Amantadine and rimantadine for preventing and treating influenza A in adults

Authors


Abstract

Background

Amantadine hydrochloride and rimantadine hydrochloride have antiviral properties, but they are not widely used due to a lack of knowledge of their potential value and concerns about possible adverse effects.

Objectives

The objective of this review was to assess the effects and safety of amantadine and rimantadine in healthy adults.

Search strategy

We searched the Cochrane Controlled Trials Register, MEDLINE, EMBASE and reference lists of articles. We also contacted manufacturers, researchers and authors.

Selection criteria

Randomised and quasi-randomised studies comparing amantadine and/or rimantadine with placebo, control antivirals or no intervention, or comparing doses or schedules of amantadine and/or rimantadine in healthy adults.

Data collection and analysis

For prevention trials the numbers of participants with clinically defined influenza, with serologically confirmed clinical influenza A and adverse effects were analysed. Analysis for treatment trials was of the mean duration of fever and adverse effects.

Main results

Amantadine prevented 23% of clinical influenza cases (95% confidence interval 11% to 34%), and 63% of serologically confirmed clinical influenza A cases (95% confidence interval 42% to 76%). Amantadine reduced duration of fever by one day (95% confidence interval 0.7 to 1.3). Rimantadine demonstrated comparable effectiveness, but there were fewer trials and the results for prevention were not statistically significant. Both amantadine and rimantadine induced significant gastrointestinal adverse effects. Adverse effects of the central nervous system and study withdrawals were significantly more common with amantadine than rimantadine.

Reviewers' conclusions

Amantadine and rimantadine have comparable effectiveness in the prevention and treatment of influenza A in healthy adults, although rimantadine induces fewer adverse effects than amantadine.

[This abstract has been prepared centrally.]

Plain language summary

Synopsis

The drugs amantadine and rimantadine can both help prevent and relieve influenza A in adults, but amantadine has more adverse effects

The flu (influenza) can be caused by many different viruses. One type is influenza A, with headaches, coughs and runny noses that can last for many days and lead to serious illnesses such as pneumonia. Amantadine and rimantadine are antiviral drugs. The review of trials found that both drugs are similarly helpful in preventing or treating influenza A in adults, but only when there is a high probability that the cause of the flu is influenza A (a known epidemic, for example). Both have adverse gastrointestinal (stomach and gut) effects, but amantadine can also have serious effects on the nervous system.

Background

Influenza is an acute respiratory illness caused by a virus of the Orthomyxoviridae family, of which three serotypes are known (A, B and C). Influenza causes an acute febrile illness with myalgia, headache and cough. Although the median duration of the acute illness is three days, cough and malaise can persist for weeks. Complications of influenza include otitis media, pneumonia, secondary bacterial pneumonia, exacerbations of chronic respiratory disease, croup and bronchiolitis in children. Additionally, influenza can cause a range of non-respiratory complications including febrile convulsions, Reye's syndrome and myocarditis (Wiselka 1994).

The influenza virus is composed of a protein envelope containing an RNA core. On the envelope are two antigens: neuraminidase (NA antigen) and haemagglutinin (HA antigen). Haemagglutinin facilitates the entry of the virus into cells of the respiratory epithelium, while neuraminidase facilitates the release of newly produced viral particles from infected cells. The influenza virus has a propensity for acquiring new HA and NA genes as well as to mutate with resulting change in its external antigenic composition to escape the hosts' immune defences. This leads to a classification of each viral subtype based on HA and NA typing (H and N for short). Additionally, strains are classified on the basis of serotype, geographical location of first isolation, serial number and year of isolation (e.g. A/Wuhan/359/95 [H3N2]).

In this century there have been four pandemics caused by different Influenza A viral subtypes:

Year Influenza A Subtype
1918-9 H1N1
1957 H2N2
1968 H3N2
1977 H1N1

Pandemics are caused by antigenic shift (a major change in H configuration and perhaps viral alteration of tissue tropism) leading to the appearance of a new subtype against which there is little natural immunity. Pandemic viruses are thought to originate in Southern China where ducks (the animal reservoir and breeding ground for new strains), pigs (which are thought to be the biological intermediate host or "mixing vessel") and humans live in very close proximity (Bonn 1997). One possible explanation of this phenomenon is that waterfowl influenza viruses are transmitted to pigs either directly or through geese. Pigs are considered plausible intermediate hosts as their respiratory epithelium cells have receptors for both avian (i.e. duck) and human viral haemagglutinins (Bonn 1997). Minor changes in viral antigenic configurations, known as "drift", cause local or more circumscribed epidemics.

The recently isolated Hong Kong avian influenza (A/HK/156/97[H5N1] virus appears to be an example of a zoonotic infection with direct spread of the avian virus to humans. The virus, first described in a Hong Kong child who died in May 1997, appears to be highly virulent in domestic duck and fowl but has low transmission potential to humans (Belshe 1998). This virus caused severe primary pneumonia and appeared highly pathogenic in infected humans (Yuen 1998). This virus has one alarming new characteristic: it appears to have crossed the avian-human barrier without prior adaptation in an intermediate host (Claas 1998).
Adaptation of avian influenza viruses to humans is probably the explanation for the high virulence of the 1918-1919 pandemic and implies that in the era of global travel new influenza viruses could cause epidemics and pandemics with little warning. Influenza pandemics are caused by viruses which have the ability to replicate and spread rapidly in humans and against which there are no antibodies in the human population (Claas 1998).

Pandemics by definition cause a very high morbidity and mortality burden (Jefferson 1998A). The 1918-19 pandemic is considered to have caused an estimated 40 million deaths world-wide.

Over the years efforts to prevent or treat influenza have had their mainstay in two separate approaches: vaccines and antivirals. The effects of influenza vaccines are the topic of a separate review. The ion channel blocking antiviral compounds have been amantadine hydrochloride and rimantadine hydrochloride (amantadine and rimantadine for short). Amantadine was introduced in the 60's as an antiviral drug and was incidentally found to have anti-Parkinsonian activity. In the USA, amantadine was licensed for the treatment and prophylaxis of influenza A infections by the FDA in 1976 and rimantadine in 1993 and physicians are invited to "consider" its use (Advisory Committee on Immunisation Practices, 1994). Recommended dose of each in the USA is 100 mg twice daily. In the USA, while amantadine is licensed for treatment and prophylaxis of adults and children aged more than one, rimantadine is licensed only for prophylaxis in children as well as for treatment and prophylaxis in adults (Advisory Committee on Immunisation Practices, 1994). In the UK amantadine only is licensed and is administered orally at a recommended does of 100 mg a day in healthy adults for five days (treatment role) or 100 mg a day as long as the risk of infection lasts (prophylaxis role). In the UK the cost of fifty-six (100 mg) tablets is £10.41. Both compounds interfere with the replication cycle of type A (but not type B) viruses (Loughlin 1993) by affecting ion channel activity through the cell membrane. They are thought to be efficacious and, given their virus-specific action, relatively free of adverse effect. Drug resistant H3N2 subtype influenza A viruses have been isolated during treatment with amantadine and rimantadine, especially in institutions, but their clinical significance is unclear (Hayden 1993).

Given both drugs' apparent efficacy in both prophylactic and therapeutic roles (if administration is started in time), their relatively scarce use is surprising (Wiselka 1994). Explanations for this finding include lack of awareness of the drugs and their properties by medical practitioners and concern over their adverse effects, which include seizures. Even more surprising is the list of indications for use of both drugs. While subjects at high risk (i.e. subjects with underlying debilitating pathologies and the elderly) are included, healthy adults, especially those working in institutions (such as healthcare workers, nursing home attendants and the military) are not. These groups are likely to greatly benefit from the use of the drugs, which could have a direct impact of length of sickness absence and diminish considerably the burden of influenza epidemics to society. Additionally a non-systematic review of the evidence of the efficacy of Rimantadine identified five small double blind placebo controlled trials of both drugs in a prophylaxis role and nine in a treatment role (Wintermeyer 1995). The largest study contained 378 individuals, indicating the need to attempt pooling data to derive more precise estimates of effect and safety for the compounds. We aim to carry out a systematic review of the effects of amantadine and rimantadine in healthy adults. As the viewpoint is that of our funder, the Ministry of Defence of the United Kingdom, children, elderly and individuals with pre-existing pathologies are excluded from our review. However, given the impact of influenza in these populations, systematic reviews of the effects of amantadine and rimantadine in children, elderly and at-risk groups are also being carried out at present.

Objectives

To identify, retrieve and assess all studies evaluating the effects of amantadine and rimantadine on influenza A in healthy adults.

To assess the effectiveness of amantadine and/or rimantadine in preventing cases of influenza A (prophylaxis) in healthy adults.

To assess the effectiveness of amantadine and/or rimantadine in shortening or reducing the severity of influenza A in healthy adults (treatment).

To estimate the frequency of adverse effects associated with amantadine and/or rimantadine administration in healthy adults.

The following hypotheses will be tested:

In comparisons between groups intended for amantadine and/or rimantadine prophylaxis or treatment compared with control/placebo groups:

There is no difference in the number of cases of influenza A and their severity (expressed as duration of symptoms, return to work, return to full activities, numbers of physician visits, numbers of complications, number of admissions to hospital and other direct or indirect indicators of severity).

There is no difference in the number and severity of adverse effects (both systemic and localised).

Criteria for considering studies for this review

Types of studies

Any randomised or quasi-randomised studies* comparing amantadine and/or rimantadine in humans with placebo, control antivirals or no intervention or comparing doses or schedules of amantadine and/or rimantadine. Only studies assessing protection or treatment from exposure to naturally occurring influenza were considered initially.

(* A study is randomised when it appears that the individuals (or other experimental units) followed in the study were definitely or possibly assigned prospectively to one of two (or more) alternative forms of health care using random allocation. A study is quasi-randomised when it appears that the individuals (or other experimental units) followed in the study were definitely or possibly assigned prospectively to one of two (or more) alternative forms of health care using some quasi-random method of allocation (such as alternation, date of birth or case record number)).

Types of participants

Apparently healthy individuals (with no known pre-existing chronic pathology known to aggravate the course of influenza) aged 14 to 60.

Types of intervention

Amantadine and/or rimantadine as prophylaxis and/or treatment for influenza, irrespective of target viral antigenic configuration.

Types of outcome measures

Clinical:

Numbers and/or severity (however defined) of influenza cases and/or deaths occurring in amantadine and/or rimantadine and placebo or control groups.

Adverse effects:

Number and seriousness of adverse effects, including cases of malaise, nausea, fever, arthralgias, rash, headache and more generalised and serious signs.

Search strategy for identification of studies

See: Unavailable search strategy

An electronic search of MEDLINE was carried out using the extended search strategy of the Cochrane Acute Respiratory Infections (ARI) Group (ARI Group 1998) with the following search terms or combined sets from 1966 to the end of 1997 in any language:
Influenza
Route (oral)
Route (parenteral)
Amantadine
Rimantadine

We read the bibliography of retrieved articles and of reviews of the topic in order to identify further trials. We also carried out a search of the Cochrane Controlled Trials Register (CCTR) and of EMBASE (1985 to 1997). In order to locate unpublished trials we wrote to the following:
- manufacturers;
- researchers active in the field;
- first or corresponding authors of studies evaluated (but not necessarily included) in the review.

The Cochrane Acute Respiratory Infections Group's trials register was searched in March 2001, and Issue 2, 2001 of The Cochrane Library was also searched for new trials.

Methods of the review

Inclusion procedure:

Two authors (VD and TOJ) read all trials retrieved in the search and applied inclusion criteria.

Trial quality assessment:

Trials fulfilling the review inclusion criteria were assessed for quality and results were analysed by VD and TOJ. Assessment of trial quality was made according to the following criteria:
1. generation of allocation schedule (defined as the methods of generation of the sequence which ensures random allocation).
2. measure(s) taken to conceal treatment allocation (defined as methods to prevent selection bias, i.e. to ensure that all participants have the same chance of being assigned to one of the arms of the trial. It protects the allocation sequence before and during allocation).
3. number of drop-outs of allocated healthcare worker participants from the analysis of the trial (defined as the exclusion of any participants for whatever reason deviation from protocol, loss to follow-up, withdrawal, discovery of ineligibility while the unbiased approach analyses all randomised participants in the originally assigned groups, regardless of compliance with protocol - known as intention to treat analysis).
4. measures taken to implement double blinding (a double blind study is one in which observer(s) and/or subjects are kept ignorant of the group to which the subjects are assigned, as in an experiment, or of the population from which the subjects come, as in a non-experimental. When both the observer and participants are kept ignorant of assignment the trial is called double-blind. Unlike allocation concealment, double blinding seeks to prevent ascertainment bias and protects the sequence after allocation).

For criteria 2, 3 and 4 there is empirical evidence that low quality in their implementation is associated with exaggerated trial results (Schulz 1995) and it is reasonable to infer a quality link between all four items. The four criteria were assessed by answering the following questions:

Generation of allocation schedule
Did the author(s) use:
a) random number tables
b) computer random-number generator
c) coin tossing
d) shuffling of allocation cards
e) any other method which appeared random

Concealment of treatment allocation
Which of the following was carried out:
a) there was some form of centralised randomization scheme where details of an enrolled participant were passed to a trial office or a pharmacy to receive the treatment group allocation.
b) treatment allocation was assigned by means of an on-site computer using a locked file which could be accessed only after inputting the details of the participant.
c) there were numbered or coded identical looking compounds which were administered sequentially to enrolled participants;
d) there were opaque envelopes which had been sealed and serially numbered utilised to assign participants to intervention(s)
e) a mixture of the above approaches including innovative schemes, provided the method appears impervious to allocation bias.
f) allocation by alternation or date of birth or case record or day of the week or presenting order or enrolment order.

[Concealment methods were described as "adequate" for (a), (b), (c), (d) or (e). Method (f) was regarded as "inadequate",as were trials using an system of random numbers or assignments. For some trials allocation was regarded as "unclear" if only terms such as "lists" or "tables" or "sealed envelopes" or "randomly assigned" were mentioned in the text.]

Exclusion of allocated participants from the analysis of the trial
a) Did the report mention explicitly the exclusion of allocated
participants from the analysis of trial results?
b) If so did the report mention the reason(s) for exclusion? (if yes, specify)

Measures to implement double blinding
a) Did the report mention explicitly measures to implement and protect double blinding?
b) Did the author(s) report on the physical aspect of amantadine/rimantadine administration - (i.e. appearances, colour, route administration)

Arbitration procedure
Where there was disagreement among TOJ and VD on the quality of a trial, JJD arbitrated.

