The Epidemiology of Vibrio Infections in Florida, 1981-1993
W. Gary Hlady and Karl C. Klontz
Epidemiology Program, Florida Department ofHealth and
Rehabilitative Services, Tallahassee; Centerfor Food Safety and
Applied Nutrition, US Food and Drug Administration, Washington, DC
The epidemiology of 690 Vibrio infections reported in Florida during 1981-1993 is described.
Most infections resulted in one of three clinical syndromes: gastroenteritis (51%), wound infections
(24%), or primary septicemia (17%). Case-fatality rates were 1% for gastroenteritis, 5% for wound
infections, and 44% for primary septicemia. While gastroenteritis had little seasonal variation, 91%
of primary septicemias and 86% of wound infections occurred from April through October, mostly
due to the seasonality of Vibrio vulnificus and Vibrio parahaemolyticus infections. Infected wounds
were largely a result of occupational activities around seawater. Some 68% of gastroenteritis cases
and 83% of the primary septicemias were associated with raw oyster consumption. Preexisting liver
disease was present in 48% of patients with primary septicemia and was associated with a fatal
outcome in both wound infections (relative risk [RR], 28.3; 95% confidence interval [CI], 6.3-127.5;
P < .0001) and primary septicemia (RR, 1.9; 95% CI, 1.2-3.1; P < .01).
Vibrio bacteria are natural inhabitants ofmarine and estuarine
environments and can cause human infections that most com-
monly present clinically as gastroenteritis, wound infections, and
septicemia [I]. Infection is generally acquired through consump-
tion of contaminated food or water or by direct invasion through
wounds. Vibrio cholerae is the best known member ofthe genus
and includes several strains classified according to 0 group. Toxi-
genic strains of0 group 1 (01) cause cholera and are rare in the
United States ; all other strains are grouped as V. cholerae
non-O1.Most illnesses in the United States due to Vibrio infection
involve V. cholerae non-O1 strains, Vibrioparahaemolyticus, and
Vibrio vulnificus .
In the United States, Vibrio infections are most common
in states bordering the Gulf of Mexico . However, Vibrio
infections may occur in persons who live some distance from
the Gulf Coast [5, 6] after they consume raw shellfish while
visiting Gulf states or that was transported interstate. Visitors
may also sustain wounds in Gulfwaters that result in infections
after they return home. Few studies have described the clinical
and epidemiologic features of Vibrio infections in a US area
over a long period [4, 7, 8]. Here we summarize more than
650 Vibrio infections reported in Florida over 13 years and
illustrate the spectrum of Vibrio species involved in human
infection, the most common risk factors, and the significant
morbidity associated with these infections.
In Florida, culture-confirmedcases of Vibrio infection have been
reportable to the Department of Health and Rehabilitative Services
Received 12 September 1995; revised 19 December 1995.
Financial support: Florida Department of Health and Rehabilitative Services;
US Food and Drug Administration.
Reprints or correspondence: Dr. W. Gary Hlady, HSDE, 1317 Winewood
Blvd., Tallahassee, FL 32399-0700.
The Journal of Infectious Diseases
© 1996 by The University of Chicago. All rights reserved.
since 1981. County public health investigators obtain clinical and
epidemiologic information using standardized Vibrio case report
forms,We reviewedallcasereportforms submittedtothe Epidemiol-
ogy Program, Departmentof Health and RehabilitativeServices,for
Vibrio infections with onset of illness during 1981-1993.
Vibrio organisms were isolated from blood, wounds, and other
infection sites by use of routine culture media. Isolation of Vibrio
organisms from stool usually required a selective medium such as
thiosulfate-citrate-bile salts-sucrose, with or without preincuba-
tion in alkaline peptone water.
We categorized Vibrio infections into syndromes on the basis
of the clinical features of illness and the specimen from which the
organism was recovered. Gastroenteritis was defined as an illness
with diarrhea, vomiting or abdominal cramps, no evidence of
wound infection, and the isolation of only Vibrio organisms from
stool. Wound infections required that the patient had been wounded
during the week before onset of illness, and vibrio had to be
isolated from the wound site or blood. The definition of primary
septicemia was an illness with fever or hypotension (systolic blood
pressure <90 mm Hg) without an apparent primary focus ofinfec-
tion and vibrio recovery from blood.
