More apocalypses than one

I tend to be pretty dismissive of predictions of the superbug apocalypse. Yes, antibiotic resistance is a real problem and causes real deaths  now and will cause even more later. But it is a pretty slow-moving sort of apocalypse and there are lots of changes and adaptations that we can make to slow it down. None of these will bring back the Golden Age, when we imagined we had mastered the bugs, but they will make antibiotic resistance just one more problem among many.

My favorite example is this figure of the 90% drop in MRSA BSI incidence in England:

From The English MRSA Miracle

There are a couple of takeaways from this. One is that there was no magic bullet, no technological or pharmacological breakthrough that caused such a drastic drop in MRSA bacteremias. There were a bunch of incremental interventions, some of which worked, some that didn’t – stuff like better communication and accountability, better hygiene, screening and awareness. I’m sure the meetings where these policies were devised and implemented were incredibly tedious and dull. But there’s no arguing with the results – thousands of lives have been saved.

The other takeaway is that few of these interventions are really specific to MRSA. They are mostly just about implementing best practices in infection control and holding people responsible. There are probably several non-MRSA infections prevented for every MRSA infection avoided.

So when I read about impending superbug apocalypses what I tend to hear is “antibiotic resistance will force us to do business in ways that are inconvenient and unaccustomed” and my response is to yawn and not care very much.

But this does not mean that we need have no worries of civilization-disrupting plagues. I just don’t think they will come from bacterial infections (with the possible exception of MDR-TB). As long as we have clean water, these plagues just aren’t going to get very far. And if we can no longer supply clean water then some other apocalypse has already hit, and bacterial infections are just one more consequence of the collapse.

But viral infections are different. They exhibit (potentially) many more of the features likely to lead to uncontrollable epidemics:

1. Airborne dissemination – a single sneeze in a subway can infect hundreds, leading to explosive rates of transmission
2. Latent infectiveness – if victims are still ambulatory while infectious, they can spread the disease fast and far.
3. High case fatality rate among otherwise healthy individuals
4. Lack of effective therapeutics.

Ebola and other hemorrhagic fever viruses fail on the first two counts – they are spread primarily by direct contact with body fluids from acutely ill patients. Insect-borne diseases like Zika can be contained through insect control measures.

Influenza is the real threat. It easily checks boxes 1 & 2, and 4 is uncertain. A truly virulent strain might be uncontrollable. The 1918 strain killed 50 million people in a world in which most people still lived on farms. Agrarian societies are slower to spread diseases and intrinsically more resilient to their disruptions.

An equivalently deadly and equally transmissible strain today could well rip through modern megacities and disrupt our highly integrated economy with its long and intricate supply chains before we could even begin to figure out how to respond. Killing or disabling just a few percent of the population (and panicking most of the rest) could quickly lead to chaos and disruption of electricity generation and food distribution. More people might die of cold and hunger than of disease.

Unless, that is, we have a really effective remedy (#4). Flu vaccines help, but tend to be 20-50% effective. That’s not nearly enough to provide herd immunity. We’d have to depend on neuraminidase inhibitors like oseltamivir (Tamiflu). How likely is that to work out?

Oseltamivir is pretty unimpressive when it comes to treating symptomatic cases of the flu, reducing symptom duration by about a day. It doesn’t reduce the risk of hospitalization or of complications such as bacterial pneumonia. Not encouraging.

But that’s for patients with mild disease. Somewhat surprisingly, oseltamivir appears to be much more effective for patients hospitalized with severe cases of flu, cutting death rates nearly in half. The resolution of this apparent paradox may be in the actual cause of death in severe influenza infections – cytokine storms. We can’t really treat cytokine storms (septic shock) much more effectively now than in 1918. But a drug that tamps down disease enough to prevent a cytokine storm should have a major impact on mortality, even if it does not cure the patient.

Even more encouraging – if we stockpile enough of the drug – is that it appears to be highly effective as a prophylactic. We can’t plausibly stockpile enough to give it to everyone, but the US does have a pretty substantial stockpile, perhaps enough to protect healthcare workers and prevent collapse of our healthcare system.

The data supporting the usefulness of oseltamivir in a flu epidemic are far from conclusive and a number of smart and well-informed people think they are unconvincing and that the drug is essentially worthless. And whatever its intrinsic efficacy, it has to be deployed in a smart and timely fashion to be effective. I wouldn’t count too heavily on that.

Antibiotic resistance isn’t going to cause civilizational collapse; a flu epidemic just might. And climate change most certainly will. I’d save the apocalypse headlines for that.

Update: Koen Poweis points out that secondary infections of S aureus pneumonia were responsible for a great deal of the mortality from the 1918 flu pandemic. Antibiotic resistance would surely exacerbate this problem. Neither vancomycin nor daptomycin, common treatments for MRSA infections, are highly effective for pneumonia, due to poor lung penetration and inhibition by surfactant. There is a report suggesting that ceftaroline is effective, but this needs more confirmation.