Impact of Xynthia Tempest on Viral Contamination of Shellfish

Abstract

Viral contamination in oyster and mussel samples was evaluated after a massive storm with hurricane wind named “Xynthia tempest” destroyed a number of sewage treatment plants in an area harboring many shellfish farms. Although up to 90% of samples were found to be contaminated 2 days after the disaster, detected viral concentrations were low. A 1-month follow-up showed a rapid decrease in the number of positive samples, even for norovirus.

TEXT

Global climate change, interfering with many complex events, may impact the hydrological cycle, altering mean meteorological measures and increasing the frequency of extreme events (i.e., excessive precipitation, storms, floods, or droughts). Disasters destroy all substructures, such as ground transportations, roads, sewage networks, and sewage treatment plants, leading to microbial contamination in coastal areas. Following hurricanes Katrina and Rita, several investigators evaluated exposure to chemical or microbial contamination originating from human and animal waste or the broader effects on algal blooms (7, 10, 25). Vibrio and Legionella concentrations were more abundant shortly after the events, and fecal indicator concentrations in offshore waters returned to prehurricane levels within 2 months (27). As shellfish are prone to microbial contamination by filtering sewage-contaminated waters, it is important to evaluate microbial quality of shellfish beds after such an event to avoid the introduction of contaminated shellfish on the market.

A massive storm with hurricane force wind, named “Xynthia tempest,” came through France during the night of 27 to 28 February 2010. At 2:30 a.m., strong wind (140 km/h), important atmospheric pressure variation (up to 2.5 hPa), and a high tide range caused major destruction on the southwestern coast of France, with a massive flood reaching more than 4 m of water depth, and claimed 51 lives. The impacted area was restricted (about 50 km of the coast and two small islands), but the flood damaged most of the sewage pipe network and sewage treatment plants (Fig. 1). As many shellfish farms are located in this area, a sanitary alert was raised and shellfish samples were collected. This study reports the follow-up of viral contamination in shellfish samples collected in this area over 1 month.

Fig 1.

Map of the area impacted by the Xynthia tempest. (A) Satellite observation of the tempest crossing the area on 28 February. (B) Detailed map of the area destroyed by the tempest (yellow diamonds, sewage treatment plants; red dots, shellfish sampling points).

Oyster (Crassostrea gigas) and mussel (Mytilus edulis) samples were collected from 2 March to 29 March 2010. Each sample consisted of at least 12 oysters or 24 mussels. Escherichia coli analysis was performed on the same samples according to a European regulation (2073/2005/EC).

For viral analysis, shellfish were shucked, and stomach and digestive tissues (DT) were removed by dissection and divided into 1.5-g portions. Mengovirus (2 × 10⁴ 50% tissue culture infective dose [TCID₅₀]) was added as an external viral control to each sample. Tissues were homogenized, extracted with chloroform-butanol, and treated with Catfloc-T (Calgon, Ellwood City, PA). Viruses were then concentrated by polyethylene glycol 6000 (Sigma, St. Quentin, France) precipitation (3).

Viral nucleic acids (NAs) were extracted with a NucliSens kit (bioMérieux, France) according to the manufacturer's instructions but with extended incubation for 30 min at 56°C for initial viral lysis. NAs were analyzed immediately or kept frozen at −80°C (15).

