Randomized Controlled Trial of Active Physician-Based Surveillance of Foodborne Illness

To the Editor: In New South Wales, Australia, physicians are obliged to report to public health authorities instances of foodborne illness in two or more cases related to a common source. This reporting of cases on a clinical basis complements laboratory-based surveillance of foodborne illness and is an essential form of surveillance in situations in which clinical samples may not be collected or in which specific diagnostic testing is not routinely available. Although cases of foodborne illness are increasing, substantial underreporting to health authorities is believed likely (1,2). However, reporting of foodborne illness and investigation of identified outbreaks are important forms of health protection (1–4). 
 
In a pilot study, we examined whether notification of single (rather than multiple) cases, active surveillance, or both would improve the reporting of foodborne illness by family physicians and thus its detection in the community. 
 
St. George Division of General Practice, one of four networks of family physicians located in the southeastern quadrant of Sydney within the jurisdiction of the South Eastern Sydney Public Health Unit, offered to participate in the study. Passive surveillance consisted of writing to all 329 members of the St. George Division asking them to report any single case of foodborne illness on a purpose-designed form that could be faxed to the Public Health Unit. Reports remained unidentifiable unless the patient gave the physician consent for Public Health Unit follow-up. The active surveillance group comprised 34 randomly selected St. George Division members who, in addition to being sent the written information, were contacted by telephone once every 3 weeks. 
 
Over the 12-week study period from August to November 1999, St. George Division physicians made 39 reports , 31 (79%) by facsimile and 8 by mail; in contrast, no reports of foodborne illlness were received from the other 900 family physicians practicing in southeastern Sydney. Of the 39 notifications, 26 were received from 12 (35%) of 34 active surveillance physicians and 13 from 8 (2.7%) of the remaining 295 physicians (odds ratio 19.6 [95% confidence intervals 6.6-59]). 
 
Consent was given for the Public Health Unit's food inspectors to follow up 13 cases, 6 of which represented multiple associated cases with possible public health implications. In one family, three members had pain, altered temperature sensation, and lower limb weakness 4 to 36 hours after eating portions of flowery cod; they were diagnosed as suffering from ciguatera poisoning. This potentially serious condition is caused by consumption of heat-stable ciguatoxin concentrated in the tissues of certain types of reef fish that have ingested toxin-producing plankton. Ciguatera poisoning has wide global distribution, including the Indo-Pacific and Caribbean regions (5); it has important public health implications because of its frequency and severity, and the fact that prompt recognition and treatment can lead to a good clinical outcome (5–7) 
 
Better ascertainment of foodborne illness is required to improve food safety in Australia, including removing suspect foods from circulation (1,3). We found that passive surveillance of single cases increased the reporting of suspected foodborne illness by family physicians, while active surveillance based on telephone contacts elicited notification of clusters of associated cases, one of which required prompt public health action. At the least, this pilot suggests vast underreporting of food poisoning and that different strategies are available to improve reporting. A large-scale study would be required to determine the feasibility, acceptability, and value to public health of this form of enhanced surveillance.


Randomized Controlled Trial of Active Physician-Based Surveillance of Foodborne Illness
To the Editor: In New South Wales, Australia, physicians are obliged to report to public health authorities instances of foodborne illness in two or more cases related to a common source.This reporting of cases on a clinical basis complements laboratory-based surveillance of foodborne illness and is an essential form of surveillance in situations in which clinical samples may not be collected or in which specific diagnostic testing is not routinely available.Although cases of foodborne illness are increasing, substantial underreporting to health authorities is believed likely (1,2).However, reporting of foodborne illness and investigation of identified outbreaks are important forms of health protection (1)(2)(3)(4).
In a pilot study, we examined whether notification of single (rather than multiple) cases, active surveillance, or both would improve the reporting of foodborne illness by family physicians and thus its detection in the community.
St. George Division of General Practice, one of four networks of family physicians located in the southeastern quadrant of Sydney within the jurisdiction of the South Eastern Sydney Public Health Unit, offered to participate in the study.Passive surveillance consisted of writing to all 329 members of the St. George Division asking them to report any single case of foodborne illness on a purpose-designed form that could be faxed to the Public Health Unit.Reports remained unidentifiable unless the patient gave the physician consent for Public Health Unit follow-up.The active surveillance group comprised 34 randomly selected St. George Division members who, in addition to being sent the written information, were contacted by telephone once every 3 weeks.
Over the 12-week study period from August to November 1999, St. George Division physicians made 39 reports , 31 (79%) by facsimile and 8 by mail; in contrast, no reports of foodborne illlness were received from the other 900 family physicians practicing in southeastern Sydney.Of the 39 notifications, 26 were received from 12 (35%) of 34 active surveillance physicians and 13 from 8 (2.7%) of the remaining 295 physicians (odds ratio 19.6 [95% confidence intervals 6.6-59]).
Consent was given for the Public Health Unit's food inspectors to follow up 13 cases, 6 of which represented multiple associated cases with possible public health implications.In one family, three members had pain, altered temperature sensation, and lower limb weakness 4 to 36 hours after eating portions of flowery cod; they were diagnosed as suffering from ciguatera poisoning.This potentially serious condition is caused by consumption of heat-stable ciguatoxin concentrated in the tissues of certain types of reef fish that have ingested toxin-producing plankton.Ciguatera poisoning has wide global distribution, including the Indo-Pacific and Caribbean regions (5); it has important public health implications because of its frequency and severity, and the fact that prompt recognition and treatment can lead to a good clinical outcome (5)(6)(7) Better ascertainment of foodborne illness is required to improve food safety in Australia, including removing suspect foods from circulation (1,3).We found that passive surveillance of single cases increased the reporting of suspected foodborne illness by family physicians, while active surveillance based on telephone contacts elicited notification of clusters of associated cases, one of which required prompt public health action.At the least, this pilot suggests vast underreporting of food poisoning and that different strategies are available to improve reporting.A large-scale study would be required to determine the feasibility, acceptability, and value to public health of this form of enhanced surveillance.

