To their surprise, Rhee and colleagues found that while DFA was a fairly accurate tool for diagnosing H1N1 in non-critical cases, it was not at all accurate for patients in the ICU. Just five of the 19 ICU patients (26 percent) had positive DFAs for H1N1 infection (four by nasopharyngeal swab, one by bronchoalveolar lavage), whereas 27 out of 33 non-ICU patients (82 percent) had a positive DFA test. The median time to first DFA was seven days in the ICU patients and three days in the non-ICU patients. Of the 31 respiratory tract samples in the ICU patients that were positive as determined by PCR, only 10 were concomitantly positive by DFA.
"For the non-ICU patients, the sensitivity of DFA was fairly good and correlated with previously published values. However, we found DFA to be significantly less sensitive in critically ill patients—those with severe respiratory distress requiring mechanical ventilation or a high degree of respiratory support in an ICU setting," said Rhee. "Interestingly, none of the DFA samples taken from the 18 endotracheal aspirates (secretions taken from the breathing tube on patients on a mechanical ventilator) were positive, despite the presence of virus detected by PCR or by bronchoalveolar lavage."
Rhee and colleagues were surprised by their findings, as they expected that severely ill patients would have a higher burden of viral disease, leading to easier detection. "We would have also expected that samples taken from endotracheal aspirates, where the secretions are coming from lower down the respiratory tract, would have a higher likelihood of being positive, but this was not the case," said Rhee.
One possible explanation for the poor performance of DFA in ICU patients is that it is an over-exuberant host inflammatory response, rather than high viral load, that is responsible for severe disease. However, it remains unclear why certain patients develop severe respiratory failure from H1N1 while others with similar risk factors develop only mild symptoms.
If confirmed by further research, these findings have important ramifications. "This study reinforces the fact that patients with suspected H1N1 influenza who are severely ill should be placed in respiratory isolation and receive antiviral treatment without delay, even if DFA testing is negative" said Rhee. "This includes patients with a negative DFA from lower respiratory tract samples. Furthermore, all critically ill patients with suspected H1N1 should have PCR testing done to confirm the diagnosis, as PCR is significantly more sensitive than DFA, though not perfect either."
"The next logical step would be analyzing data from a much larger pool of patients from different institutions to confirm these findings," said Rhee.