Clinician-initiated research on treating the host response to pandemic influenza

ABSTRACT

To prepare for the next influenza pandemic and other emerging virus diseases, scientists and health officials are focused on developing new vaccines and treatments that target these viruses. Ideally, these interventions could be highly effective, but for many practical reasons these “top down” efforts are unlikely to provide clinicians with what they will need to manage their patients. As a “bottom up” alternative, combinations of generic drugs like statins and angiotensin receptor blockers (ARBs) might be used to treat the host response to infection. These drugs counteract endothelial dysfunction, a central abnormality in these diseases. Observational studies in patients with influenza, pneumonia, sepsis and Ebola suggest they might work.

During the 1918 influenza pandemic, children were infected more frequently than adults, but their mortality rate was much lower. Their survival was probably due to better tolerance (reduced pathogen damage), not greater resistance (reduced pathogen burden). The same pattern of susceptibility characterizes other infectious diseases, and it probably reflects the heritage of human evolution. Drugs like statins and ARBs can metabolically reprogram the host response and improve tolerance to infection.

Treating the host response is not on the research agendas of international agencies responsible for pandemic preparedness. Consequently, clinicians might have to undertake clinical trials in patients hospitalised with seasonal influenza, pneumonia and sepsis in order to show convincingly whether treating the host response would work. Most candidate generic drugs are inexpensive, widely available, known to be safe and used by clinicians every day. Demonstrating their efficacy would mean that patients in all countries would have access to treatment on the first pandemic or epidemic day.

KEYWORDS:

• angiotensin receptor blockers
• Ebola
• host response
• pandemic influenza
• statins

Introduction

In 2018, we will commemorate the 100^th anniversary of the 1918 influenza pandemic, an event that caused an estimated 50–100 million deaths worldwide. Several years ago, a physician who was a medical student in Philadelphia in 1918 described what it was like to care for patients during the pandemic.¹ All clinicians should read his article: it is harrowing. Because clinicians today might experience something similar with the next influenza pandemic, it is also deeply disturbing.

The challenge of responding to the next influenza pandemic

A few years ago, the question was asked, “How will physicians respond to the next influenza pandemic?”² There is still no practical answer to this question.

In an interview with The Washington Post published on January 16, 2017, Thomas Frieden, who was stepping down as Director of the Centers for Disease Control and Prevention (CDC), was asked, “What scares you the most? What keeps you awake at night?”³ He replied, “The biggest concern is always for an influenza pandemic. … (It) really is the worst-case scenario. If you have something that spreads to a third of the population and can kill a significant proportion of those it affects, you have the makings of a major disaster.” In commenting on the US response to the 2009 influenza A(H1N1) pandemic, he added, “when the vaccine became available, we were able to send out over 300,000 doses to more than 70,000 places without a glitch … you don't hear about that because it went so well. The point is, it was an everyday system. And because it could be scaled up, it went well.”³

In 2009, pandemic vaccination was clearly effective for individuals. A comprehensive systematic review and meta-analysis has shown that monovalent pH1N1 vaccination prevented laboratory-confirmed clinical illness.⁴ Vaccination effectiveness (VE) was better for adjuvanted than for nonadjuvanted vaccine (VE = 80% vs. 60%), and it was also effective in preventing laboratory-confirmed hospitalization (overall pooled adjusted VE = 61%).⁴ Adjuvanted vaccines were more effective in children than in adults for both outcomes.⁴ Vaccination effectiveness in reducing influenza-related mortality could not be determined.

Unfortunately, the impact of pandemic vaccination on populations was negligible. In his interview, Frieden neglected to mention that in the US, no pH1N1 vaccines were available during the first pandemic wave and for most of the second wave. According to CDC investigators, pandemic vaccination affected only 2–4% of all pandemic cases, hospitalizations and deaths nationwide.⁵ Millions of doses arrived too late to do much good and had to be destroyed. All other countries had similar experiences, or worse.²

The US Department of Health and Human Services recently released an update of its Pandemic Influenza Plan.⁶ Federal officials estimate that a severe pandemic would cause almost six million hospital admissions and 770,000 deaths (Table 1).

Table 1. Estimated numbers of hospitalizations, ICU admissions and deaths due to a future influenza pandemic in the US.

Pandemic severity	Hospitalizations	ICU admissions	Deaths
Moderate	1,200,000	240,000	72,000
Severe	5,800,000	1,700,000	770,000
Very severe	11,500,000	3,500,000	1,930,000

Adapted from Table A.1. in reference.⁶ The estimates assume that 30% of the US population will develop clinical illness. The estimates do not account for any public health interventions that might be implemented during a pandemic. By comparison, in 2014–2015, when seasonal influenza vaccination was not very effective, influenza A(H3N2) caused an estimated 924,206 hospital admissions in the US.

