|Home | About | Journals | Submit | Contact Us | Français|
Large clinical trials have demonstrated the therapeutic efficacy of oseltamivir against influenza. Here we assessed its indirect effectiveness in reducing household secondary transmission in an incident cohort of influenza index cases and their household members.
We recruited index outpatients whose rapid tests for influenza were positive in 2007 and 2008. Household contacts were followed for 7–10 days during 3–4 home visits to monitor symptoms. Nose and throat swabs were collected and tested for influenza by reverse transcription polymerase chain reaction (RT-PCR) or viral culture.
We followed 384 index cases and their household contacts. Index cases who took oseltamivir within 24 hours of symptom onset halved the time to symptom alleviation (adjusted acceleration factor (AF) 0.56; 95% CI: 0.42, 0.76). Oseltamivir treatment was not associated with statistically significant reduction in the duration of viral shedding. Household contacts of index cases who had taken oseltamivir within 24 hours of onset had a non-statistically significant lower risk of developing laboratory-confirmed infection (adjusted odds ratio (OR) 0.54; 95% CI: 0.11, 2.57) and a marginally statistically significant lower risk of clinical illness (adjusted OR 0.52; 95% CI: 0.25, 1.08) compared to contacts of index cases who did not take oseltamivir.
Oseltamivir treatment is effective in reducing the duration of symptoms but evidence for household reduction in transmission of influenza virus was inconclusive.
Oseltamivir is a neuraminidase inhibitor used for the treatment of influenza virus infection, including 2009 pandemic influenza A(H1N1) virus which emerged in 2009 [1–3], with well established antiviral efficacies in reducing clinical illness and complications . Large randomized controlled trials have confirmed that oseltamivir is also effective for chemoprophylaxis [5, 6]. However, less is known about the indirect benefit of oseltamivir treatment in reducing infectiousness . Household transmission is thought to be responsible for around one third of all influenza virus transmission in the community[8, 9]. A better understanding of the role of oseltamivir treatment in reducing household transmission will provide important insights into the value of antiviral use during pandemics as well as during seasonal epidemics. We assessed the effectiveness of oseltamivir treatment of household index cases on reducing duration of clinical illness and viral shedding, and transmission to household contacts in studies conducted in Hong Kong in the 2007 and 2008 influenza seasons.
The present study was a secondary analysis of data collected in a community-based randomized controlled trial of the use of face masks and enhanced hand hygiene to prevent household transmission of influenza virus [10, 11]. We enrolled index subjects who reported at least 2 symptoms of acute respiratory illness (ARI) with symptom onset within 48 hours, and lived with at least 2 other individuals, none of whom had reported ARI symptoms during the previous 14 days. Subjects were recruited from February through September 2007 and January through September 2008 from 45 public and private outpatient clinics in Hong Kong. Index cases who were positive for influenza A or B virus infection by the QuickVue Influenza A+B test (Quidel Corp., San Diego, CA), and their household members, were followed up for 7–10 days. Medications prescribed to index cases, including oseltamivir and drugs for symptomatic relief, were recorded at recruitment sites and confirmed at subsequent home visits. Oseltamivir treatment decisions were made at the discretion of the treating physician.
Participating households were visited within 48 hours of index case recruitment, and further 2 or 3 visits were scheduled over the subsequent 7 days in 2007 and 10 days in 2008. A digital tympanic thermometer was provided and all household members were requested to record their body temperature and any systemic and respiratory signs and symptoms on symptom diaries once daily until the final home visit. During each home visit, nasal and throat swabs were collected from all household members (regardless of illness) for laboratory testing by viral culture (in 2007) or reverse transcription polymerase chain reaction (RT-PCR) (in 2008). Further technical details of the laboratory methods have been published elsewhere [10, 11].
Written consent was obtained from all participants aged 18 years or older and proxy written consent for participants under 18 years old was obtained from their parents or legal guardians. Additional verbal assent was obtained from participants aged between 8 and 17 years. The study protocol was approved by the Institutional Review Board for the University of Hong Kong/Hospital Authority Hong Kong West Cluster.
