Subjects

A confirmed patient was defined as a person with compatible signs and symptoms (ulcerated lesions in the hands and regional lymphadenopathies with or without fever and general discomfort) and an accompanying positive laboratory result (seroconversion or single antibody titer >1:128 to F. tularensis as measured by microagglutination or a positive polymerase chain reaction [PCR] result).

Epidemiologic Study

Several visits were made to the epidemic site. Fishing areas were identified by photographs that were subsequently shown to patients and controls. Interviews were conducted with local persons to obtain relevant information and with experts in red swamp crayfish ecology. Mean monthly dam water levels for 1998 were obtained.

An active search was made for compatible cases and for persons reporting possible contact with any potential reservoir of the disease. Health-care centers were alerted. We searched reports of emergency visits (for January 1 to August 31, 1998) and discharges (for January 1 to September 30, 1998) at the provincial hospital for patients who met the case definition or had tularemia-compatible symptoms.

Patients were interviewed to evaluate the probability of any contact with other tularemia reservoirs, pinpoint the fishing site, and define risk factors linked to crayfish fishing. We selected 20 controls who did not meet the case definition, had been in contact with crayfish caught on the same dates and in the same places, and generally resided in the same towns as patients. All patients and controls were interviewed from October 8 through December 7, 1998.

The crude odds ratio (OR), its 95% confidence intervals (CI), as well as adjusted OR, were determined by logistic regression.

Bacterial Strains and Culture

The live vaccine strain of F. tularensis type B (NCTC 10857) was used for serologic diagnosis and as a positive control for PCR. The strain was grown in modified Thayer-Martin (Difco Laboratories, Detroit, MI) chocolate, supplemented with Isovitalex (10 mg/L; BBL Microbiology Systems, Cockeysville, MD) and 1% L-Cysteine (Sigma-Aldrich Co., Alcobendas, Madrid, Spain) at 37°C in 5% CO₂ (3,13).

Environmental Samples

Eight water samples were collected at several points on the river where all patients had fished. Three water samples from a sewage plant located 15 km upstream were also collected. Water sampling was repeated 1 month later. Three crayfish captured by one patient had been frozen and were available for study (batch A). Twenty more crayfish were collected from two stretches of the river where the outbreak had started (batch B). Sixteen hares (Lepus spp.) (12 alive; 4 dead), 7 common shrews (Sorex araneus), 26 field mice (Apodemus sylvaticus), 2 gudgeons (Gobio gobio), 3 rats (Rattus rattus), 1 badger (Meles meles), and 1 common vole (Microtus arvalis), as well as sera from 60 sheep from the epidemic area were also studied. Wild mammals were euthanized in a CO₂ chamber and both liver and spleen cultured (13).

Intestines of batch A crayfish were not available for study as they were removed in situ and arrived at the laboratory after being extensively washed in tap water. Consequently, branchial tissue from three crayfish, one stomach, two gonads, and one hepatopancreas were aseptically collected and thinly ground with a scalpel. The carcasses were immersed in autoclaved distilled H₂O, shaken at 33°C for 30 min, centrifuged at 4,000 x g for 15 min at 4ºC, and the pellet (named wash) resuspended in 100 µL of autoclaved distilled H₂O. The rest of the process was as described above. For batch B crayfish, intestines were also dissected and studied. All fractions were plated in the same media as above.

Water samples (10 mL) were concentrated by centrifugation at 8,000 x g for 15 min and the pellet resuspended in 1 mL of the original sample and subjected to acidic shock, as described for Legionella pneumophila (15). Briefly, 0.1 mL of each concentrated sample was mixed with 0.9 mL of buffer HCl/KCl pH 2.2, and after an incubation of 5 min at room temperature, the samples were plated. This method was based on the assumption that the survival capacities of these microorganisms in the intracellular environment could account for resistance to acidic solutions and reduce the level of contaminants. Individual colonies that looked similar to F. tularensis live vaccine strains were pooled in groups of 10 in 100 µL of autoclaved distilled H₂O, carefully mixed, and half the volume was reserved for PCR. The rest of these samples were kept at 4°C.

Microaglutination Test

Agglutinating antibodies against whole F. tularensis live vaccine strain were determined as described (16,17). Titers >1:64 were considered positive.

PCR Detection of F. tularensis

A lymph node aspirate from one of the patients was subjected to guanidine thyocianate lysis for DNA extraction as described (18). All reagents were purchased from Sigma-Aldrich, except for glycogen and proteinase K (Boehringer Mannheim, Roche Molecular Systems, Inc., Branchburg, NJ).