Data collection

The following data was extracted, checked and recorded:
Characteristics of participants:
- Number of participants
- Age, gender, ethnic group, risk category
Characteristics of interventions:
- Type of antiviral, type of placebo, dose, treatment or prophylaxis schedule, length of follow-up (in days).
Characteristics of outcome measures
- Number and severity of influenza cases and deaths in amantadine/rimantadine and placebo groups
- Adverse effects. Four categories were used:
GI symptoms (nausea, vomiting, dyspepsia, diarrhoea and constipation) .
Increased CNS activity (insomnia, restlessness, light-headedness, nervousness and concentration problems).
Decreased CNS activity (malaise, depression, fatigue, vertigo and feeling drunk).
Dermatological changes (urticaria and rash).
[Adverse effect data was collected as the number of participants experiencing each (or any) adverse effect].
Number of withdrawals due to adverse effects.
Date of trial;
Location of trial;
Sponsor of trial (specified, known or unknown);
Publication status;

Data synthesis

The relative risks of events (cases of influenza, deaths, and adverse effects) comparing treatment and placebo/control groups from the individual trials were combined using Mantel-Haenszel (fixed effect) meta-analytical techniques. We did not combine estimates from treatment and prophylactic trials, as these were conducted to answer different study questions. Between-trial variability in results was examined and incorporated into the estimates of uncertainty of treatment effect using DerSimmonian and Laird random effects models where significant (P<0.05) heterogeneity was noted. In treatment trials the choice of methods for combining the estimates of severity of influenza depended on the format in which the data was presented. Where possible, comparisons were made between the mean duration of symptoms in the two groups, and methods for combining differences in means were used. Specifically, where the data were presented as the number of subjects with duration of symptoms beyond a cut-off time period these were presented as "Cases with fever at 48 hours". The bewildering array of outcomes used in the treatment trials (see Results section) prevented us from using more than the "cases with fever" outcome. Readers should also note that included trials did not contain sufficient information to enable us to assess the number of cases with no documented fever at entry into the trial.

Description of studies

PREVENTIVE TRIALS

We identified 18 reports of 19 preventive and safety trials fulfilling our inclusion criteria. We were unable to identify any unpublished trials, despite receiving 9 letters and 3 electronic communications from manufacturers, authors and researchers.

The mean amantadine arm size was 302 individuals (median 140, 25th percentile 75.5, 75th percentile 305), the mean rimantadine arm size was 87 individuals (median 102, 25th percentile 63, 75th percentile 114) and the mean placebo arm size was 285 individuals (median 109, 25th percentile 82.5, 75th percentile 234). The mean total population was 546 individuals (median 297, 25th percentile 199, 75th percentile 506). Mean length of follow up was 30 days (median 30 days, 25th percentile 16.5 days, 75th percentile 42 days). The duration of the epidemic was specified by only one trial (Kantor) and was of 49 days.

Identified trials are listed below using the name of the first author:
Brady
Callmander
Dolin
Hayden 1981
Kantor
Millet
Monto
Muldoon
Oker-Blom
Payler
Peckinpaugh/1
Peckinpaugh/2
Pettersson
Plesnik
Quarles
Reuman
Schapira
Smorodintsev
Wendel

Brady reports the conduct and results of a preventive randomised double-blind controlled trial of rimantadine during an epidemic of influenza A/Leningrad/87 [H3N2] virus. The trial took place between 8 February and 14 April 1988 in Baltimore, Maryland and Columbus, Ohio, USA. Two hundred and twenty-eight healthy not previously vaccinated adult volunteers aged 18 to 55 were randomised to receive either rimantadine 100 milligrams daily or placebo for six weeks. Paired sera were taken from all participants at the beginning and the end of the study. Within-trial surveillance was carried out on a weekly basis and cases were defined on the basis of seroconversion and a pre-defined list of symptoms and signs. Viral isolation took place by nasal washout. There were four dropouts in the rimantadine arm and two from the placebo arm, although the reasons for these are not given, the data for these participants are included in the assessment of adverse effects, but not of efficacy. This is the reason why different numbers of participants in the rimantadine and placebo arms are reported for adverse effects (n=114 in both arms in Table 2) and for efficacy (n=112 and 110 respectively in Table 3). However, as the discrepancy in dropouts and denominators between text and tables cannot be explained, we have included table denominator data in our meta-analysis. We doubt that the small discrepancy will make any difference to our findings. As populations from the two sites were homogeneous, the authors present pooled data for both clinical efficacy (Table 3) and adverse effects (Table 2). We included in our meta-analysis both data on adverse effects " probably related" (defined by the authors as "symptoms related to administration of the study drug for which another aetiology was more likely") and "possibly related" to rimantadine and placebo administration in order to present conservative estimates of toxicity. Efficacy data is presented both for clinically defined cases and laboratory confirmed cases and we have included both in our meta-analysis, although readers should take care not to attempt to add effect estimates for these two sub-groups, lest double-counting introduces an overestimated estimate of effect. Table 3 reports that out of 112 rimantadine arm participants seven had serologically confirmed influenza A, of which only 1 had symptoms, as opposite to 20 and 7 in the placebo arm. The corresponding figures for influenza-like illness (clinical case definition only) are 19 and 21 respectively. Table 2 presents both aggregate ad symptom-specific data for adverse effects. We have included both in our review. The authors conclude that low dose rimantadine is safe and effective in preventing influenza A in healthy adults. Brady is a clearly written and well-reported trial (with the exception of the minor discrepancy between text and tables on the affiliation of drop-outs). Randomisation was computer generated and allocation was concealed with a centralised scheme. Additionally, intention to treat analysis is clearly stated in the text.

Callmander is a report of a randomised controlled trial conducted in Sweden in February to March 1965 and published three years later. The setting is that of a community, including some military personnel, and the age range of the 96 volunteer participants is 20 to 60 years (46 male and 50 female). During the period of the trial there was considerable influenza A2 (Leningrad) activity. The design was a 30-day follow-up randomised double-blind and placebo-controlled preventive trial, with the intervention arm receiving 100 mg of amantadine hydrochloride twice daily and the control arm a no better described placebo. The practices of randomisation, allocation and concealment are no further defined, making it impossible to assess methodological rigour although as the distribution of sex and age was checked and found to be similar, randomisation is likely to have been satisfactory. Ninety-four of the ninety-six participants completed the trial (forty-seven in each arm). The reasons for the two drop-outs are given as dizziness (probably attributable to the respiratory ailment) and admission for a hernia operation, although the text seems to imply that this is caused by a lack of compliance with the schedule. Outcomes are expressed in numbers of clinically defined cases (from a list of symptoms) in each arm and a symptom score (reported in Table 1 without an indication of time of intensity). Additionally surveillance for adverse effects (systemic) was carried out. A list of symptoms (without a denominator) is reported at Table 2. Overall thirty-one participants in the placebo arm became ill, compared with thirty-three in the intervention arm. Although the symptoms were classed as "respiratory" an in-trial serological survey of seven out of seven individuals from the amantadine arm showed rises in anti-virus A rises compared to five out of seven samples in the placebo arm. All of these 14 participants became ill during the trial. The authors conclude that although amantadine reduced hoarseness and fever in sick persons, it did not reduce the incidence of respiratory ailments in the study population.

Dolin reports the results of a preventive randomised double-blind placebo controlled trial carried out in Burlington Vermont, USA. The trial was commenced on 10 January 1981 during an outbreak of influenza A/Bangkok/1/79H3N2 and A/Brazil/11/78H1N1 detected by surveillance (shown in Figure 1). Participants were healthy non-vaccinated volunteers aged 18 to 45 (mean age 25.6 + 0.45 years). Four hundred and fifty participants were randomly assigned to receive either amantadine 200 milligrams daily or rimantadine 200 milligrams or placebo. Participants were followed up weekly for six weeks and a sample of 38 participants (21 placebo, eight from the rimantadine arm and nine from the amantadine arm) had their urine tested for amantadine as an indication of compliance. Eight subjects out of the nine (89%) had detectable urine amantadine levels. Assays for rimantadine were not carried out for unspecified reasons. Ten volunteers dropped out of the trial prior to assuming any medication, the remaining 450 were randomised as follows: 148 (placebo), 147 (rimantadine), 145 (amantadine). Case definition was based on a list of symptoms plus virus isolation or a rise in serum antibody titres to influenza A. Table 1 presents both clinically-defined cases and cases defined on the basis of laboratory confirmation. Both isolates and antibody titres demonstrated that the majority of infections were caused by A/Brazil/11/78H1N1-like viruses. Additionally, data for cases of influenza A among the study population (irrespective of illness) are also presented in Table 1. Sixty-two participants dropped out from the study, 47 because of adverse effects (significantly higher in the amantadine arm) and 15 for causes unrelated to adverse effects (Table 2). Adverse effects included insomnia, difficulty in concentrating and jitteriness. We have included the adverse effect data from dropouts in our review, as the text of the trial specifies that, less serious, other adverse effects had not been significantly different in the arms of the study (page 582). The final total of participants was 378 (132 in the placebo arm, 133 in the rimantadine arm and 113 in the amantadine arm). The authors conclude that both antivirals are efficacious and safe, with amantadine reducing the rate of influenza-like illness by 78% and rimantadine by 65% and by 91% and 85% when laboratory documented cases are taken into account. Although a well-written report, no real information is given on random allocation, blinding and concealment. Intention to treat analysis was not carried out.

Hayden 1981 is a toxicity study reporting a randomised control trial undertaken in an unspecified period in USA and published in 1981. The setting is that of a state farm insurance company and the participants were 251 adult volunteers, aged between 18-65 (mean age of 32). Influenza A virus was isolated. Two trials were carried out simultaneously, both involving Rimantadine and Amantadine. One was a low dose (200mg daily of each drug, n=52) trial and the other a higher dose (300mg daily of each drug, n=199). The low dose trial however has been excluded due to the absence of any "cases" data, and the lack of outcomes. The only data included were "ability tests". The high dose trial has been included. This had a four and a half day follow-up and was a double-blind, placebo controlled ("inert placebo") preventive trial. There were 3 dropouts (one person from each group) which were stated as being due to unacceptable adverse effects (not specified). Due to this being a toxicity study, outcomes were very patchy and no cases were recorded. Adverse effects (systemic only with no other classification) were noted, although not specified. Overall 41 out of 67 (61%) in the Amantadine arm, 13 out of 66 in the placebo arm (20%), and 18 out of 63 in the Rimantadine arm (29%) experienced adverse effects. The authors concluded that rimantadine was better tolerated than amantadine in healthy young workers with significantly more CNS effects than placebo recipients. The practices of randomisation, allocation and concealment are no further defined, although all doses were stated as being administered by a project nurse. This is a poorly reported trial as no detailed classification of adverse effects is given, which is a strange practice for a toxicity study. Additionally, data reported in the text are not consistent with that in Table 1.

Kantor is the report of a preventive double-blind randomised controlled trial of the efficacy and safety of oral amantadine compared to a (not further defined) placebo. The trial took place over the period 20 February - 7 March 1978 in the military barracks at Fort Sam Houston (FSH), Texas and the target serotype was A/USSR/77. The results were published over two years later. Trial participants were 139 healthy paramedic recruits (mean age 22 years) who were randomised to receive either amantadine 100 milligram tablets twice daily (n = 64) or placebo (n=62). The text does not explain the discrepancy in total denominators compared to denominators in each arm of the trial. Follow-up lasted three weeks and participants were supervised through unit parades and weekly visits at which symptoms recorded in the questionnaires were discussed. Case definition consisted of a list of recognised symptoms and serological confirmation and adverse effects were recorded in the questionnaires. Paired sera were obtained at the beginning of the study and five weeks later. Cases were also presented by severity. Results are presented (Table 2) by duration of preventive treatment (less than 1 week, 1 week, two weeks). 64 participants took the medication or placebo for two weeks, 29 for one week and 46 for less than a week. Compliance with the study protocol was poor, as seventy-three participants admitted discontinuing the tablets before the end of the two week period (but the numbers do not coincide with the totals in table 2 (n=48) of participants taking medication or placebo for less than two weeks). The reasons for lack of compliance were given as fear of adverse effects, loss of interest in the study and forgetting to take the tablets. This is probably a reflection of the high-pressure environment at FSH. Additionally, data for 29 participants are not presented. An additional discrepancy appears comparing the numbers of influenza cases at Table 2 with the seroconversion-confirmed cases (Table 4), although it is possible that some seroconvertors were not clinically ill. Overall we decided to include in our review only illness data from Table 2 (regardless of severity) which appears more conservative (assuming that seroconversion data at table 4 refers to the same cases as those in table 4 and that the data for participants by period of drug taking are mutually exclusive). Insufficiently detailed adverse effect data are presented (but cognitive impairment scores are in Table 5). The text indicates that seven individuals had difficulty in concentrating, two had insomnia and one impotence. The authors conclude that amantadine is a safe drug, but because of relatively little influenza activity no conclusions can be drawn as to its effectiveness (although it is possible that some seroconverting recruits had contracted influenza prior to joining FSH, thus diluting the attack rate). Although the trial was randomised and allocation concealment was protected through a system of sealed envelopes, the difficulty in reconciling figures and understanding what actually happened during the trial makes this a very poorly reported study. No reason for the loss of 20 participants to follow up is given.

Millet reports the conduct of a randomised double blind safety trial conducted between October 1979 and January 1980 on 52 healthy adult volunteers recruited from the University of California at Los Angeles. The study assessed the safety profile of amantadine alone or in combination with chlorpheniramine (an anti-histaminic compound), rimantadine alone or chlorpheniramine. Volunteers were randomly assigned to taking either amantadine 100 mg and chlorpheniramine placebo (10 volunteers), amantadine 100 mags and chlorpheniramine 4 mags (11 volunteers), chlorpheniramine 4 mags and antiviral placebo (11 volunteers), rimantadine 100 mags and chlorpheniramine placebo (10 volunteers) or antiviral placebo and chlorpheniramine placebo (10 volunteers) twice daily for 3 to 4 days. Outcomes assessed included subjective side effects (from a predefined list) with a grading of 1 (mild) to 3 (severe) and performance testing. Moderate to severe subjective side effects are presented in Table 2 (decreased concentration, dizziness, headache, fatigue, drowsiness, dry mouth) with frequencies and 4-day aggregate severity scores. As no breakdown into severity is possible we have included all side-effect data for amantadine and rimantadine and their respective placebo groups. As the text does not present separate data for the placebo arm, we have included no subjective side effect data for placebo recipients in order to present the most conservative estimate of safety of the antivirals. The study was randomised and certainly double-blind, with centralised preparation of active and placebo tablets. Antiviral tablets and their placebos were salmon coloured whereas chlorpheniramine tablets and their placebos were white coloured. We have excluded data on the safety of chlorpheniramine either with amantadine or on its own and performance testing outcome data. The authors conclude that side effects were comparable and minimal but highest in the combined amantadine-chlorpheniramine arm. The conclusions can be partly explained by the relatively low daily amantadine dose (200 mags) and short administration period.