Until recently, the influenza A(H5N1) virus has been regarded as the primary threat to cause the next pandemic, but now an influenza A(H7N9) pandemic is considered to be more likely. A recent report from China described 1220 patients who had been hospitalized with H7N9 influenza over the past five years.⁷ Approximately half of them had no high-risk condition and 55% required ICU admission. Although 74% of these patients were treated with antiviral drugs, their overall mortality was 40%. Many experts believe a similar mortality rate will be seen with the next pandemic. Influenza virologists have reported that only three mutations would be required to switch the H7N9 virus to human-type receptor specificity.⁸ This would bring us even closer to the next pandemic.

Pandemic vaccination alone will be inadequate

Plans to confront the next influenza pandemic have focused on developing new vaccines and antiviral agents.^6,9 However, producing a pandemic vaccine will be complex and difficult (see reference;⁹ especially Annex 3. Timelines of pandemic vaccine production). A candidate vaccine virus (CVV) must be generated, and even if everything goes well this will still take several weeks. Because the CVV must be produced in embryonated eggs, the antigen yield must be sufficiently high. The availability of reagents, regulatory requirements for immunogenicity and safety trials, regulatory approval for an antigen-sparing adjuvant and the need for one or two doses must all be considered. Purchasing contracts must be negotiated and the logistics for vaccine distribution and administration worked out. In 2009 all of these factors contributed to the almost six-month delay in launching pandemic vaccination programs.

The updated (2017) US Pandemic Influenza Plan calls for the first doses of vaccine to be produced within three months, followed by rapid scale up of production, distribution and administration.⁶ This goal is aspirational, and it almost certainly won't be achieved. According to the World Health Organization (WHO), it will still take approximately six months after the declaration of a pandemic before the first doses of pandemic vaccine can be produced.⁹

New influenza viruses are not the only threats to cause epidemics and pandemics. WHO has identified 11 emerging viruses (including Ebola virus) for which there is an urgent need for research.¹⁰ The Coalition for Epidemic Preparedness Innovations (CEPI) was recently established to “stimulate, finance and coordinate the development of vaccines against epidemic infectious diseases, especially in cases in which market incentives alone are insufficient”.¹⁰ Among other things, CEPI will focus “its initial investments on essential gaps in product development, especially (phase 2 trials)” and support “technical and institutional platforms that can be used for rapid vaccine development”.¹⁰ CEPI's budget calls for $1 billion, which will be spent on developing vaccines against two or three high-priority pathogens that could be entered into clinical trials during the initials stages of a new outbreak. Yet the obstacles that currently compromise rapid pandemic influenza vaccine production will likely affect vaccines for any emerging virus disease. Furthermore, no one can know in advance whether CEPI's chosen vaccine candidates will match the next emerging virus.

When the next pandemic influenza virus emerges, clinicians who want to improve patient outcomes will have to rely on something other than vaccination. This will be especially important for those who live and work in low- and middle-income countries where seasonal influenza vaccines are seldom used.¹¹ Although antiviral drugs specific for influenza will be available in developed countries, they will be largely unavailable in the rest of the world. Unless given early in the course of illness, they may have little effect on pandemic mortality.¹² Moreover, antiviral resistance will always be a threat. Because a very large number of patients might have to be treated at any one time, strategies that depend on the availability of convalescent plasma, mechanical ventilators and extra-corporeal membrane oxygenation will be impractical. For other emerging virus diseases, the limitations of vaccination, antiviral agents and other interventions will be even more severe.

For the next influenza pandemic, clinicians will have no alternative but to use whatever drugs are available on the very first day. Because the next pandemic might arise in a low- or middle-income country, it will be important for these drugs to be inexpensive generics that are widely available and familiar to clinicians.¹³ Some of these drugs might target the host response to infection, not the pandemic virus.¹⁴

The scientific rationale for treating the host response

There is an enormous experimental and clinical literature on the cellular and molecular pathogenesis of influenza, acute respiratory distress syndrome (ARDS), and other forms of acute critical illness, including sepsis. Changes in pro- and anti-inflammatory cytokines and other biomarkers, along with alterations in redox metabolism, autophagy, apoptosis and mitochondrial function, are among the many factors that have been explored. There is no consensus on which of these changes are critical determinants of fatal vs. nonfatal disease. However, influenza virus loads in mice with fatal and nonfatal disease are similar,¹⁵ suggesting that host factors must be important determinants of who lives and who dies.

