(Photo by Bastian via Flickr)
When Paul Manley awoke in the hospital, he had one clear conviction running through his groggy mind. “I thought, ‘this is what death feels like, I’m going to die’,” he says. The 62-year-old retired Air Force Colonel had been in a coma for eight days, not because he had been in a car accident, nor because he’d had a stroke or heart attack. What had landed him in the intensive care unit was sepsis, a complication of infection that is still not fully understood.
Infections are a mundane part of life, and in most cases, our body’s immune system shakes them off with little effort. But for more than three quarters of a million people in the US each year – and many millions globally – infections turn into life-threatening sepsis. One in ten people admitted to intensive care units in the US have sepsis, and startlingly, it is the culprit in up to half of all hospital deaths.
Common and deadly though sepsis is, the tools we have to fight it have barely changed over the past three decades. Hundreds of clinical trials during that time have enrolled tens of thousands of patients at a cost of many millions of dollars. Yet all that effort has yielded not a single new drug to treat the condition.
Many in the sepsis community believe it’s time for a re-think. “It’s hard for people to deny that the current strategy […] hasn’t been successful,” says veteran sepsis researcher Jonathan Cohen from the University of Sussex in the UK. As researchers grapple with how to move the science of sepsis forward, arguments are erupting over the fundamentals – how to define the condition, and whether science’s treasured randomized controlled trial should be abandoned.
Sepsis is a baffling condition. What starts out as a burst appendix, or an infection following surgery, or even something as innocuous as a bug bite or urinary tract infection, can silently progress to a body-wide assault. Sepsis was long thought to be an invasive systemic infection, or ‘blood poisoning’, where bacteria overrun the body. But the widespread introduction of antibiotics in the latter half of the 20th century made it clear that infection is only half the story – many septic patients die even after the infecting bacteria are successfully eradicated. In sepsis, scientists discovered, an infection is aided and abetted by the person’s own immune system. “The immune response becomes totally unbalanced and this is what causes a lot of harm in patients,” says researcher Tom van der Poll from the University of Amsterdam.
Van der Poll is a member of an international task force charged with tightening the definition of sepsis. The third and most recent iteration – Sepsis-3 – was released in February this year. The new definition describes sepsis, in lay terms, as “a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs.” The definition singles out organ damage, drawing a clear line of demarcation between sepsis and a straight infection. In reality, identifying when a person crosses that line is far from easy.
Like most sepsis survivors, Manley has no idea when his infection became septic. He felt fine when he returned home from his 40th Air Force Academy class reunion in Colorado, noticing only that his urine was darker than normal, which he put down to dehydration. But over the next few days, fever set in. It was a Saturday when he called a nurse practitioner friend, who told him to go to the emergency room for antibiotics, rather than wait until Monday to see a doctor. Despite his stoic nature, he took the advice. “Luckily I did,” he says, “or I probably would have died that weekend, at home in bed.”
Manley managed, only just, to drive himself to hospital – he hit the curb a number of times along the way as his sickness worsened. Staff at the emergency department promptly administered intravenous antibiotics, but Manley recalls little after that. One of his last memories was of a doctor explaining to him that he was going into septic shock. “He said, ‘I want to be straight with you. Your prospects are not good for living’,” recalls Manley. “That was kind of a shocker.”
During septic shock – sepsis’s most critical point – the body is in a death spiral. Blood pressure continues to plummet, cells throughout the body begin to self-destruct, and organs shut down. Some patients go into cardiac arrest; others sustain devastating damage to their kidneys, liver or brain. It’s not uncommon for tissue in the extremities – hands, feet, arms, legs – to die, necessitating amputation. Chances of death from septic shock are around 50-50.
Over the eight days that Manley was in a coma, his lungs, kidney and liver all started to fail. At one point, his family was instructed to say their farewells, doctors fearing he wouldn’t make it through another night. A ventilator breathed for him, a dialysis machine filtered his blood – work his kidneys could no longer manage – and blood tests revealed that his liver was in trouble. All the while, intravenous lines delivered antibiotics to rid his body of infection, along with fluids and drugs to keep his blood circulating.
