble 1. Twenty-four clinically abnormal Her-ring Gulls were collected from the Blekingearchipelago from 7 to 16 June 2004. Gullswere collected from islands (n522) or theadjacent mainland (n52), and often more thanone affected bird was collected from eachisland. Following observation (described be-low), birds were captured manually, placed inindividual boxes, and transported to themainland for euthanasia and necropsy. Birdswere held in these boxes for up to 10 hr fromthe time of capture and were killed bydecapitation. Blood was collected from thejugular vein following euthanasia.

Blood was collected from additional HerringGulls euthanized from the Blekinge archipel-ago on 22 June (n54), 5 July (n51), and 6 July(n58) 2004. Nine of these samples werecollected within the coordinates providedabove, and four were collected on islands east

of Ronneby to longitude E15u419. Blood froma single affected Great Black-backed Gull(Larus marinus) was collected in the samemanner on 6 July 2004 from an island withinthe study site.

For unaffected controls, seven HerringGulls were shot at the Ronneby landfill site(56u119N, 15u199E), approximately 7 km fromthe eastern edge of the study area; blood wascollected from five of these birds. Tenadditional normal Herring Gulls were trappedas part of unrelated study in Iceland (64u49N,21u589W) on 14 July 2004 using net traps. Livebirds were transported by air to Sweden andwere held in individual boxes for a maximumof 48 hr from the time of trapping toeuthanasia. Birds were killed by cervicaldislocation and blood was collected prior toeuthanasia. Control birds were examined bynecropsy and sampled in the same manner asbirds collected from Blekinge.

Based on plumage (Mullarney et al., 1999),birds collected at Blekinge during 2004 con-sisted of 20 adults and four were 2–3-yr-oldbirds. The control birds collected at RonnebyIsland consisted of five adults and two 2–3-yr-old birds; all 10 birds collected in Iceland wereadults.

Clinical observations

Before capture, affected gulls were ob-served in situ, recorded on videotape, andexamined clinically.

Clinical and anatomic pathology

Blood was collected into sterile tubes withand without EDTA anticoagulant (Becton-Dickinson Vacutainer Systems Europe, Ply-mouth, UK). Blood smears were made imme-diately with freshly drawn blood. Sampleswithout anticoagulant were centrifuged andserum was frozen at 220 C. Hematology andbiochemistry analyses were performed by theDivision of Diagnostic Imaging and ClinicalPathology, Department of BiomedicalSciences and Veterinary Public Health, Fac-ulty of Veterinary Medicine and AnimalScience, Swedish University of AgriculturalSciences, Uppsala (Table 1).

Complete necropsies were performed on 23of the 24 affected gulls and on all 17 controlbirds; tissues were collected in 10% neutralbuffered formalin (Table 1). Formalin-fixedtissues from 20 affected birds and all controlbirds were processed routinely for histologicexamination. Briefly, tissues were dehydratedin graded alcohol, cleared in xylene, andembedded in paraffin. Sections (5 mm) werestained using Mayer’s hematoxylin and eosin

FIGURE 1. A map of southern Sweden depictingthe Blekinge archipelago. Insert: Blekinge archipel-ago (Modified map projection from ArcMap version9.1 in ArcGIS 9# 1999–2005, EnvironmentalSystems Research Institute, Inc.)

328 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007

TA

BL

E1.

Asu

mm

ary

of

speci

men

sam

ple

ssi

zes

for

test

sca

rrie

dou

ton

gu

lls

inS

wed

en

,2000–04.

So

urc

eY

ear

Sam

ple

nB

eh

avio

ur

ob

serv

ed

Hem

a-to

logya

Bio

chem

-is

tryb

Necr

op

syH

isto

logyc

Bac

teri

alcu

ltu

re

AIV

and

AP

MV

-1se

rolo

gyd

Vir

us

iso

lati

on

Bo

tuli

sm:

mo

use

ino

cula

tio

nB

otu

lism

:E

LIS

A

Aff

ecte

d

Ble

kin

ge

2004

Bir

d24

24

17

16

23

20

20

56

13

e18

Ble

kin

ge

2004

Seru

m14

––

––

––

––

314

Ble

kin

ge

2000–2003

Seru

m41

––

––

––

––

16

31

Gav

le2002

Seru

m1

––

––

––

––

10

Con

trol

Ron

neb

y2004

Bir

d7

75

47

75

––

–5

Icela

nd

2004

Bir

d10

10

4–

10

10

––

––

3

aIn

clu

ded

hem

oglo

bin

con

cen

trat

ion

,h

em

ato

crit

,an

dto

tal

and

dif

fere

nti

alle

uko

cyte

cou

nts

(au

tom

ated

hem

ato

logy

anal

yzer

Cell

-Dyn

,A

bb

ott

Lab

ora

tori

es,

Ch

icag

o,

Illi

no

is,

US

A,

inth

ere

sist

ant-

RB

C-m

od

eu

sin

gV

ET

2.3

soft

war

e).