Recent studies have focused attention on changes in metabolic programming that are associated with the host response to infection.¹⁶ In addition, for influenza, sepsis and other infectious diseases, a major abnormality is endothelial dysfunction and the loss of vascular barrier integrity.^17,18 Endothelial dysfunction is also seen with all of the emerging virus diseases identified as threats by WHO.¹⁷ Among other things, endothelial dysfunction is associated with alterations in the ACE2/angiopoietin-(1–7)/Mas and angiopoietin/Tie2 signaling axes.^17-21 Treatment with immunomodulatory drugs such as statins and angiotensin receptor blockers (ARBs) can modify these signaling axes and restore vascular barrier integrity.^17,19 These findings suggest that treating the host response could be an effective way to reduce the morbidity and mortality of pandemic influenza and other emerging virus diseases.

Observational studies suggest treating the host response could be beneficial

More than a decade ago, it was suggested that statins might be used in the treatment and prophylaxis of pandemic influenza.²² Since then, numerous clinical studies, most of them observational, have provided further (although not yet convincing) support for treating the host response to seasonal influenza, sepsis, ARDS and other forms of acute critical illness.^13,14,17

Two observational studies of patients hospitalised with laboratory-confirmed seasonal influenza have reported that inpatient statin treatment was associated with reductions in 30-day all-cause mortality (Table 2).^23,24 The two studies were conducted using the same database. The investigators who conducted the second study confirmed the results of the first study and obtained better results. However, they concluded that because of hidden bias, their findings did not support “using statins as adjunct treatment for preventing death among persons hospitalized for influenza”.²⁴ More recently, a secondary analysis of a case test-negative observational study of influenza vaccination effectiveness showed that outpatient statin treatment reduced VE in preventing laboratory-confirmed influenza from 46% to 25% (discussed in reference²⁵). Paradoxically, in persons who had not been vaccinated, statin treatment was protective, reducing the occurrence of influenza by 37%.²⁵

Table 2. Inpatient statin treatment of patients hospitalised with laboratory-confirmed influenza.

Study (reference)	Method	No. of subjects cases/controls	Adjusted OR/HR (95% CI)	Reduction in 30-day all-cause mortality
Vandermeer, et al.^23	Retrospective cohort	1030/2013	OR = 0.59 (0.38–0.92)	41%
Laidler, et al.^24	Case control	670/670	HR = 0.41 (0.25–0.68)	59%

From reference,¹⁷ with permission. The two studies were performed using the same database. Both used propensity scoring and the analyses adjusted for age, sex, underlying high-risk conditions and previous influenza vaccination and antiviral treatment. OR = odds ratio; HR = hazard ratio; CI = confidence interval. See references^23,24 for details.

A large observational study showed that in patients hospitalised with community-acquired pneumonia (CAP), inpatient treatment with statins, ACE inhibitors and ARBs were each associated with a reduction in 30-day all-cause mortality.²⁶ Perhaps more important, outpatient treatment of CAP patients with both a statin and an ARB was associated with a 59% reduction in mortality, almost twice that of single drug treatment.^17,27 This is not surprising; cardiovascular investigators have known for many years that a statin/ARB combination is more effective than treatment with either drug alone.^17,28

Treating the host response might also benefit patients with other emerging virus diseases.¹⁷ For example, in patients with Ebola virus disease, combination treatment with a statin (atorvastatin) and an ARB (irbesartan) appeared to dramatically improve patient survival.^17,29 In Sierra Leone, local clinicians treated approximately 100 Ebola patients with this combination, and only three patients are known to have died. There was no logistical or financial support for a proper clinical trial, and Ebola scientists and international health officials (including those at WHO) dismissed or actively opposed this approach to treatment; serendipity and a $25,000 private donation made it possible.²⁹ Unfortunately, the clinicians who treated these Ebola patients refused to release information on their experience, although their letters and memoranda documented “remarkable improvement” in their patients.^17,29

Most reports of the “lessons learned” from clinical research during the Ebola outbreak in West Africa have ignored what happened in Sierra Leone.¹⁷ One report concluded, “… it is impossible to draw any meaningful conclusion” from this experience.³⁰ Yet, in contrast to the statin/ARB experience, several clinical trials were undertaken of experimental treatments that target the Ebola virus. All of the trials received generous international support, yet none of the treatments led to meaningful improvement in patient survival.^30,31

