Exclusion of Bird Flu Clusters and Transmissions in Vietnam
Recombinomics Commentary
March 20, 2005
>> Within an hour or so the pudding was ready and he, his mother, his father and his two sisters, aged 14 and 27, sat down to consume several bowls. Five days later, on 13 February, Sy Tuan began to feel unwell. He had a headache and a slight fever. Thinking he had common or garden flu, his family dosed him with aspirin and tucked him up in bed. But two days later, on 15 February, his fever shot up to 40C. On 20 February, he started coughing and complaining of chest pains. By the time Sy Tuan was transferred from the local Thai Binh hospital to Hanoi's Institute for Clinical Research into Tropical Diseases two days later, the damage had been done……..
To my surprise, Sy Tuan was not alone. In the bed next to him, watching with wide black eyes, was his 14-year-old sister. Two days after her brother had been rushed to hospital she'd also started running a high fever and had to be admitted. Like her brother she would also test positive for H5N1.
On 9 January, Hung's brother died, the disease having spread from his lungs to his kidneys and liver. The following day, back in Thai Binh for his brother's funeral, Hung also began to feel unwell. On his return to Hanoi, now running a high fever, he made an appointment at another hospital, where his wife worked as a nurse. There an X-ray showed a small white shadow on his lung, which was misdiagnosed as tuberculosis. His fever soaring, Hung decided to check himself into the Tropical Diseases Institute……
During Tet 2003, Dr Van treated a similar multiple infection involving two sisters, also from Thai Binh, whose brother had died of a mysterious pneumonia after slaughtering chickens for a wedding banquet. The brother's body was subsequently cremated, so Dr Van was unable to confirm he'd had H5N1, but 10 days later the sisters fell ill and were admitted to her ward with raging fevers.
she says she was shocked by how quickly the sisters went into decline. Seven days after she admitted the sisters, Dr Van was forced to intubate. Three days later they were both dead, expiring within an hour of one another on the eve of Tet. Tests confirmed they both had H5N1.
But the strongest evidence for human-to-human transmission so far has come from Thailand. Last September, an 11-year-old girl from a remote village, Kamphaeng Phet, fell ill with fever. A few days earlier all her pet chickens had died suddenly. Her aunt rushed her to hospital, where she was joined by the girl's mother, who lived near Bangkok. Within 16 hours the girl was dead and, in keeping with Thai custom, her body was cremated. Two weeks later her mother, who had had no exposure to the chickens, was admitted to hospital in Bangkok with the same symptoms. She also died. Then, days later, the aunt also developed a cough and fever. Tests showed that, like her sister, she had H5N1. Unlike her sister, however, she survived.
The WHO concluded both had almost certainly caught the virus while nursing the sick girl. 'The worrying thing about the Thai case,' Peter Horby told me in Hanoi, 'is the girl and her mother lived several miles apart, so they could not have been exposed to a common poultry source. The hospital bedside was the most likely contact.' <<
The four familial clusters described provide additional specifics on the bimodal distribution of disease onset dates and provide clear evidence for human-to-human transmission. Although a great deal of detail has been provided for these cases, there are now over a dozen familial clusters which have virtually identical bimodal distributions. The secondary patients cared for the index case, or had contact at hospital visits and/ or the funeral. The onset dates are 1-2 weeks after the index case, and are inconsistent with a common exposure by the index case and secondary case(s).
In the first case cited, even the onset date for the index case makes the blood pudding an unlikely source of infections. The 14-year-old sister developed symptoms after her brother's temperature spiked, which was two weeks after the meal. None of the other family members who had multiple helpings of the blood pudding developed symptoms.
Although the H5N1 testing on the clusters was not described in detail, the test results show why the clinical picture and disease onset are a much better predictor of H5N1 infection and transmission than the PCR test, which is used to exclude approximately 80% of the H5N1 cases in Vietnam. In the most recent Thai Binh case, the sister was reported to be negative for one of the early tests. In the Thai Binh cluster in January, 2005 the index case was negative twice and was only positive at autopsy. His brother was initially negative and was diagnosed as having tuberculosis. In the Thai Binh cluster in January 2004, both sisters were initially inconclusive and no sample was collected from the index case. In the cluster in Thailand the index case was misdiagnosed as having dengue hemorrhagic fever. Swabs were not collected from her or her mother and her aunt initially tested as negative.
Thus, for the ten H5N1 cases in the four clusters, only one tested positive initially. Swabs were not collected for three, two were initially inconclusive, and four were initially negative. These poor test results were not limited to the clusters. Testing in Tokyo of samples collected this season in Ho Chi Minh City indicated that 78% of the H5N1 infected patients were incorrectly classified as H5N1 negative patients based on PCR results. Details on 7 false negatives have been made public. Four were positive on retesting of the negative samples indicating that the test was not properly run. The resulting negative data of the initial test was used to declare these patients as negative for H5N1. The other three repeated as negatives in Ho Chi Minh City, indicating that even when run properly, the test lacks sensitivity on 3 of 7 cases.
Therefore, the cluster highlights the many false negative test results on patients who eventually test positive. However, it would appear that the individual cases are not test tested multiple times, and the false negative PCR data is used to exclude the patients from the official list of H5N1 cases. Moreover, even multiple tests on one of the highest profile cases, the second Thai Binh nurse with bird flu symptoms, failed to detect H5N1.
The above examples clearly demonstrate human-to-human transmission in Vietnam and Thailand, and highlight the gross deficiencies in the testing program which the WHO uses to exclude clear-cut H5N1 positive cases and clusters.
Media link