CONTENTS

• Preamble:  what this chapter is about
• Clinical presentation
• Differential diagnosis:  post-flu pneumonia
• Diagnostic panel
• Treatment
  • Resuscitation
  • Antibiotics
  • Anti-viral therapy
  • Respiratory support
  • Adjunctive therapy
  • Virus-associated hemophagocytic syndrome (VAHS)
• Approach to treatment failure
• Algorithm
• Podcast
• Questions & Discussions
• Pitfalls

---

what this chapter is about

(back to contents)

---

This chapter is about the patient presenting with suspected or definite influenza pneumonia and respiratory failure requiring critical care.  For a broader discussion of community-onset pneumonia, please see this chapter.

Influenza is one of the most dangerous epidemic infectious diseases in modern history, causing ~75 million deaths in the 1918 H1N1 influenza pandemic.  Different strains of influenza return annually, claiming ~25,000 lives per year in the United States alone.  Given that our healthcare systems routinely run near maximal capacity in the winter, it's only a matter of time until influenza out-strips available resources.

---

clinical presentation

(back to contents)

---

influenza pneumonia

• Initial flu-like clinical syndrome deteriorates steadily, with development of dyspnea.
• Some sputum production may occur, which can be bloody.
• This can look a lot like uncomplicated influenza;  hypoxemia and tachypnea are the key distinguishing features.

imaging

• Chest ultrasonography will usually show bilateral B-lines in a patchy distribution (e.g. one interspace will look normal, but the adjacent interspace will have dense B-lines).
• Chest X-ray
  • Will often show patchy bilateral infiltrates.
  • Films may not appear impressive; they will often under-estimate how ill the patient is.
• CT scan
  • Common features include ground glass opacities and multi-focal areas of consolidation.
  • Cavitation or lobar consolidation suggests an alternative diagnosis (e.g. superimposed bacterial pneumonia).

---

differential diagnosis: post-flu pneumonia

(back to contents)

---

Influenza alters the host immune system, predisposing to bacterial pneumonia.  Post-influenza pneumonia is classically associated with infection by streptococcus pneumoniae, group A streptococcus, or staph aureus (including MRSA).

clinical features suggestive of post-flu bacterial pneumonia may include:  

1. Patient develops flu, starts getting better for a few days, then deteriorates.
2. Copious sputum production (not generally a feature of influenza pneumonia).
3. Radiographic features may suggest bacterial pneumonia (e.g. dense lobar consolidation, cavitation).

---

diagnostic panel

(back to contents)

---

diagnostic studies for influenza

• Nasopharyngeal PCR is the front-line test with ~90% sensitivity.  However, if swabbing isn't performed deeply enough, sensitivity may be lower.
• If a nasopharyngeal PCR is negative and yet suspicion for influenza remains:
  • For a non-intubated patient, it may be reasonable to repeat the nasopharyngeal PCR (especially if the sample quality of the first test is unclear).  Sputum may also be tested using PCR.
  • For an intubated patient, tracheal aspirate or bronchoalveolar lavage fluid should be sent for PCR.¹
  • Expanded PCR testing for multiple viruses should be considered (e.g. influenza, RSV, and metapneumovirus).  Other viruses may mimic influenza pneumonia.

other studies to evaluate for alternative (or superimposed) diagnoses:

• Blood cultures x2
• Sputum for gram stain & culture
• Swab nares for MRSA PCR
• Urine streptococcal & legionella antigens
• Procalcitonin
  • Influenza doesn't generally increase procalcitonin.  Thus, procalcitonin may be used to exclude bacterial pneumonia here, the same way it would be in other contexts.²
• Ferritin, liver function tests
• Chest X-ray
• Chest CT scan may be considered (e.g. if chest X-ray suggests nodular infiltrates or cavitation).

---

resuscitation

(back to contents)

---

avoid large-volume resuscitation

• Patients with influenza pneumonia usually die from ARDS, which may be aggravated by large-volume resuscitation.
• Whenever reasonable, fluid administration should be avoided.  For example:
  • For a mildly hypotensive patient, low-dose vasopressors may be preferable to large-volume resuscitation.
  • For a normotensive patient, DO NOT give the Surviving Sepsis Campaign mandated 30 cc/kg fluid bolus.
  • The value of cycling lactate is dubious, and shouldn't be used as a stimulus to give fluid.

