Journal Club FLUshing Out the Evidence for Neuraminidase Inhibitors in the Treatment of InFLUenza Manish Khullar BSc.(Pharm) Sukhjinder Sidhu BSc.(Pharm) Interior Health Pharmacy Residents Infectious Disease Rotation May 1, 2014
Dec 31, 2015
Journal ClubFLUshing Out the Evidence for Neuraminidase
Inhibitors in the Treatment of InFLUenza
Manish Khullar BSc.(Pharm)Sukhjinder Sidhu BSc.(Pharm)
Interior Health Pharmacy ResidentsInfectious Disease Rotation
May 1, 2014
2014
The Controversy
Influenza• Signs/symptoms– Fever– Sore throat– Rhinitis– Nonproductive cough– Myalgias– Malaise
• Complications – Secondary infections (bacterial pneumonia, otitis media)– Hospitalizations– Death
How does Oseltamivir Work?
Our PICODesign Meta-analysisPopulation Hospitalized patients with confirmed or suspected exposure
to influenza
Intervention Oseltamivir 75 mg PO bid x 5 daysComparator PlaceboOutcome Reducing mortality
Reducing morbidity (i.e. pneumonia)Reducing length of hospital stayMinimizing adverse events
Cochrane PICOD Meta-analysisP Hospitalized patients with confirmed or suspected exposure to
natural occurring influenza
I Oseltamivir- 20 RCTs
- 11 treatment (adults)- 4 treatment (children)- 5 prophylaxis
- F/U 6-42 days
Zanamivir- 26 RCTs
- 14 treatment (adults)- 2 treatment (children)- 10 prophylaxis
- F/U 5-35 daysC PlaceboO Many outcomes for prophylaxis, treatment, on-treatment and off-
treatment- Not mortality
Title
• Title does not state it’s a meta-analysis or systematic review
• Authors are Cochrane Collaboration Group – Can imply it’s at least a systematic review
Introduction
Rationale• Influenza antivirals are commonly used and stockpiled• Previous reviews have risk of reporting bias• On list of WHO essential drugsObjectives• potential benefits and harms of NIs for influenza in all
age groups… all clinical study reports of published and unpublished R, PC trials and regulatory comments
MethodsEligibility Criteria• Studies– NIs RCTs for prophylaxis, post-exposure prophylaxis
and treatment of influenza• Published and unpublished trials• Manufacturer-funded and non-manufacturer funded
clinical trials
– No specific length of follow-up considered– Exclusion criteria not identified
THOUGHTS?
Methods
Eligibility Criteria• Participants– Previously health children and adults– Exposed to naturally occurring influenza with or
without symptoms– Excluded people with illnesses with more
significant effects on the immune system (i.e. malignancy or HIV infection)
Methods
Information Sources• Electronic databases– CENTRAL, MEDLINE, MEDLINE (Ovid), EMBASE,
PubMed (NOT MEDLINE), DARE, NHSEED, HEED– January 2010 – July 2013
• Clinical study reports– Extensive searches conducted
• Regulatory information searches– Extensive searches conducted
MethodsSearch Strategy
Methods
Study Selection• 2 authors reviewed title & abstracts• 4 authors independently read all data– definitely include; definitely exclude; need more
information• 2 more authors reviewed for inclusion in Stage 1
MethodsStudy Selection• Stage 1 assessing the reliability and completeness
of trial data– authors discussed face-to-face each trial with comments
and other information from regulatory sources– decision to whether move trial to stage 2 via consensus
• Stage 2 satisfied following criteria:– completeness – CONSORT-specified methods &
specified results– internal consistency – external consistency
MethodsData Collection Process• Utilized a modified CONSORT statement-based
extraction template – 2 authors each searched oseltamivir and zanamivir trials– Disagreements were resolved amongst each other in
oseltamivir group and by a 3rd author in the zanamivir group
• For clinical study reports, complete list of trials were sent to manufacturers asking them to check accuracy and completeness of their list
Methods
Risk of Bias in Individual Studies• Used Cochrane “Risk of bias” tool• Review author judged risk of bias• Bias assessed at outcome and study level– Studies included had high risk of bias– Some outcomes from studies were poorly
documented/collected
Methods
Synthesis of Results• Chi2: test for heterogeneity • I2: level of statistical heterogeneity– Threshold for significance unknown
• Tau2: estimate of between-study variance• Combining data using random-effects
approachWHAT SHOULD THEIR THRESHOLD BE?