For the water samples, 1 mL of each was centrifuged at 15,000 x g for 10 min, and the pellet was resuspended in 100 µL of the same sample and clarified by a 5-min centrifugation at 3,000 x g to eliminate insoluble material. The supernatant was then collected, lysed as above, and DNA precipitated and resuspended with 10 µL of autoclaved distilled H₂O.

For the crayfish, all fractions studied by culture were also digested as above and subjected to PCR by using the 16S ribosomal DNA-based universal F1-R13 and Francisella-specific F5-F11 sets of primers described by Forsman et al. (19) in a nested PCR (Table 1).

The F5-F11 amplicons obtained from water, lymph node aspirates, and crayfish samples were sequenced by PCR with the BigDye Terminator Cycle Sequencing kit on an ABI 377 automated DNA sequencer (Perkin-Elmer Corp, Foster City, CA) following the manufacturer's instructions. Sequence homology was searched by using the BLAST 2.0 queueing system (20).

Patients and Serology

Nineteen patients who fulfilled laboratory criteria for tularemia were identified. All had had contact with river-caught crayfish in the period July 13-31. Overall, 11 (58%) required hospitalization, with a mean stay of 9.4 days (Table 2). Symptoms included adenopathies on the elbow, armpit, or both (100%); cutaneous lesions (89%); fever (80%); discomfort at the injury site (42%); vomiting (32%); diarrhea (5%); and general symptoms (84%) such as headache, arthralgia-myalgia, malaise, anorexia, dysuria, and bad taste in the mouth.

All cutaneous lesions were on hands, generally the fingers; no specific finger or hand predominated. In six cases, ulcers were present. The remaining patients reported inflammation of the finger or phalanx to which they had received the crayfish-related injury. All ulcers were at the injury site.

Forty-one serial serum samples from all the patients were collected (Table 2). All 19 patients reached F. tularensis specific microagglutination titers >1:256. Increases in serum antibody titers were observed in five cases (Table 2). For the remaining patients, high antibody titers in the presence of compatible clinical symptoms and full recovery after antibiotic treatment were interpreted as confirmatory criteria.

Serum samples from wild animals and sheep were negative.

Case Survey and Risk Factors

Patients included 11 men and 8 women, mean age 59.1 years (range 38-75); the occupations of housewife (7 patients) and pensioner (5) predominated. None had professions entailing risk practices for tularemia. Most had been catching and cleaning crayfish for many years.

Figure 1

Figure 1. . Epidemic curve of tularemia cases, central Spain, by week. One case had unknown symptom-onset date.

Cases were concentrated (by date of symptom onset) within a 3-week period (July 16-August 5; Table 2, Figure 1); the greatest concentration of cases occurred during July 17-26. All patients had been in contact with crayfish caught during July 13-30. Fishing had been officially permitted on June 1 but was prohibited on August 6 because of the outbreak. The estimated incubation period was 3.5 days (range 1-8 days).

Patients resided in six towns situated around the Mayor River fishing area. Two patients had been in contact with other tularemia reservoirs (hares). No insect bites were reported for any patients.

Figure 2

Figure 2. . Fishing areas where crayfish handled by patients were caught. (Data were unavailable for one patient.)

All crayfish caught were of the red swamp species (P. clarkii). The section of the river authorized for fishing is 14 km, from the town of Huete to the beginning of the Buendía Dam into which the river flows. There are designated hunting (including hare-hunting) areas on both river banks. Patients had been fishing a ±5-km stretch, particularly the area at the neck of the dam, where access to the river was simplified by the proximity of the road and two footbridges (Figure 2). This terrain is alluvial, so crayfish can easily dig their underground burrows (22); it is prone to periodic variations in the water level and marked by reeds. The abundance of crayfish favored fishing by hand (an outlawed practice) rather than with a crayfish net.

Our case survey showed that 113 patients had caught crayfish; at least 9 of these patients had also cleaned them. Six patients had cleaned crayfish but had neither caught them nor been in the fishing area; five of these six had cleaned crayfish the day after the catch, and one had done so 5 days afterwards. In all, 15 patients had cleaned crayfish, and at least 17 of the total 19 had also eaten crayfish. Cooking methods used were all low risk.

At least 17 patients had incurred crayfish-related scratches, cuts, and abrasions while fishing, emptying the net, or cleaning the catch; 8 had only one injury. One patient, who had been injured when fishing and cleaning, also grazed against a riverside reed; the cutaneous lesion subsequently appeared at the site of the scratch.

Water used for washing crayfish was from the residential drinking water supply; chlorine levels were checked and shown to be acceptable. In 9 (75%) of 12 respondents who cleaned crayfish, the washing method consisted of leaving the crayfish to soak and changing the water several times. During this operation, a considerable quantity of mud and dirt was released, which the cleaner touched. In the remaining three cases, crayfish were rinsed in running tap water. These two washing operations were performed by 58% (7/12) of patients before they gutted the crayfish. All but one of the 11 respondents reported washing their hands after handling the crayfish, but 5 of those 10 used no soap or disinfectant.