bIn

clu

ded

ure

a,cr

eat

inin

e,

calc

ium

,m

agn

esi

um

,p

ho

sph

oru

s,so

diu

m,

po

tass

ium

,ch

lori

de,

cop

per,

zin

c,ir

on

,p

rote

in,

and

tota

lb

ilir

ub

inco

nce

ntr

atio

ns,

and

creat

inin

ekin

ase,

asp

arta

team

ino

tran

sfe

rase

,al

anin

eam

ino

tran

sfera

se,

alkal

ine

ph

osp

hat

ase,

gam

ma

glu

tam

yltr

ansf

era

se,

and

glu

tam

ate

deh

ydro

gen

ase

acti

viti

es

(Ko

nela

b3

0au

toan

alyz

er,

Th

erm

oE

lect

ron

Oy,

Van

taa,

Fin

lan

d,

usi

ng

the

man

ufa

ctu

rer’

sre

agen

ts).

cB

rain

,sp

inal

cord

,p

eri

ph

era

ln

erv

es,

skele

tal

mu

scle

(pect

ora

l,w

ing,

and

leg

mu

scle

s),

lun

g,

hear

t,kid

ney,

sple

en

,li

ver,

adre

nal

gla

nd

,go

nad

s,eso

ph

agu

s,p

rove

ntr

icu

lus,

ven

tric

ulu

s,d

uo

den

um

,je

jun

um

,il

eu

m,

cecu

m,

colo

n,

clo

aca,

pan

creas

,th

yro

idgla

nd

,tr

ach

ea,

bo

ne

mar

row

,an

dth

eb

urs

ao

fF

abri

ciu

san

dth

ymu

s,w

hen

avai

lab

le.

dA

IV5

Avi

anin

flu

en

zavi

rus;

AP

MV

-15

Avi

anP

aram

yxo

viru

s-1

,e

Eig

ht

sam

ple

sw

ere

test

ed

inS

wed

en

,fi

vew

ere

test

ed

inC

anad

a,an

dtw

osa

mp

les

were

test

ed

inb

oth

lab

ora

tori

es.

All

oth

er

mo

use

ino

cula

tio

nte

sts

were

perf

orm

ed

inC

anad

a.

NEIMANIS ET AL.—BOTULISM TYPE C IN HERRING GULLS IN SOUTHEASTERN SWEDEN 329

(Luna, 1968). Additional sections were stainedwith Perl’s method for iron and Bennhold’smethod for amyloid (Congo Red) to investi-gate observed histologic abnormalities (Luna,1968). Investigation of parasitism was confinedto gross and microscopic examination oftissues.

Bacteriology and virology

Fresh liver samples from 20 affected andfive unaffected local control Herring Gullswere inoculated onto blood agar plates con-taining 5% horse blood and onto eosinmethylene blue agar plates at the WildlifeDivision, National Veterinary Institute, Up-psala, Sweden. Plates were incubated at 37 Cin aerobic conditions and inspected after 24 hrand 48 hr for growth. Plates were held atambient temperature (approximately 20 C) forfive additional days and inspected for growthat the end of seven days.

Sera from five severely affected birds weretested for antibodies to avian influenza virusand avian paramyxovirus-1 (APMV-1) at theVirology Laboratory, National Veterinary In-stitute, Uppsala, Sweden, using agar gelimmunodiffusion tests and blocking ELISAtests, respectively. For avian influenza, anH9N2 reagent supplied by an OIE referenceLaboratory for Avian Influenza (InstitutoZooprofilattico Sperimentale Delle Venezie,Venice, Italy) was used as antigen andmethods followed those of the OIE (2004).The blocking ELISA for APMV-1 was per-formed using commercially available test kits(SvanovirTM, Svanova Biotech AB, Uppsala,Sweden). Virus isolation was attempted frompooled tissues (spleen, liver, kidney, lung, andbrain) from six moderately to severely affectedgulls. Tissue homogenates were inoculatedinto embryonated chicken eggs and passagedthree times at the Virology Laboratory.

