Nov 22, 2006 (CIDRAP News) – Two new reports on human cases of H5N1 avian influenza that occurred in Turkey and Indonesia last year show that the illness proved difficult to diagnose, with many tests yielding false-negative results.

A report on eight cases in Turkey and a similar report on eight Indonesian cases, published in the Nov 23 New England Journal of Medicine, says many of the patients tested negative the first time around, even on polymerase chain reaction (PCR) tests.

The Turkish report describes 8 of the 12 confirmed cases the country has had so far, all of which occurred in late December 2005 and January 2006. All eight patients—four of whom survived—were treated at a hospital in Van in eastern Turkey. The patients included sets of three and two siblings, and seven lived in the same community.

The hospital tested a total of 290 patients for avian flu during the outbreak, using a rapid influenza test, enzyme-linked immunosorbent assay (ELISA), and real-time PCR. All the rapid and ELISA test results were negative, the report says.

Four of the eight case-patients initially had negative results on all three tests, using nasopharyngeal specimens. Because the patients were severely ill, the tests were repeated using tracheal aspirate specimens, and the real-time PCR tests then were positive. Tests of nasopharyngeal specimens established the diagnosis in the other four patients.

"Before H5N1 infection was diagnosed in the eight patients, a total of 8 real-time PCR tests, 12 rapid influenza tests, and 12 ELISA tests were negative," the article states.

The Indonesian report covers three family clusters of H5N1 cases that occurred between June and October of 2005 and included the country´s first cases. The clusters consisted of three cases in Tangerang, Java; two cases in Bintaro, Java; and three in Lampung, Sumatra. Four of the eight patients died.

All rapid tests on the patients were negative, and many reverse-transcriptase PCR tests were negative, particularly with nasal specimens, according to the report. Throat swabs were more likely to test positive on RT-PCR than nasal swabs were.

The Turkish authors write, "Because of the difficulties in detecting H5N1 infection, repeated testing from nasopharyngeal swabs or deep tracheal-aspiration samples in patients who are strongly suspected of having H5N1 infection should be performed even if tests of initial nasophyaryngeal swabs are negative."

In other findings, the Indonesian report says that the source of infection for the first patient in two of the clusters was never identified. The three patients in the first cluster reported no contact with birds, other animals, or sick people other than family members before they fell ill. In the second cluster, patients reported no contact with birds, other animals, or sick people, but the index patient used fertilizer containing chicken droppings that tested positive for H5N1.

The report says limited person-to-person transmission could not be ruled out in the first two Indonesian clusters, since the patients had no other known exposures to the virus.

Three of the four Indonesian patients who recovered were children (aged 4, 5, and 9) who had mild disease. This resembles the pattern in Hong Kong´s 1997 H5N1 outbreak, in which most children who were infected had relatively mild disease, the article says.

The Indonesian authors write that the clusters in Indonesia and Turkey, as well as others in Hong Kong, Vietnam, Thailand, China, and Azerbaijan, "raise questions as to whether genetic or other factors may predispose some persons to H5N1 virus infection or to severe disease."

They add that more investigation is needed to understand "the role of mild cases in the epidemiology of this disease and whether genetic, behavioral, immunologic, and environmental factors may contribute to case clustering."