Monto reports the results of preventive randomised double-blind placebo controlled trial carried out among healthy student volunteers at the University of Michigan, USA. The trial began in February 1978, during an outbreak of influenza A/USSR-like (H1N1) and lasted seven weeks (the when the outbreak was over as confirmed by clinical and serological surveillance). 286 students aged 18 to 24 were randomly assigned to receive either amantadine 100 milligrams twice daily or placebo. Participants were followed up weekly for seven weeks, the remaining duration of the outbreak of influenza. Respiratory and adverse effect symptoms were recorded in weekly calendar-diaries and paired sera samples were taken at the beginning and end of the trial (including 5 dropouts from the trial because of clinical influenza or major adverse effects) from 281 of the participants. Six participants who dropped out in the first week of the study were not included in the analysis because of insufficient time to observe illness. Viral isolation from secretions was attempted whenever participants presented with at least two or more respiratory symptoms. Results are presented in the text at 2, 4 and 6 weeks for both clinically defined (Table 1) and serologically and clinically defined (in the text) cases. Analysis of the results was only carried out at weeks two and four as at week six, when the outbreak was reaching its natural conclusion, there was insufficient time for an antibody titre to rise. Serologically confirmed cases were 8/136 in the amantadine arm and 28/139 in the placebo arm. We decided to include in our review data from both sets of case-definitions at four weeks of follow-up (for which the report provides both sets of data) to test specificity of clinical influenza case definition. 15 participants from the study dropped out of the study because of adverse effects (twelve from the amantadine and three from the placebo arm). Adverse effects included: difficulty in concentrating, mental depression, insomnia and dizziness in both arms. Although in the text it is not clear what relationship these data have with that at Table 2, we have included data from Table 2 in our review. The trial is well written and appears well conducted with only minor discrepancies in total numbers. The practice of random allocation and concealment appear adequate. Analysis was not carried on an intention to treat basis.

Muldoon is a report of a crossover preventive controlled trial conducted in a school near Chicago (USA) between 2 December and 20 December 1968 (period 1) and published eight years later. The age range of the 105 unvaccinated volunteer student participants was around 21. At the beginning and the end of period 1 blood samples were taken from participants. During the eighteen days of period 1 the 53 individuals in the intervention arm received 100 mg of amantadine hydrochloride twice daily and the control arm no treatment. After a "holiday" period of 16 days (period 2) the participants were crossed over to assess whether withdrawal of amantadine rendered participants more or less susceptible to infection. The procedure was inverted and students were followed up for a further 10 days (period 3). Outcomes are expressed in numbers of clinically defined cases from a symptom score coupled with virus-specific antibody rise. This allowed the authors to distinguish non-influenza related illness during periods 1 and 2 (reported in Table 1 without an indication of intensity of symptoms). During period 1 (Table 2) 1 student had influenza with serological confirmation and 5 had antibody rise but no illness out of the 53 in the intervention arm and 8 (10 with serological confirmation) out of the 52 in the placebo arm. We were uncertain as the relationship of the data presented in Tables 1 and 2 and decide to use only the data from Table 2, as a conservative estimate. Throughout the period of the trial there was considerable influenza A/Hong Kong/1/68 (H3N2) activity (throughout the 2 months of the trial the average seroconversion rate was 8%). Adverse effects are reported as "insignificant". Only data from the first period of the cross over trial is included in our review: data from subsequent periods are excluded due to carry- over effects. The authors conclude that oral amantadine significantly prevents influenza and its withdrawal leads to an increased incidence of the illness. The practices of randomisation, allocation and concealment are not defined, making it impossible to assess methodological rigour, although the similar numbers of participants in each arm led us to consider this a randomised trial. No drop outs are reported but lack of clarity as to the meaning of the case definitions made this trial a very difficult read.

Oker-Blom reports the results of preventive randomised double-blind placebo controlled trial begun on 20 January 1969 in Helsinki, Finland and lasting for 30 days. Participants were 391 healthy medical students (132 females and 259 males) aged on average 22 years. Influenza A2 Hong Kong was the main circulating strain at the time. Participants were randomised to receive either amantadine 100 mg or placebo (tablets containing calcium lactate) twice daily. Additionally the authors analysed age-stratified results for each arm (although the results are not included in the text). Cases were defined on the basis of self-reported (by questionnaire) clinical symptoms and serological titres. Adverse effects were recorded too. Compliance was reasonable with 141 and 152 questionnaires returned in the active and placebo arms respectively. We have included in the review data for those cases which had both serological and clinical evidence of influenza (16/141 and 41/152 for the active and placebo arms) and those who self reported influenza-like illness (62/141 and 88/152 respectively). Adverse effects were mild and self-limiting, with participants in the amantadine arm reporting a variety of symptoms (insomnia being the most prevalent). The authors conclude that amantadine is safe and effective in preventing influenza. Randomisation and allocation are not sufficiently described whereas the placebo tablets are described as "identical". No reason is given for the two dropouts from the study. It is perhaps debatable whether some of the self-reported influenza cases really were such.

Payler is the report of a preventive randomised open trial of 536 members of Malvern Boys School in the UK. Boys, 13 to 19 had been previously immunised in the autumn of 1982 (viral strain not reported). Boys were randomised to receive either amantadine 100 milligrams daily (n=267 or no treatment n=269) for fourteen days. The trial started on 18 February 1983, as soon as the cases of laboratory-confirmed influenza were described. Case definition is unclear but from the results (Table 2) it would appear that clinical definitions and laboratory isolates were used. A/H1N1 was the prevalent viral strain. One case of urticaria as adverse effect (in the amantadine arm) only is reported. Seventy participants were excluded (Table 1) because of day boarding, influenza before the beginning of the trial and failure to take amantadine. The authors conclude that adding amantadine to vaccination could be an effective and safe (at a lower dose of 100 milligrams a day) way to prevent outbreaks. They remark that efficacy may have been facilitated by vaccination, although this remark is difficult to interpret in the light of universal vaccination of the school population. Although the trial was randomised, the design is open and the analysis was not carried out on intention to treat. Additionally, length of follow up is not reported.

Peckinpaugh reports the results of three studies carried out on military personnel of the Naval Research Unit in Illinois, USA, at different periods during 1967 and 1968. The first study is a double-blind randomised controlled cross-over toxicity study of amantadine and rimantadine in Navy recruits. As no outcome data are presented in the report, the study was not considered further.

The second study (called Peckinpaugh/1) was a randomised factorial double-blind controlled trial conducted between 5 December 1967 and 1 June 1968 comparing the effects of oral amantadine (200 milligrams daily) on 2079 individuals and oral placebo on 2104 individuals with and without the addition of vaccine. Administration of amantadine or placebo continued for 28 days. Individuals in each arm, three days after commencing treatment were assigned to be injected with either influenza vaccine (2059 individuals) or subcutaneous saline placebo (2124 individuals). Serological assessment was carried out on one fourth of participants on prior to immunisation, at termination of amantadine administration and on completion of the 9 week recruit training period. Clinical outcomes are presented in histogram form (Figure 3) by intervention arm (amantadine and influenza vaccine - 1039 individuals, influenza vaccine and oral placebo - 1020 individuals, amantadine and subcutaneous placebo - 1040 and subcutaneous and oral placebo - 1084 individuals). The text implies that 25% fewer individuals in the amantadine arm seroconverted but no detailed serological outcomes are presented. Outcomes in the figure are: % of participants who did not seek treatment, sought treatment as outpatients or were admitted to local hospital. We estimated the numerator of each outcome from these percentages. No adverse effects were reported and the authors conclude that amantadine is safe and acceptable for influenza prevention. The study was definitely randomised and arrangements to conceal allocation appear adequate. The standard of reporting is poor with outcomes reported in histogram form, making estimation of their incidence necessary.

Peckinpaugh/2 reports the results of a study conducted from December 1968 to May 1969. Successive intakes of recruits were randomised to receive either amantadine 200 milligrams daily (n=1329) or placebo (n=1321) orally for 20 days. As in Peckinpaugh/1, serological assessment was carried out on one fourth of participants (416 sera pairs). Outcome definition and presentation (Figure 6) were similar to those in Peckinpaugh 1, as is our transformation of them. The authors remark on the poor influenza A2 activity level which explains the lack of difference in outcomes found between the two arms. As this statement contrasts with the very high level of "acute respiratory disease" reported in Figure 6, we deduce that these must be clinically defined cases a proportion of which are unlikely to be influenza A. We have thus classified and entered the outcomes from this trial as "clinically defined". The study was definitely randomised and arrangements to conceal allocation appear adequate. The standard of reporting is poor with outcomes reported in histogram form, making estimation of their incidence necessary. Additionally contradictory statements in the text (concerning the definition of cases and serological outcomes) make interpretation of the text difficult.

Pettersson is the report of preventive randomised placebo- controlled double-blind trial in a military training centre in Finland commencing on 17 January 1978. At the time an epidemic of A H1N1 was underway in the surrounding community. Data relating to two other arms, comprising elderly and hospital patients are excluded from the analysis. 192 participants (military recruits aged 21 average) were randomly assigned to receive either amantadine 200 milligrams daily of calcium lactate placebo tablets. Participants were followed up for three weeks and cases were defined on the basis of retrospective analysis of questionnaires and antibody titre rise in paired sera. Infection with influenza A H3N2 virus also was verified serologically. Dropouts (all due to adverse effects) were 22 out of 117 in the amantadine arm and 11 out of 108 in the placebo arm. The authors conclude that protection against the two viral strains was low, possibly because of late start of the amantadine schedule (less than 24 hours before contracting influenza). Withdrawal due to adverse effects were markedly more prevalent in the amantadine arm. Although the study report is well written and random allocation was carried out on the basis of random number tables, allocation concealment was unspecified and intention to treat analysis did not take place.

Plesnik is the report of a preventive randomised double-blind controlled trial of amantadine during an epidemic of influenza A/Victoria/42/75 [H3N2] virus. The trial took place in the middle of March 1976 in the former Czechoslovakia and lasted 20 days. Participants were 1133 healthy students, industrial workers and elderly aged 10 to 69. The age group breakdown (Table 1) is within that specified in the inclusion criteria assuming that half participants reported in the age group 10 -19 are aged 15-19. No results broken down by age are presented. Participants were randomised to receive either amantadine 100 milligrams once or twice daily (n=574) or placebo (n=559), one or two tablets daily. Assumption of medication was supervised, except at weekends. Case definition was based on influenza symptoms and antibody titre rise from paired sera and viral isolation (although from the translated text it is not clear whether this practice covered the whole study population). Results show only clinically defined cases and adverse affects (Table 2), although the text appears to imply that 40 participants out of 574 in the amantadine arm (not broken down by dose) and 49 out of 559 in the placebo arm had serologically confirmed influenza we have not included these data in our meta-analysis as we do not know whether these are patients with clinical symptoms. Adverse effects (reported in detail in the trial) were serious enough to cause withdrawal from the trial in eight adults (in the amantadine arm) and four adults and one adolescent (in the placebo arm). Results are presented by dose schedule for adverse effects and cases of respiratory illness. The authors conclude that amantadine is efficacious in reducing cases of influenza but its dosage at 200 milligrams daily is associated with a higher incidence of adverse effects, especially drowsiness. Although randomisation, allocation and analysis on an intention to treat are not described, concealment is described as a centralised scheme with coded tablets and is therefore assessed as adequate.

Quarles reports the conduct and results of a preventive randomised double blind placebo-controlled trial carried out at the University of Texas. Participants were 444 healthy students aged 18 to 24 who were followed up for six weeks on the basis weekly of self-reporting. The trial was commenced on week 7 (possibly mid February) 1978 during an epidemic of influenza A/USSR/90/77 virus. Students were randomised to receive either amantadine (100 milligrams), or rimantadine (100 milligrams), or placebo (not specified) tablets twice daily. A random blood sample to assess schedule compliance was also taken at the end of the second week from 54 randomly selected volunteers. Cases were defined on the basis of paired sera and clinical symptoms. Viral isolation from throat swabs was additionally carried out. Only 308 students out of 444 (70%) returned completed questionnaires and at least two blood specimens. Although the text states that dropouts were equally distributed in the three arms, no reasons for the dropouts are given other than on 12 students who dropped out because of adverse effects (six in the amantadine, two in the rimantadine four in the placebo arms). Table 1 reports numbers of cases of influenza illness by seroconversion status in the three arms of trial. In our review we included only cases which had both seroconverted and showed signs of clinical illness. We excluded the cases of asymptomatic seroconversion as these do not answer the study question of estimating the clinical efficacy of the compounds. Usable (presented by arm of the trial) data on adverse effects is presented in the text grouped into the "one or more adverse effect" category, although in the description these are qualified a nausea and vomiting for rimantadine, insomnia for amantadine and dizziness for both drugs. Most adverse effects took place in the first 48 hours. The study report also contains an analysis of age-related efficacy and overall concludes that both compounds are effective in preventing influenza. The practices of randomisation, allocation and concealment are not defined, making it impossible to assess methodological rigour. Analysis on the basis of intention to treat was not carried out and the overall quality of reporting is poor (with no complete details of reasons for dropout).This study was probably commenced at of just after the peak of influenza transmission with a consequent possible underestimation of the efficacy of preventive interventions (due to late commencement of prophylaxis).