Although the statin/ARB treatment experience in Sierra Leone was poorly documented, it would be a mistake to dismiss it out of hand. A case series “may be the ‘lowest’ or the ‘weakest’ level of evidence”, but it is often “the ‘first line of evidence’ ”.³² Moreover, a case series can be especially valuable when it uncovers a surprising finding,³² something that rarely happens in a formal clinical trial.³³ Furthermore, whenever outcomes are 5–10 times better in treated patients compared with historical controls, one must assume an “implausibly large” level of confounding in order to conclude that the results are not meaningful.³³ Finally, according to a report from the National Academies of Science, Engineering and Medicine, there are five “special situations” in which a non-randomized trial could be justified (Table 3).³⁴ All of these special situations were met by statin/ARB treatment in Sierra Leone.^17,29,35

Table 3. Special situations in which an uncontrolled Phase 3 clinical trial of an Ebola treatment might be considered.

	Situation met by
Special situations	Statins	Angiotensin receptor blockers
1. No other treatment is appropriate to use as a control	Yes	Yes
2. Patients not receiving treatment have a uniformly poor prognosis	Yes	Yes
3. No potential side effects would compromise the potential benefit of treatment	Yes	Yes
4. Expected benefit will be sufficiently large to make interpretation of the results unambiguous	Yes^*	Yes^*
5. Scientific rationale is sufficiently strong so a positive result would be widely accepted	Yes^**	Yes^**

From reference,³⁵ with permission.

* The observed benefit of treatment in Sierra Leone was sufficiently large so that interpretation of the results should have been viewed as unambiguous.²⁹

** The scientific rationale for treatment is sufficiently strong but it has not been well understood by Ebola scientists and health officials, so the results of Ebola treatment in Sierra Leone have been dismissed.¹⁷

Lower mortality in children compared with adults reflects tolerance, human evolution and suggests an approach to treatment

During the influenza pandemic in 1918, children were infected more frequently and had much lower mortality than adults, leading some to observe that, “for reasons that are as mysterious today as they were in 1918, they were able to cope with the disease much better than their adult counterparts”.³⁶ The same “pattern of susceptibility” has been seen in a wide range of other infections,^14,36 and the “change in susceptibility occurs around the time of puberty”.³⁶ This pattern of susceptibility has been replicated in mice: influenza mortality was much lower in pre-pubertal mice than it was in mice that were post-pubertal.³⁷

Lower mortality from infection in children could reflect either enhanced resistance (i.e., a greater ability to reduce the pathogen burden) or better tolerance (i.e., a greater ability to reduce its damaging effects).³⁸ In the mouse model of influenza described above, levels of virus replication in the pre- and post-pubertal mice were the same, a strong indication that pre-pubertal mice were more tolerant, not more resistant. Better tolerance before puberty might be due to different patterns of metabolic reprogramming following infection.^16,38 For example, metabolic reprogramming in macrophages is associated with the anti-inflammatory effects of IL-10.¹⁶ More than a decade ago, a study of human macrophages obtained from children and adults showed that children generated greater IL-10 responses to an inflammatory stimulus.³⁹ Thus, patterns of metabolic reprogramming found in children but not adults might explain why children are more tolerant to a wide range of infectious diseases.^14,35

Statins and ARBs (and several other drugs, including glitazones and metformin) have known immunometabolic effects.^12,16 The clinical benefits seen with their use in influenza,^22-24 community-acquired pneumonia^25,26 and Ebola^16,28 might reflect metabolic reprograming that leads to better tolerance of infection. No reports indicate this has been studied experimentally in both children and adults, with possibly one exception. Almost a decade ago, surgeons who were involved in liver transplantation in children and adults sought to better understand hepatic inflammation by studying a mouse model of hepatic ischemia-reperfusion injury.⁴⁰ They showed that “children” had much milder inflammatory changes compared with “adults”. Adult mice were then pre-treated with rosiglitazone, a drug with known anti-inflammatory (immunometabolic) activities. Instead of developing extensive inflammation, treatment was associated with milder inflammation like that seen in younger mice. In other words, adult mice became more tolerant. Since the immunometabolic changes associated with puberty appear to have been hardwired by evolution, the investigators in effect “rolled back” evolution.^11,12 This study suggests the possibility that immunomodulatory drugs might be used to “roll back’ the harmful inflammatory changes seen in patients who become critically ill with pandemic influenza and many other emerging virus diseases.^12,16,34

Compelling arguments have been made for both the scientific promise and the enormous practicality of treatment that target the host response.^14,17 Yet, leading scientists and health officials who support their work (and count on their advice) apparently have shown little or no interest in this idea, and in many instances they have actively opposed it.¹⁷ The reasons for their opposition may have less to do with science itself, and more to do with social bias and herding behaviour among scientists themselves.⁴¹ For many, the reputational and financial costs of adopting this new idea may simply be too great.

