To the Editor—Unlike severe influenza A, which has characteristic findings, mild or moderate influenza A is clinically indistinguishable from influenza-like illness [1, 2]. Early in the swine influenza (H1N1) pandemic (April and May 2009), our emergency department was inundated with 856 patient requests for rapid influenza A testing and/or evaluation.
The H1N1 pandemic began at the end of the annual human influenza A season, and other respiratory viruses that cause influenza-like illness were present in the community as well. Although the rapid flu test (Quick-S Influ A/B kit; Denka Seiken) has ∼20% false negative results for influenza A, it is used in our emergency department because it is rapid and inexpensive [3, 4]. In April and May, because the local health department was overwhelmed with requests for H1N1 testing, reverse transcription polymerase chain reaction (RT-PCR) testing was performed if the rapid influenza A test was positive and/or the patient reported recent travel to Mexico or close contact with New York City epicenter cases.
Previously, relative lymphopenia has been found to be an early and reliable laboratory finding of adult influenza A [5-9]. Early in the pandemic, we reviewed complete blood counts in admitted emergency department patients with influenza-like illness for relative lymphopenia (lymphocytes ⩽21% of white blood cells) [5, 6]. It quickly became apparent that the demand for testing exceeded the emergency department's ability. We reasoned that, like adult human influenza A, relative lymphopenia might be a marker for H1N1 and thus could also be used to prioritize H1N1 PCR testing if the emergency department's ability was exceeded.
Of the 856 patients admitted to the emergency department with influenza-like illness for rapid influenza testing in April and May, 229 had positive results for influenza A and 13 had positive results for influenza B (Table 1). Because the rapid influenza test for influenza A may be associated with false negative results but not false positive results, we believe 25 adults with positive test results for influenza A with complete blood counts had H1N1.
Complete blood counts were obtained in all 25 adults and 16 children who tested positive for influenza A with probable H1N1. Relative lymphopenia (⩽21% of white blood cells) was present in 23 of these 25 adults (it was not evaluated in 2). These 23 adults had no other disorders associated with relative lymphopenia . Thrombocytopenia (⩽160 × 103platelets/mm3) was present in 7 of the 25 adults with probable H1N1 (range, [132-158] × 103platelets/mm3). Leukopenia (⩽3.9 × 103cells/mm3) was not present in any of the 25 adults (Table 2). Atypical lymphocytes were present in 3 of 16 children but no adults with influenza A and in no patients with influenza B (Table 3).
Like human seasonal influenza A, relative lymphopenia appears to be a laboratory marker of H1N1. In adults with positive test results for H1N1, relative lymphopenia with or without thrombocytopenia was common, but leukopenia was not present. Our first adult patient with severe human seasonal influenza A, admitted to the hospital and requiring mechanical ventilation, had profound and prolonged relative lymphopenia with thrombocytopenia but not leukopenia.
In children with influenza A, relative lymphopenia was uncommon (in 3 of 16 children). In adults who test positive for influenza A by the rapid influenza test, relative lymphopenia appears to be a marker to identify those likely to have H1N1 and thus to merit specific RT-PCR testing.
Potential conflicts of interest. All authors: no conflicts.