Rate of use and effectiveness of oseltamivir in the treatment of influenza illness in high‐risk populations: A systematic review and meta‐analysis

Abstract Background Oseltamivir is recommended in the treatment of influenza illness in high‐risk populations, including those with chronic heart and lung diseases. Objectives We conducted a systematic review and meta‐analysis to determine the rate of use and effectiveness of oseltamivir in these groups of patients. Methods The protocol for the systematic review was registered on PROSPERO (CRD42019125998). Medline, EMBASE, Cochrane CENTRAL, and CINAHL were searched for observational studies and randomized controlled trials published up to 16 February 2020. Quality appraisal of final studies was conducted using GRADE guidelines. Data were extracted using a predeveloped template. Main outcomes measured included the rate of use of oseltamivir for influenza‐like‐illness and its effectiveness in reducing disease severity in patients with cardiopulmonary diseases. Outcomes measured for effectiveness were influenza‐related complications (respiratory infections and asthma exacerbations), hospitalization rates, and time to freedom from illness. Risk of bias was assessed using Cochrane's Risk of Bias 2.0 tool for randomized trials and Cochrane's Risk of Bias in nonrandomized Studies of Interventions tool for nonrandomized trials. Where data were available, pooled analyses were conducted. Dichotomous variables were evaluated using the Mantel‐Hansel method. A random effect model was applied. Summary measures were reported as risk ratios where relevant. Results Our systematic review identified nine studies. Oseltamivir use ranged from 25% to 100%. When oseltamivir group was compared to placebo, rates of respiratory tract infections reduced by 28% (RR = 0.72, 95% CI = 0.59‐0.90), hospitalization reduced by 52% (RR = 0.48, 95% CI = 0.28‐0.80) and median time to illness alleviation decreased by 10.4 to 120 hours. There was no significant reduction in asthma exacerbation rates. Conclusions Our systematic review suggests that the use of oseltamivir is beneficial in reducing disease severity, however, its use in high‐risk population remains suboptimal.

Conclusions: Our systematic review suggests that the use of oseltamivir is beneficial in reducing disease severity, however, its use in high-risk population remains suboptimal.

K E Y W O R D S
heart diseases, human, influenza, lung diseases, oseltamivir

| INTRODUCTION
Influenza is a significant contributor to the global burden of disease causing severe illness in an estimated 3 to 5 million people 1 and 291 243 to 645 832 respiratory deaths globally annually due to influenza-related complications. 2 Certain populations are at increased risk of complications due to influenza infection, including individuals under the age of 5, over the age of 65, people with medical conditions, including cardiac and respiratory disease, and pregnant women. 3,4 Patients with chronic respiratory conditions are particularly at increased risk of influenza-related hospitalizations, need for intensive care unit (ICU) admission and ventilation when compared to those without such conditions. 5 Influenza infection has also been associated with exacerbation of underlying respiratory diseases such as asthma and cystic fibrosis. 6,7 Furthermore, in patients with cardiac disease, there is at least two to five times increased risk of mortality from influenza. 5 Currently, antiviral medications are the only treatment available for influenza infection. There are three classes of antiviral drugs that target influenza: the adamantanes (matrix-2 [M2]-ion channel inhibitors), neuraminidase inhibitors, and most recently, the selective inhibitor of influenza-cap dependent endonuclease which is currently only approved for use in the United States of America (USA) and Japan.
The adamantanes are no longer first-line treatment for influenza due to the increasing development of resistance to these antivirals. 8 Neuraminidase inhibitors such as zanamivir (Relenza) and oseltamivir (Tamiflu) are more widely prescribed in the treatment of influenza.
Although in 2017, the World Health Organization (WHO) downgraded oseltamivir from a "core" drug to a "complementary" drug in its list of essential medicines 9 based on the unclear evidence around the effectiveness of oseltamivir, it is recommended that oseltamivir or zanamivir is used empirically in high-risk populations with influenzalike illness (ILI), even when presenting with uncomplicated disease. 3 Oseltamivir is the drug of choice for the treatment of influenza for people aged ≥1 year due to its easy oral administration whereas zanamivir is recommended for people aged >5 years which is administered through intravenous route or inhalation. 3 Oseltamivir is generally well tolerated and safe for use in both adults and children, with some side effects. 10 Oseltamivir use is recommended within 48 hours of symptom onset in the patient, however, multiple studies have reinforced that earlier administration of oseltamivir results in better outcomes. [11][12][13] Despite being recommended for use in ILI, rates of use of oseltamivir often remain suboptimal. Prior to the 2009 H1N1 influenza pandemic, antiviral prescribing rates in hospitalized patients were less than 30%. 14 However, during the 2009 H1N1 influenza outbreak, prescribing rates exceeded 80% in hospitalized patients. 14 More recent data from the 2012 to 2013 influenza season from outpatient care settings showed that <20% of high-risk patients for whom antiviral treatment was appropriate were actually prescribed antiviral medication, with particularly low prescription rates in children. 15 Previous studies on the effectiveness of oseltamivir have generally focused on healthy adults and children, with very few studies in high-risk populations, especially children. Given the increased susceptibility of patients with cardiopulmonary conditions to influenza complications, research on the effectiveness of oseltamivir in this specific population may help in guiding clinical practice regarding the use of oseltamivir in this population presenting with ILI.
We conducted a systematic review and meta-analysis to ascertain the rate of oseltamivir use and its effectiveness in people of all ages with cardiopulmonary conditions.