Reuman is the report of two double-blind preventive randomised controlled trials. The second trial assesses preventive efficacy of amantadine in protecting participants from a challenge dose of influenza virus and will not be considered further as it does not fulfil study inclusion criteria. The first study was conducted in Cincinnati Christ Hospital, Ohio, USA during the 1985-86 influenza season and published three years later. Participants were 476 healthy hospital personnel aged 18 to 55. There were three arms to the trial
a) 159 subjects randomly allocated to receive amantadine 100 milligram/day
b) 159 subjects randomly allocated to receive amantadine 200 milligram/day
c) 158 subjects randomly allocated to receive identical looking placebo capsules each day.
Participants were followed up weekly for six weeks through symptom diaries and visits and all those reporting with influenza- like illness (pre-defined from a list of signs and symptoms) were seen on two consecutive days for examination and collection of nasal washouts. Adverse effects were also recorded. To test compliance with the schedule random blood test were carried out on 48 subjects (all results were consistent with compliance). Paired sera were collected at the beginning and end of the trial and tested for antibodies against A/Chile/1/83(H1N1), A/Philippines/1/82(H3N2) and B/USSR/100/83 the three prevalent viruses during the study period. Viral assays were performed from nasal washouts. Five participants were infected in the placebo group, two subjects in the 100 milligram group and one in the 200 mg group. The number of participants with at least one (systemic) adverse effect was 49 in the placebo arm, 47 in the 100 mg arm and 71 in the 200 mg arm, indicating a clear dose-adverse effect relationship (Table 1). The authors conclude that influenza activity was too low to be detected by the study (which was powered for a 30% incidence). The study is of good quality with computer-generated randomisation, checks on allocation concealment and good blinding procedures. No dropouts are reported.

Schapira is the report of a placebo-controlled trial of amantadine conducted in the Northern English village of Keighley by one general practitioner during the winter of 1968-69. 297 healthy volunteers from firms, schools, a newspaper, a bank and families were allocated to receive either amantadine 100 milligrams every twelve hours (157 participants) or an oil-based placebo capsule (140 participants), identified only by a code. Participants' mean age was evenly matched at 38 years (active arm) and 38.5 years (placebo arm). Follow up was ten days. All participants' sera were screened prior to commencement and those who complained of symptoms of influenza underwent a second antibody titre estimation against A2/Hong Kong/1/68, the prevalent strain at the time. As it likely that some participants were already infected prior to commencement the authors point out that the trial is both preventive and curative. 49 out of the 157 volunteers in the treatment arm suffered influenza symptoms compared to 39 out 140 on the placebo arm. However only 8 and 15 cases respectively were serologically confirmed as influenza (Table 3). Twelve paired sera were unavailable for the 49 symptomatic volunteers in the treatment arm and three paired sera were unavailable although the authors do not explain the reason for the drop outs. Serological follow-up was incomplete for participants in the amantadine arm. Adverse effects are not reported. The authors conclude that the trial provides "marginal" evidence of preventive effect of amantadine. The practices of randomisation, and allocation are not described, making it impossible to assess methodological rigour, although the similar numbers of participants in each arm led us to consider this a randomised trial. Allocation concealment appeared adequate.

Smorodintsev is a preventive randomised double-blind placebo controlled trial carried out in eight engineering schools around Leningrad, in the former USSR. The design comprised an "internal" (day students) and "external" (i.e. boarders) do-nothing control groups which will not be further taken into consideration as they took no medication. Participants were all males, aged 18-30 and were randomised to receive amantadine 100 milligrams daily (n=5092) or placebo tablets (n=3175). The study commenced around January 1969 with staggered starts, according to the results of school specific surveillance against A2/Hong Kong/68 virus. Differential follow-up is also recorded between placebo and amantadine arms (30 days) and controls (12 days). Case definition was based on a list of symptoms plus a four-fold rise in serum antibody titres to influenza A. Cases defined on the basis of clinical diagnosis only are reported in Table 1 and are included in our analysis. However, the denominators reported in association with laboratory-confirmed cases are not reconcilable with other denominator data in the text and were excluded from our analysis. Overall there were 1894 dropouts, 748 of which has influenza before the start of the trial, 776 in the amantadine and 370 in the placebo arms failed to comply with drug schedule. Only 1825 out of 6383 "non-ill" participants were interviewed at the end of the trial for adverse effects but denominators are uneven (1313 in the amantadine arm and 512 in the placebo arm) presumably as interviews were carried out while interviewers were still blinded. The authors conclude that amantadine influenza prophylaxis is effective and safe. Although randomisation occurred and concealment is described in an adequate manner, an unknown number of participants were given amantadine syrup. As no reference to syrup placebo is made in the text (only to tablet form), robustness of blinding and concealment is questionable. Additionally the numerous discrepancies in the text make this a poorly reported trial.

Wendel is the report of a preventive and treatment randomised double blind placebo-controlled trial in a prison community in Philadelphia, USA in January 1963. Participants were prison inmates aged 17-54 exposed to naturally occurring influenza A2Japan305/57 virus. Placebo was a lactose tablet. In the preventive trial 794 inmates were randomised to receive amantadine 200 milligrams daily (n=439) or placebo (n=355). Follow-up length was ten days. Cases were defined both clinically and serologically but as we could not reconcile data in the text we only included data for clinically defined cases. The only adverse effect reported upon was insomnia (2 cases in the amantadine arm and five cases in the placebo arm). No dropouts were allowed. The authors conclude that amantadine administration causes a 30% reduction in cases of influenza. In the treatment trial, 55 cases with laboratory diagnosis were randomly allocated to similar interventions and followed-up for two weeks. Adverse effects (insomnia) outcomes only are reported. The trials are well reported although randomisation was not good (as confirmed by uneven numbers in the arms - 20% excess in the amantadine arm) and allocation concealment is adequate.

TREATMENT TRIALS

We identified 10 published treatment trials. We were unable to identify any unpublished trials. The mean amantadine arm size was 91 individuals (median 72, 25th percentile 15, 75th percentile 110), the mean rimantadine arm size was 61 (median 56, 25th percentile 15, 75th percentile 104 individuals) and the mean placebo arm size was 77 individuals (median 76, 25th percentile 14, 75th percentile 99). The mean total population was 161 individuals (median 153, 25th percentile 30, 75th percentile 225). Mean length of follow up was 25 days (median 25.5 days, 25th percentile 16 days, 75th percentile 33 days).

Identified trials are listed below using the name of the first author and year of publication in the case of there being more than one trial by the same author:

Galbraith 1971*
Hayden 1980
Hayden 1986
Hornick
Kitamoto 1968
Kitamoto 1971
Knight*
Younkin*
van Voris
Wingfield

One trial (Hornick) was broken down further into four sub-trials (see below for explanation).

Galbraith 1971* is a treatment trial carried out in December 1969 January 1970, at the time of an epidemic caused by a variant of A2/Hong Kong/68. Participants were unvaccinated family members aged more than 2 years recruited by fifty-seven family doctors in the United Kingdom. 153 participants with laboratory-confirmed diagnosis of influenza A2 were randomised to receive either doses appropriate to their ages, for adults amantadine 200 milligrams a day (n=72, mean age 37.4 years) or placebo (n=81, mean age 39.1). Treatment was commenced within 48 hours of symptoms and continued for seven days. Although twenty three percent of participants were outside our target age group, this proportion still fulfills our inclusion criteria. As the outcome data are not presented by age group, we have included aggregate data in our meta-analysis (Table 3). There were 9 participants (4 in the amantadine and 5 in the placebo group) lost to follow-up and whose data are not included in the analysis. The reasons for this finding are not given in the text. Outcomes are clinical (Tables 2 and 3) and serological (Table 4 and 5). In our meta-analysis, we have included the time of duration of fever (in days after commencement of treatment (Table 2) approximating the standard deviation of duration (not reported in the text) from the P value reported in the table. No adverse effect is mentioned or reported in the text. The authors report that there was a significant shortening of the length of fever in the amantadine arm but no other noticeable effects. Although randomisation is clearly mentioned, no detailed description of allocation and concealment is given, making its assessment impossible.

Hayden 1980 reports the conduct and results of a randomised double-blind, placebo controlled treatment trial of inhaled (20 milligrams daily) amantadine. The trial took place during January-February 1978 and 20 participants were randomised to receive either amantadine (n=9, mean 19.1 years) or diluted water placebo (n=11, mean age 20.3 years) within 48 hours of developing symptoms for a duration of four days. One dropout took place in the placebo arm because of lack of confirmation of influenza. A further 3 patients were lost to follow up (no reason given). Influenza A/Texas/77[H3N2] and influenza A/USSR/77[H1N1] caused infection in the participants. Cases were ascertained clinically and immunologically and outcomes in all cases are presented as scores at day 2 of follow up for "respiratory illness" and "constitutional illness" which does not include flu-like symptoms (Figures 1 and 2). Adverse effects reported in Table 2 are all local and due to the aerosol. We only included nasal burning as the most significant. The authors conclude that there were no important differences between groups but amantadine aerosol recipients experienced significantly more local adverse effects. The trial was clearly randomised, but no description of allocation and concealment is given making its assessment impossible. Additionally the rationale for distinguishing between constitutional and respiratory illness is unclear, results of outcomes are not clearly reported (mean scores only are given) and reasons for dropouts are not explained. Randomisation and allocation concealment appeared adequate. No data is presented on the 12 participants who did not have laboratory-confirmed influenza.

Hayden 1986 is the report of a randomised double-blind, placebo controlled treatment trial of oral rimantadine. The trial took place in the universities of Virginia and Michigan in 1983, when 14 adults with confirmed A/Bangkok/1/79(H3N2) influenza were randomised to receive either rimantadine 200 milligrams once daily (mean age 28 years) or placebo (mean age 23 years) for five days. Treatment started within 48 hours of symptom onset. Outcomes measured were nasal virus shedding, duration of fever (in hours) and symptom scores (presented broken down into systemic headache, chills, malaise etc and respiratory). Average duration of fever in the rimantadine arm was 31 hours (SD 22 hours) and 68 (SD 8 hours) in the placebo group. Adverse effects were mild with one participant in each arm reporting dizziness (classified as "other" in our meta-analysis) and one in each arm reporting insomnia. A further subject had unexplained galactorrhea. No dropouts took place. The authors conclude that a daily oral dose of rimantadine is safe and effective in reducing the symptoms of influenza. Although the trial is extremely clearly reported, no description of allocation and concealment is given.

Hornick* reports the conduct and results of a placebo controlled double-blind treatment comparison of amantadine 100 milligrams with lactose placebo twice daily for ten days. Participants were 153 inmates of four prisons: Richmond, Wynne, Walls and Jessup (full data were not presented for another two prison sites). Influenza diagnosis was made on the basis of clinical and laboratory findings. Ninety-four inmates were randomised to receive amantadine and 103 placebo in January 1968, during an epidemic of influenza A2. Overall results show that participants could be divided into "rapid resolvers" to treatment (whose illness resolved within 36 hours or less), medium resolvers (whose illness resolved within 24-36 hours) and slow resolvers (whose illness resolved in more than 36 hours) in both arms of the trial. Table 2 reports the mean duration of fever in the two arms of the trial. As duration of fever data for the three sites are entered separately in the text (Table 2) we included it in our meta-analysis by calculating the SD of duration of fever from initiation of treatment from the p value. We assumed denominators in each of the three prison sites (in the text arm denominators are not reported by trial sites) as follows:
Richmond site (renamed Hornick/Richmond)
amantadine N=21 mean duration 60.9 hours,
placebo n=28, mean duration 80.1 hours,
duration SD= 33 hours (for both arms).
Wynne site (renamed Hornick/Wynne)
amantadine N=17 mean duration 49.8 hours,
placebo n=17, mean duration 82.1 hours,
duration SD=39 hours (for both arms);
Walls site (renamed Hornick/Walls)
amantadine N=23 mean duration 65.1 hours,
placebo n=20, mean duration 88.3 hours,
duration SD=28 hours (for both arms);
Jessup site (renamed Jessup/Maryland)
amantadine N=15 mean duration 66 hours,
placebo n=15, duration 92 hours,
duration SD=35 hours (for both arms).
We transformed the duration data into 24 hour days. No adverse effects were observed in the volunteers. The authors conclude that amantadine is effective against influenza. The word "randomised" in not visible in the text, however denominators in each of the arms are highly suggestive of randomisation. No mention of the allocation procedure is made in the text, nor are drop-outs mentioned.

Kitamoto 1968 is the report of a randomised double-blind, placebo controlled treatment trial of oral amantadine. The trial took place during the 1967-68 influenza season throughout Japan and the 355 participants were randomised to receive either amantadine (n=182) (200 milligrams/day for adults) or placebo (n=173) within 48 hours of developing symptoms. Case definition was based on symptoms and serological confirmation. During the period of the trial A2/Kumamoto/1/65 virus and its drifted variants were circulating in the community. Although the trial report also contains data for children and for non-serologically confirmed cases of respiratory disease, we have included only the data for confirmed adult influenza cases with duration of fever (in hours) outcome (Table 2). These were 14 in the amantadine arm and 23 in the placebo arm. However, the numbers of serologically confirmed cases in Table 2 are lower (missing 7 cases in the amantadine arm and 6 in the placebo arm) than those given in Table 1. No explanation for this discrepancy is given. Adverse effects are reported in Table 9. The authors conclude that amantadine significantly shortened fever and no severe adverse effects were observed, making amantadine safe and effective. Although the trial is adequately reported and clearly randomised, no description of allocation and concealment is given making its assessment impossible.

Kitamoto 1971 is the report of a randomised double-blind, placebo controlled treatment trial of oral amantadine. The trial took place during the 1968-69 influenza season throughout Japan and results are reported for 155 participants with serologically confirmed A/Hong Kong influenza. The authors exclude from the report data relating to 582 participants with either influenza B (n=158) or A who had commenced treatment later than 48 after symptoms began (n=71). No mention is made of the fate of the remaining 353 participants. Seventy nine participants were randomised to receive amantadine (200 milligrams/day for adults) and seventy six placebo for seven days within 48 hours of developing symptoms. Case definition was thus based on symptoms and serological confirmation of infection. Outcomes were duration of fever (in days), defined as a temperature of more that 37 degrees centigrade and there were no dropouts from the study. We have included in our review only data relating to adults (38 in the amantadine and 46 in the placebo arm) who had no concomitant administration of antipyretics (Table 1) as data on other subjects are not given. Adverse effects were recorded for 120 adults in the amantadine arm and 149 adults in the placebo arm of the trial, these denominators obviously include participants infected with Influenza B and those who did not start treatment within 48 hours of symptoms becoming evident. Although these data are not presented separately for the 38 adults in the amantadine and 46 adults in the placebo arm who had no concomitant administration of antipyretics we have included them in our meta-analysis. This assumes that participants in the two arms did not have a different adverse effect profile caused the use of antipyretics. Adverse effect incidences are reported in Table 5. The authors conclude that amantadine administration within 2 days after the onset of illness resulted in significant shortening of the duration of fever, and suppression of high fever with no significant adverse effects. Although randomisation was clearly applied and no dropouts are reported, no description of allocation and concealment is given making its assessment impossible. In addition the fate to follow up of 353 participants is unclear.

