Treatment of SARS with human interferons.

Sir

We agree with J Cinatl and colleagues (July 26, p 293) that effective antiviral agents are urgently needed to treat severe acute respiratory syndrome (SARS).

On the basis of their results, the authors state that interferons inhibit replication of SARS-associated coronavirus (SARS-CoV) in vitro, with interferon beta being the most potent of those interferons tested. However, we are concerned that shortcomings in the methods used to calculate and interpret their results could have led to misleading conclusions.

First, when comparing the antiviral action of different preparations of interferons, the use of antiviral units of measurement, including international units (IU), might be inappropriate. Different preparations can have different specific activities—ie, IU/mg protein—as in the case of interferon beta (32×106 IU/mg protein) and interferon alfa (2·0–2·4×108 IU/mg protein). Therefore, for instance, the inhibitor concentration (EC50) value in Vero cells of interferon beta is not 62—ie, 6500/105 IU—times higher than the EC50 of interferon alfa, as stated, but only nine times—ie, 29·5/3·2 ng; this difference could be clinically relevant in pharmacokinetic and pharmacodynamic terms.

Second, Cinatl and colleagues calculated the selectivity index, a parameter of fundamental importance from a therapeutic viewpoint, without knowing the cytotoxic concentration (CC50) values of the interferons used. In their calculation, a value of more than 10 000 was assumed. However, when the therapeutic efficacy of different drugs is compared, this assumption might be incorrect: higher than 10 000 might mean 10 001 IU, for example, for interferon beta and 100 000 IU, for example, for interferon alfa. Although such wide variations in the values are highly unlikely, they would imply indirectly that interferon alfa, which has a lower antiviral activity, is more interesting from a therapeutic viewpoint than interferon beta because the selectivity index for interferon alfa is higher.

Finally, the antiviral action of interferons against a specific virus is usually, historically, measured by back titration of the viral yields when the interferon is added some 18–24 h before virus adsorption. The addition of interferon before and after virus infection does not allow a direct comparison of the sensitivity of SARSCoV with that of other animal viruses, including human coronaviruses.

Cinatl and colleagues have undoubted merit in having addressed promptly the issue of antiviral action of interferons against SARS-CoV. We consider, however, that their calculations could have been made and their general conclusion—that only interferon beta can be used as an antiviral agent after infection—might have been drawn with undue haste, which has led to errors.