Mouse inoculation

Sera from six affected birds and pooled serafrom two additional affected birds were testedfor botulinum toxins C and E at the Bacteri-ology Laboratory, National Veterinary Insti-tute, Uppsala, Sweden in July 2004. Wherepossible, testing was performed on individualsthat also were tested for viruses and bacteria.Young adult Swiss white mice weighingapproximately 20 g were used. For each test,two mice were injected intraperitoneally (IP)with 0.5 ml of gull serum and observed for upto four days for signs of botulism (Wobeser,1997a). Mice were immediately euthanized ifsigns of botulism developed. Sera that causedclinical signs in the mice, and for which there

was adequate volume, were diluted 1:5 and1:10, and 0.5 ml of each concentration wasinjected IP into two additional mice. Formouse protection assays, two mice wereinjected IP with antitoxin to type C botulinumtoxin (developed by the Bacteriology Labora-tory, National Veterinary Institute, Uppsala,Sweden) and two mice were injected IP withantitoxin to type E botulinum toxin (BAB-TUSE, National Institute for Biological Stan-dards and Controls, Hertfordshire, UK).Thirty minutes later, each mouse was injectedIP with 0.5 ml of undiluted gull serum andobserved for up to four days.

Sera from all remaining affected gulls thatwere examined from 2004, for which adequateserum volumes were available (n55), weretested for type C botulinum toxin at theWestern College of Veterinary Medicine(WCVM), University of Saskatchewan, Saska-toon, Saskatchewan, Canada. Sera from threeadditional affected birds collected from 22June 2004 to 6 July 2004 (Table 1) and fromtwo affected gulls which tested negative fortoxin during tests in Sweden also were testedat WCVM. For each test, two young Swisswhite mice (approximately 30 g) were used.One was injected IP with 0.1 ml antitoxin totype C botulinum toxin (developed by theNational Wildlife Health Center, Madison,Wisconsin, USA). Thirty minutes later, bothmice were injected IP with 1 ml of undilutedgull serum. Observations and test interpreta-tion followed methods described above. InSweden and Canada, mice that were founddead or that were euthanized after showingsigns of botulism were examined by necropsyto rule out peritonitis or septicemia.

To determine if botulism contributed to theseabird mortality observed in previous years,16 serum samples collected in previous out-breaks in Blekinge from 2000 to 2003 and oneserum sample from an outbreak reported inGavle, northern Sweden (60u369N, 19u579E)in 2002 were inoculated into mice at WCVMas described above. These sera were obtainedfrom 15 Herring Gulls and two Great Black-backed Gulls (Larus marinus) with clinicalsigns identical to those of affected gulls in2004.

Enzyme-linked immunosorbent assay (ELISA)

Seventy-one serum samples were tested fortype C toxin by ELISA at the National WildlifeHealth Center, Madison, Wisconsin. Theseincluded 31 affected Herring Gulls and oneGreat Black-backed Gull from 2004, fiveunaffected gulls from Blekinge and threeunaffected Herring Gulls from Iceland from

330 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007

2004, and 27 affected Herring Gulls and fouraffected Great Black-backed Gulls from theBlekinge archipelago collected from 2000 to2003 (Table 1). A capture ELISA method wasfollowed as outlined in Rocke et al. (1998).

Data analysis

Hematology results for affected gulls, un-affected gulls from Sweden, and unaffectedgulls from Iceland were compared with theKruskal-Wallis test (Zar, 1984). Post-hoccomparison of means using Wilcoxon RankSum tests was performed for parameters thatwere significantly different (Zolman, 1993).Biochemical parameters for affected versusunaffected gulls also were compared using theWilcoxon Rank Sum test. The proportion ofgulls infected with each parasite species wascompared between affected and unaffectedgulls from Sweden using Fisher’s exact test.Analyses were performed at an alpha level of0.05 using Statistix 7 for Windows# 2000(Analytical Software, Tallahassee, Florida,USA).