Knight* reports the conduct and results of a treatment trial of amantadine compared to placebo in 29 unvaccinated male inmates aged 22 to 42 years of two units of the Texas Department of Corrections. The study, a randomised double blind comparison, took place during a A2/Hong Kong influenza outbreak in January 1969. Only participants with both serological and culture (nasal shedding of the virus) confirmation of illness took part, thus of the 37 original participants only 29 were included. Illness had become manifest an average of 42 hours before for amantadine-treated patient and 36 hours for placebo (the authors report that the differences were not significant). Participants were randomised to receive either amantadine 100 milligrams (n=13) or placebo capsules (n=16) twice daily for at least 6 days. Outcomes were clinical (duration of fever and symptom score) or serological and laboratory-based (antibody titre and viral shedding). Length of duration of fever is reported in the text (page 202) either at the 84 hour cut-off point or as a mean time to defervescence (44.5 hours in the amantadine arm and 71.3 in the placebo arm) and we estimated the SD (35 hours) from the p value. No adverse effects of the medication are reported. The authors conclude that amantadine has a limited therapeutic effect (the differences barely reaching significance), although no comments can be made on its effect in preventing or treating pneumonia, the major cause of influenza-related deaths. Not surprisingly, participants who had the longest standing duration of illness defervesced more rapidly. Although randomisation was clearly applied and no dropouts are reported, no description of allocation and concealment is given making its assessment impossible. The conclusions of the trial may be due to its lack of power.

Younkin* reports the conduct and results of randomised double blind comparison of amantadine with aspirin. The trial took place during an outbreak of influenza A/Brazil/78[H1N1] in the winter of 1981. The 48 participants in the trial were otherwise healthy college students aged 17 to 20 from the University of Rochester, New York, USA. Participants all had clinical and (later) laboratory confirmed diagnosis of influenza. They were randomised to receive either amantadine 100 milligrams a day (n=16), 200 milligrams a day (n=14) or aspirin 3.25 grams a day (n=17). Outcomes are presented as both symptoms scores (Figure 1) and duration of fever (in the text). We have included in our meta-analysis only data for the group of 29 participants who had fever at randomisation and the amantadine arm data are not presented by dose as that data are not available in the text. SD for duration of fever was calculated from the P value as 9 hours for both arms. Adverse effects are reported in the text, but numbers are not given clearly. Compliance was 100% at 3 days but thereafter a number of subjects did not take medication. Although the numbers are not reported the authors report that a significantly greater number of participants in the aspirin group dropped out because of adverse effects. The authors report that although there was no difference in defervescence between the two arms, amantadine provided a significant improvement in systemic symptom scores. Randomisation and allocation concealment are well detailed in the text and concealment in particular appears very robust with an elaborate system of envelopes. To uniformly make up the tablets number, the aspirin arm had two placebo tablets included. Compliance was equally tested by daily questioning.

van Voris reports the conduct and results of a treatment trial carried out in February 1978 in University of Rochester, USA on 45 students with laboratory confirmed influenza A/USSR/77[H1N1]. Students were randomised in a double-blind design to receive either amantadine 200 milligrams daily (N=14) or rimantadine 200 milligrams daily (N=14) or inert placebo capsules (N=12) and followed up for five days. Outcomes reported were symptom scores (Figure 1), duration of temperature (Figure 2 and in the text at page 1130). We used the outcome "number of participants with temperatures more than 37.5 degrees centigrade", 48 hours after commencing medication (a dichotomous outcome). Data on mean duration of fever were not given. There were no dropouts from the trial and "difficulty in concentrating" and light headedness effects were reported in one third of amantadine recipients and one placebo recipient. The authors conclude that both drugs have a notable therapeutic effect in shortening the illness and possibly the spread of the virus. The trial is well reported and definitely randomised. Allocation concealment is described and appears to be adequate.

Wingfield is the report of a randomised double blind placebo-controlled treatment trial which took place in the Virginia State Penitentiary, USA, during an outbreak of influenza A2/Rockville/1/65 in January-February 1968. Participants were inmates whose symptoms developed in the previous 24 hours. Ninety-five inmates were randomised to receive either amantadine 100 milligrams (n=23), or rimantadine 150 milligrams (n=24) or placebo lactose capsules (n=48) twice daily. Medication was commenced within 24 hours of developing symptoms and continued for ten days. Only data relating to participants with temperatures of 37.5 degrees centigrade or more on the first day of the study were analysed of the duration of fever outcome and one participant was later found not to have a rise at the second antibody titre estimation. The final denominators for which data are provided in the report are:
Placebo n=39
Amantadine n=20
Rimantadine=20
We have included duration of fever as the clinical outcome in our meta-analysis. We calculated a standard deviation from the significance test results given and we transformed the outcome into number of participants with temperatures more than 37.5 degrees centigrade, 48 hours after commencing medication (data from Figure 2, right hand graph) in each arm. No adverse effects are reported as having occurred in participants. The authors conclude that administration of both amantadine and rimantadine led to significantly rate of defervescence of temperature and improvement tin clinical symptoms than placebo. Randomisation and allocation concealment are detailed in the text with individual coded "vials" (which we assume to mean containers) system. Intention to treat analysis appears to have taken place on individuals who fulfilled clinical and serological criteria.

Methodological quality

Two reviewers (VD/TJ) assessed allocation method, allocation concealment, blinding and completeness of follow-up.

There were 28 trials in all, 26 of which considered either amantadine and/or rimantadine efficacy and 2 (Hayden 1981; Millet) which considered adverse effects only. Twelve preventive trials and seven treatment trials reported sufficient data on adverse effects. The quality of preventive and treatment trials will be discussed separately.

The quality of the preventive trials was relatively good, considering their age. Among the 18 preventive trials, 15 stated that the allocation method was randomisation, although only 4 mentioned a particular method and two did not mention random allocation at all. These 2 trials have therefore been classified as controlled clinical trials (CCT) rather than RCT. All preventive trials were stated to be double blind with the exception of Payler which was open and had no placebo group (the comparison group was no intervention other than influenza vaccine at the beginning of the season).

Among the 10 treatment trials, 9 stated that the allocation method was randomisation and no trials mentioned a particular method and one (Hornick) did not mention random allocation at all.

Major flaws in the reporting of trials lay in the following:
- Lack of information on the completeness of follow-up. In many trials there was a large difference between the number randomized and the number who actually participated.
- Lack of detailed description of methods to conceal allocation with many trials just describing a "double blind" procedure.
- Frequent inconsistencies in the reporting of numerators and denominators in various arms of trials.
- In the treatment trials, the use of a bewildering variety of outcomes, such as severity scores, of which none are alike.

Results

We carried out six main comparisons:

Comparison A - oral amantadine compared to placebo in the prevention of influenza.
Comparison B - oral rimantadine compared to placebo in the prevention of influenza.
Comparison C - oral amantadine compared to oral rimantadine in the prevention of influenza.
Comparison D - oral amantadine compared to placebo in the treatment of influenza.
Comparison E - oral rimantadine compared to placebo in the treatment of influenza.
Comparison F - oral amantadine compared to oral rimantadine in the treatment of influenza.

For comparisons A, B and C we analysed the effects on "cases", stratified either on the basis of clinical-laboratory criteria (a defined set of signs and symptoms backed up by serological confirmation and/or isolation of influenza virus from nasal fluids) or clinical criteria alone. We did not assess the effects of amantadine/rimantadine administration on asymptomatic cases (defined only by serology or viral isolation), as data on these were scarce. If more data become available we will consider its analysis in future updates of our review. We stratified comparisons on the basis of whether participants had received vaccination or not.

Additionally we assessed adverse effects in both comparisons.

In Comparisons A and B significant heterogeneity between the trial results was evident for both types of influenza analyses, so all results quoted are based on random effects models.

In Comparison A, amantadine prevented 63% (95% CI: 42% to 76%) of influenza cases defined clinically and with laboratory tests and 23% (95% CI: 11% to 34%) of clinically defined influenza cases. Both of these results are highly statistically significant (P<0.001).

In Comparison B rimantadine prevented 72% (95% CI: -8% to 92%)of the influenza cases defined clinically and with laboratory tests and 35% (95% CI: -20% to 65%) of clinically defined cases. Whilst these results are conventionally not statistically significant (P=0.07 and P=0.17 respectively), the estimates are based on only 688 individuals, and are of a very similar magnitude to those for amantadine.

In Comparison C there is no evidence of a difference in efficacy between amantadine and rimantadine, although the confidence interval is quite wide (RR amantadine vs rimantadine 0.88. 95% CI: 0.48 to 1.63).

The "all adverse effects" category includes all types and was derived from those trials which either did not report sufficient information to allow a more detailed classification or that presented aggregate data. Adverse effects incidence is reported in our meta-analysis as the number of participants with at least one event, thus the incidence of individual adverse effects cannot be summed to give the total with any adverse effect as more than one adverse event is likely to have taken place in the same individual during the trial.

In Comparison A all categories of adverse effects were significantly more common in participants who received amantadine than placebo, except for dermatological changes and decreased CNS activity. Nearly twice as many amantadine recipients experienced increased CNS effects, and more than twice as many withdrew from the trials due to adverse effects.

In Comparison B, rimantadine recipients were also more likely to experience adverse effects than placebo recipients. However, there was no evidence of an increase in CNS-related effects with rimantadine, and withdrawal rates were similar in both groups.

The direct comparison of amantadine with rimantadine (Comparison C) confirmed that CNS adverse effects and withdrawal from trials were more significantly common among amantadine recipients than rimantadine recipients (CNS effects; RR 2.58, 95% CI: 1.54 to 4.33; withdrawals RR 2.30, 95% CI: 1.23 to 4.30).

Thus rimantadine appears no less efficacious but safer than amantadine in preventing cases of influenza in healthy adults. Readers should bear in mind that the study sizes of the safety trials of rimantadine are considerably smaller than those of amantadine, so that the conclusions that can be drawn for rimantadine are somewhat less certain than those for amantadine.

We considered meta-analysing symptoms outcome data to further inform the assessment of the effects of amantadine and rimantadine in the treatment role. When we tabulated the outcome typology we discovered that such a meta-analysis would be impossible as can be seen from the following table:

Trial Symptom outcomes used

Galbraith 1971 Average time to clearance of symptoms
Hayden 1980 Aggregate scores of systemic and respiratory symptoms
Hayden 1986 Aggregate scores of systemic and respiratory symptoms
Hornick/Jessup Percentage of patients in three symptoms clearance time periods
Kitamoto 1968 No symptoms
Kitamoto 1971 No symptoms
Knight Between arms symptoms concordance. Aggregate data only
Togo* Percentage of patients in three symptoms clearance time periods
Younkin* Significance of the difference of symptoms scores
van Voris Percentage of improvement of symptom scores at different time periods
Wingfield Significance of difference of proportions of patients in three symptoms clearance time periods

We resorted to using duration of fever (defined as a temperature greater than 37 degrees centigrade) as the only common outcome. One obvious cost of this approach is the possible confounding effect of the presence of fever for a variable length of time prior and after entry to the study (and hence at the moment of commencement of treatment). However if random allocation had been properly carried out, this effect should disappear.

In Comparison D amantadine significantly shortened duration of fever compared to placebo (by 1.00 days: 95% CI: 0.73 to 1.29). The meta-analysis is based on 506 subjects (230 in the amantadine and 276 in the placebo arm). Where time to fever clearance data were not available (as in van Voris and Wingfield), a dichotomous outcome was used (cases with fever at 48 hours). This comparison showed that amantadine was significantly better than placebo (relative risk 0.21: 95% CI: 0.07 to 0.65).

In Comparison E, rimantadine too significantly shortened duration of fever compared to placebo (by 1.27 days. 95% CI: 0.77 to 1.77). There were a significantly higher number of afebrile cases 48 hours after commencing rimantadine treatment (RR=0.17; 95% CI: 0.04 to 0.74).

The little data available directly comparing amantadine and rimantadine for treatment (Comparison F), showed that the efficacy of the two drugs was comparable, although the confidence intervals are very wide.

In contrast to the increased adverse effect rates for prevention, there was no evidence that amantadine recipients had increased adverse effect rates than placebo recipients (Comparison D). There were very little data available for the assessment of adverse effects of rimantadine for treatment (45 participants) or the direct comparison between amantadine and rimantadine (33 participants).

Readers of this review should bear in mind that the difference in incidence of adverse effects is of importance, rather than the estimated incidence itself, as the adverse effects reported with these drugs are very similar to the clinical manifestations of influenza infection.

Overall both drugs appear to be effective and well tolerated, but readers should bear in mind that the evaluation of the effects rimantadine was carried out on a very small population.

In all comparisons duration of action and protection appeared directly related to duration of prophylaxis or treatment with amantadine and rimantadine. This finding is in keeping both with the half-life of the compounds which are excreted by the kidneys (at the rate of 6.4 ml/min/kg for amantadine and 1.2 ml/min/kg for rimantadine - Aoki 1998) and the self-limiting duration of the illness. No trials assessed onset of resistance to the drugs although this is known to have a relatively short induction time (10-27% of patients secrete drug-resistant virus within 4-5 days of commencing treatment - Aoki 1998).

We carried out two minor comparisons (G and H) each based on the results of single trials.

In Comparison G, based on Younkin, aspirin was significantly more effective than amantadine in reducing the length of fever (by 0.47 days. 95% CI: 0.17 to 0.76). This observation is based on 29 individuals. Aspirin is well known for being a very effective antipyretic and anti-inflammatory drug, however it does not inhibit viral replication and as such remains a symptomatic remedy and is likely to have no effect of other clinical or laboratory outcomes not included in our review .