---

anti-bacterial therapy

(back to contents)

---

It's generally impossible to immediately differentiate influenza pneumonia from bacterial pneumonia upon clinical grounds (influenza PCR may be positive in both scenarios).  Therefore, empirical coverage of bacterial coinfection should be provided initially (until tmicrobiologic studies return).¹

#1)  macrolide therapy

• Macrolide therapy plays up to three roles here:
  • 1) Coverage for atypical pneumonia.
  • 2) Clarithromycin has direct anti-viral activity against influenza (including both H1N1 and H3N2 types).^3–5
  • 3) Anti-inflammatory effects might limit pneumonitis.
• Clarithromycin is the preferred agent due to its superior activity against influenza.  An appropriate dose may be 500 mg twice daily.
• Azithromycin should be used for patients who cannot receive oral medications (500 mg IV daily).
• A negative procalcitonin cannot exclude atypical pneumonia.  Thus, it's reasonable to continue atypical coverage for a short course even if the procalcitonin is negative (e.g. 5 days).

#2)  beta-lactam 

• Beta-lactam should be used to cover typical bacterial pathogens initially (usually ceftriaxone 1 gram IV daily).
• An anti-pseudomonal beta-lactam (piperacillin-tazobactam or cefepime) may be used in patients with risk factors for pseudomonas, for example:
  • Septic shock due to pneumonia.
  • Structural lung disease (e.g. bronchiectasis or advanced COPD with frequent exacerbations).
  • Broad-spectrum antibiotics for >7 days within past month.
  • Hospitalization for >1 day within past three months.
  • Immunocompromise (e.g. chemotherapy, chronic use of >10 mg prednisone daily).
  • Nursing home resident with poor functional status.
• The need for ongoing beta-lactam therapy may be determined based on procalcitonin, urinary pneumococcal antigen, and sputum culture results.  For uncomplicated influenza pneumonia, the beta-lactam can generally be discontinued within <48 hours.

#3) MRSA coverage

• MRSA is classically associated with post-flu bacterial pneumonia.  MRSA coverage should be used if the clinical history suggests a post-flu bacterial pneumonia.¹
• Linezolid is arguably the best agent here, given evidence of superiority for MRSA pneumonia and reduced risk of nephrotoxicity.^6–16  If there are contraindications to linezolid, vancomycin may be used.
• MRSA coverage may be discontinued if procalcitonin is negative (<0.5 ng/ml) or if no laboratory evidence of MRSA emerges (e.g. nares PCR, blood, and sputum cultures are negative).  Overall MRSA therapy should generally be stopped within <48hr unless there is some laboratory evidence of MRSA.

---

anti-viral therapy

(back to contents)

---

1st line = oseltamavir

• Overall the utility of oseltamavir has been greatly overblown.  However, critically ill patients with known influenza pneumonia should receive oseltamavir, regardless of illness duration (75 mg PO twice daily, with dose reduction in renal failure).¹
• A minimum of five days of therapy should be provided.  Longer courses of therapy can be considered in critically ill patients, in whom influenza viral replication is often protracted.  Repeat virologic testing may help guide treatment duration.¹
• In pregnancy, oseltamavir is preferred therapy.  Increasing the dose to 105 mg BID may be considered given increased renal excretion in pregnancy.¹
• Side-effects of oseltamavir include nausea/vomiting, delirium, and very rarely Stevens-Johnson syndrome.
• More: Medscape monograph on oseltamavir.

2nd line = peramivir

• Peramivir is an intravenous neuraminidase inhibitor with the same mechanism of action as oseltamavir.  This is preferred therapy for patients without enteral access who cannot receive oseltamavir.
• Among outpatients, a single dose of 600 mg IV is used.  For critically ill patients, consideration should be given to administer a multi-day regimen (but the optimal dosing is unknown).¹
• More:  Medscape monograph on peramavir.

unclear role:  zanamivir

• Zanamivir is an inhaled neuraminidase inhibitor.  It may have an advantage against strains of influenza which are resistant to oseltamavir and peramavir.  Currently oseltamavir resistance isn't a problem (with ~98% sensitivity).¹⁷
• Contraindicated in intubated patients.¹
• Doesn't appear useful currently (only available from FDA for compassionate use.)
• More:  Medscape monograph on zanamivir.

unclear role:  baloxavir

• Recently demonstrated to be effective among outpatients.¹⁸
• Still not widely available.  Currently no evidence regarding its use in ICU.
• Baloxavir sounds like a promising agent, hopefully we will hear more about it in the future.
• More:  Medscape monograph on baloxavir.