RESULTS

Clinical signs

Affected birds displayed a range ofprogressively developing neurologic signs,with limb paresis to complete paralysisbeing the most consistent finding. Initially,gulls showed delayed or uncoordinatedflight compared to normal conspecificswhen approached. Two such individualswere observed in the field, but escaped

capture. Mildly affected individuals wereincapable of flight, but could stand andrun and were mildly ataxic. Moderatelyaffected birds were found in sternalrecumbency with the wings dropped atthe shoulder and fallen away from thebody (Fig. 2). Intermittently, these birdsstood and walked a few steps. Often, birdscould not stand up completely and rested,and even walked, on their tarsometatarsi(Fig. 3). More severely affected gullscould not walk, and dragged themselvesforward using their wings and beak.Moribund birds displayed open-mouthed,slow, labored breathing, occasional bob-bing or ventroflexion of the head, andrarely, intermittent nystagmus. Their eyesoften were partially closed, and rarely, thenictitating membranes were prolapsed orrapid constriction and dilation of thepupils was observed. With the exceptionof moribund individuals which were dulland poorly responsive, all birds were alert,responsive, aggressive, and had mobileheads. Wing and leg tone and beakstrength decreased with increasing sever-ity of clinical signs. Frequently, there wasstaining of the vent feathers with uratesand bright green feces. None of theunaffected birds from the Blekinge landfillsite and Iceland showed any of theseclinical signs.

Clinical and anatomic pathology

Affected birds had significantly higherhemoglobin (158612.2 vs. 126.5611.5g/l),

FIGURE 2. A Herring Gull (Larus argentatus)from Blekinge, Sweden with clinical signs ofbotulism. The wings are dropped at the shoulderand have fallen away from the body. The bird is alert.

FIGURE 3. A Herring Gull (Larus argentatus)from Blekinge, Sweden with clinical signs ofbotulism. The bird was unable to stand fully uprightand stood and walked on its tarsometarsi.

NEIMANIS ET AL.—BOTULISM TYPE C IN HERRING GULLS IN SOUTHEASTERN SWEDEN 331

hematocrit (0.6260.04 vs. 0.4360.01),urea (6.4364.95 vs. 1.5560.34 mmol/l),sodium (160.0667.32 vs. 149.5064.12mmol/l), and total bilirubin values(15.57634.63 vs. 3.4361.20 mmol/l) thanlocal control birds (mean 6 standarddeviation; P,0.01 for all comparisons). Theaffected gulls and the control birds fromIceland had significantly higher total leuko-cyte counts (9.7965.81 and 7.0862.573109/l,respectively) than did the local controlbirds (2.1361.963109/l; P50.04). Meanvalues of all other hematologic and bio-chemical parameters for affected gulls werenot significantly different from controls.

At post-mortem examination, affectedgulls had dry, tacky subcutaneous tissues.The proventriculus and ventriculus werebile-stained and empty, the gall bladderwas full, and the cloaca was distended withurates and feces. All affected and controlbirds were in fair to good nutritionalcondition based on the presence ofmoderate to abundant fat stores. Localcontrol birds had full gastrointestinaltracts, but tracts in control birds fromIceland were empty. Occasional, inciden-tal lesions were found in individual birdsfrom all three groups (e.g., scatteredgranulomas, mild air sacculitis, and shotpellets).

Histologically, amyloid (Congo Redstain) was seen in vessel walls of thespleen and occasional other organs in 29%

to 35% of birds from all three groups.Affected gulls often had abundant hemo-siderin (Perl’s stain) in Kupffer cells, andtwo affected birds had moderate, multifo-cal distension of renal tubules with uratesand the tubular epithelium was attenuat-ed. Five out of 10 (50%) control gulls fromIceland had mild to severe, subacuteskeletal muscle necrosis in wing and legmuscles. Mild, acute skeletal myocytedegeneration in limb and pectoral muscleswas seen in 3/20 (15%) of affected gullsand mild to moderate, subacute musclenecrosis in limbs was seen in 2/20 (10%) ofaffected gulls. With the exception of mildto moderate pathologic changes associated

with parasitism, no other significant le-sions were seen in any of the tissuesexamined.

Numerous nematodes, trematodes, ces-todes, and coccidia were seen in all groupsof birds during microscopic examination,but for some parasites, the proportion ofinfected gulls appeared to differ betweengroups. Affected gulls more frequentlyhad detectable schistosome infection bylight microscopy (17/20 or 85%) than localunaffected gulls (3/7 or 43%; P50.049).No control birds from Iceland had detect-able infection with schistosomes. Althoughthere was a tendency for a higher pro-portion of unaffected local gulls to haveesophageal Capillaria sp. in tissue sectionsthan affected gulls (5/7 or 71% vs. 6/20 or30%), this difference was not statisticallysignificant (P50.084).

Bacteriology and virology

No pathogenic bacteria were cultured.No antibodies to avian influenza virus orAPMV-1 were detected and no viruseswere isolated from pooled tissue samples.