| Protocol and registration
The protocol for the systematic review was registered on PROSPERO (CRD42019125998). Study intervention: The intervention was the use of oseltamivir for influenza or ILI in people with chronic lung or cardiac diseases.

| Eligibility criteria
The comparison group was people with chronic lung or cardiac diseases who did not receive oseltamivir for influenza or ILI.

| Outcome measures
Outcomes measured were the rate of use and the effectiveness of the intervention. We determined the rate of use as the rate of prescription of oseltamivir in patients with cardiopulmonary conditions. Effectiveness of oseltamivir was determined as the effect on severity of illness (measured as area under the symptom curve), rates of hospitalization, asthma exacerbations, respiratory tract complications such as tracheitis, bronchitis, pneumonia, nasosinusitis, and pharyngitis, and time to alleviation of illness measured in hours. Time to alleviation was defined differently in different studies as median or mean time to resolution of fever (temperature <37.2 C) and symptoms (chills and myalgia).

| Study selection
Records generated by the literature search were managed using End-Note X9. 16 Once duplicates were removed, secondary articles such as systematic reviews, literature reviews, meta-analyses, case reports, and conference abstracts were excluded based on screening of the titles. Abstracts of all remaining articles were then screened based on PICO criteria. The full text of the remaining articles was assessed based on predetermined eligibility criteria.
The reference lists of pertinent systematic reviews and metaanalyses identified in the search as well as reference lists of articles were also searched for relevant studies.
One reviewer (SS) conducted the initial search and screening of articles. Any ambiguities in study selection were resolved by discussion with another reviewer (NH).

| Data collection process
Data from the systematic review were extracted using predeveloped data extraction template (Table A2). The following information was extracted from eligible studies: study duration, study design, number of participants in the study with cardiopulmo-  The quality of evidence across the different outcomes assessed in the systematic review was graded using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) guidelines. 19 The certainty of the evidence behind each outcome was assessed based on the study designs, risk of bias, inconsistency, indirectness, imprecision, publication bias, and participant size. Based on these factors, the overall quality of evidence was deemed very low, low, moderate, or high. The GRADEpro guideline development tool software was used to assist the synthesis of this data. 20 Risk of publication bias was assessed qualitatively as part of the quality of evidence assessment. Quantitative assessment of publication bias was also carried out using funnel plots.

| Synthesis of results and meta-analysis
The main findings were summarized in a tabular format and a qualitative narrative synthesis of the results was undertaken. Meta-analyses were performed to pool data from studies with sufficient information of the same outcome measures. Dichotomous data were analyzed using risk ratio (RR) and continuous outcomes were analyzed using weighted mean difference (MD). DerSimonian and Laird's random effects model was used to estimate the overall effect size with 95% confidence interval (CI). For continuous data, missing standard deviations were estimated from other summary statistics such as confidence intervals, standard errors, t values or P values. In studies where these values were missing, the corresponding authors were contacted and values reported as ranges. All statistical analyses were carried out using Review Manager (RevMan) software version 5.3.

| Study selection
The search generated 330 citations. After removing duplicates and studies that did not meet the eligibility criteria, 29 studies were assessed in full length. One study was excluded based on English language restriction. 21  contacted to obtain subset data specific to oseltamivir effectiveness in our high-risk populations, however, these data could not be obtained. 22,23 Nine studies were included in the final analysis. One study had data on both effectiveness and rate of use of oseltamivir ( Figure 1).