NA extracts were screened by real-time reverse transcription (RT)-PCR (rRT-PCR) with previously published primers and probes for mengovirus (21), norovirus (NoV) (26), sapovirus (SaV) (19), hepatitis A virus (HAV) (5), hepatitis E virus (HEV) (11), Aichi virus (AiV) (14), enterovirus (EV) (18), and rotavirus (RV) (20). Positive controls constituted by plasmids (NoV, SaV, HAV), French positive stool (HEV), or cultured viruses (AiV, EV, RV) were included in each run. rRT-PCR was performed using the RNA UltraSense one-step quantitative RT-PCR system (Invitrogen, France) with adjusted concentrations of primers and probes and thermal conditions as described previously (15). To avoid possible false-negative results due to PCR inhibitors, all samples were analyzed in duplicate by using 5 μl of undiluted or 10-fold-diluted RNA extracts. Negative amplification controls (water) were included in each amplification series, and precautions (filter tips and separate rooms) were taken to prevent false-positive results. The cycle threshold (C_T) was defined as the cycle at which a significant increase in fluorescence occurred. To be considered positive, a sample had to yield a C_T value of ≤41. The efficiency of virus extraction procedures was determined for each sample based on mengovirus recovery (15). For samples presenting an extraction efficiency above 10%, quantification was performed for NoV and SaV considering the NA volume analyzed and the weight of DT extracted (1.5 g). If the extraction efficiency was less than 10%, extraction was repeated. If the extraction efficiency percentage was not improved, the sample was considered positive but excluded for quantification.

All concentrations obtained were log transformed, and geometric mean concentrations were calculated. Mean concentrations were compared by using the Student t test, and a P value of <0.05 was considered significant (Statgraphics Centurion XV).

The tempest impacted two production areas located in two bays separated by an island (Fig. 1B, areas 1 and 2). Twenty-two samples were collected from area 1 and 24 samples were collected from area 2, representing 28 oyster and 18 mussel samples. On 2 to 3 March, all 8 samples collected from area 1 displayed fewer than 230 E. coli CFU/100 g of shellfish meat (class A area according to European regulation 854/2004/EC). Among the 9 samples collected from area 2, the mean concentration was 446 E. coli CFU/100 g of shellfish meat, with three samples having fewer than 230 E. coli CFU/100 g of shellfish meat. All samples collected later except one sample collected on 29 March from area 2 (240 E. coli CFU/100 g) met European regulation class A requirements. The extraction efficiency was considered acceptable (>10%) for 40 samples and varied from 26% to 40% over the sampling period. Despite repeated extractions, 6 samples showed extraction efficiencies below 10%: one sample collected on 2 March (area 2), one on 14 March (area 1), three on 18 March (all 3 from area 1), and one on 29 March (area 2). Six samples were positive for RV, 7 for EV, 15 for NoV, and 26 for SaV (Table 1). None of the samples was positive for HEV, HAV, or AiV. Multiple contaminations were observed more frequently at the beginning of the month, and one sample, collected on 2 March from area 2, was found to be contaminated by at least 4 different enteric viruses. However, most of the samples were contaminated by only one type of enteric virus, and the number of samples with concentrations lower than the sensitivity threshold of the method (about 50 RNA copies/g of DT) increased over time (Fig. 2). The 15 samples positive for NoV were found to be contaminated by GII strains, and none were found to be contaminated by NoV GI or GIV. More samples were found to be contaminated on 2 to 3 March (59%) than on 29 March (25%); however, average concentrations stayed in the same range (Table 1). SaVs were detected in 26 samples. On 2 to 3 March, SaVs were detected in 76% of samples, and on 29 March, they were detected in 25% of samples, with comparable average concentrations (Table 1).

Table 1.

Detection of human enteric viruses in shellfish samples

Sampling dates in March	No. of samples	Avg extraction efficiency (%)	No. of positive samples and mean concn (RNA copies/g of DT) for each virus				No. of positive samples for each virus
			NoV		SaV		EV	RV	AiV	HAV	HEV
			No. of positive samples	Mean concn	No. of positive samples	Mean concn
2 to 3	17	34.91	10	134.86	13a	1,976.88	6	3	0	0	0
11 to 14	8	38.25	2	160.09	5a	2,978.24	0	1	0	0	0
17 to 18	13	26.35	1a		6b	1,334.41	1	1	0	0	0
29	8	40.46	2	97.40	2	1,342.28	0	1	0	0	0

^aOne positive sample not considered for quantification due to an extraction efficiency of <10%.