---

respiratory support

(back to contents)

---

high-flow nasal cannula (HFNC)

• HFNC is the front-line mode of noninvasive respiratory support in most patients with influenza pneumonia.
• BiPAP may be preferred among patients with underlying COPD or asthma, who are experiencing an exacerbation of airway obstruction (i.e. COPD + influenza).  Some patients may benefit from alternating BiPAP and HFNC (e.g. BiPAP may support the work of breathing better, but periods of HFNC may be needed to promote secretion clearance).
• Early institution of noninvasive support might prevent the development of diaphragmatic fatigue with eventual frank respiratory failure.  As such, HFNC may be considered in any patient who has substantial work of breathing (e.g. severe dyspnea or respiratory rate >30 b/m).
• More on the selection of HFNC vs. BiPAP here.

intubation

• Intubation may be indicated for progressive desaturation or worsening dyspnea, despite optimization of noninvasive respiratory support.
• Influenza can impair surfactant production in animal models, thereby leading to lung derecruitment.¹⁹  This supports an early trial of airway pressure release ventilation (APRV), preferably prior to paralysis/proning.^20,21
  • The concept of an early APRV trial to sort out true ARDS vs. derecruitment (pseudoARDS) is explored further here.
  • Evidence regarding APRV in ARDS is explored here.
  • Nuts & bolts guide to APRV is here.

---

adjunctive therapy

(back to contents)

---

steroid

• No randomized controlled trials are available on this.
• Several retrospective time-to-intervention studies have demonstrated a correlation between early steroid administration and mortality.^22–24  However, this doesn't prove a causal relationship.  Early steroid administration may merely be a marker of sicker patients who were treated more aggressively.  Attempts to perform adjusted regression analysis to eliminate confounding variables have yielded conflicting results (unfortunately, regression analysis is limited because it can remove only confounders which are numerically recorded in the dataset).
• Overall, current evidence doesn't support the routine administration of steroid to patients with severe influenza.¹⁷  However, this isn't necessarily a contraindication to steroid if it is indicated for another reason (e.g. COPD or asthma).

ascorbic acid

• Evidence supporting the use of ascorbic acid for ALI/ARDS includes the following:
  • In mouse models of influenza, ascorbic acid improves survival and reduces lung inflammation.^25–27
  • In an RCT involving burn patients, Tanaka 2000 showed that high-dose IV vitamin C improved hemodynamics, improved oxygenation, and decreased the duration of time on mechanical ventilation.²⁸
  • A recent RCT suggested that the combination of IV ascorbic acid (1.5g IV q6), thiamine (200 mg IV q12) and hydrocortisone (50 mg q6) reduced mortality among patients with severe pneumonia.²⁹
  • The CITRIS-ALI trial is ongoing in the United States, involving rather high doses of IV ascorbic acid for treatment of acute lung injury (200 mg/kg/day divided in four doses).
• Currently, administration of moderate-dose IV ascorbic acid is reasonable (e.g. 1.5 grams IV q6hr for up to four days).  If ascorbic acid is administered, co-administration of thiamine (200 mg IV q12hr) may improve efficacy and safety.³⁰
• Equipoise exists and the results from CITRIS-ALI are awaited.

naproxen ?