| Participant characteristics
All the studies (n = 4) included to determine the rate of use of oseltamivir were conducted in patients aged <21 years. In the studies (n = 6) included to determine the effectiveness of oseltamivir, two were in children ranging from 1 to 17 years old, 11 one did not state age range, 25 and the remaining three studies were in participants older than 12 years old.
All the studies in our systematic review included broad populations, and so subset data on people with cardio-pulmonary conditions were extracted according to our predetermined inclusion criteria. The subset data from three studies were from patients with asthma, 11,26,27 and two were from those with cardio-pulmonary conditions. 25,28,29 Two studies defined their high-risk population as those with chronic cardio-pulmonary conditions and/or the elderly and as individual data were not available, the whole population was included. 30,31 In one study which included patients with any medical conditions, subset data could not be extracted and thus all participants were included as the majority of this population (85.0%) had cardio-pulmonary conditions. 29

| Study characteristics
All four studies on the use of oseltamivir were observational studies, three being retrospective [27][28][29] and one being prospective 23 (Table 1).
One study included ambulatory patients, 29 two included hospitalized patients 26,27 and one included patients admitted to ICUs. 28 Two of the studies only included patients prescribed oseltamivir within F I G U R E 1 Study selection flowchart based on preferred reporting items for systematic reviews and meta-analyses (PRISMA) 24 24 hours of symptoms onset, 28,29 one study was within 48 hours of symptoms onset 27 and the fourth study did not state treatment initiation in relation to symptom onset. 26 All four studies were in patients with confirmed influenza infection. Influenza infection was confirmed through reverse transcriptase polymerase chain reaction (RT-PCR) 26,27 or based on ICD-9-CM diagnostic codes of influenza. 28,29 Four of the six studies on effectiveness were randomized, double-blind, placebo-controlled multicentre trials 11,22,30,31 (Table 2).
One study was a randomized open-label trial. 25 A retrospective study included to evaluate oseltamivir's effectiveness was also used to determine its rate of use. 29 One study included patients prescribed oseltamivir within 24 hours of influenza diagnosis, 30 two studies included patients prescribed oseltamivir within 48 hours of symptom onset 11,25 and three studies were patients within 36 hours of symptom onset. 22,30,31 All studies on effectiveness were also in patients with confirmed influenza infection. Five studies confirmed influenza infection based on virus isolation from patient swabs and/or rises in serum influenza antibody titers 11,22,25,30,31 while one study confirmed influenza based on ICD-9-CM diagnostic codes for influenza. 29

| Rates of use of oseltamivir
Rates of oseltamivir use varied between the four different studies from 25% to 100% (Table A1). Two studies were based on data from the USA and had prescribing rates of 25% 28 and 31%. 29 Subset data for people with chronic cardiac or respiratory disease were taken from these studies as their study populations were broader than required by our study. The study conducted in hospitals in Spain had 41% usage rate 27 while Saudi Arabia had 100%. 26 The subset data used from these two studies relevant to our inclusion criteria were patients with asthma.

| Effectiveness of oseltamivir
Meta-analysis of the data suggests administration of oseltamivir for the treatment of confirmed influenza infection in patients with cardiopulmonary conditions compared to placebo reduced hospitalization rates significantly (RR = 0.48, 95% CI = 0.28-0.80, I2 = 0%, Figure 2A).

F I G U R E 2
Forest plots comparing effectiveness of oseltamivir in reducing rates of, A, hospitalization, B, respiratory complications, C, asthma exacerbations in people with cardiopulmonary disease who were prescribed oseltamivir compared to those who were not Rates of respiratory complications were also significantly less likely when comparing the two groups (RR = 0.72, 95% CI = 0.59-0.90, 6908 patients, I2 = 44%, Figure 2B). There was no significant difference in rates of asthma exacerbations between the treatment and control group (RR = 0.63, 95% CI = 0.35-1.12, 680 patients, I2 = 0%, Figure 2C). The absolute values, however, suggest a trend favoring oseltamivir when compared to control.
Due to heterogeneity of data and inability to obtain subset data, a pooled analysis to determine time to alleviation of illness and severity of illness could not be performed. However, the absolute values indicated a reduction in the time taken to illness alleviation and reduced severity of illness. 11,22,25,31 Time to alleviation of illness ranged between 37.9 to 148.8 hours in the oseltamivir group and 40.8 to 268.8 hours in the placebo group ( Figure 3A). The severity of illness was based on the area under curve symptom score in one study 25

| Risk of bias within studies
There were five randomized controlled studies of which three had high overall risk of bias and two had unclear risk of bias ( Figure A1).
Four nonrandomized studies had overall unclear risk of bias ( Figure A2). Risk of publication bias was deemed high based on qualitative assessment (Table A3) and quantitative evaluation of funnel plots ( Figure A3).

| Overall quality of evidence
The quality of evidence (Table A3) for the duration of illness and respiratory complications outcomes was judged to be high. The quality of evidence for hospitalisation and asthma exacerbation rate were deemed moderate while the certainty assessment for severity of illness outcome was low.