In comparison H (based on Hayden's 1980 trial) inhaled amantadine was no more efficacious than placebo in bringing down the respiratory or constitutional symptom score (Weighted Mean Difference: 1.0. 95% CI: 3.64 to 1.64 and 2.0, 95% CI: 16.9 to 12.9 respectively). This comparison too is based on small numbers of participants (20). Not surprisingly, amantadine caused significantly more nasal irritation (RR 6.11: 95% CI: 0.86 to 43.3). Inhaled amantadine does not appear to be particularly effective but has a high incidence of local adverse effects which would make compliance difficult.

The interpretation of Comparisons G and H is made difficult by the small numbers involved and the absence of multiple trials.

All trials tested the effects of amantadine and rimantadine on a wide variety of influenza A viruses. None tested the effects on influenza B, on which the molecules are known to be ineffective.

No trial tested the role of the compounds on workplace outbreak control, which is a pity considering the trial settings (prisons, factories, schools, barracks).

Some trials are likely to have included individuals who took aspirin to relive symptoms (especially in the treatment trials). However the effects of this potential confounder should have been eliminated by the process of randomisation.

All trials commenced administration of the compounds within a reasonable time lapse. Treatment started at the latest 48 hours after positive identification of the first case in the population and prevention when the results of surveillance made it reasonable to do so.

No trials assessed onset of resistance, but data in one study demonstrated that 10 to 27% of patients treated with amantadine secreted drug-resistant virus within 4 to 5 days of commencing treatment (Aoki 1998).

Finally, we considered carrying out sub-analysis by dose (100, 200, 300 mags daily), but decided against this given the small size of the resulting meta-analysis. We will re-consider this policy if any further data become available.

Discussion

The results of our review show that both amantadine and rimantadine are efficacious and relatively safe in prevention and treatment of cases of influenza A. The role of amantadine in prophylaxis (63% effective ) and treatment (shortens duration of illness by 1 day) is beyond question and does not need to be investigated further compared to placebo. Rimantadine appears equally efficacious in prevention (72%), but in direct comparison with placebo, when a random effects model is applied the lower bound of the 95% confidence does not achieve statistical significance.

There are two explanations for this difference in the significance of the findings. The first is that the rimantadine result is a "false negative". This idea is supported by noting that its average efficacy is both large and similar to amantadine, and that there have been many fewer participants in rimantadine trials than amantadine trials (there are clinical data for approximately only 700 rimantadine compared to 2,500 amantadine recipients in the review).

The second explanation is centred on between trial heterogeneity. If a fixed effect analysis is used (effectively ignoring the heterogeneity) then the difference between rimantadine and placebo for the prevention of clinically and serologically defined influenza cases is highly significant (P<0.001 for both outcomes). All of the analyses of influenza outcomes demonstrated excessive variation in the results of the trials. Such a pattern has been noted in other reviews in preventive procedures, such as influenza and cholera vaccination, and may reflect differences between the trial populations to natural exposure and immunity to influenza A and other similar viruses. We have not been able to explain this heterogeneity in this systematic review.

Rimantadine was also seen to be equally therapeutically efficacious, shortening duration of fever by just over one day. However, again, this observation is based on 54 patients only.

There do not appear to be significant differences in effectiveness in either role between the two compounds, although again our comparisons are based on small numbers with large confidence intervals.

The safety profile of the two drugs appears however significantly different with rimantadine causing significantly fewer CNS adverse effects than amantadine and fewer withdrawals from the trials. Although these observations are based on smaller numbers of rimantadine recipients, amantadine definitely causes signs of significantly increased and decreased CNS activity, an effect which is not easily acceptable by healthy adults, especially in employment which require concentration and mental fitness. Rimantadine has a different pharmacokinetic profile from amantadine, reaching prophylactic concentration in the nasal mucus at much lower plasma concentrations than amantadine. This finding would explain the effectiveness of rimantadine at less toxic concentrations than amantadine (Hayden 1983; Hayden 1985). There was a tendency for lower doses of amantadine (100 milligrams daily) to cause fewer adverse effects then higher doses at the cost of lower effectiveness (data not shown).

We conclude that from available evidence rimantadine appears the better choice for prevention.

Significantly decreased CNS activity is caused by amantadine in a treatment role, but this appears to be the only significant adverse effect of both compounds in a treatment role, although numbers of patients are again very low.

Low numbers make the interpretation of the other two comparisons equally difficult.

In future more attention should be paid to the assessment of adverse events of the two compounds, particularly those of rimantadine which at present are based on relatively small numbers.

The quality of the trials was not good with significant numbers of studies failing to give adequate descriptions of methods and of results. This may be in part due to the number of older trials in the review. Both quality of trial conducting and reporting should be improved and adverse effects and case outcome definitions should be standardised. Finally, the bewildering array of outcome definitions used in treatment studies made the task of meta-analysis difficult and led to a great loss of information.

Reviewers' conclusions

Implications for practice

Both drugs can be used effectively for the prevention and treatment of influenza, both in vaccinated and unvaccinated populations. However, the higher incidence of CNS adverse effects of amantadine (especially at higher doses) makes rimantadine the current ion-channel blocker antiviral of choice against influenza.

Implications for research

There is a requirement for better reported trials, particularly those comparing amantadine and rimantadine directly in both prevention and treatment roles. Attention should be paid to the assessment of adverse events of the two compounds, particularly those of rimantadine which at present are based on relatively small numbers. Quality of trial conducting and reporting should be improved and adverse effects and case outcome definitions should be standardised. Investigation of reasons for the heterogeneity seen between the trials may be of some value.

Potential conflict of interest

None known.

Acknowledgements

The authors would like to thank Drs Aoki, Couch, Hayden and Monto for helpful comments and Mrs Carol Hobbs for assistance with trial retrieval.

Characteristics of included studies

Study Brady
MethodsPreventive randomised double-blind controlled trial of rimantadine during an epidemic of influenza A/Leningrad/87 [H3N2] virus.
ParticipantsTwo hundred and twenty-eight healthy not previously vaccinated adult volunteers aged 18 to 55
InterventionsParticipants were randomised to receive either rimantadine 100 milligrams daily or placebo for six weeks.
OutcomesPaired sera were taken from all participants at the beginning and the end of the study. Within-trial surveillance was carried out on a weekly basis and cases were defined on the basis of seroconversion and a pre-defined list of symptoms and signs. Viral isolation took place by nasal washout.
NotesBrady is a clearly written and well-reported trial (with the exception of the minor discrepancy between text and tables on the affiliation of drop-outs). Randomisation was computer generated and allocation was concealed with a centralised scheme. Additionally, intention to treat analysis is clearly stated in the text.
Allocation concealmentA
Study Callmander
MethodsRandomised controlled trial conducted in a community, including some military personnel. During the period of the trial there was considerable influenza A2 (Leningrad) activity.
ParticipantsThe age range of the 94 volunteer participants is 20 to 60 years (44 male and 50 female).
InterventionsThe intervention arm received 100 mg of amantadine hydrochloride twice daily and the control arm a no better described placebo.
OutcomesOutcomes are expressed in numbers of clinically defined cases (from a symptoms list) in each arm and a symptom score (reported in Table 1 without an indication of time of intensity). Surveillance for adverse effects (systemic) was carried out. A list of symptoms (without a denominator) is reported at Table 2.
NotesThe practices of randomisation, allocation and concealment are notfurther defined, making it impossible to assess methodological rigour although as the distribution of sex and age was checked and found to be similar, randomisation is likely to have been satisfactory.
Allocation concealmentB
Study Dolin
MethodsPreventive randomised double-blind placebo controlled trial carried out in Burlington Vermont, USA. The trial was commenced on 10 January 1981 during an outbreak of influenza A/Bangkok/1/79H3N2 and A/Brazil/11/78H1N1 detected by surveillance (see Figure 1).
ParticipantsParticipants initially were 450 healthy non-vaccinated volunteers aged 18 to 45 (mean age 25.6 + 0.45 years).The final total of participants was 378 (132 in the placebo arm, 133 in the rimantadine arm and 113 in the amantadine arm).
InterventionsAmantadine 200 milligrams daily or rimantadine 200 milligrams or placebo.
OutcomesCase definition was based on a list of symptoms plus virus isolation or a rise in serum antibody titres to influenza A. Table 1 presents both clinically-defined cases and cases defined on the basis of laboratory confirmation.
NotesAlthough a well-written report, no real information is given on random allocation, blinding and concealment. Intention to treat analysis was not carried out.
Allocation concealmentB
Study Galbraith 1971
MethodsTreatment trial carried out in December 1969 January 1970, at the time of an epidemic caused by a variant of A2/Hong Kong/68.
ParticipantsParticipants were unvaccinated family members aged more than 2 years recruited by fifty-seven family doctors in the United Kingdom.
Interventions153 participants with laboratory-confirmed diagnosis of influenza A2 were randomised to receive either doses appropriate to their ages, for adults amantadine 200 milligrams a day (n=72, mean age 37.4 years) or placebo (n=81, mean age 39.1). Treatment was commenced within 48 hours of symptoms and continued for seven days.
OutcomesOutcomes are clinical (Tables 2 and 3) and serological (Table 4 and 5). In our meta-analysis, we have included the time of duration of fever (in days after commencement of treatment (Table 2) approximating the standard deviation of duration (not reported in the text) from the P value reported in the table. No adverse effect is mentioned or reported in the text.
NotesAlthough randomisation is clearly mentioned, no detailed description of allocation and concealment is given, making its assessment impossible.
Allocation concealmentB
Study Hayden 1980
MethodsRandomised double-blind, placebo controlled treatment trial of inhaled (20 milligrams daily) amantadine.
Participants20 participants
InterventionsParticipants were randomised to receive either amantadine (n=9, mean 19.1 years) or diluted water placebo (n=11, mean age 20.3 years) within 48 hours of developing symptoms for a duration of four days.
OutcomesInfluenza A/Texas/77[H3N2] and influenza A/USSR/77[H1N1] caused infection in the participants.
Cases were ascertained clinically and immunologically and outcomes in all cases are presented as scores at day 2 of follow up for "respiratory illness" and "constitutional illness" which does not include flu-like symptoms (Figures 1 and 2). Adverse effects reported in Table 2 are all local and due to the aerosol. We only included nasal burning as the most significant.
NotesThe trial was clearly randomised, but no description of allocation and concealment is given making its assessment impossible. Additionally the rationale for distinguishing between constitutional and respiratory illness is unclear, results of outcomes are not clearly reported (mean scores only are given) and reasons for dropouts are not explained.
Allocation concealmentB
Study Hayden 1981
MethodsA toxicity study reporting a randomised control trial undertaken in an unspecified period in USA and published in 1981.
ParticipantsThe setting is that of a state farm insurance company and the participants were 251 adult volunteers, aged between 18-65 (mean age of 32).
InterventionsTwo trials were carried out simultaneously, both involving Rimantadine and Amantadine. One was a low dose (200mg daily of each drug, n=52) trial and the other a higher dose (300mg daily of each drug, n=199). The low dose trial however has been excluded due to the absence of any "cases" data, and the lack of outcomes. Overall 41 out of 67 (61%) in the Amantadine arm, 13 out of 66 in the placebo arm (20%), and 18 out of 63 in the Rimantadine arm (29%) experienced adverse effects.
OutcomesAdverse effects (systemic only with no other classification) were noted, although not specified.
NotesThe practices of randomisation, allocation and concealment are notfurther defined, although all doses were stated as being administered by a project nurse. This is a poorly reported trial as no detailed classification of adverse effects is given, which is a strange practice for a toxicity study. Additionally, data reported in the text is not consistent with that in Table 1.c
Allocation concealmentB
Study Hayden 1986
MethodsRandomised double-blind, placebo controlled treatment trial of oral rimantadine.
The trial took place in the universities of Virginia and Michigan in 1983.
Participants14 adults with confirmed A/Bangkok/1/79(H3N2) influenza.
InterventionsParticipants were randomised to receive either rimantadine 200 milligrams once daily (mean age 28 years) or placebo (mean age 23 years) for five days. Treatment started within 48 hours of symptom onset.
OutcomesOutcomes measured were nasal virus shedding, duration of fever (in hours) and symptom scores (presented broken down into systemic headache, chills, malaise etc and respiratory). Average duration of fever in the rimantadine arm was 31 hours (SD 22 hours) and 68 (SD 8 hours) in the placebo group.
NotesAlthough the trial is extremely clearly reported, no description of allocation and concealment is given.
Allocation concealmentB
Study Hornick/Jessup
MethodsPlacebo controlled double-blind treatment comparison of amantadine 100 milligrams with lactose placebo twice daily for ten days.
ParticipantsParticipants were 153 inmates of four prisons: Richmond, Wynne, Walls and Jessup.
InterventionsInfluenza diagnosis was made on the basis of clinical and laboratory findings. Ninety-four inmates were randomised to receive amantadine and 103 placebo in January 1968, during an epidemic of influenza A2.
OutcomesOverall results show that participants could be divided into "rapid resolvers" to treatment (whose illness resolved within 36 hours or less), medium resolvers (whose illness resolved within 24-36 hours) and slow resolvers (whose illness resolved in more than 36 hours) in both arms of the trial.
Jessup site (renamed Jessup/Maryland) amantadine N=15 mean duration 66 hours, placebo n=15, duration 92 hours, duration SD=35 hours (for both arms). We transformed the duration data into 24 hour days.
NotesThe word "randomised" in not visible in the text, however denominators in each of the arms are highly suggestive of randomisation. No mention of the allocation procedure is made in the text, nor are drop-outs mentioned.
Allocation concealmentB
Study Hornick/Richmond
MethodsPlacebo controlled double-blind treatment comparison of amantadine 100 milligrams with lactose placebo twice daily for ten days.
ParticipantsParticipants were 153 inmates of four prisons: Richmond, Wynne, Walls and Jessup.
InterventionsRichmond site (renamed Hornick/Richmond) amantadine N=21 mean duration 60.9 hours, placebo n=28, mean duration 80.1 hours, duration SD= 33 hours (for both arms).
OutcomesInfluenza diagnosis was made on the basis of clinical and laboratory findings. Ninety-four inmates were randomised to receive amantadine and 103 placebo in January 1968, during an epidemic of influenza A2.
NotesThe word "randomised" in not visible in the text, however denominators in each of the arms are highly suggestive of randomisation. No mention of the allocation procedure is made in the text, nor are drop-outs mentioned.
Allocation concealmentB
Study Hornick/Walls
MethodsPlacebo controlled double-blind treatment comparison of amantadine 100 milligrams with lactose placebo twice daily for ten days.
ParticipantsParticipants were 153 inmates of four prisons: Richmond, Wynne, Walls and Jessup.
InterventionsWalls site (remaned Hornick/Walls) amantadine N=23 mean duration 65.1 hours, placebo n=20, mean duration 88.3 hours, duration SD=28 hours (for both arms);
OutcomesInfluenza diagnosis was made on the basis of clinical and laboratory findings. Ninety-four inmates were randomised to receive amantadine and 103 placebo in January 1968, during an epidemic of influenza A2.
NotesThe word "randomised" in not visible in the text, however denominators in each of the arms are highly suggestive of randomisation. No mention of the allocation procedure is made in the text, nor are drop-outs mentioned.
Allocation concealmentB
Study Hornick/Wynne
MethodsPlacebo controlled double-blind treatment comparison of amantadine 100 milligrams with lactose placebo twice daily for ten days.
ParticipantsParticipants were 153 inmates of four prisons: Richmond, Wynne, Walls and Jessup.
InterventionsWynne site (renamed Hornick/Wynne) amantadine N=17 mean duration 49.8 hours, placebo n=17, mean duration 82.1 hours, duration SD=39 hours (for both arms);
OutcomesInfluenza diagnosis was made on the basis of clinical and laboratory findings. Ninety-four inmates were randomised to receive amantadine and 103 placebo in January 1968, during an epidemic of influenza A2.
NotesThe word "randomised" in not visible in the text, however denominators in each of the arms are highly suggestive of randomisation. No mention of the allocation procedure is made in the text, nor are drop-outs mentioned.
Allocation concealmentB
Study Kantor
MethodsPreventive double-blind randomised controlled trial of the efficacy and safety of oral amantadine compared to a (not further defined) placebo. The trial took place over the period 20 February - 7 March 1978 in the military barracks at Fort Sam Houston (FSH), Texas and the target serotype was A/USSR/77.
ParticipantsTrial participants were 139 healthy paramedic recruits (mean age 22 years)
InterventionsParticipants were randomised to receive either amantadine 100 milligram tablets twice daily (n = 64) or placebo (n=62).
OutcomesCase definition consisted of a list of recognised symptoms and serological confirmation and adverse effects were recorded in the questionnaires. Paired sera were obtained at the beginning of the study and five weeks later. Cases were also presented by severity
NotesAlthough the trial was randomised and allocation concealment was protected through a system of sealed envelopes, the difficulty in reconciling figures and understanding what actually happened during the trial makes this a very poorly reported study. No reason for the loss of 20 participants to follow up is given.
Allocation concealmentA
Study Kitamoto 1968
MethodsRandomised double-blind, placebo controlled treatment trial of oral amantadine. The trial took place during the 1967-68 influenza season throughout Japan.
ParticipantsNot specified
Interventions355 participants were randomised to receive either amantadine (n=182) (200 milligrams/day for adults) or placebo (n=173) within 48 hours of developing symptoms.
OutcomesCase definition was based on symptoms and serological confirmation. During the period of the trial A2/Kumamoto/1/65 virus and its drifted variants were circulating in the community.
NotesAlthough the trial is adequately reported and clearly randomised, no description of allocation and concealment is given making its assessment impossible.
Allocation concealmentB
Study Kitamoto 1971
MethodsRandomised double-blind, placebo controlled treatment trial of oral amantadine. The trial took place during the 1968-69 influenza season throughout Japan.
ParticipantsSeventy nine participants were randomised to receive amantadine (200 milligrams/day for adults) and seventy six placebo for seven days within 48 hours of developing symptoms.
InterventionsAmantadine and placebo in a treatment role.
OutcomesCase definition was based on symptoms and serological confirmation of infection. Outcomes were duration of fever (in days), defined as a temperature of more that 37 degrees centrigrade and there were no dropouts from the study.
NotesAlthough randomisation was clearly applied and no dropouts are reported, no description of allocation and concealment is given making its assessment impossible.
In addition the fate to follow up of 353 participants is unclear.
Allocation concealmentB
Study Knight
MethodsRandomised double blind comparison, took place during a A2/Hong Kong influenza outbreak in January 1969. This is a treatment trial
Participants29 unvaccinated male inmates aged 22 to 42 years of two units of the Texas Department of Corrections. Only participants with both serological and culture (nasal shedding of the virus) confirmation of illness took part, thus of the 37 original participants only 29 were included.