Average annual incidence rates were calculated using population
data from the Florida Office of Vital Statistics  and estimates
of seafood consumption and liver disease prevalence from the 1988
Florida Behavioral Risk Factor Survey . AIDS prevalence was
obtained from routine surveillance information and used to calcu-
late an incidence rate among patients with AIDS. Data were ana-
lyzed using Epi Info 6.0 software . Ninety-five percent confi-
dence intervals (CI) for single proportions were calculated
according to the method of Fleiss .
A total of690 Vibrio infections in 675 persons were reported
during 1981-1993 (4.3patients/l,000,000 population/year); 14
patients were simultaneously infected with multiple Vibrio spe-
cies. Simultaneous infection with V parahaemolyticus and Vib-
by guest on July 13, 2011
JID 1996; 173 (May)
Vibrio Infections in Florida
Table 1. Clinical syndromes associated with Florida Vibrio infections, 1981-1993.
V. cholerae non-Ol
V. cholerae 01
* Urinary tract infection.
t 2 pulmonary infections of drowning victims, 1 ear infection, 1 urinary tract infection.
t 2 pulmonary infections of drowning victims, 2 ear infections, 2 urinary tract infections.
§ Gall bladder infections.
IITotal exceeds cases (675): 13 patients were simultaneously infected with 2 Vibrio species and 1 patient with 3
Data are no. (%).
rio fluvialis was found in 3 patients with gastroenteritis and in
.1 with wound infection. Both V.parahaemolyticus and Vibrio
damsela were found in 2 patients with wound infections. One
patient with gastroenteritis and 1 with wound infection were
infected with V. parahaemolyticus and V. vulnificus. The 4
patients with wound infections were infected with one of the
following combinations: Vibrio alginolyticus-V. damsela, 'v.
alginolyticus-V.parahaemolyticus, V. cholerae non-Ol-Vib-
rio mimicus, and V. parahaemolyticus- V. cholerae non-OI-
V. vulnificus. One patient with gastroenteritis was infected with
V.parahaemolyticus and V.mimicus, and another was infected
with V. parahaemolyticus and V. cholerae non-O1.
The largest proportion (51%) ofpatients presented with gas-
troenteritis, followed by wound infections (24%) and primary
septicemia (17%) (table 1). The mean age of patients with
Vibrio septicemia (60 years; range, 18- 91) was greater than
for patients with gastroenteritis or wound infection (40 years
[range, 3 weeks-91 years] and 43 years [range, 4-92 years],
respectively; P < 10-6) . Ofnote, 5 patients with Vibrio gastro-
enteritis were < 1 year old (3 were <2 months old).
Gastroenteritis.Vibrio species that most often presented
clinically as gastroenteritis (n = 355) included V. parahaemo-
lyticus, V. cholerae (non-O1 and 01), Vibrio hollisae, V. mim-
icus, and V.fluvialis (table 1). The male-to-female ratio was
1.7:1. Symptoms of Vibrio gastroenteritis included diarrhea
(98% of cases), abdominal cramping (84%), nausea (73%),
fever (58%), and vomiting (54%). The mean duration ofillness
was 8.1 days (median, 7; range, 1-60) and did not differ by
Vibrio species. The proportion ofpersons hospitalized differed
by species, with the highest proportion (71%) occurring among
patients infected with V. cholerae 01 and the lowest (29%)
among patients infected with V. mimicus (table 2). The mean
hospitalization for patients with Vibrio gastroenteritis was 5.4
days (median, 4; range, 1-30).
Three cases (1%) were fatal. One patient (described in ),
a 54-year-old man, was hospitalized after experiencing vom-
iting and diarrhea for 4 days and anuria for 1 day. He ate raw
oysters during the week before becoming ill. A stool specimen
grew V.fluvialis but no other pathogens. A second patient, a
64-year-old woman with a history ofalcoholic liver disease and
gastric surgery, died after experiencing diarrhea for 2 weeks. V.
mimicus was recovered from a stool specimen; she had denied
eating seafood during the week before becoming ill. The third
patient, an 89-year-old woman with diabetes mellitus, had a
hemicolectomy 1month before death. Several days after eating
raw oysters, she became ill with diarrhea and was hospitalized
2 days before death; V. vulnificus was recovered from a stool
The 7 patients with V. cholerae 01 infection included 4
who had recently been in cholera-epidemic areas in South and
Central America. Of the 3 infections acquired in Florida, 2
were due to nontoxigenic strains and all were associated with
raw oyster consumption. The origin of oysters associated with
the single Florida-acquired case of toxigenic V. cholerae 01
infection in 1986 is unknown.
