BACKGROUND: Interpreting rRT-PCR results for human avian influenza A virus (AIV) detection in contaminated settings like live bird markets (LBMs) without serology or viral culture poses a challenge. METHODS: During February-March 2012 and November 2012-February 2013, we screened workers at nine LBMs in Dhaka, Bangladesh to confirm molecular detections of AIV RNA in respiratory specimens with serology. We tested nasopharyngeal (NP) and throat swabs from workers with influenza-like-illness (ILI) and NP, throat, and arm swabs from asymptomatic workers for influenza virus by rRT-PCR and sera for seroconversion and antibodies against HPAI A(H5N1) and A(H9N2) viruses. RESULTS: Among 1,273 ILI cases, 34 (2.6%) had A(H5), 56 (4%) had A(H9), and 6 (0.4%) had both A(H5) and A(H9) detected by rRT-PCR. Of 192 asymptomatic workers, A(H5) was detected in 8 (4%) NP and 38 (20%) arm swabs. Of 28 ILI cases with A(H5) or A(H9) detected, none had evidence of seroconversion, but 1 (3.5%) and 12 (43%), were seropositive for A(H5) and A(H9), respectively. CONCLUSION: Detection of AIV RNA in respiratory specimens from symptomatic and asymptomatic LBM workers without evidence of seroconversion or virus isolation suggests environmental contamination, emphasizing caution in interpreting rRT-PCR results in high viral load settings.