Time to alleviation of influenza symptoms and signs in index subjects was determined from analysis of symptom diaries. The respiratory symptom score was the sum of the presence (versus absence) of sore throat, cough, coryza, and phlegm and ranged from 0 to 4. The total symptoms score comprised the respiratory symptom score plus the presence (versus absence) of fever ≥37.8°C, headache, and myalgia, and ranged from 0 to 7. We defined the time to alleviation of respiratory (total) symptoms as the time from symptom onset to the first day on which the respiratory (total) symptom score was 0. The time to alleviation was right-censored at the last follow-up date if symptoms persisted throughout follow-up. We analyzed the time to alleviation of symptoms using Kaplan-Meier estimates and lognormal and Weibull accelerated failure time regression models which allow explanatory variables (including oseltamivir use, age, sex, vaccination history, baseline symptom score, type of influenza virus, underlying chronic health conditions, use of antibiotic, antipyretic and steroid, and study year) to proportionally increase (decelerate) or decrease (accelerate) the time to the endpoint . We used the Akaike Information Criterion (AIC) to compare alternative regression models . In analyses of time to symptom resolution, we only included subjects who reported the relevant symptoms at recruitment.
The time at which viral shedding ceased was defined to occur within the interval between the last positive laboratory result and the first negative laboratory result. The time from symptom onset to cessation of viral shedding was analyzed using non-parametric estimates , and lognormal and Weibull accelerated failure time regression models, allowing for the interval censoring [15, 16]. Explanatory variables included oseltamivir use, age, sex, vaccination history, baseline symptom score and type of influenza virus. Regression models were compared with the AIC. Analyses of time to cessation of viral shedding were stratified by year to allow for potential differences in duration of shedding detected by culture (2007) and RT-PCR (2008).
We defined the secondary attack ratio (SAR) as the proportion of household contacts that developed influenza virus infection during the follow-up period. We used a laboratory definition of one or more nose and throat swabs positive for influenza viruses by viral culture or RT-PCR. We also used a clinical definition of influenza in household contacts which was at least 2 signs or symptoms of fever ≥ 37.8°C, cough, headache, sore throat, aches or pains in muscles or joints on one or more days during follow-up [10, 11]. We calculated 95% confidence intervals for the crude SARs using the exact binomial method. We estimated adjusted odds ratios of laboratory-confirmed and clinical influenza in household contacts using logistic regression with generalized estimating equations adjusting for index and household contact characteristics including age, sex, vaccination history, type of influenza virus, and household characteristics including use of non pharmaceutical interventions, study year, and within household clustering [17, 18].
Households where one or more household contacts had laboratory-confirmed influenza virus infection by viral culture or RT-PCR at the baseline home visit (i.e. a potential co-index case) were excluded from these analyses.
All statistical analyses were performed in R version 2.7.2 (R Development Core Team, Vienna, Austria). Raw data from the study and R syntax to permit reproducible statistical analyses are available on the authors’ website at http://www.hku.hk/bcowling/HK_NPI_study.htm.
In 2007 and 2008, 3695 outpatients were recruited, and 568 (15%) had a positive QuickVue rapid test result. We were able to successfully schedule home visits for 450 index subjects and their household members. Influenza virus infection could not be confirmed by RT-PCR or viral culture in 53 index subjects. 13 index cases were prescribed antiviral medications other than oseltamivir (10 amantadine, 2 zanamivir, 1 ribavirin). These 66 index cases were excluded, and 384 remain for analysis of cessation of symptoms and viral shedding. Among them 90/384 (23%) index cases were treated with oseltamivir.
Table 1 shows the characteristics of index cases. The baseline total symptoms score of index subjects were comparable between the oseltamivir and no antiviral group. There were more febrile cases and fewer cases prescribed antihistamines or antipyretics in the oseltamivir group. There were more males in the oseltamivir group, while other index characteristics were not statistically significantly different. Household contacts of the two groups were generally similar (Appendix Table 1).
The Kaplan Meier curves in Figure 1 show the proportions of index subjects with unresolved signs or symptoms through the follow-up period. The median duration of illness was 9 days for the oseltamivir group and 11 days for those who did not receive antiviral treatment. A persistently smaller proportion of index cases in the oseltamivir group had unresolved respiratory symptoms and total symptoms from day 1 onwards. Index subjects who had taken oseltamivir within 24 hours of symptom onset experienced a 44% reduction in time to alleviation of all signs and symptoms, with adjusted acceleration factor (AF) of alleviation of 0.56 (95% CI: 0.42, 0.76) compared to index cases who did not take any antiviral (Table 2). Results were similar for time to alleviation of fever and time to alleviation of respiratory symptoms (Table 2). Analysis based on the lognormal distribution yielded similar estimates to those based on the Weibull distribution while the lognormal fit was inferior according to the AIC for duration of all symptoms and respiratory symptoms, and vice versa for fever (data not shown).