Clinician-initiated research on treating the host response

Preparations for pandemic influenza and other emerging virus diseases involve scientists and institutions that are focused on developing new vaccines and drugs that target these viruses.^6,9,10,17 Research progress will be slow, regulatory hurdles and obstacles to production will inevitably arise, and costs will be high. Moreover, the identity of the next emerging virus cannot be predicted. When clinicians are faced with the next emerging virus, the “top down” efforts of scientists and institutions are unlikely to have provided them with what they will need to manage their patients. They will need a “bottom up” approach to treatment that is familiar and universally available. This will be especially important for clinicians who practice in low- and middle-income countries, for they will be especially hard pressed.

Recommendations for research on treating the host response cannot be found in the agendas of institutions responsible for pandemic and epidemic preparedness.^{6,13,14,17,42} The 2017 update of the US Pandemic Influenza Plan says that developing “host-targeted therapeutic approaches may mitigate the emergence of antiviral drug resistance ” (page 23 in reference⁶). Clearly, these new treatments are to be developed for their effects on the virus, not for how they might modify the host response of the patient who is infected.

In 2014, clinicians were encouraged to ask, “why influenza scientists and health officials who support their work have not undertaken pragmatically focused laboratory and clinical research to see if statins and other promising immunomodulatory agents could be used to reduce influenza-related mortality. There is no guarantee that any of these drugs will work, but physicians will never know unless those responsible for pandemic preparedness recognize and act on the extraordinary possibility that these agents might save lives”.² Clinicians still need an answer to this question.

It is becoming increasingly evident that if any research is to be undertaken on treating the host response, clinicians might have to do it on their own (Table 4). In many instances, they will need guidance from local experts (e.g., intensive care specialists, statisticians). They will also need to call upon the knowledge of scientists in other disciplines;⁴³ for example, vascular biology, mitochondrial biogenesis, and immunometabolism.^14,17 Logistical and financial support will be essential. This will have to be obtained locally or from public agencies or non-governmental organizations because companies will not support research on off-patent drugs in which they have no commercial interest. Initial clinical trials could be conducted in patients hospitalized with everyday illnesses like seasonal influenza, sepsis and pneumonia. If case fatality rates are high, sample sizes for some of these studies need not be large.^13,17,33,35

Table 4. Considerations for clinicians who will undertake research on treating the host response to pandemic influenza and other emerging virus diseases.

Choose drugs that are:

– known to modify mechanisms involved in the host response to infection

– safe in patients with critical illness

– produced as inexpensive generics

– widely available in low- and middle-income countries

– familiar to practicing physicians

– likely to affect important outcomes (e.g., 28-day all-cause mortality)

Before the emergence of a new virus:

– undertake observational studies and prospective clinical trials of treatment in patients hospitalised with every day illnesses; e.g., seasonal influenza, community-acquired pneumonia, sepsis

– study outcomes in children and adults

– evaluate outcomes in hospitalized patients who are treated with these drugs, both individually and in combination

– plan observational and prospective studies to be undertaken when a new virus emerges

Prepare for clinical studies to be undertaken when a new influenza or other virus emerges:

– plan observational studies of treatment effectiveness

– consult with scientists who understand the biology of the host response (e.g., vascular biology, mitochondrial biogenesis, immunometabolism)

– choose two or three drugs for clinical trials, especially in combinations

– prepare clinical trial protocols for children and adults

– consult with statisticians on a study design and a plan for evaluating results

– involve a local ethics review committee

– organize a data safety monitoring board

– obtain logistical and financial support

– if needed, assemble networks of physicians who will participate in multi-center trials

Plan what to do with the results:

– identify local sources of supply for potentially efficacious drugs, quantities usually supplied, capacities for surge production and distribution, need for stockpiling, and logistics for delivery

– determine drug costs for public programs

– prepare plans to communicate trial results to physicians, public health officials, and the public

Adapted from similar tables in references.^13,14,17

Clinicians in low- and middle-income countries can and should become involved in this research. They will share with their colleagues in developed countries the risks of pandemic influenza and many emerging virus diseases, but they will also share similar access to inexpensive generic drugs. If, for example, one of these drugs (or drug combinations) could be shown to be effective in reducing morbidity and mortality due to seasonal influenza, it would probably be effective during the next pandemic.

Most important, this approach to treatment would be available to people in any country with a basic healthcare system on the first pandemic or epidemic day. This opens up the extraordinary possibility that treating the host response in this way could contribute significantly to global health, global equity and global security.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.