• Naproxen has in vitro activity which prevents replication of H1N1 and H3N2 influenza.³¹
• It's possible that the combination of naproxen with clarithromycin and oseltamavir could create an effective triple-therapy against influenza.  One single-center unblinded RCT found that triple therapy improved mortality compared to monotherapy with oseltamavir alone.³²
  • Their cocktail was oseltamavir 75 mg BID, naproxen 200 mg BID, and clarithromycin 500 mg BID.  Naproxen and clarithromycin were used only for the first two days, whereas oseltamavir was continued for a five-day course.
• The use of naproxen for influenza remains unclear.  Naproxen may be considered if the following criteria are met:
  • Symptom duration <96 hours.
  • Laboratory-confirmed influenza (ideally H3N2 or H1N1 strains).
  • Pulmonary infiltrates on chest X-ray or CT scan.
  • Creatinine clearance >30 ml/min, no acute kidney injury or hypoperfusion, no other nephrotoxic medications.
• If naproxen is used, it should be limited to a two-day course.

---

virus-associated hemophatocytic syndrome (VAHS)

(back to contents)

---

what is VAHS ?

• Hemophagocytic lymphohistiocytosis (HLH) is a state of immune hyperactivation wherein macrophages phagocytize blood cells (including erythrocytes and neutrophils).  It can be caused by a wide variety of underlying disorders (e.g. malignancy, autoimmune disorders, or hematologic malignancy).
  • Virus-associated hemophagocytic syndrome (VAHS) refers to hemophagocytic lymphohistiocytosis due to viral infection.
• A background discussion on the role of hemophagocytic lymphohistiocytosis in critical illness is located here.

epidemiology

• VAHS appears to occur predominantly with H1N1 and H5N1 types, but it can occur with other types as well (H3N2).^33,34
• One prospective case series detected VAHS in 1/3 of critically ill adults with H1N1.³⁵
• Among autopsies of patients who died from H1N1 influenza, histological evidence of hemophagocytosis is present in 68/87 cases (78%).^36–38
• Usually VAHS seems to be a delayed complication, arising ~1-2 weeks after ICU admission.  However, several reports describe patients presenting to the hospital with VAHS present upon admission.^39–41

diagnostic features

• Diagnosis of VAHS is challenging, because it overlaps with the features of severe infection.  Features suggesting VAHS include the following:
  • Fever
  • Splenomegaly
  • Cytopenias in at least two cell lines (anemia, thrombocytopenia, neutropenia)
  • Fibrinogen <150 mg/dL, disseminated intravascular coagulation
  • Ferritin >>500 ng/mL
  • LDH elevation >1,000 U/L
  • Liver function test abnormalities (may be severe)
  • Hypertriglyceridemia (>265 mg/dL or >3 mM/L)
• In suspect cases, serial monitoring of HLH labs may facilitate prompt diagnosis (ferritin, triglycerides, fibrinogen, and liver function tests).

diagnostic criteria

• The traditional definition of HLH is based upon the HLH-2004 criteria shown above.  HLH-2004 criteria are unhelpful in critically ill adults for several reasons:⁴²
  • 1) These criteria were developed for use in a clinical trial of pediatric HLH.
  • 2) They rely on tests which are impossible to obtain rapidly (or ever).
  • 3) By the time HLH-2004 criteria are officially “positive” the patient may be moribund and beyond the point of optimal intervention.
• The modified 2009 criteria listed above appear superior to the traditional HLH-2004 criteria.  An autopsy series of patients dying from H1N1 influenza found the modified HLH criteria more likely to be satisfied than HLH-2004 criteria (table below).^38,43
• Determining the H-score might be a preferable approach.⁴⁴  This is easily obtained using an online calculator (units conversion may be needed for triglycerides:  1.5mM = 133 mg/dL; 4mM = 354 mg/dL).  The optimal cutoff value is unclear (with potential cutoffs ranging between 138-169).⁴⁵
• Some clinicians are moving past rigid criteria, with initiation of therapy before diagnostic criteria are met.⁴⁶