Wound infections. Vibrio wound infections, some with
multiple species, were reported for 159 patients. With the ex-
ception of V. cholerae 01, all Vibrio species during the study
period were recovered from a wound at least once; however,
V. parahaemolyticus, V. vulnificus, and V. alginolyticus collec-
tively accounted for 86% (137/159) of the wound infections
(table 1). Fever was reported in 45% ofthe cases. Although not
by guest on July 13, 2011
Hlady and Klontz
JID 1996; 173 (May)
Deaths and number and percentages of patients with Vibrio infections hospitalized in Florida, 1981-1993, by diagnosis.
Vibrio species Hospitalized
V cholerae non-Ol
V cholerae 01
3 (1) 110 (93)52 (44)
* All reported Vibrio infections, including those in other sites and those with insufficient information for clinical classification.
Data are no. (%). Nos. may not add to total because of infections caused by multiple species.
a query on the surveillance form, cellulitis was often reported
anecdotally. The mean duration of illness was 8.3 days (me-
dian, 6.5; range, 1-43). Eighty patients (50%) with Vibrio
wound infections were hospitalized, and 8 (5%) died (table 2).
The mean hospital stay was 9.9 days (median, 7.5; range, 1-
Exposure histories were available for 143 of the patients
with Vibrio wound infections: 126 (88%) had been exposed to
seawater inthe 7days preceding illness. Ofthe patients exposed
to seawater, 98% acquired their wounds while they were in or
around the water. Anecdotally, patients were often employed
as dock workers, fishermen, or oyster shuckers, reflecting the
occupational nature of such wounds and explaining, in part,
the high male-to-female patient ratio (6:1). All 8 patients with
fatal wound infections were males (table 2).
Vibrio organisms were recovered from blood specimens of
24 patients with wound infections (15%): V vulnificus (15), V
parahaemolyticus (6), and V cholerae non-O1, V. hollisae, and
V alginolyticus (1 each). The case-fatality ratio for patients
with vibrio recovered from a blood specimen was 32%; for
those without vibrio recovered from blood, the case-fatality
ratio was 1% (risk ratio [RR], 39.5; 95% CI, 5.1-305.2; P <
10-5) . Likewise, the presence of preexisting liver disease was
a strong predictor of death, with a case-fatality ratio of 50%
among patients with liver disease versus 2% among patients
without liver disease (RR, 28.3; 95% CI, 6.3-127.5; P <
Most cases of primary septicemia (n
== 118) were due to infection with V vulnificus, V. cholerae
non-O1,or V parahaemolyticus (table 1).Presenting symptoms
included fever (84%), diarrhea (60%), nausea (55%), abdomi-
nal cramping (55%), and vomiting (53%). Subsequent symp-
toms typically included swollen painful lower extremities with
hemorrhagic bullae. The male-to-female ratio was 4.4:1. The
mean duration of illness in patients with Vibrio septicemia was
9.6 days (median, 6; range, 1-46). The mean hospital stay was
8.2 days (median, 4; range, 1-75).
The case-fatality ratio differed depending on the Vibrio spe-
cies associated with primary septicemia (table 2): The highest
ratio (56%) was among patients infected with V vulnificus. In
all fatal cases, the mean time from hospitalization to death was
4.3 days (median, 2; range, 1-20). The case-fatality ratio
among septicemic patients with preexisting liver disease was
58%; in those without preexisting liver disease, it was 30%
(RR, 1.9; 95% CI, 1.2-3.1; P < .01). Ofpatients with primary
septicemia, 83% reported eating raw oysters during the week
before onset of illness (median, 2 days; range, 1-7).