The median duration of viral shedding after symptom onset was 6 days and viral shedding resolved sooner in individuals prescribed oseltamivir within 24 hours of onset (Figure 2). Some of the index cases who had not taken any antiviral medication continued to shed virus 12 days after illness onset. After adjustment in a lognormal model with stratification by year, index cases who took oseltamivir within 48 hours of onset had non-statistically significant reductions in duration of viral shedding in year 2007 (AF 0.76; 95% CI: 0.51, 1.14) and 2008 (AF 0.99; 95%CI: 0.83, 1.17) compared with index cases who did not take any antiviral (Table 3). Analysis based on the Weibull distribution yielded similar estimates to those based on the lognormal distribution, while the Weibull fit was inferior according to the AIC (data not shown).
We identified 53 households that included one or more household contacts with detectable viral shedding by RT-PCR or viral culture at the initial home visit [10, 11]. These households with potential co-index cases were excluded from our analysis of household transmission. The baseline characteristics of index subjects and their household contacts included in the analysis of household transmission were similar to these in the main analyses (Appendix Table 1).
Among the 331 households with a single index case, influenza virus infection was confirmed by RT-PCR or viral culture in 80 of 989 household contacts, corresponding to an overall SAR of 8.1% (95% CI: 6.5%, 10%). Stratifying by groups, the SAR was 4.7% (95% CI: 1.0%, 13%), 6.0% (95% CI: 2.5%, 12%), 7.0% (95% CI: 1.5, 19%), in household contacts of index cases who had taken oseltamivir within 24 hours, between 24 to 48 hours and more than 48 hours after symptom onset, compared to 8.7% (95% CI: 6.8%, 11%) in the no antiviral group (p-value for trend: <0.01). The unadjusted protective effectiveness (1-SAR1/SAR0)of oseltamivir treatment initiated within 24 hours or between 24 and 48 hours after symptom onset was 46% and 31% respectively.
Table 4 shows that household contacts of index cases who took oseltamivir within 24 hours of first symptoms had a non-statistically significant lower risk of developing influenza virus infection confirmed by RT-PCR or viral culture (adjusted odds ratio (OR) 0.54; 95% CI: 0.11, 2.57), clinical influenza (adjusted OR 0.52, 95% CI: 0.25, 1.08) and clinical influenza confirmed by RT-PCR or viral culture (adjusted OR 0.47; 95%CI: 0.05, 4.03). The risk reduction was attenuated for those with index who had taken oseltamivir later than 24 hours after symptom onset (p-value for trend in risk of transmission for oseltamivir administered by time since onset: laboratory-confirmed influenza 0.09; clinical influenza 0.41). Household contacts were at lower risk of illness from influenza virus infection if they had been vaccinated, or if they were older, or if their corresponding index case was older. Results for the year 2008 separately were similar (Appendix Table 2), while the sample size was insufficient to analyse data from 2007 separately.
Our results confirm the effectiveness of early oseltamivir treatment of uncomplicated influenza virus infections in terms of reducing duration of illness. When treatment was given within 24 hours of symptom onset, our results were also suggestive of an indirect benefit to household members. A study by Halloran et al. investigated the indirect effect of oseltamivir treatment in reducing transmission to household contacts by analyzing pooled data from 2 randomized controlled trials of prophylaxis, one of which provided oseltamivir treatment to all index cases within 48 hours of onset and another which did not, and estimated the unadjusted antiviral efficacy to be 16% . However, other differences between the two trials might have confounded the comparison. In our study, we found that the earlier after illness onset oseltamivir was taken, the quicker the resolution of symptoms, and by implication infectiousness. This is consistent with the trends in odds of household transmission, where the point estimates for the odds of transmission were lowest when treatment was given earliest.
In 2007, influenza A(H3N2) predominated, while in 2008 influenza A(H1N1), A(H3N2) and B co-circulated. While oseltamivir resistance among seasonal H1N1 strains became a concern in 2008/9, during our study period in Hong Kong resistance was rare in 2007 and a small proportion (12.5%) of influenza A/H1N1 virus isolates were resistant to oseltamivir in 2007/8 . While no resistance has yet been reported in influenza B virus or other subtypes of human influenza A viruses, and only sporadically in novel influenza A/H1N1 to date, early oseltamivir treatment is likely to continue to be effective in reducing the duration of clinical illness and viral shedding associated with influenza virus infections .
