<p><b>OBJECTIVE</b>To investigate the prevalence of influenza virus infections in children in 2006 using the real-time PCR method.</p><p><b>METHODS</b>(1) Consulting the most conserved sequence NP gene of influenza virus, after comparing with the NP gene sequences of influenza virus in GenBank, one pair of specific primers and one TaqMan probe were designed for each subtype of influenza virus by the software Primer Express. The sensitivity of influenza was evaluated by testing known positive samples which had been two-fold diluted. The specificity of real-time PCR for influenza virus detection was assessed by cross testing 60 isolates of influenza A, 16 isolates of influenza B, and by testing a variety of other respiratory viruses positive samples; (2) 281 nasopharyngeal aspirate samples were detected by real-time PCR and virus isolation; (3) the 12 301 specimens from the patients of Guangzhou Children's Hospital were tested by using the real-time PCR method. Furthermore, the real-time PCR reagent was evaluated by comparing with the result of virus isolation.</p><p><b>RESULTS</b>(1) The sensitivity of real-time PCR developed in this study for influenza A detection was 1:2(22) and for influenza B was 1:2(20) in two-fold serially diluted way. (2) No positive results were found in cross testing of other viruses positive specimens. (3) Influenza virus was detected from 1687 cases (13.71%) out of the 12 301 cases, including 773 cases (45.8%) positive for subtype A and 914 cases (54.2%) positive for subtype B; 455 out of 525 (86.7%) of influenza B positive specimens and 70 out of 525 (13.3%) of influenza A (H1N1) positive specimens were from patients seen during January to April; 419 out of 1118 (37.5%) specimens positive for influenza B and 699 out of 1118 (62.5%) specimens positive for influenza A (H1N1) were from patients seen from May to August. Influenza virus could be identified from 1380 samples by the methods of virus isolation, accounting for 81.80% of the 1687 positive samples detected by real-time PCR. All the influenza virus subtype A was H1N1.</p><p><b>CONCLUSION</b>The real-time PCR method developed in this study was sensitive and specific for detecting influenza A and B in clinical specimens. During 2006, influenza A and influenza B co-circulated. The predominant virus was influenza B from January to April, peaking in April. Influenza A (H1N1) prevailed from May to August, with the peak in June.</p>