How effective were neuraminidase inhibitors during the 2009 pandemic?Archived

Muthuri SG, Myles PJ, Venkatesan S, Leonardi-Bee J, Nguyen-Van-Tam J. Impact of neuraminidase inhibitor treatment on outcomes of public health importance during the 2009-10 influenza A (H1N1) pandemic: a systematic review and meta-analysis in hospitalised patients. J Infect Dis. (2012) doi: 10.1093/infdis/jis726

This systematic review and meta-analysis of observational data from an established group of specialists assessed treatment outcomes for all three neuraminidase inhibitors (NAIs); oseltamivir, zanamivir and peramivir given to hospitalised patients with influenza A(H1N1)pdm09 during the 2009 pandemic.

A total of 107 articles met pre-established criteria for inclusion, 53 focused on mortality; 59 explored severe outcomes, classified in this case as either critical care admission or death; while 14 focused on A(H1N1)pdm09 pneumonias.

Endpoint Mortality. Forty four studies were analysed that assessed the relationship between NAI treatment and mortality. When looking at intervention compared to no intervention, pooled results showed a decline in mortality risk but it did not quite reach statistical significance (OR 0.72; 95% CI, 0.51-1.01). The same was true for pre-admission NAI treatment versus no pre-admission treatment (OR 0.59, 95% CI 0.21-1.71). However for studies comparing early versus late intervention with NAI, as well as early versus no intervention there was a significant decrease in mortality, with OR 0.37; 95% CI, 0.27-0.53 and OR 0.35; 95% CI 0.18-0.71 respectively.

Endpoint all severe outcomes. For intervention versus no intervention, there was a significantly increased risk of severe outcomes associated with NAI treatment (OR 1.76; 95% CI 1.22-2.54). However pre-admission NAI treatment was associated with a significant decrease in severe outcomes (OR 0.51; 95% CI 0.29-0.89) compared to no pre-admission treatment. Another statistically significant decrease was noted in terms of prospect of developing severe outcome with early NAI administration (OR 0.41; 95% CI, 0.30-0.56) when compared with late administration.

Endpoint Pneumonia. There was a statistically significant increase in risk of developing pneumonia associated with NAI treatment (OR 2.29; 95% CI, 1.16-4.53). Early versus late treatment found a decrease in the risk of pneumonia (OR 0.35; 95% CI, 0.24-0.50) but there was no significant reduction in risk of pneumonia with early treatment compared to none.

ECDC Comment (16 December 2012):

Evidence from various randomised control trials have always found that NAI were able to prevent infection when given as prophylaxis against mild influenza and speed up relief of symptoms when given for treatment. (1) In contrast controversy has constantly arisen over whether NAIs are able to reduce the risk of more significant disease and premature death, especially among the higher risk groups (older people and those with chronic ill-health).

The conclusions from this review are in accordance with pre-pandemic observational studies on seasonal influenza and the best meta-analyses of trial data for oseltamivir, namely that there is a significant benefit of treatment, especially early treatment. (2-4) The meta-analysis was an independent reanalysis of an earlier controversial study sponsored by industry. (5)

At the same time the article demonstrates some of the difficulties of undertaking observational studies. It seems inconsistent that the risks of severe outcomes and pneumonia were raised in association with use of NAIs. One reason for that could be more severe cases are simply more likely to be treated. Hence the particular attraction of analyses comparing early versus late treatment.

Currently there is renewed interest in revisiting the old oseltamivir trial data (much of it from the 1990s) and making all and every trial data available. (6) This is defensible in itself. Though it should be recalled that repeated analyses to date have shown benefit and that should be the default position. (3,5) However, how relevant are these date for clinical and public health practise nearly two decades on Influenza viruses constantly change.(7) The influenza (H3N2) viruses circulating now have significant differences from what was around in the 1990s(8) and the old A(H1N1), which latterly developed oseltamivir resistance has been totally replaced by the pandemic A(H1N1). (7) It is difficult to envisage new randomised placebo-controlled trials happening among risks groups since finding like these which would presumably make them unethical,(2) though there would be value in stepped wedge designs where antivirals are being introduced.(9) Rather practice is going to have to be informed by careful observational studies and analyses like in this article.