The Sepsis-3 definition was designed, in part, to help healthcare staff readily identify patients like Manley who turn up at the emergency department. Faster than normal breathing, lower than normal blood pressure, and any sign of confusion (“altered mentation”) are clear signs a person could have sepsis. These clinical criteria are all easy to measure at the bedside, and should elicit a swift response from medical staff without the need to wait for time-consuming lab tests.
The new definition also brings much needed consistency to the field. “Standardization is lacking,” says Christopher Seymour, an intensive care physician and sepsis researcher at the University of Pittsburgh. An analysis conducted in conjunction with Sepsis-3 found that researchers use a mixed bag of cutoffs and combinations for blood pressure, fluid resuscitation, and other measures to define who has septic shock and who doesn’t. A septic shock patient in one study or hospital might not meet the criteria for septic shock in another. A single, broad definition helps to identify the whole spectrum of sepsis patients that hospitals encounter.
A catch-all term is also helpful for getting an accurate picture of how big of a problem sepsis is, and whether measures to improve outcomes are working. According to Sachin Yende, an epidemiologist from the University of Pittsburgh, “there are a lot of inaccuracies” in the way that sepsis rates are measured. Measure sepsis one way, and rates could be double what they are if you measure it another way. And without a clear definition, projects like the Global Burdens of Disease Project, which tallies worldwide mortality and disability from everything from cancer and cardiovascular disease, to road accidents and drownings, ignore sepsis altogether.
But questions have been raised about whether the new definitions for sepsis and septic shock, as well as the signs proposed to identify them, are now too broad. “The simpler that you make it, essentially it’s just, ‘is this a sick patient’?” says Cohen. “Many people feel that that’s not particularly helpful.”
At best, ‘sepsis’ has always been an umbrella term. The difficulty in finding one clear definition for the condition is that sepsis is itself quite nebulous. “It’s not one single disease,” says Yende.
The situation is reminiscent of another of life’s great killers – cancer. But unlike sepsis, cancer is now treated in a highly targeted fashion. There is no single treatment for all cancers, or even a single treatment for all breast cancers. “Oncologists treat all kinds of subgroups of breast cancer depending on their histological and immunological profile,” says van der Poll. “We are way behind oncology.”
One of the key sources of variability in sepsis, and one that has only recently come to light, is in the way our immune system responds. For years, sepsis was thought of as a problem of unbridled inflammation. The immune system goes into overdrive in an attempt to rid the body of a pathogen, producing a fever and inflammatory molecules that end up damaging tissues in a case of auto-immune friendly fire.
Sepsis is now known to involve both inflammation and its antithesis, immune suppression. The tricky part is that nobody really knows which septic patients will have a problem with hyper-inflammation, and which with immune suppression. Indeed, both processes can occur in the same patient at different times, or even in the same patient at the same time, but in different tissues.
Added to the variability of immune response at the molecular level are arguably more obvious factors that differ between septic patients. What organism is the person infected with? Where did the infection start – was it pneumonia in the lungs, or a skin infection? Was the patient healthy before they became ill, or do they have cancer or diabetes?
This variability, or ‘heterogeneity’ – in type of infection, type of patient, type of immune response – is at the heart of why the research community has struggled to come up with effective treatments, despite enormous efforts. By lumping together a college student who has meningococcal disease, with a diabetic octogenarian with hospital-acquired pneumonia, and an otherwise healthy 62-year-old like Manley into what Cohen calls ‘all-comers’ trials, researchers have potentially undermined their own efforts to find new sepsis treatments. “What you finish up with is a signal-noise problem,” he says. The drug being tested may be effective in a subset of patients, but there are so many others for whom the drug is either useless or harmful, that there’s no way of telling.