Botulism testing

Type C botulinum toxin was detectedby mouse inoculation in 11 of 16 (69%)and 5 of 16 (31%) serum samples collect-ed from clinically affected gulls in theBlekinge archipelago in 2004 and 2000–03, respectively. This included one of theGreat Black-backed Gulls collected in2001. The single sample tested from Gavlehad no detectable toxin. Of the threepositive serum samples for which a dilutionseries was performed, only the pooledsample produced signs typical for botulismin mice at a dilution of 1:5, none of theserum sample induced detectable signs inmice at a 1:10 dilution. One serum sampletested in both Sweden and Canada causedclinical signs in a mouse injected with1.0 ml, but not in a mouse injected with0.5 ml. None of the eight samples from2004 tested in Sweden contained type Ebotulinum toxin. No evidence of septice-mia or peritonitis was seen in mice that

332 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007

died or were euthanized during botulismtesting.

Nonspecific reactions were observed in10 affected gulls and all five of the controlsthat were collected at Ronneby; theseresults probably resulted from bacterialcontamination of serum and were consid-ered inconclusive. Type C toxin wasdetected by ELISA in nine of 23 (39%)and six of 30 (20%) samples from affectedbirds collected in 2004 and from 2000–03,respectively. Sera from control birds fromIceland and from affected Great Black-backed Gulls tested negative for Type Ctoxin by ELISA.

DISCUSSION

The clinical signs, lack of consistent andsignificant morphologic lesions, negativebacteriologic and virologic test results, andpositive results for botulism type C allsupport a diagnosis of avian botulism typeC in Herring Gulls examined in 2004.Botulism occurs following the ingestion ofvarious types of preformed botulinumtoxin produced by the bacterium Clostrid-ium botulinum (Quinn et al., 1999). Theseneurotoxins cause flaccid paralysis byinterfering with acetylcholine release atthe neuromuscular junction (Simpson,1986). Birds are susceptible to intoxicationby botulinum toxin type C and type E(Rosen, 1971). Botulism type C most oftenaffects birds of the Order Anseriformes,but it also has been documented to causemorbidity and mortality in gulls, shore-birds, herons, cormorants, and possibly infish (Jensen and Price, 1987; Rocke et al.,2004).

In this outbreak, other possible etiolo-gies such as avian cholera, viral infections(avian influenza virus, avian paramyxovi-rus-1, West Nile virus and other viralencephalitides), and avian vacuolar myeli-nopathy were eliminated based on patho-logic, bacterial, and viral assessments.West Nile virus, Usutu virus, or otherrelated virus were not detected by poly-merase chain reaction in samples from

birds collected from 2000 to 2003 (Morneret al., 2005). Algal blooms were notreported during the 2004 study period,and tests of water in previous outbreaksdid not detect toxic algae (Morner et al.,2005). Trace mineral imbalances or heavymetal intoxication were not detected incarcasses collected from outbreaks in 2000to 2003 (Morner et al., 2005).

Clinical signs in affected Herring Gullsin this outbreak were consistent with thosedescribed in botulism outbreaks in Her-ring Gulls in Scotland (MacDonald andStandring, 1978) and Ireland (Quinn andCrinion, 1984). General flaccid paralysis isdescribed for all affected avian species(Rosen, 1971). In gulls (this study) andother species (Borland et al., 1977), wingmuscles appear to be affected first.Drooping wings were observed in allaffected gulls examined in this investiga-tion. In addition, affected gulls in thisstudy frequently were observed standingand walking on their tarsometatarsi. De-spite progressive flaccid paralysis, gullsremained alert and aggressive until theywere moribund; this is also typical of typeC botulism in gulls (MacDonald andStandring, 1978; Quinn and Crinion,1984).

Pathologic findings were minimal andreflected debilitation, stress, and handling.Clinically affected gulls were dehydrated,as was evidenced by necropsy findings andblood parameters (Thrall et al., 2004).Paresis and paralysis likely precludedthese animals from eating and drinking.The leukogram and vascular amyloidosisare nonspecific indicators of stress (Hoff-man and Leighton, 1985). The higher totalleukocyte counts in affected gulls andunaffected birds from Iceland are consis-tent with stress from intoxication and/orcapture and transport. The subacutemuscle necrosis of limbs in gulls formIceland suggests capture myopathy duringtrapping and transport. Similarly, theacute muscle damage in limbs and pecto-ral muscles of a few affected gulls could berelated to repeated attempts to ambulate,

NEIMANIS ET AL.—BOTULISM TYPE C IN HERRING GULLS IN SOUTHEASTERN SWEDEN 333

prolonged sternal recumbency, capture,and transport.