Fifty-nine children (9%; ages, 21 days
to 17years) had Vibrioinfections: wound infections, 30;gastro-
enteritis, 15; ear infections, 2 (V. alginolyticus in boys ages 6
and 8 years); urinary tract infection, 1 (a 6-year-old girl with
V parahaemolyticus); and 1drowning victim with Valginoly-
ticus in sputum. In 10cases, there was insufficient information
for classification. There were no fatal infections. Ten children
with wound infections were hospitalized (mean, 5.8 days; me-
dian, 3.5; range, 1-12). Five children with Vibrio gastroenteri-
tis were hospitalized a mean of3.8 days (median, 4; range, 2-
5); all were < 1 year old. Anecdotally, we found that raw
oysters were likely to have transmitted Vibrioinfection in some
of the children. Four of 15children < 18 years old with gastro-
enteritis ate raw oysters before becoming ill. Of note, the father
of a 9-month-old child with V parahaemolyticus gastroenteritis
reported feeding the infant raw oysters before the onset of
of selected preexisting medical conditions among patients with
the three principal syndromes ofVibrio infection. Prevalence of
Table 3 summarizes the prevalence
by guest on July 13, 2011
JID 1996; 173 (May)
Vibrio Infections in Florida
syndrome in patients with Vibrio infections in Florida, 1981-1993.
Prevalence of selected preexisting conditions by clinical
(n = 349)
(n = 159)
(n = 118)
Peptic ulcer disease
Any of above
were counted as not having preexisting condition; thus, % represents minimum
* Conditions are not mutually exclusive.
t Includes AIDS, human immunodeficiency virus infection, leukemia,
lymphoma, systemic corticosteroid therapy, and cancer chemotherapy.
Data are no. (%). Patients for whom information was not available
each condition was much higher among patients with primary
septicemia than in those with either gastroenteritis or wound
One patient, a 40-year-old alcoholic man with fatal V. vul-
nificus septicemia, had hemochromatosis and liver disease. A
33-year-old woman with V. hollisae gastroenteritis was preg-
nant (20 weeks) and had no other preexisting conditions. Also
not shown in table 3 is a patient with fatal V. vulnificus septice-
mia whose only preexisting condition was pernicious anemia.
Two patients with Vibrio infection had AIDS. One had gas-
troenteritis due to V. cholerae non-O1; he had no other preex-
isting conditions and an unknown history ofseafood consump-
tion. The other had gastroenteritis due to V.fluvialis, no other
preexisting conditions, and no history ofseafood consumption.
The observed incidence rate of Vibrio infection was 2.6 per
1,000,000 people with AIDS per year.
Two patients had human immunodeficiency virus (HIV) in-
fection without AIDS when infected with vibrio. One had gas-
troenteritis due to V.parahaemolyticus; he had no other preex-
isting condition, and the only seafood he had eaten during the
week before onset of illness was cooked shark. The other pa-
tient had a fatal V.parahaemolyticus wound infection; he also
had preexisting liver disease. Because HIV infection without
AIDS is not reportable in Florida, we could not calculate an
Figure 1 shows the monthly distribution of
Vibrio infections associated with each of the major clinical
syndromes. A seasonal pattern was most evident for wound
infections and septicemia: 91% (145/159) ofwound infections
and 86% (102/118) ofsepticemias occurred from April through
October. Peak incidence during warm weather months was
most evident for infections caused by V. vulnificus and V.
parahaemolyticus: Much less seasonal variation was observed
for infections with other Vibrio species (figure 2).
common in the week preceding onset ofillness in patients with
Vibrio gastroenteritis or septicemia (table 4). Among patients
with Vibrio wound infections, recent seafood consumption was
reported by only 5% percent, and none had eaten raw seafood.
The most common seafood associated with infection was raw
oysters. In all but 2 cases, the raw oysters were served on the
half shell. One case of gastroenteritis due to V. mimicus and 1
fatal case of septicemia due to V. vulnificus were associated
with consumption ofraw oysters purchased as shucked product.
The observed average annual incidence of Vibrio infection
was 5.1 per 1,000,000 adults>17 years old. The estimated
average annual incidence among raw oyster-consuming adults
was 10.3 per 1,000,000 (95% CI, 8.8-12.5). Among raw
oyster-consuming adults with known liver disease, the esti-
mated average annual incidence was 70.2 per 1,000,000 per-
sons (95% CI, 42.1-1684.5).