Case-Control Study

Demographic information on patients was obtained (Table 3). Although there was a higher proportion of women and a lower proportion of crayfish catchers among controls, these differences were not statistically significant.

The most important result (Table 4) was that those persons who were injured during the handling (catching or cleaning) of crayfish had a significantly higher risk for tularemia (94% of exposed subjects among cases versus 30% among controls; OR = 39.7; 95% CI 4.3-369.7). Furthermore, once data were reciprocally adjusted for the catching- and cleaning-related injury variables, merely incurring an injury (regardless of whether it was received while catching crayfish [OR adjusted for cleaning-related-injury = 29.1; 95% CI 2.6-330.9] or cleaning crayfish [OR adjusted for catching-related-injury = = 38.8; 95% CI 3.5-427.6]) led to a statistically significant rise in the risk for tularemia.

Environmental Study

Figure 3

Figure 3. . Buendía Dam water levels, April-September 1998. OP = outbreak period.

Local environmental variations detected over the previous 2 years were higher rainfall, with correspondingly higher river levels; abundant crayfish; and increased murkiness of the river water. In July 1998, release of water through the sluice gates led to an estimated 1.8-m drop in the overall dam level, leading in turn to a 200- to 400-m narrowing in the shoreline at the neck of the dam, i.e., the fishing area (Figure 3). This coincided with a mass movement of crayfish to the surface.

Culture and PCR

Attempts to culture the organism from water, patients, crayfish, and wild animals were unsuccessful. A positive PCR result was obtained from one of the pools of 10 individual colonies taken from the culture of a water sample from the sewage plant, but we were unable to recover any F. tularensis in culture from the rest of the same sample kept at 4°C.

Figure 4

Figure 4. . Polymerase chain reaction results in selected samples from water, crayfish, patients, and controls. Lanes 1 and 2: positive control (first [1550-bp] and second [950-bp] rounds of amplification); lanes 3 and...

The PCR conditions best adapted to our system were an annealing temperature of 62°C for the first round of cycles and 60°C for the second. These conditions allowed us to amplify DNA from 1 CFU of spiked distilled water, which yielded a faint band only in the second amplification run (Figure 4), as did the lymph node aspirate from a patient (Figure 4, Table 1).

Two of three samples from the sewage plant yielded amplicons, although those samples were negative in culture. Stomach and hepatopancreas samples from the crayfish specimens collected at a patient's house were positive, and branchial tissue, gonad, and the washed pellet were negative. All water samples collected from different parts of the river were negative. PCR results of three additional water samples collected 1 month later from the same locations yielded the same results. Intestinal, stomach, and hepatopancreas samples from crayfish collected from the river after the outbreak started were negative (Figure 4, Table 1).

Sequencing of 16S rDNA

Figure 5

Figure 5. . A) Sequence alignment of a 75-nt fragment of the 16S rRNA gene of different bacterial species (nucleotide position 1113 to 1188 as for Escherichia coli). The signature nucleotide (nt 1153),...

An identical fragment of the 16S rDNA gene (nucleotide position 145-1290 as for Escherichia coli) was sequenced from the water, crayfish, and lymph node aspirate amplicons and analyzed. The homology in this region between the lymph node aspirate amplicon and the published sequences for F. tularensis tularensis, F. tularensis palaearctica, F. novicida, F. phylomiragia, Salmonella Typhimurium, Legionalla pneumophila, and E. coli was determined (Table 5). The lymph node aspirate sequence shared 100% of homology with the sequence L26086 from F. tularensis palaearctica (99.2% with F. tularensis tularensis and F. novicida, and 97.7% with F. philomiragia). The homology with S. Typhimurium, E. coli, and L. pneumophila was 81.0%, 81.1%, and 80.8%, respectively. When a shorter fragment was analyzed, between nucleotide positions 1113 and 1188, the sequence of the PCR products obtained had 100% homology with F. tularensis palaearctica, 98.6% with F. tularensis tularensis, 99.3% with F. novicida, and 98.9% with F. philomiragia (data not shown). In this region, we found 15 of 75 nt positions specific for Francisella spp., which implies a divergence of 20% compared with S. Typhimurium, L. pneumophila, and E. coli (Figure 5A). All sequences generated shared the nucleotide signature that differentiates F. tularensis palaearctica from other members of the genus (an A at nucleotide position 1153, as described for F. tularensis palaearctica, compared with a G, as described for F. tularensis type A, F. novicida, and F. phylomiragia) (23) (Figure 5A)].