In this study, type C botulinum toxinwas detected by mouse inoculation in 69%

of affected gull sera collected in 2004. Thislevel of sensitivity of toxin detection istypical of other outbreaks of avian botu-lism. For example, Rocke et al. (1998)reported that mouse inoculation detectedtoxin in 79% of clinical avian botulismcases, and in a botulism outbreak inIreland, toxin was detected in 64% ofaffected gulls (Quinn and Crinion, 1984).

Mouse inoculation currently is the mostsensitive test available for the diagnosis ofbotulism, and is highly specific, but it isnot sensitive to low levels of toxin(Thomas, 1991) that often are present inclinically affected birds (Wobeser, 1997a).Birds affected by botulism generally haveless than 20 mouse lethal doses per ml ofserum, and in many cases, they have none(Smith, 1987). Botulinum toxin bindsirreversibly to receptors on the nerve atthe neuromuscular junction (Rossetto andMontecucco, 2003). False negative resultsin mouse inoculations can occur if themajority of ingested toxin is bound at theneuromuscular junction, rather than cir-culating in blood (Swerczek, 1980).

As has been described in other cases(Smith, 1987), the concentration of toxinin the serum of birds sampled in this studywas low. Only one sample for whicha dilution series was performed had a highenough toxin concentration to induceclinical signs in mice at a dilution of 1:5and no samples induced signs in mice ata 1:10 dilution. One serum sample causedclinical signs in a mouse injected with1.0 ml of serum, but not in a mouseinjected with 0.5 ml.

It is most likely that epidemics in 2000to 2003 also were caused by type Cbotulism, because Type C botulinum toxinwas detected in sera from some of theseaffected gulls and clinical signs in thesebirds were identical to those observed in2004 (Morner et al., 2005). The possiblelower prevalence of detectable toxin in

samples from 2000 to 2003 might reflectloss of toxin activity during storage.Hubalek and Halouzka (1988) reportedthat only 6% of the toxicity of type Cbotulinum toxin persisted after 5 yr instorage at 220 C.

In this study, mouse inoculation wasused sparingly out of ethical considera-tions, and testing efforts were augmentedthrough ELISA testing. Type C toxin wasdetected by ELISA, but in a lower pro-portion of birds. Samples tested by ELISAexperienced additional cycles of freezingand thawing, bacterial overgrowth, andprolonged storage. These factors areknown to decrease sample quality (Huba-lek and Halouzka, 1988; Mei et al., 2001),and in some cases, samples used forELISA were of inadequate volume (Rockeet al., 1998). Mouse inoculations per-formed at the NWHC to confirm ELISAresults detected toxin in an additionalthree birds from 2004 and two birds from2000–03 (data not shown).

The birds examined in this investigationsuffered from type C botulism. Theecology and ultimate cause of this out-break, however, are not known. An out-break of this magnitude over such a largegeographic scale has not been documen-ted in the Baltic Sea region and does notfit the most-studied scenario of type Cbotulism in which waterfowl are theprimary species affected in outbreaksperpetuated by a carcass-maggot cycle.However, this outbreak does bear strikingsimilarities to outbreaks described inBritain and Ireland in the 1970s and1980s (Lloyd et al., 1976; Quinn andCrinion, 1984). In these outbreaks, nu-merous avian species along a large stretchof coastline were affected by botulism typeC. In various locations in Scotland, Wales,England (Lloyd et al., 1976), and inIreland (Quinn and Crinion, 1984), thevast majority of affected birds also wereHerring Gulls. The source of the toxin inthese outbreaks never was elucidated.

In the Swedish outbreak, gulls mighthave become intoxicated simply by scav-

334 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007

enging toxin-laden carcasses. Increasednumbers of vertebrate carcasses, for anyreason, can increase the probability ofa botulism outbreak, provided that car-casses contain C. botulinum spores at thetime of death (Wobeser, 1997b). Alter-nately, invertebrate or vertebrate vectorscould have facilitated toxin transfer to thegulls. Further investigation into the diet ofHerring Gulls in the Blekinge archipelagois critical to the determination of the toxinsource in this outbreak.

It is unclear what factors precipitated thelarge-scale epidemic in Blekinge in 2000and in each subsequent year to 2004. TheBaltic ecosystem has deteriorated over thepast half-century (Jansson and Dahlberg,1999). Large areas of anoxia dominate thesea bottom, species of fish populations havechanged (Jansson and Dahlberg, 1999),and mean temperature and precipitation insouthern Sweden recently have increased(Lindstrom and Alexandersson, 2004). Ifconditions in the Baltic Sea have favoredincreased proliferation of C. botulinum,seabirds might be coming into increasingcontact with spores and/or toxin throughthe marine food web. Habitat partitioningcould explain why gulls collected at a nearbylandfill site in Blekinge were unaffected.Differences in environmentally acquiredparasites such as schistosomes (Horak etal., 2002) provide further support for thedifferential use of environments by affectedand unaffected birds.