Being able to test whether a patient is predominantly in a state of hyper-inflammation versus immune suppression, could also better direct drugs to the right patients. As an illustration, Cohen points to trials of drugs targeting a particular immune molecule called Tumour Necrosis Factor, or TNF. Patients were enrolled without first being tested for whether their TNF levels were even elevated. “In retrospect, it was completely daft,” he says.
It’s not an isolated case, and even recent trials suffer from similar flaws, opting to measure overall mortality rather the specific molecular changes that could enlighten researchers about the drug’s effects. A trial of the drug Eritoran, which targets a molecule called Toll-like Receptor 4 (TLR4) ended in 2011, having failed to decrease mortality for those taking the drug. This is perhaps not surprising, according to van der Poll. Patients weren’t selected on the basis of TLR4 levels, and no measurement was taken to check whether TLR4 levels changed – if indeed they did – in people who were given the drug. “You don’t know if you treat the right patients, and you don’t know if the treatment that you’re giving is really doing what it’s supposed to do,” says van der Poll. “It’s basically a shot in the dark.”
Everyone in the sepsis community now recognises the need to differentiate classes of patient, but some of the tools to do so don’t yet exist. There’s no rapid biochemical test to separate patients with hyper-inflammation from patients with immune suppression, or those with a mix of the two. Not only would such a test improve clinical trials, it would also improve diagnosis, ensuring that septic patients are recognised and treated as early as possible.
Even without such a test, one question Cohen and others are asking: should we be doing trials of sepsis at all? Or, for that matter, using a single over-arching definition of sepsis? A better way forward could be to conduct trials on tightly defined sub-groups of sepsis patients – those with hospital-acquired pneumonia, say, or young adults with meningococcal disease – rather than on a rag-tag population of patients who all meet the broad definition of sepsis as it currently stands.
A different approach that has been gaining support does away altogether with the randomized controlled trial – the gold standard in medical research. “The future is really not in traditional clinical trials,” says Seymour.
One tweak to trial design would be to make trials adaptive. Instead of randomly assigning patients 50:50 to either receive an experimental drug or not, the ratio would change as the trial progresses based on an algorithm that continuously monitors data as it accumulates. Patients would still be randomly allocated to either receive treatment or not, but the die would be loaded in favor of patients receiving whichever treatment was performing better as the trial goes on. Trials could also compare several treatments or drug regimes in the one trial, directing different patients to the appropriate treatment arm. In this way, trials would more closely resemble a quality improvement program, where treatments are constantly evaluated and refined as time goes on. Such trials would potentially cost less, require fewer participants, and deliver results that could be translated into practice faster, according to Seymour. Yende and others are already embarking on adaptive trials for sepsis treatments, but results are still a number of years away.
As long as new drug treatments for sepsis remain elusive, some in the sepsis community argue that our focus should first and foremost be on the basics – early treatment and awareness. These measures have already seen overall mortality from sepsis fall over the past 15 years in countries with well-resourced intensive care units.
“It’s about having a culture of awareness,” says Ron Daniels, CEO of the Global Sepsis Alliance and the charity, UK Sepsis Trust. And it’s not just about awareness in the emergency department, he says. Part of the onus should be on the public to be alert to signs, such as confusion, breathlessness, or mottled skin, that indicate an infection might be turning septic. Daniels has designed a simple card for patients to carry in their wallet that reminds them what to look out for. Catching sepsis early could bring mortality down into single figures, says Daniels. “In the UK, that would be saving 30,000 lives a year.”
For Manley, awareness came at a high price. After regaining consciousness, it took him nine gruelling months to gain the 30 pounds he had shed and to shake off a secondary fungal bone infection, a not uncommon complication of sepsis. He remains bewildered that someone as healthy as he was could end up at death’s door, but counts himself lucky. He says, “I’m always flexing my arms and my legs and appreciating them, that I still have them.”