Illness had become manifest an average of 42 hours before for amantadine-treated patient and 36 hours for placebo (the authors report that the differences were not significant).
InterventionsAmantadine compared to placebo. Participants were randomised to receive either amantadine 100 milligrams (n=13) or placebo capsules (n=16) twice daily for at least 6 days.
OutcomesOutcomes were clinical (duration of fever and symptom score) or serological and laboratory-based (antibody titre and viral shedding).
NotesAlthough randomisation was clearly applied and no dropouts are reported, no description of allocation and concealment is given making its assessment impossible.
Allocation concealmentB
Study Millet
MethodsRandomised double blind safety trial of amantadine alone or in combination with chlorpheniramine (an anti-histaminic compound), rimantadine alone or chlorpheniramine. The trial was conducted between October 1979 and January 1980
Participants52 healthy adult volunteers recruited from the University of California at Los Angeles (USA)
InterventionsAmantadine 100 mgs and chlorpheniramine placebo (10 volunteers), amantadine 100 mgs and chlorpheniramine 4 mgs (11 volunteers), chlorpheniramine 4 mgs and antiviral placebo (11 volunteers), rimantadine 100 mgs and chlorpheniramine placebo (10 volunteers) or antiviral placebo and chlorpheniramine placebo (10 volunteers) twice daily for 3 to 4 days.
OutcomesSubjective side effects (from a predefined list) with a grading of 1 (mild) to 3 (severe) and performance testing.
NotesThe study was randomised and certainly double-blind, with centralised preparation of active and placebo tablets.
Allocation concealmentA
Study Monto
MethodsPreventive randomised double-blind placebo controlled trial carried out among healthy student volunteers at the University of Michigan, USA. The trial began in February 1978, during an outbreak of influenza A/USSR-like (H1N1) and lasted seven weeks (the when the outbreak was over as confirmed by clinical and serological surveillance).
ParticipantsUniversity students
Interventions286 students aged 18 to 24 were randomly assigned to receive either amantadine 100 milligrams twice daily or placebo. Participants were followed up weekly for seven weeks, the remaining duration of the outbreak of influenza.
OutcomesRespiratory and adverse effect symptoms were recorded in weekly calendar-diaries and paired sera samples were taken at the beginning and end of the trial (including 5 dropouts from the trial because of clinical influenza or major adverse effects) from 281 of the participants.
NotesThe trial is well written and appears well conducted with only minor discrepancies in total numbers. The practice of random allocation and concealment appear adequate. Analysis was not carried on an intention to treat basis.
Allocation concealmentA
Study Muldoon
MethodsCrossover preventive controlled trial conducted in a school near Chicago (USA) between 2 December and 20 December 1968 (period 1).
Participants105 unvaccinated volunteer students aged around 21 took part.
InterventionsAt the beginning and the end of period 1 blood samples were taken from participants. During the eighteen days of period 1 the 53 individuals in the intervention arm received 100 mg of amantadine hydrochloride twice daily and the control arm a no better described oral placebo. After a "holiday" period of 16 days (period 2) the participants were crossed over to assess whether withdrawal of amantadine rendered participants more or less susceptible to infection. The procedure was inverted and students were followed up for a further 10 days (period 3)
OutcomesOutcomes are expressed in numbers of clinically defined cases from a symptom score coupled with virus-specific antibody rise. This allowed the authors to distinguish non-influenza related illness during periods 1 and 2 (reported in Table 1 without an indication of intensity of symptoms). During period 1 (Table 2) six students had influenza (5 with serological confirmation) out of the 53 in the intervention arm and 10 (2 with serological confirmation) out of the 52 in the placebo arm.
NotesThe practices of randomisation, allocation and concealment are not defined, making it impossible to assess methodological rigour, although the similar numbers of participants in each arm led us to consider this a randomised trial. No drop outs are reported.
Allocation concealmentB
Study Oker-Blom
MethodsPreventive randomised double-blind placebo controlled trial begun on 20 January 1969 in Helsinki, Finland and lasting for 30 days.
ParticipantsParticipants were 391 healthy medical students (132 females and 259 males) aged on average 22 years. Influenza A2 Hong Kong was the main circulating strain at the time.
InterventionsParticipants were randomised to receive either amantadine 100 mg or placebo (tablets containing calcium lactate) twice daily. Additionally the authors analysed age-stratified results for each arm (although the results are not included in the text).
OutcomesCases were defined on the basis of self-reported (by questionnaire) clinical symptoms and serological titres. Adverse effects were also recorded.
NotesRandomisation and allocation are not sufficiently described whereas the placebo tablets are described as "identical". No reason is given for the two dropouts from the study. It is perhaps debatable whether some of the self-reported influenza cases really were such.
Allocation concealmentB
Study Payler
MethodsPreventive randomised open trial of 536 members of Malvern Boys School in the UK.
ParticipantsBoys, 13 to 19 had been previously immunised in the autumn of 1982 (viral strain not reported).
InterventionsBoys were randomised to receive either amantadine 100 milligrams daily (n=267 or no treatment n=269) for fourteen days. The trial started on 18 February 1983, as soon as the cases of laboratory-confirmed influenza were described.
OutcomesCase definition is unclear but from the results (Table 2) it would appear that clinical definitions and laboratory isolates were used. A/H1N1 was the prevalent viral strain.
NotesAlthough the trial was randomised, the design is open and the analysis was not carried out on intention to treat.
Additionally, length of follow up is not reported.
Allocation concealmentC
Study Peckinpaugh/1
MethodsComparison of the effects of oral amantadine (200 milligrams daily) on 2079 individuals and oral placebo on 2104 individuals with and without the addition of vaccine. Administration of amantadine or placebo continued for 28 days. Individuals in each arm, three days after commencing treatment were assigned to be injected with either influenza vaccine (2059 individuals) or subcutaneous saline placebo (2124 individuals). Peckinpaugh/1 was a randomised factorial double-blind controlled trial conducted between 5 December 1967 and 1 June 1968.
ParticipantsUS Naval trainees
InterventionsComparison of the effects of oral amantadine (200 milligrams daily) on 2079 individuals and oral placebo on 2104 individuals with and without the addition of vaccine. Administration of amantadine or placebo continued for 28 days. Individuals in each arm, three days after commencing treatment were assigned to be injected with either influenza vaccine (2059 individuals) or subcutaneous saline placebo (2124 individuals).
OutcomesCinical outcomes are presented in histogram form (Figure 3) by intervention arm (amantadine and influenza vaccine - 1039 individuals, influenza vaccine and oral placebo - 1020 individuals, amantadine and subcutaneous placebo - 1040 and subcutaneous and oral placebo - 1084 individuals). Outcomes in the figure are: % of participants who did not seek treatment, sought treatment as outpatients or were admitted to local hospital.
NotesThe study was definitely randomised and arrangements to conceal allocation appear adequate. The standard of reporting is poor with outcomes reported in histogram form, making estimation of their incidence necessary.
Allocation concealmentA
Study Peckinpaugh/2
MethodsPeckinpaugh 2 reports the results of a study conducted from December 1968 to May 1969.
ParticipantsUS Naval recruits
InterventionsRecruits were randomised to receive either amantadine 200 milligrams daily (n=1329) or placebo (n=1321) orally for 20 days.
OutcomesAs in Peckinpaugh 1, serological assessment was carried out on one fourth of participants (416 sera pairs). Outcome definition and presentation (Figure 6) were similar to those in Peckinpaugh 1, as is our transformation of them.
NotesThe study was definitely randomised and arrangements to conceal allocation appear adequate. The standard of reporting is poor with outcomes reported in histogram form, making estimation of their incidence necessary. Additionally contradictory statements in the text (concerning the definition of cases and serological outcomes) make interpretation of the text difficult.
Allocation concealmentA
Study Pettersson
MethodsPreventive randomised placebo- controlled double-blind trial in a military training centre in Finland commencing on 17 January 1978. At the time an epidemic of A H1N1 was underway in the surrounding community.
Participants192 participants (military recruits aged 21 average).
InterventionsParticipants were randomly assigned to receive either amantadine 200 milligrams daily of calcium lactate placebo tablets.
OutcomesParticipants were followed up for three weeks and cases were defined on the basis of retrospective analysis of questionnaires and antibody titre rise in paired sera. Infection with influenza A H3N2 virus also was verified serologically.
NotesAlthough the study report is well written and random allocation was carried out on the basis of random number tables, allocation concealment was unspecified and intention to treat analysis did not take place.
Allocation concealmentB
Study Plesnik
MethodsPreventive randomised double-blind controlled trial of amantadine during an epidemic of influenza A/Victoria/42/75 [H3N2] virus. The trial took place in the middle of March 1976 in the former Czechoslovakia and lasted 20 days.
ParticipantsParticipants were 1133 healthy students, industrial workers and elderly aged 10 to 69.
InterventionsParticipants were randomised to receive either amantadine 100 milligrams once or twice daily (n=574) or placebo (n=559), one or two tablets daily.
OutcomesCase definition was based on influenza symptoms and antibody titre rise from paired sera and viral isolation (although from the translated text it is not clear whether this practice covered the whole study population). Adverse effects ar reported in detail in the trial.
NotesAlthough randomisation, allocation and analysis on an intention to treat are not described, concealment is described as a centralised scheme with coded tablets and is therefore assessed as adequate.
Allocation concealmentA
Study Quarles
MethodsPreventive randomised double blind placebo-controlled trial carried out at the University of Texas.
ParticipantsParticipants were 444 healthy students aged 18 to 24 who were followed up for six weeks on the basis weekly of self-reporting. The trial was commenced on week 7 (? mid February) 1978 during an epidemic of influenza A/USSR/90/77 virus.
InterventionsStudents were randomised to receive either amantadine (100 milligrams), or rimantadine (100 milligrams), or placebo (not specified) tablets twice daily.
OutcomesCases were defined on the basis of paired sera and clinical symptoms. Viral isolation from throat swabs was additionally carried out.
NotesThe practices of randomisation, allocation and concealment are not defined, making it impossible to assess methodological rigour. Analysis on the basis of intention to treat was not carried out and the overall quality of reporting is poor (with no complete details of reasons for dropout). This study was probably commenced at of just after the peak of influenza transmission with a consequent possible underestimation of the efficacy of preventive interventions (due to late commencement of prophylaxis).
Allocation concealmentB
Study Reuman
MethodsReuman is the report of a double-blind preventive randomised controlled trial. The study was conducted in Cincinnati Christ Hospital, Ohio, USA during the 1985-86 influenza season. Participants were followed up weekly for six weeks through symptom diaries and visits and all those reporting with influenza- like illness (pre-defined from a list of signs and symptoms) were seen on two consecutive days for examination and collection of nasal washouts. Adverse effects were also recorded.