Seafood consumption was very
The clinical and epidemiologic features of domestically ac-
quired Vibrio infections are being described in increasing detail
. In addition to offering new information about the incidence
of and risk factors for infection, reports describe previously
unrecognized Vibrio species implicated in human illness (e.g.,
Vibrio cincinnatiensis  and Vibrio carchariae ). Several
studies report the results of multiyear surveillance for Vibrio
infections in communities , metropolitan areas , and for
a I-year period in Gulf Coast states . To our knowledge,
the Vibrio case series summarized here represents the largest
domestic series described to date. Earlier descriptions of sub-
sets of this series have included 333 cases reported during
1981-1988  and problems associated with V. vulnificus
Our findings corroborate previous observations that most
Vibrio infections in the United States present clinically as gas-
troenteritis, wounds, or primary septicemia. In Florida, V.para-
haemolyticus is the species most often associated with human
illness, although infection with V. vulnificus and V. cholerae
non-O1 are also common. Other surveillance studies have also
found these three species to be the most commonly reported
[4, 7, 8]. We found several species were primarily reported in
the clinical setting ofdiarrheal disease (V. mimicus, V. hollisae,
V.fluvialis, and V. cholerae 01), others with wound infection
(V. damsela and V. alginolyticus), and some with diarrheal
disease, wound infection, and septicemia (V. vulnificus, V.chol-
erae non-Ol, and V. parahaemolyticus). However, on rare oc-
casions, species such as V. hollisae, V. mimicus, and V.fluvialis,
were recovered from blood specimens.
We found that several Vibrio species previously infrequently
associated with major roles in US morbidity contributed sig-
nificantly to illness. V. mimicus, V. hollisae, and V. fluvialis
collectively accounted for 20% of reported Vibrio infections,
by guest on July 13, 2011
1180Hlady and Klontz
JID 1996; 173 (May)
number of patients
- - wound infection
. - - .. - . - . - - - , - _.. - - - .. - . - .. - - - - .. - - - - . - .. - - - - - - - - -
tions in Florida by clinical syndrome, 1981-1993.
Monthly distribution of Vibrio infec-
, \.. ...
- -,--- ..•
- - - - - - - - - -. - - - - . - -,- ,- -. - - . -.-. _.~
. - - - - - . .': - . - - . - . - - -. - -
- - - . - - - . - - - -, -,,
month of onset-
and 42% ofpatients infected with these organisms were hospi-
talized. This proportion is higher than for many other foodbome
enteric infectious diseases  but should be considered a
maximum estimate, as infections in hospitalized patients may
have been more likely to be reported. It is also notable that
most patients infected with these 3 Vibrio species were young
(median age, 36 years) and healthy (80%reported no underly-
ing medical condition).
Even with the knowledge that hospitalizedpatients with Vibrio
infection may have been more likely to be reported, the overall
53% case-hospitalization rate we observed is impressive. On aver-
age, Vibrio infections(including those with atypical presentations)
accounted for 218 patient hospital days per year in Florida during
our study period (1.8 days/l00,000 population/year).
Seafood consumption in the week before onset of illness
was very common for patients with Vibrio gastroenteritis or
primary septicemia (87% and 88%, respectively; table 4). Sea-
food, particularly molluscan shellfish, can harbor various Vib-
rio species [21-23]. We suspect that some persons with Vibrio
infections who reported eating only cooked seafood before
becoming ill may have eaten seafood that became cross-con-
taminated with the drippings from raw seafood after cooking.
Raw oysterswereingestedby 72% ofpatientswith Vibriogastro-
enteritisor primary septicemiawithin 7 days beforeonsetofillness.
A strong role for raw oysters in facilitating Vibrio infections has
been reported [1, 3, 10, 21, 24, 25]. Consumption of raw clams
and other raw seafoods was less common among patients in our
study. The prominent role of raw oyster consumption in the trans-
mission of Vibrio infection in Florida is not surprising given that
this filter-feeding mollusk is harvested from Gulf Coast marine
waters replete with Vibrios and that, according to a 1988 Florida
survey, one-quarter of the state's adult population ate raw oysters
at least once during the preceding year. The seasonal occurrence
of Vibrio infections,especiallythose resultingin septicemia,largely
reflects the seasonal pattern of isolation of V vulnificusfrom Gulf
Coast oysters. Virtually 100% of Gulf Coast oysters harvested
during warm weather months contain this organism , no safe
level has been established, and >90% of raw oyster-associated
V vulnificus infectionsoccur from April through October [16, 18,
While a seasonal pattern of infection was most evident for
V. vulnificus, infection with V. parahaemolyticus, V. cholerae
non-Ol, and V. alginolyticus was also more common during
warm weather months. Infection with other Vibrio species did
by guest on July 13, 2011