The Baltic ecosystem historically hashad low species diversity and is thusparticularly sensitive to environmentaldisturbances (Jansson and Dahlberg,1999). One of the consistent signs ofstressed ecosystems is changing patternsof disease (Rapport and Whitford, 1999).Epidemic botulism in Herring Gulls andother seabirds could be a signal ofa broader environmental perturbation inthe Baltic. Investigation into the source oftoxin and initiating conditions of theseoutbreaks is needed to explore the poten-tial roles of environmental change and toidentify possible intervention strategies.

ACKNOWLEDGMENTS

We are indebted to Le Carlsson, Ronneby,Sweden for his dedication, knowledge, andfield support during the investigation of thisoutbreak. R. Mattsson and others in theDepartment of Wildlife, Fish and Environ-ment, National Veterinary Institute, Uppsala,Sweden, provided technical and field support.S. Smith, National Wildlife Health Center,Madison, WI, USA tested sera for botulismtype C using ELISA. C. Waldner, WCVM,Saskatoon, SK, Canada provided statisticaladvice. This manuscript was improved bycomments from M. Cattet and A. Allen,WCVM, Saskatoon, SK, Canada. All work inSweden was carried out under animal care andwelfare permits held by T. Morner and theDepartment of Wildlife, Fish and Environ-ment, National Veterinary Institute, Uppsala,Sweden. Botulism testing in Canada wasperformed under an animal care permit issuedto T. Bollinger. Financial support was pro-vided by the Department of Wildlife, Fish andEnvironment, National Veterinary Institute,Uppsala, Sweden, the Western College ofVeterinary Medicine Interprovincial GraduateStudent Fund and the Canadian Co-operativeWildlife Health Centre.

LITERATURE CITED

BORLAND, E. D., C. J. MORYSON, AND G. R. SMITH.1977. Avian botulism and the high prevalence ofClostridium botulinum in the Norfolk Broads.Veterinary Record 100: 106–109.

HOFFMAN, A. M., AND F. A. LEIGHTON. 1985.Hemograms and microscopic lesions of HerringGulls during captivity. Journal of the AmericanVeterinary Medical Association 187: 1125–1128.

HORAK, P., L. KOLAROVA, AND C. M. ADEMA. 2002.Biology of the schistosome genus Trichobilhar-zia. Advances in Parasitology. 52: 155–234.

HUBALEK, Z., AND J. HALOUZKA. 1988. Thermalsensitivity of Clostridium botulinum type Ctoxin. Epidemiology and Infection 101: 321–325.

JANSSON, B. O., AND K. DAHLBERG. 1999. Theenvironmental status of the Baltic Sea in the1940s, today, and in the future. Ambio 28: 312–317.

JENSEN, W. I., AND J. I. PRICE. 1987. The globalimportance of type C botulism in wild birds. InAvian botulism. An international perspective, M.W. Eklund and V. R. Dowell, Jr. (eds.).Charles C Thomas Publisher, Springfield, Illi-nois, pp. 33–54.

KULLENBER, G. G. 1981. Physical oceanography. InThe Baltic Sea, A. Voipio (ed.). ElsevierScientific Publishing Company, New York,New York, pp. 135–181.

LINDSTROM, G., AND H. ALEXANDERSSON. 2004. Recent

NEIMANIS ET AL.—BOTULISM TYPE C IN HERRING GULLS IN SOUTHEASTERN SWEDEN 335

mild and wet years in relation to long observa-tion records and future climate change inSweden. Ambio 33: 183–186.

LLOYD, C. S., G. J. THOMAS, J. W. MACDONALD, E. D.BORLAND, K. T. STANDRING, AND J. L. SMART.1976. Wild bird mortality caused by botulism inBritain, 1975. Biological Conservation 10: 119–129.

LUNA, L. G. (ED.). 1968. Manual of histologic stainingmethods of the Armed Forces Institute ofPathology, 3rd edition. The Blakiston Division,McGraw-Hill Inc., Toronto, Ontario, Canada,pp. 258.

MACDONALD, J. W., AND K. T. STANDRING. 1978. Anoutbreak of botulism in gulls on the Firth ofForth, Scotland. Biological Conservation 14:149–155.