^bTwo positive samples not considered for quantification due to extraction efficiencies of <10%.

Fig 2.

Multiple contaminations observed for shellfish samples over time. Black bars indicate two or more different enteric viruses detected per sample. Gray bars indicate one virus detected per sample. White bars indicate no virus detected. The x axis shows the sampling time, and the y axis shows the percentage of positive samples.

No statistical differences were observed between areas 1 and 2 in comparing the numbers of samples positive for NoV (P = 0.603), SaV (P = 0.393), EV (P = 0.157), or RV (P = 0.429) or the NoV and SaV concentrations (P = 0.958 and P = 0.217, respectively) (Table 2). A large diversity of human enteric viruses may be detected in human sewage, with some being detected frequently (for example, NoV and RV) and some sporadically (HAV, HEV, AiV), based on local epidemiology (8, 9, 12, 13, 24). Raw sewage may contain high viral concentrations, especially during cold months, the period of the winter time gastroenteritis epidemic in many countries (2, 26). Thus, direct discharge of raw water may have an important impact on shellfish contamination (16).

Table 2.

Distribution of viral contamination in the impacted areas

Area	No. of samples	Avg extraction efficiencya (%)	No. of positive samples and mean concn (RNA copies/g of DT) for each virus				No. of positive samples for each virus
			NoV		SaV		EV	RV
			No. of positive samples	Mean concn	No. of positive samples	Mean concn
1	22	31.98	8b	130.75	11c	2,707.30	5	2
2	24	32.73	7	161.35	15b	1,570.41	2	4

^aThere were four samples from area 1 with extraction efficiencies of <10%. There were two samples from area 2 with extraction efficiencies of <10%.

^bOne positive sample not considered for quantification due to an extraction efficiency of <10%.

^cThree positive samples not considered for quantification due to extraction efficiencies of <10%.

Clearly, this tempest had an impact on shellfish quality, as 2 days after the event, up to 90% of samples were found to be contaminated. No sample collected prior to the event was available, as this area has never been implicated in a shellfish-related outbreak in France, suggesting that such a high number of positive results is unlikely to represent the normal situation. Nevertheless, the diversity of detected viruses was low. Controls included in the method made us confident that these samples were truly negative. This observation may be explained by the low prevalence of some viruses in the French population or, in the case of HEV, the absence of pig farms in this area (1, 6, 22, 23). In contrast, NoV and SaV, which are frequently detected in French sewage, were detected in the two impacted areas (4, 26).

If the rapid decrease of E. coli was expected, it was more surprising to observe that after 10 days, only 20% of samples were found contaminated by two different enteric viruses, since viruses, particularly NoVs, are known to persist in oyster tissues for several weeks. This may be explained by the low concentrations detected per gram of digestive tissues (which then rapidly reached the sensitivity limit of detection of the method) and the fact that only GII strains were detected. During winter epidemic outbreaks, GII.4 strains are the most prevalent strains in human cases, suggesting that sewage waters may contain mostly those strains. We previously demonstrated that GII.4 is less efficiently concentrated by oysters (17) and that the decrease of NoV GII concentrations in oyster may be faster than that of NoV GI (14).

Both bacterial and viral concentrations detected in shellfish tissues were low. A few years ago, an important rainfall event in southern France led to a massive shellfish contamination, with high E. coli concentrations and a large diversity of human enteric viruses being detected at high concentrations (using the same detection method) (14). However, this contamination occurred in a lagoon and over a longer period of time. Here, the impacted area was open to the ocean, submitted to marine currents and tide. In addition, we hypothesize that the phenomenon's abruptness (wind, atmospheric pressure, and large volumes of fresh water) stressed the shellfish, hampering their filtration activity for a few hours. These observations highlight the role of environmental parameters that may contribute to the probability of shellfish contamination. Indeed, in cases of natural disasters, it is important to react rapidly to protect the consumers but also to protect shellfish producers' business.