treatment

• There is no consensus regarding the optimal treatment for infection-induced HLH.  Steroid is a mainstay of treatment for HLH and should be given.^35,47  Four basic treatment strategies are as follows:
• #1:  Moderate dose steroids alone
  • A moderate dose of steroid may consist of  dexamethasone 10 mg/m2 body surface area daily or ~125 mg methylprednisolone daily (based on the HLH-2004 protocol).⁴⁸
  • This might be adequate for early or impending HLH, but not fulminant HLH.
• #2:  Pulse-dose steroid
  • Some case reports describe success from pulse-dose steroid in HLH due to influenza H1N1 (e.g., 500-1000 mg methylprednisolone daily for three days).^39,41
  • Advantages include that this is inexpensive, readily available, and has a reasonable side-effect profile (particularly at 125 mg IV q6hr, a dose commonly used for obstructive lung disease).
• #3:  Moderate dose steroid plus Anakinra
  • Anakinra is a fairly safe agent which reduces inflammation via inhibition of IL-1 receptors.  Re-analysis of a multi-center RCT of septic adults with features of HLH detected mortality benefit from Anakinra.⁴⁹
  • This regimen may be the best in terms of combining safety and efficacy; the main limitation is availability of anakinra.
•  #4:  Moderate dose steroid plus Etoposide
  • Etoposide is a chemotherapeutic agent, which has been used for pediatric forms of congenital HLH.  Case reports describe the successful use of etoposide for influenza-induced HLH.^34,50
  • Etoposide's side effect profile isn't as benign as biologic agents such as anakinra.  However, only one or two doses may be needed to facilitate recovery, with a low cumulative exposure.⁵⁰  Compared to biologic therapies, etoposide is more widely available and more affordable.
• More complex regimens are possible as well, for example the strategy shown below.⁴⁶  A recent case series of adults with acquired HLH reported reasonably successful use of anakinra combined with pulse-dose steroid and IVIG.⁴²  IVIG is generally safe and sometimes added to multimodal regimens.^40,42,51  Ruxolitinib is an extremely promising option which could become a front-line therapy in the near future.⁵²

---

approach to non-responding patient

(back to contents)

---

definition of non-response?

• Unclear.  Ideally patients will show some improvement within 2-3 days.¹  However, critically ill patients may take longer to recover.
• If the patient improves but then deteriorates, this is a signal that something is amiss.

differential diagnosis may include:  

• Virus-associated hemophagocytic syndrome (discussed above)
• Progression of influenza pneumonia into organizing pneumonia^53,54
• Superimposed bacterial pneumonia (either community- or hospital-acquired)
• Invasive aspergillosis^55,56
• Ventilator-induced lung injury with progressive ARDS
• Influenza-induced myocarditis
• Superimposed pulmonary embolism (not uncommon in severe influenza with profound inflammation)
• Exacerbation of chronic disease (e.g. COPD, asthma, or heart failure)
• Pleural effusion(s)
• Post-influenza Guillain-Barre Syndrome
• Ascorbic acid deficiency⁵⁷

evaluation may include:

• Repeat infectious workup (e.g. cultures, bronchoscopy)
• Echocardiogram to evaluate for myocarditis or underlying heart failure
• Bedside ultrasonography to exclude pleural effusion
• CT chest with contrast to evaluate for PE or other pulmonary complications
• Hemophagocytic lymphohistiocytosis labs (ferritin, liver function tests, fibrinogen, triglyceride level)

---

algorithm

(back to contents)

---

---

podcast

(back to contents)

---

Follow us on iTunes

---

questions & discussion

(back to contents)

---

To keep this page small and fast, questions & discussion about this post can be found on another page here.

• Avoid high-volume resuscitation if at all possible.  Fluid isn't the answer for severe influenza.
• Don't miss post-flu pneumonia.  Simply because the patient has PCR-positive influenza doesn't necessarily exclude the presence of a concomitant bacterial pneumonia.
• Have a high index of suspicion for MRSA pneumonia among patients with post-flu pneumonia.
• Among ICU patients who aren't responding well to therapy, consider unusual entities such as influenza-induced hemophagocytic lymphohistiocytosis and influenza-induced organizing pneumonia.  These diseases are highly treatable if diagnosed promptly.

CDC resources on influenza

• Diagnostic testing
• Antiviral therapy
• Epidemiologic & outbreak data

Related

• Influenza (Chris Nickson, LITFL)
• Naproxen/oseltamavir/clarithromycin triple therapy (PulmCrit)
• Understanding sepsis-HLH overlap syndrome (PulmCrit).

Acknowledgement:  Thanks for an enlightening conversation with Samuel Merrill MD PhD, HLH specialist at Johns Hopkins.

References