Skip navigation

Evidence on zanamivir (Relenza)


The National Institute of Clinical Excellence (NICE) has issued guidance to the NHS concerning the use of zanamivir in the treatment of influenza. It has also posted on its Internet site ( ) a summary of the evidence used in making its decision ( ).


NICE conducted a widespread search, and examined studies that were randomised, compared zanamivir with placebo or current therapy in adults with influenza A or B infections. Excluded were studies addressing prophylaxis, which did not use the licensed dosing and formulation or which looked at experimentally induced influenza.

This left three studies, one of which was published in full [1]. These three studies had 813 patients given zanamivir and 775 given placebo. The high risk population (patients with chronic respiratory, cardiovascular or metabolic disorders, or who were immunocompromised or older than 65 years) numbered 217 patients. Another study [2] was not included, though the reason for its exclusion is unclear (at time of writing).


The primary outcome was time for symptoms to be alleviated. A secondary outcome was the number of patients with complications, most commonly pulmonary disorders including bronchitis, pneumonia and chest infections, though these are not described in any detail.


Overall, about 73% of patients had influenza infection confirmed by laboratory tests. The median reduction in the number of days of illness was one day (Table 1) for all patients. It was 1.5 days (95% CI 1 to 2 days) in patients with confirmed influenza infection and 2.5 days in the high risk group, though this was not statistically better than placebo. A similar reduction was seen in the additional trial [2] not considered by NICE (Table 1).

Table 1: Intention to treat analysis showing primary outcome of median days of illness with influenza with placebo and zanamivir, for three phase III studies examined by NICE, plus one additional study.

    Median illness days  
Trial Number of patients Placebo Zanamivir Reduction in illness days (95%CI)
3001 455 6.5 5.0 1.5 (0.5 to 2.3)
3002 777 6.0 5.5 0.5 (-0.5 to 1.0)
3003 356 7.5 5.0 2.5 (0.8 to 3.5)
Overall 1588 6.0 5.0 1 .0 (0.5 to 1.5)
Hayden et al, 1997 276 6.0 5.3 0.7 (0 to 1.4)
Intention to treat analysis

Rates of complications of influenza infection were lower with zanamivir than with placebo (Table 2), for all patients (intention to treat analysis) and for those with influenza infection and the high risk group. The reduction in high risk patients was not significantly different from placebo. The number of patients needed to be treated to prevent one complication of influenza infection was 14 (95% CI 9 to 33).

Table 2: Intention to treat analysis showing secondary outcome of complications of illness with influenza (mainly pulmonary disorders like bronchitis, pneumonia and chest infections) with placebo and zanamivir, for three phase III studies examined by NICE.

  Complications with    
Analysis Placebo (%) Zanamivir (%) Relative benefit (95%CI) NNT (95%CI)
Intention to treat 211/775 (27%) 163/813 (20%) 0.74 (0.62 to 0.89) 14 (9 to 33)
Influenza positive 152/558 (27%) 119/609 20%) 0.73 (0.59 to 0.90) 13 (8 to 35)
High risk 46/117 (39%) 26/99 (26%) 0.68 (0.46 to 1.02) 8 (4 to 150)


Access to the information upon which NICE based its judgements is terrific, and the NICE Internet site is one to be bookmarked for the future. Having a brief (only five printed pages without references) outline of the current evidence on a new product is just what the doctor ordered. It is what we want for every new product, especially when so much material we need is in press, and we have to feed off scraps of conference abstracts and data on file. The need for a systematic review is at launch, not years later. There is clearly a tension between companies which want information available early, and journals, which want the exclusivity of the original papers. But that's not our problem. If we all said that we want a systematic review at launch, and we want it NOW, then Bandolier guesses it might somehow be done.

It is hard, on the evidence presented, to argue with NICE's conclusions on Relenza. For a fast track appraisal done in a very short time, this is good stuff and NICE should be congratulated. Is there anything missing? There is an outline of the health economic arguments given, but not in enough depth to get to grips with. The most cogent health economic argument probably lies in reduction of complications of influenza in patients at high risk, but the small numbers and lack of statistical benefit makes this an impossible line of argument to follow. Some modelling to give us an idea of how much benefit would be needed to make a clinical and cost difference is an obvious next step for someone.

It is sad, perhaps, that this first NICE appraisal is negative about a product. Conflict between the NHS and healthcare industries is in the best interests of neither. There are lessons to be learned here, about outcomes for patients, professionals and the NHS and industry, and about the general benefits of cooperation.


  1. MIST study group. Randomised trial of efficacy and safety of inhaled zanamivir in treatment of influenza A and B virus infections. Lancet 1998 352: 1877-81.
  2. FG Hayden et al. Efficacy and safety of the neuraminidase inhibitor zanamivir in the treatment of influenzavirus infections. New England Journal of Medicine 1997 337: 874-80.
previous or next story in this issue