| DISCUSSION
Our systematic review demonstrated that rates of oseltamivir use in people with cardio-pulmonary conditions with influenza are suboptimal. The two studies from the USA from before the 2009 H1N1 influenza pandemic conducted in high-risk children in both inpatient 28 and outpatient 29 settings, had similar usage rates (25% and 31%). In contrast, the study which included hospitalized children with asthma in Saudi Arabia in the 2009 influenza season had an oseltamivir prescription rate of 100%. 26 The only study with post-2009 pandemic data included in the systematic review had prescription rates of 41% in children with asthma in Spain (2010-2012). 27 As the rate of use ranged between 20% and 40% outside of the pandemic years, it is possible that the high usage rate during the 2009 influenza 26 season may have been due to the perceived risk associated with pandemic influenza leading to the implementation of pandemic strategies, making that year an outlier compared to usual prescribing practices.
Low rate of use could be associated with laboratory testing for influenza. Influenza is not routinely tested for and the lack of confirmatory laboratory data may impact the decision in prescribing oseltamivir despite its recommendation for empirical use in ILI. 3,23 The low usage rate may also be due to lack of data on its effectiveness, particularly regarding the target populations who would benefit most from treatment. 23 Recent systematic reviews in both adult 32 and paediatric 33 non high-risk populations have indicated modest benefit with oseltamivir. However, these studies were not in high-risk populations for whom oseltamivir use is recommended by the WHO and for whom treatment may yield a greater benefit when compared to the general population. On the other hand, our systematic review which included only high-risk patients suggests that oseltamivir was F I G U R E 3 Effectiveness of oseltamivir in reducing, A, time to alleviation of illness (hours), b, severity of illness in people with cardiopulmonary disease who were prescribed oseltamivir compared to those who were not effective in reducing rates of hospitalization and respiratory complications in people with chronic cardiopulmonary disease who were prescribed oseltamivir when compared to those who were not. Our analysis suggests that despite the sub-optimal use, oseltamivir is effective in improving health outcomes in this high-risk population.
Our study demonstrated a 52% reduction in hospitalization rates and a 28% reduction in rates of respiratory tract infections in oseltamivir treated group when compared to placebo in patients with chronic cardiopulmonary disease. We also found that there was a trend suggesting that oseltamivir is effective in reducing the likelihood of asthma exacerbation. These three outcomes are generally indicators of a more severe clinical picture due to influenza infection, suggesting that oseltamivir use in these situations should be indicated to reduce morbidity in high-risk patients.
Other studies have reported variable rates of effectiveness across these outcomes, such as a 34% reduction reported in otitis media incidence in children. 33 This is in contrast to another systematic review which found no significant reduction in hospitalization rates, bronchitis, sinusitis, and otitis media in adults and children. 32 However, these studies again did not focus on the high-risk population for whom use is particularly recommended and could have higher beneficial impact.
Due to a lack of specific data in the studies which we required to conduct analysis, we could not perform meta-analysis for two outcomes: duration of illness and severity of illness. Based on the available data, however, our systematic review found oseltamivir reduced duration of illness by 10.4 to 120 hours in the chronic cardiopulmonary population which suggests that there was a trend favoring a reduced illness duration. Other systematic reviews in the pediatric population without chronic conditions reported a similar reduction of 17.6 hours (CI 95% = 0.62-34.7 hours) 33 and 29 hours (CI 95% = 12-47 hours). 32 Similarly, we found that there was a trend in our data to show that the severity of illness was reduced in the oseltamivir group when compared with placebo.

| Strengths and limitations
Our study is a comprehensive updated synthesis of available data on the use and effectiveness of oseltamivir in high-risk population. This review evaluated the effectiveness of oseltamivir across multiple health outcomes relevant to decision making factors for clinicians.
Our study was limited to studies published in English, however, there was only one study excluded due to this language restriction. 21 Despite extensive search strategy, there were only a few publications eligible for analysis, limiting the power of synthesized results. This highlights the need for more research in this specific population.
There was heterogeneity across the studies in terms of the study populations and study setting which limited pooled analysis of two of the health outcomes. In addition, the studies included in our systematic review had varied definitions of respiratory complications, we combined the upper and lower respiratory tract infections to standardize the outcome, which previous studies have also done. 34 Despite each study having different parameters to define influenza-related complications, oseltamivir demonstrated benefits over no treatment in reducing influenza associated complications including pneumonia and otitis media. 32,33 While some studies were focused on specific populations, others were very broad and subset data could not be obtained to quantitatively assess effectiveness for these studies. For instance, Johnston 11 was focused on children with asthma, while Kaiser's 30 "at-risk" population included both elderly patients and/or people with cardiopulmonary disease, while Piedra 29 included children with all chronic medical conditions in analyses for effectiveness. However, despite these variances, oseltamivir had improved outcomes when compared to control, suggesting a trend toward effectiveness.
In conclusion, our systematic review and meta-analysis suggest oseltamivir is effective in reducing the severity of influenza illness in people with chronic cardio-pulmonary disease, however, the rates of

CONFLICT OF INTEREST
The authors have no conflict of interest to declare.