Efficient Human to Human Transmission of H5N1 in SE Asia
Recombinomics Commentary
February 4, 2005
>> There have been probable cases of human-to-human transmission before but this is the first in which the person infected - the mother - contracted severe illness and died. It proves the virus can be passed from person to person without losing its lethality. <<
The above comment, with regard to the case cluster in Thailand, published in the New England Journal of Medicine, is not true. The current H5N1 pandemic, which began in December, 2003 has generated at least 9 familial clusters of cases, resulting in 11 likely human to human transmissions, and 8 of the 11 have died (one was in critical condition Jan 22). Unfortunately, the efficiency of human to human transmission of fatal H5N1 influenza is much higher than transmission of H5N1 from birds to humans.
The misconception quoted above, comes from repeated comments from WHO that human to human transmission of H5N1 is very rare. These comments are supported by a flawed database. Collecting and testing of samples, especially for index cases of familial clusters, is poor. Of the 9 clusters, no sample was collected from the index case in 4 instances, including the cluster published in the New England Journal of Medicine. However, the clinical presentation is quite clear and in each cluster a relative has been laboratory confirmed, so there is little doubt that the fatal cases of children or young adults were due to H5N1. However, WHO excludes these cases, thereby eliminating the cluster. These clusters however, answer many questions about human to human transmission of H5N1 this season and last, in Vietnam, Thailand, and Cambodia.
The clusters have been described previously. Three are from Vietnam from the beginning of last season. Two are from last summer (one from Thailand and one from Vietnam). Four are from this season (1 from Cambodia and 3 from Vietnam). In all 9 clusters the index cases died with bird flu symptoms and in each cluster there is at least one confirmed H5N1 case. Remarkably, all 9 clusters have a bimodal distribution of disease onset dates, separated by a week or more. The 9 clusters are composed of 21 cases. Of the 21 cases 18 have died (1 was in critical condition Jan22) and 15 have been laboratory confirmed as H5N1 positive. Of the 12 patients who were not index cases, only 1 developed symptoms on or about the same time as the index case. The other 11 developed symptoms a week or more later, strongly suggesting that they are human to human transmissions.
This transmission is far more efficient than transmission from birds to people. As has been noted previously, such a process is extremely inefficient. Last season over 100 million birds were culled, but there were only 44 official cases in Vietnam and Thailand. Adding in the missed cases as well as those members of a positive cluster keeps the total far below 100, so there were over 1 million birds culled for every reported human H5N1 case. This season, the culling has been limited, but since 1.1 million birds have been culled and there have been only 11 cases not tied to human transmission, the ratio is still 100,000 birds culled for each reported case. Thus, the efficiency of human to human transmission, 11 cases from less than 100 human cases, was over 1000 fold more efficient than bird to human.
Since the virus that is transmitted bird to human is virtually identical to the virus transmitted human to human, the difference in efficiencies is due to interactions between the hosts. Most of the culled birds had limited interactions with people. Many were culled by a limited number of workers, and none of the workers were reported as being infected. In contrast, the infected humans were related to an index case and in general they were caregivers and had unprotected close contact with the infected family member.
The transmission chain was limited because the disease was so severe. Most of the index cases were initially not diagnosed with bird flu so much of the contact with other family members was unprotected. However, after the index case died, the suspicion index was raised significantly when the next family member developed symptoms, so the H5N1 was not transmitted further because family members used more care to avoid infection and / or brought the subsequent case to the hospital where they were put into isolation.
However, the virus was quite lethal. All 10 cluster members infected at the time of the index case died, but 8 of the 11 infected by a family member also died (1 was in critical condition Jan 22), so the failure to transmit further was limited by better infection control, not by significantly reduced lethality of the transmitted virus.
Thus, it would be useful to issue a warning that H5N1 can be transmitted to close family members quite easily, especially if unprotected care is given. This is quite clear from the cluster data, when analyzed properly.
Media link