A meta-analysis of randomized controlled trials showed that early oseltamivir treatment reduced time to alleviation of symptoms by 1 to 2 days . Our study provides additional information on how other factors affect the duration of illness and viral shedding. Our results indicate that those who were administered oseltamivir within 24 hours of symptom onset, took antipyretics, were aged above 18 years old, were vaccinated against influenza in the past year and were infected with influenza A experienced shorter duration of illness. Use of antibiotics, steroids and antihistamines did not significantly shorten the duration of illness (Table 2).
Symptoms and signs typically lasted for longer than viral shedding and around half of the subjects continued to show symptoms 10 days after onset, while fever had subsided in all subjects by day 6. After adjustment, those who had taken oseltamivir within 48 hours experienced a non-statistically significant 24% reduction in time to cessation of viral shedding in 2007 (Table 3). Children shed virus for almost 40–80% longer than adults depending on their age and year of study (Table 3). The most likely explanation for the lower effectiveness against viral shedding in 2008 is that RT-PCR is more sensitive than viral culture particularly later in illness , while RT-PCR measures RNA load and may not distinguish between viable and inactivated influenza virions. One other possibility is the presence of oseltamivir-resistant influenza A/H1N1 virus infections in 2008, when around 50% of influenza A virus infections were associated with the H1N1 subtype and resistance had begun to emerge [19, 22].
After adjusting for potential confounding factors, the effectiveness of oseltamivir treatment in reducing secondary infection of household contacts approached 50% if treatment was initiated within 24 hours of onset, and approached the null if treatment was initiated later, although with wide confidence intervals indicating substantial uncertainty particularly for laboratory-confirmed secondary infection. Household contacts of index cases below 18 years old, and unvaccinated contacts had higher risk of laboratory-confirmed secondary infection. Our findings suggest that treating index cases with oseltamivir may confer some degree of protection to their household contacts. While previous studies on the cost-effectiveness of oseltamivir treatment have focused on reduction of morbidity in the treated individuals, a re-evaluation which incorporates the potential indirect benefits of protection to household contacts may be warranted .
Our study has several limitations. Firstly, our evaluation of the effects of oseltamivir treatment is based on non-randomized allocation to treatment, and our findings may be affected by unobserved confounders or insufficient adjustment for observed confounders. Our study was not planned in advance in our study protocol as we had not anticipated such substantial use of oseltamivir in local outpatient care. The sample size of our study was relatively small, although larger than that of the previous study on effects of oseltamivir treatment in reducing household transmission . Households with individuals concurrently infected with influenza as confirmed by RT-PCR or culture at first home visit were excluded from the study. This may lead to exclusion of subjects who were infected with more transmissible strains that secondary infection might have had occurred before the first visit.
Findings from this study suggest that early oseltamivir treatment of index cases is effective in reducing duration of illness, and could provide some degree of protection to their household contacts. Our results may be useful when considering the allocation of oseltamivir in mitigating influenza epidemics and pandemics. Pandemic plans should consider the feasibility and logistics required to efficiently distribute antiviral treatment to cases, either by ensuring that sufficient stockpiles are available at local clinics or pharmacies, or perhaps by pre-dispensing antivirals to households that include individuals at higher risk of severe illness . Our estimation of the plausible size of indirect benefits may also be useful for planning further studies of antiviral treatment .
We thank all the doctors, nurses, and staff of participating centers for facilitating recruitment; and Rita Fung, Lai-Ming Ho, Conrad Lam, Joey Sin, Winnie Wai and Eileen Yeung for research support.
This work received financial support from the US Centers for Disease Control and Prevention (grant no. 1 U01 CI000439-02), the Research Fund for the Control of Infectious Disease, Food and Health Bureau, Government of the Hong Kong SAR (grant no. 08070632), US National Institutes of Health (cooperative agreement no. 5 U01 GM076497, Models of Infectious Disease Agent Study, ML), and the Area of Excellence Scheme of the Hong Kong University Grants Committee (grant no. AoE/M-12/06). The funding agencies had no role in data collection and analysis, or the decision to publish, but the CDC was involved in study design and preparation of the manuscript. This work represents the views of the authors and not their institutions, including the Centers for Disease Control and Prevention.
The authors report no conflict of interest.
Oseltamivir is effective for treatment of influenza virus infections. We found that oseltamivir treatment may also provide indirect benefits by reducing household transmission of influenza virus.