MethodsRisk of Bias Across Studies• Included unpublished trials
MethodsAdditional analyses• Subgroup analysis to investigate high estimates of
heterogeneity• Meta-regression to investigate pneumonia
heterogeneity • Sensitivity analysis– Fixed-effect method of Mantel and Haenszel to
supplement primary analyses using random-effects method
– Peto’s method used when sparse data and borderline sensitivity WHAT ANALYSES WOULD BE BENEFICIAL?
ResultsStudy Selection 208 studies identified
form various sources
123 studies excluded19 studies awaiting classification
66 studies for which clinical study reports requested
53 studies met eligibility5 trials excluded due to incompleteness2 trials excluded as didn’t fit inclusion criteria 46 trials included
- 20 oseltamivir- 26 zanamivir
ResultsRisk of Bias within Studies• Presented selection, attrition, reporting,
performance and detection bias– To address issue of reporting bias, they ignored
published trial reports if clinical study reports and regulatory information were available
– Random sequence generation missing in many trials– Blinding may have been affected in many trials– Placebos may have contained active substances
Results
• No analysis conducted on mortality outcome– Discussed deaths in the oseltamivir and zanamivir
arms, but no statistical analyses completed
THOUGHTS?
Time to First Alleviation of Symptoms in Adult Treatment (ITT Population)
Oseltamivir vs. Placebo
Clinically significant reduction?When to use random vs. fixed effects?
Time to First Alleviation of Symptoms in Adult Treatment (ITT Population)
Oseltamivir vs. Placebo
Complication: PneumoniaOseltamivir vs. Placebo
Complication: PneumoniaOseltamivir vs. Placebo
NNT = 100
Nausea in Adult Treatment (On-Treatment)Oseltamivir vs. Placebo
Nausea in Adult Treatment (On-Treatment)Oseltamivir vs. Placebo
NNH = 28
Vomiting in Adult Treatment (On-Treatment)Oseltamivir vs. Placebo
Vomiting in Adult Treatment (On-Treatment)Oseltamivir vs. Placebo
NNH = 22
Results
Risk of Bias Across Studies• No funnel plot or Egger’s test done to rule out
publication bias – Made every effort to minimize publication bias by
including non-published trials• No discussion on bias across studies, only bias
within studies
SHOULD WE BE WORRIED ABOUT PUBLICATION BIAS?
ResultsAdditional Analysis• Subgroup analysis– Time to first alleviation of symptoms in adults by
infection status– Pneumonia (diagnosis vs. non-diagnosis)
Discussion
Summary of Evidence• NIs have small, non-specific effects on reducing
time to alleviation of influenza-like illness symptoms in adults
• Treatment trials… do not settle the question of whether complications of influenza, such as pneumonia are reduced
• Use of oseltamivir increases the risk of adverse events such as nausea, vomiting…
Discussion
Conclusions• … appears to be no evidence for patients,
clinicians or policy-makers to use these drugs (NIs) to prevent serious outcomes
• Implications for future research– More effective preventative measures– Early identification of complications
Limitations• Did not formally assess mortality outcome• Effect of complications was based on unclear and
potentially unreliable definitions• Included many trials with high risk of bias– Affects validity of the results
• Authors do not summarize results in the context of observational study results that are driving standard of care in influenza treatment
• A generalized conclusion is made without taking severity of illness into account
Our Conclusions
• Population studied was broad– Not generalizable to specific groups– Doesn’t address the question of whether to give
NIs in high risk patients• More reviews needed to draw definite
conclusions about mortality and reductions in complications
• More reviews needed in severe influenza (i.e. critical care patients)