Comparison of RT-PCR With Other Diagnostic Assays for Rapid Detection of Influenza Viruses Fabrizio Pregliasco,* Carolina Mensi, Laura Camorali, and Giovanni Anselmi Istituto di Virologia, Universita ` degli Studi di Milano, Milan, Italy To compare the effectiveness of reverse tran- scription-polymerase chain reaction (RT-PCR), shell vial culture and cytospin assay as labora- tory techniques for rapid diagnosis of influenza infections, a retrospective study was carried out on 270 aliquots of oropharyngeal swabs col- lected from October 1993 to March 1996 and already characterized by standard isolation pro- cedures, and a prospective study in which 65 clinical samples taken from patients with influ- enza-like syndrome between October 1996 and March 1997 were tested. In the retrospective study, using conventional isolation as the gold standard, the sensitivity of RT-PCR and cytospin assay for virus A was 100% (95% confidence in- terval (CI), 89.1–100) and for virus B it was 100% (95% CI, 56.1–100) compared with 77.5% (95% CI, 61.1–88.6) and 71.4% (95% CI, 30.3–94.9) for shell vial culture. The specificity of all the three assays was 100% (95% CI, 98.0–100) for virus A and 100% (95% CI, 98.2–100) for virus B. In the prospective study the sensitivity of RT-PCR was greater than that of the other tests considered, both rapid and standard. It is suggested that RT- PCR should be employed in combination with conventional culture techniques in routine diag- nosis of influenza infections in order to obtain results more rapidly and to improve virus detec- tion even in circumstances in which standard isolation could be problematic. J. Med. Virol. 56: 168–173, 1998. © 1998 Wiley-Liss, Inc. KEY WORDS: influenza detection; rapid as- says; RT-PCR INTRODUCTION Influenza viruses are a major cause of acute respira- tory infection (ARI) and are responsible for high mor- bidity and mortality worldwide. Although in most pa- tients influenza infection results only in short absences from work or school, there are groups at high risk for which it may constitute a serious illness. During epi- demics it can cause 10,000–20,000 excess deaths in the elderly and in patients with chronic cardiovascular and pulmonary diseases [Pachucki, 1992]. The clinical symptoms of influenza are very similar to those asso- ciated with other respiratory viruses, often circulating in the community at the same time, so that laboratory confirmation of influenza infection is essential for sur- veying influenza outbreaks and for assessing the effi- cacy of vaccines and specific antiviral therapies. The conventional laboratory diagnosis of influenza is based on virus isolation, either in embryonated chicken eggs or in cell cultures, and subsequent identification of the virus by hemagglutination inhibition or hemad- sorption. The problems of these standard reference methods are the time needed (7–20 days for a definite diagnosis) and the difficulty of storing the pathologic specimen correctly in order to preserve viral infectivity. Hence, alternative, rapid diagnostic methods must be developed to enable prompt treatment and prophylaxis with specific antiviral drugs such as amantadine, rimantadine (for influenza A), and neuraminidase in- hibitors, since when given within 48 hours of onset of illness, these agents decrease the severity of clinical symptoms [Cherian et al., 1994; Scott and Huang, 1996]. The timely identification of influenza viruses in a community and of new emerging strains would also allow efficient epidemiological surveillance of influenza to be carried out. A variety of techniques involving a molecular biology have been suggested [Zhang and Evans, 1991; Claas et al., 1992; Donofrio et al., 1992; Atmar et al., 1996], or immunocapture ELISA [Chomel, 1989]. The aim of the present study was to optimize laboratory assays for rapid detection of influenza vi- ruses and to evaluate the appropriateness of introduc- ing reverse transcription-polymerase chain reaction (RT-PCR) for routine diagnosis of influenza infections. MATERIALS AND METHODS Specimens This study was divided into two parts: retrospective and prospective. Retrospective study. Aliquots of original oropha- ryngeal swabs (270) collected from October 1993 to *Correspondence to: Dr. Fabrizio Pregliasco, Istituto di Virolo- gia, Universita ` di Milano, via Pascal 38, 20133 Milano, Italy. E-mail: Fabrizio.Pregliasco@unimi.it Accepted 20 April 1998 Journal of Medical Virology 56:168–173 (1998) © 1998 WILEY-LISS, INC.