MEI, J. V., J. R. ALEXANDER, B. W. ADAM, AND W. H.HANNON. 2001. Use of filter paper for thecollection and analysis of human whole bloodspecimens. Journal of Nutrition 131: 1631S–1636S.

MORNER, T., H. UHLHORN, E. AGREN, K. BERNODT, R.MATTSSON, C. BROJER, K. NILSSON, D. GAVIER-WIDEN, D. S. JANSSON, D. A. CHRISTENSSON, L.PETERSSON, AND A. NEIMANIS. 2005. Onormalthog dodlighet bland trutar och annan sjofagel iSverige (Mass mortality in waterfowl in Sweden).Svensk Veterinartidning (Swedish VeterinaryJournal) 8–9: 11–18.

MULLARNEY, K., L. SVENSSON, D. ZETTERSTROM, AND P.J. GRANT. 1999. Fagelguiden. Europas ochMedelhavsomradets faglar i falt. Albert BonniersForlag, Stockholm, Sweden, pp. 400.

OIE, 2004. Manual of Diagnostic Tests and Va-ccines for Terrestrial Animals 2004. www.oie.int/eng/normes/mmanual/A_00037.htm. Accessed De-cember 2005.

QUINN, P. J., AND R. A. P. CRINION. 1984. A two yearstudy of botulism in gulls in the vicinity ofDublin Bay. Irish Veterinary Journal 38: 214–219.

———, M. E. CARTER, B. K. MARKEY, AND G. R.CARTER. 1999. Clinical Veterinary Microbiology.Mosby, Toronto, Ontario, Canada, pp. 648.

RAPPORT, D. J., AND W. G. WHITFORD. 1999. Howecosystems respond to stress. Bioscience 49:193–203.

ROCKE, T. E., S. R. SMITH, AND S. W. NASHOLD. 1998.Preliminary evaluation of a simple in vitro test

for the diagnosis of type C botulism in wildbirds. Journal of Wildlife Diseases 34: 744–751.

———, P. NOL, C. PELIZZA, AND K. K. STURM. 2004.Type C botulism in pelicans and other fish-eating birds at the Salton Sea. Studies in AvianBiology 27: 136–140.

ROSEN, M. N. 1971. Botulism. In Infectious andparasitic diseases of wild birds, J. W. Davis, R. C.Anderson, L. Karstad, and D. O. Trainer (eds.).Iowa State University Press, Ames, Iowa,pp. 100–117.

ROSSETTO, O., AND C. MONTECUCCO. 2003. Howbotulinum toxins work. In Hand book ofbotulinum toxin treatment, 2nd edition, P.Moore and M. Naumann (eds.). BlackwellScience Ltd., Malden, Massachusetts, pp. 9–27.

SIMPSON, L. L. 1986. Molecular pharmacology ofbotulinum toxin and tetanus toxin. Annual Re-view Pharmacology and Toxicology 26: 427–453.

SMITH, G. R. 1987. Botulism in water birds and itsrelation to comparative medicine. In Avianbotulism. An international perspective, M. W.Eklund and V. R. Dowell Jr. (eds.). Charles C.Thomas Publisher, Springfield, Illinois, pp. 73–86.

SWERCZEK, T. W. 1980. Toxicoinfectious botulism infoals and adult horses. Journal of the AmericanVeterinary Medical Association 176: 217–220.

THOMAS, R. J. 1991. Detection of Clostridiumbotulinum types C and D toxin by ELISA.Australian Veterinary Journal 68: 111–113.

THRALL, M. A., D. C. BAKER, T. W. CAMPBELL, D.DENICOLA, M. J. FETTMAN, E. D. LASSEN, A.REBAR, AND G. WEISER. 2004. Clinical chemistryof birds. In Veterinary hematology and clinicalchemistry. Lippincott Williams & Wilkins, Balti-more, Maryland, pp. 479–492.

WOBESER, G. 1997a. Botulism. In Diseases of wildwaterfowl. Plenum Press, New York, New York,pp. 149–161.

———. 1997b. Avian botulism—Another perspec-tive. Journal of Wildlife Diseases 33: 181–186.

ZAR, J. H. 1984. Biostatistical analysis, 2nd edition.Prentice-Hall, Englewood Cliffs, New Jersey,pp. 718.

ZOLMAN, J. F. 1993. Biostatistics. Experimentaldesign and statistical inference. Oxford Univer-sity Press, New York, New York, 368 pp.

Received for publication 6 April 2006.

336 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007