First Reported Case of Imported Hantavirus Pulmonary Syndrome in Europe
To the Editor: We report the first imported case of hantavirus pulmonary syndrome (HPS) in Europe.The patient, a 58-year-old man, traveled in Chile and Argentina from February 2 to March 2, 2001.A keen amateur botanist, he spent several days (February 17-23) on foot in Cohaique, Puerto Monte, and the surrounding rural areas collecting plant material.On February 24 and 25, he traveled by bus to Bariloche, Argentina.On March 3, shortly after he returned to France, he had a high fever (40°C to 41°C) with myalgia and headache.On March 18, he had dyspnea, which increased during the next 2 days.On March 21, he was hospitalized in the intensive care unit of the Centre Hospitalier de Compiègne, France.
On admission, the patient's fever was still high (40°C), there was severe hypoxemia with bilateral diffuse pulmonary infiltrates, a tachyarrhythmia with auricular fibrillation and gallop, and conjunctival injection.Laboratory results indicated mild renal insufficiency (urea 12.5 mmol/L; creatininemia 180 µmol/ L), hepatic cytolysis (serum glutamicoxaloacetic transaminase 236 and serum glutamic-pyruvic transaminase 72), a moderate thrombocytopenia (platelet counts 86 000/mm 3 ), an inflammatory syndrome (C-reactive protein 272 mg/L), and a capillary leak syndrome (hematocrit 49%; albuminemia 20 g/L).On the night after admission, an aggravation of the cardiac function with myocarditis developed; it responded quickly to symptomatic treatment.The patient's condition improved steadily on the following days with a reduction of the pulmonary manifestations, and he was discharged on April 2.
Blood samples obtained on March 21 and March 29 were tested for the presence of antibodies to hantaviruses (Puumala, Hantaan, Sin Nombre, [SNV] Seoul, and Laguna Negra) by immunoglobulin (Ig) M-capture and IgG enzyme-linked immunosorbent assay.IgM antibodies were detected for all these antigens on the first sample, but there was no increase on the second sample.A substantial increase in IgG titer for SNV and Laguna Negra antigens was observed from the first to the second sample, but not for the other antigens.The virus could not be detected either by reverse transcription-polymerase chain reaction or by inoculation into cell culture (three passages).Since the identification in 1993 of SNV as the cause of HPS (1), numerous cases of this disease have been confirmed in various regions of North and South America.The first HPS cases associated with Andes virus in Argentina (2) were observed in 1995.Since then, more than 500 HPS cases have been reported in six countries of South America (Argentina, Bolivia, Brazil, Chile, Paraguay, and Uruguay), with mortality rates ranging from 30% to 70%.
Hantaviruses are rodent-borne, and each is associated with a specific rodent.Sigmodontine rodents are the vectors of hantaviruses associated with HPS.Infections are most frequently transmitted by inhalation of virus-contaminated aerosols of rodent excreta, but human-tohuman transmission has also been described (3).
The patient described here was probably infected in Chile and more likely in the Puerto Monte area, where HPS cases were reported in 2001.Unfortunately, virus could not be detected because the first blood sample was obtained 2 weeks after onset of fever.

Efficacy of Interferon Alpha-2b and Ribavirin against West Nile Virus In Vitro
To the Editor: West Nile virus (WNV) infected humans in the Western Hemisphere for the first time in the late summer of 1999.During 1999 and 2000, nine deaths occurred among 80 patients with meningitis or encephalitis in New York City; Westchester County, New York; New Jersey; and Connecticut (1)(2)(3).Effective antiviral agents are unknown for infections caused by WNV.Odelola (4) described 83% survival of WNV-infected mice and eradication of virus from brain when 1.5 mg. of ribavirin was administered by intraperitoneal injection after virus inoculation.Survival of controls was 25%.More recently, Jordan et al. have reported inhibition of WNV by a relatively high concentration of ribavirin (200 µM) given after infection of human oligodendroglial cells in vitro (5).Shahar et al. (6) reported protection of fetal mouse spinal cord tissues with mouse alpha and beta interferon before inoculation with WNV.We tested human recombinant interferon alpha-2b and ribavirin for activity against WNV in a primate cell system similar to that used to measure the effect of these agents on Bovine viral diarrhea virus, a cultivatable, closely related surrogate for Hepatitis C virus.Vero cells were cultured at 37 o and 5% CO 2 in a 96-well microtiter plate.Approximately 13,000 cells were seeded in each well 24 hours before specific concentrations of ribavirin or interferon alpha-2b were added.Approximately 2.5 X 10 3 PFU of WNV isolated from Culex pipiens (7) was added 1.5-2 hours after or before the addition of interferon alpha-2b or ribavirin to Vero cells.Forty-four hours after treatment, a colorimetric proliferation assay was used to measure viable cells in each treated well according to the protocol of Promega (Madison, WI).Cells exposed to specific concentrations of antiviral compounds, but without WNV, were used as negative controls.
Interferon alpha-2b was protective and therapeutic.Interferon alpha-2b inhibited viral cytotoxicity at low dosage when applied before or after infection of cells with WNV.For example, viral protection was observed from 3,000 U/mL to 188 U/mL 2 hours before infection of cells with WNV.Interferon alpha-2b was also therapeutic when applied after cells were infected with WNV.Virusinduced cytotoxicity was inhibited by concentrations of >5.9 U/mL when added 1.5 hours after infection (Figure).The optical density 490 values in these tests were significantly different (p<0.05,using Tukey HSD multiple