To test compliance with the schedule random blood test were carried out on 48 subjects (all results were consistent with compliance) . Paired sera were collected at the beginning and end of the trial and tested for antibodies against A/Chile/1/83(H1N1), A/Philippines/1/82(H3N2) and B/USSR/100/83 the three prevalent viruses during the study period. Viral assays were performed from nasal washouts.
ParticipantsParticipants were 476 healthy hospital personnel aged 18 to 55.
InterventionsThere were three arms to the trial:
a) 159 subjects randomly allocated to receive amantadine 100 milligram/day
b) 159 subjects randomly allocated to receive amantadine 200 milligram/day
c) 158 subjects randomly allocated to receive identical looking placebo capsules each day.
OutcomesFive participants were infected in the placebo group, two subjects in the 100 mg group and one in the 200 mg group. The number of participants with at least one (systemic) adverse effect was 49 in the placebo arm, 47 in the 100 mg arm and 71 in the 200 mg arm, indicating a clear dose-adverse effect relationship (Table 1).
NotesThe authors conclude that influenza activity was too low to be detected by the study (which was powered for a 30% incidence). The study is of good quality with computer-generated randomisation, checks on allocation concealment and good blinding procedures. No dropouts are reported.
Allocation concealmentA
Study Schapira
MethodsThe study is a placebo-controlled trial of amantadine conducted in the Northern English village of Keighly by one general practitioner during the winter of 1968-69.
Participants297 healthy volunteers from firms, schools, a newspaper, a bank and families. Participants' mean age was evenly matched at 38 years (active arm) and 38.5 years (placebo arm).
InterventionsParticipants were allocated to receive either amantadine 100 milligrams every twelve hours (157 participants) or an oil-based placebo capsule (140 participants), identified only by a code. Follow up was ten days. All participants' sera were screened prior to commencement and those who complained of symptoms of influenza underwent a second antibody titre estimation against A2/Hong Kong/1/68, the prevalent strain at the time.

As it likely that some participants were already infected prior to commencement the authors point out that the trial is both preventive and curative.
Outcomes49 out of the 157 volunteers in the treatment arm suffered influenza symptoms compared to 39 out 140 on the placebo arm. However only 8 and 15 cases respectively were serologically confirmed as influenza (Table 3). Twelve paired sera were unavailable for the 49 symptomatic volunteers in the treatment arm and three paired sera were unavailable although the authors do not explain the reason for the drop outs. Adverse effects are not reported.
NotesThe authors conclude that the trial provides "marginal" evidence of preventive effect of amantadine. The practices of randomisation, and allocation are not described, making it impossible to assess methodological rigour, although the similar numbers of participants in each arm led us to consider this a randomised trial. Allocation concealment appeared adequate.
Allocation concealmentB
Study Smorodintsev
MethodsPreventive randomised double-blind placebo controlled trial carried out in eight engineering schools around Leningrad, in the former USSR. The design comprised an "internal" (day students) and "external" (i.e. boarders) do-nothing control groups which will not be further taken into consideration as they took no medication.
ParticipantsParticipants were all males, aged 18-30.
InterventionsParticipants were randomised to receive amantadine 100 milligrams daily (n=5092) or placebo tablets (n=3175). The study commenced around January 1969 with staggered starts, according to the results of school specific surveillance against A2/Hong Kong/68 virus. Differential follow-up is also recorded between placebo and amantadine arms (30 days) and controls (12 days).
OutcomesCase definition was based on a list of symptoms plus a four-fold rise in serum antibody titres to influenza A.
NotesAlthough randomisation occurred and concealment is described in an adequate manner, an unknown number of participants were given amantadine syrup. As no reference to syrup placebo is made in the text (only to tablet form), robustness of blinding and concealment is questionable. Additionally the numerous discrepancies in the text make this a poorly reported trial.
Allocation concealmentA
Study Wendel
MethodsPreventive and treatment randomised double blind placebo-controlled trial in a prison community in Philadelphia, USA in January 1963.
ParticipantsParticipants were prison inmates aged 17-54 exposed to naturally occurring influenza A2Japan305/57 virus.
InterventionsPlacebo was a lactose tablet. In the preventive trial 794 inmates were randomised to receive amantadine 200 milligrams daily (n=439) or placebo (n=355). Follow-up length was ten days.
OutcomesCases were defined both clinically and serologically.
NotesThe trials are well reported although randomisation was not good (as confirmed by uneven numbers in the arms - 20% excess in the amantadine arm) and allocation concealment is adequate.
Allocation concealmentA
Study Wingfield
MethodsRandomised double blind placebo-controlled treatment trial which took place in the Virginia State Penitentiary, USA, during an outbreak of influenza A2/Rockville/1/65 in January-February 1968.
ParticipantsParticipants were 95 inmates whose symptoms developed in the previous 24 hours.
InterventionsNinety-five inmates were randomised to receive either amantadine 100 milligrams (n=23), or rimantadine 150 milligrams (n=24) or placebo lactose capsules (n=48) twice daily. Medication was commenced within 24 hours of developing symptoms and continued for ten days.
The final denominators for which data are provided in the report are:
Placebo n=39
Amantadine n=20
Rimantadine=20
OutcomesOnly data relating to participants with temperatures of 37.5 degrees centigrade or more on the first day of the study were analysed of the duration of fever outcome and one participant was later found not to have a rise at the second antibody titre estimation. We have included duration of fever as the clinical outcome in our meta-analysis
NotesRandomisation and allocation concealment are detailed in the text with individual coded "vials" (which we assume to mean containers) system. Intention to treat analysis appears to have taken place on individuals who fulfilled clinical and serological criteria.
Allocation concealmentA
Study Younkin
MethodsRandomised double blind comparison of amantadine with aspirin. The trial took place during an outbreak of influenza A/Brazil/78[H1N1] in the winter of 1981.
ParticipantsThe 48 participants in the trial were otherwise healthy college students aged 17 to 20 from the University of Rochester, New York, USA. Participants all had clinical and (later) laboratory confirmed diagnosis of influenza.
InterventionsParticipants were randomised to receive either amantadine 100 milligrams a day (n=16), 200 milligrams a day (n=14) or aspirin 3.25 grams a day (n=17).
OutcomesOutcomes are presented as both symptoms scores (Figure 1) and duration of fever (in the text).
NotesRandomisation and allocation concealment are well detailed in the text and concealment in particular appears very robust with an elaborate system of envelopes. To uniformly make up the tablets number, the aspirin arm had two placebo tablets included. Compliance was equally tested by daily questioning.
Allocation concealmentA
Study van Voris
MethodsRandomised in a double-blind design.
ParticipantsThis treatment trial was carried out in February 1978 in University of Rochester, USA on 45 students with laboratory confirmed influenza A/USSR/77[H1N1].
InterventionsParticipants received either amantadine 200 milligrams daily (N=14) or rimantadine 200 milligrams daily (N=14) or inert placebo capsules (N=12) and followed up for five days.
OutcomesOutcomes reported were symptom scores (Figure 1), duration of temperature (Figure 2 and in the text at page 1130). We used the outcome "number of participants with temperatures more than 37.5 degrees centigrade", 48 hours after commencing medication (a dichotomous outcome). Data on mean duration of fever were not given.
NotesThe trial is well reported and definitely randomised. Allocation concealment is described and appears to be adequate.
Allocation concealmentA

Characteristics of excluded studies

StudyReason for exclusion
Aoki 1986Contains insufficient data
AtmarThis study reports the results of safety and pharmacokinetic studies of inhaled rimantadine aerosol in both healhty and ill (with confirmed influenza A and B) volunteers. Although the text reports that inhaled rimantadine appears efficacious against influenza, no data is presented and the text goes on to qualify the effectiveness as not statistically significant against placebo.
BakerInsufficient data are presented
BrysonInsufficient data presented
CrawfordAges of participants were outside protocol age range (majority of participants are aged under 15 years).
Galbraith40% of trial population are outside the protocol's age inclusion criteria.
Galbraith 1973Population is outside (target) age range. Does not fulfil study inclusion criteria
Hayden 1983Toxicity study presenting further analysis of the data contained in Hayden 1981.
Hayden 1985Crossover study of the pharmacokinetics of amantadine and rimantadine in 12 adults. There are no clinical or safety outcomes presented.
Hayden 1989This trial assesses efficacy, rate of transmission and resistance to rimantadine from index cases to contacts within families . The trial is excluded because 42% of contacts in families with serologically documented influenza A are children (Table 1).
Hayden 1991The study is a treatment RCT of both families and 12 adults. No age breakdown is given for the outcome data.
O'DonoghueSemi-randomised allocation. In part, patients were admitted to hospital, therefore do not fit in inclusion criteria for "healthly adults". Additionally, patient ages are outside those stipulated in inclusion criteria.
RoseNon randomised, non placebo-controlled study.
SmorodinstevPresents duplicate data of included study.
Smorodintsev 1970/2Does not fulfil study inclusion criteria - trial of preventive effeicacy in artificially induced influenza.
TogoDuplicate of Hornick (identical sample sizes, same sites, same period, not cross-referenced)

Graphs

Comparison 01. Oral amantadine vs placebo (prevention)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Influenza cases (clinically and serologically defined)92438Relative Risk (Random) 95% CI0.37 [0.24, 0.58]
Influenza cases (clinically defined)1417281Relative Risk (Random) 95% CI0.77 [0.66, 0.89]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 1.

Figure 1.

Influenza cases (clinically and serologically defined)

Figure 2.

Figure 2.

Influenza cases (clinically defined)

Figure 3.

Figure 3.

Adverse effects

Comparison 02. Oral rimantadine vs placebo (prevention)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Influenza cases (clinically and serologically defined)3688Relative Risk (Random) 95% CI0.28 [0.08, 1.08]
Influenza cases (clinically defined)3688Relative Risk (Random) 95% CI0.65 [0.35, 1.20]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 4.

Figure 4.

Influenza cases (clinically and serologically defined)

Figure 5.

Figure 5.

Influenza cases (clinically defined)

Figure 6.

Figure 6.

Adverse effects

Comparison 03. Oral amantadine vs oral rimantadine (prevention)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Influenza cases (clinically and serologically defined)2455Relative Risk (Random) 95% CI0.89 [0.48, 1.65]
Influenza cases (clinically defined)2455Relative Risk (Random) 95% CI0.88 [0.57, 1.35]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 7.

Figure 7.

Influenza cases (clinically and serologically defined)

Figure 8.

Figure 8.

Influenza cases (clinically defined)

Figure 9.

Figure 9.

Adverse effects

Comparison 04. Oral amantadine vs placebo (treatment)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Duration of fever (37 degrees centigrade or more) in days9506Weighted Mean Difference (Random) 95% CI-1.01 [-1.29, -0.73]
Cases with fever at 48 hours285Relative Risk (Random) 95% CI0.21 [0.07, 0.66]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 10.

Figure 10.

Duration of fever (37 degrees centigrade or more) in days

Figure 11.

Figure 11.

Cases with fever at 48 hours

Figure 12.

Figure 12.

Adverse effects

Comparison 05. Oral rimantadine vs placebo (treatment)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Duration of fever (37 degrees centigrade or more) in days264Weighted Mean Difference (Random) 95% CI-1.27 [-1.77, -0.77]
Cases with fever at 48 hours3104Relative Risk (Random) 95% CI0.17 [0.04, 0.74]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 13.

Figure 13.

Duration of fever (37 degrees centigrade or more) in days

Figure 14.

Figure 14.

Cases with fever at 48 hours

Figure 15.

Figure 15.

Adverse effects

Comparison 06. Oral amantadine vs oral rimantadine (treatment)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Duration of fever (37 degrees centigrade or more) in days140Weighted Mean Difference (Random) 95% CI0.20 [-0.56, 0.96]
Cases with fever at 48 hours273Relative Risk (Random) 95% CI0.99 [0.23, 4.37]
Adverse effects  Relative Risk (Random) 95% CISubtotals only
Figure 16.

Figure 16.

Duration of fever (37 degrees centigrade or more) in days

Figure 17.

Figure 17.

Cases with fever at 48 hours

Figure 18.

Figure 18.

Adverse effects

Comparison 07. Oral amantadine vs aspirin (treatment)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Duration of fever (37 degrees centigrade or more) in days129Weighted Mean Difference (Random) 95% CI0.47 [0.18, 0.76]
Adverse effect - insomnia147Relative Risk (Random) 95% CI0.94 [0.42, 2.14]
Figure 19.

Figure 19.

Duration of fever (37 degrees centigrade or more) in days

Figure 20.

Figure 20.

Adverse effect - insomnia

Comparison 08. Inhaled amantadine vs placebo (treatment)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
Symptoms score in confirmed cases (respiratory illness)120Weighted Mean Difference (Random) 95% CI-1.00 [-3.64, 1.64]
Symptoms score in confirmed cases (constitutional illness)120Weighted Mean Difference (Random) 95% CI-2.00 [-16.98, 12.98]
Adverse effects - local - nasal irritation120Relative Risk (Random) 95% CI6.11 [0.86, 43.30]
Figure 21.

Figure 21.

Symptoms score in confirmed cases (respiratory illness)

Figure 22.

Figure 22.

Symptoms score in confirmed cases (constitutional illness)

Figure 23.

Figure 23.

Adverse effects - local - nasal irritation

Comments and criticisms

Missing study?

Summary

I recently stuck upon a paper that may be relevant for this review, either to include (am not sure whether the trial was randomized, but it was placebo-controlled) or as 'excluded study'.

Here are the details:

Máté J, Simon M, Juvancz I, et al. Prophylactic use of amantadine during Hong Kong influenza epidemic.

Acta Microbiol Acad Sci Hung 1970;17: 285-296.

Can send a copy if necessary.

I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter of my criticisms.

Author's reply

Contributors

Johannes C van der Wouden

Sources of support

External sources of support

  • No sources of support supplied

Internal sources of support

  • Ministry of Defence UK

Ancillary