The Best Australian Science Writing 2015

by Heidi Norman

  ‘The critical point was that, for the first time, an analysis of the effect of statins in people without disease but at high, moderate, low and very low risk was presented – it showed consistent benefits across all these groups,’ explained Ebrahim.

  Overall, the risk of serious side effects was low. One in 10 000 reported a rare form of muscle breakdown known as rhabdomyolysis. There also appeared to be a 9 per cent increased risk for developing diabetes. But overall, the patients gained by avoiding cardiovascular events. That was reflected in the final figures – overall a slight reduction (9 per cent in the risk of death from all causes. And for every unit of cholesterol lowered, there were 11 fewer cardiovascular events per 1000 people. In an accompanying commentary in the Lancet in April 2012, Ebrahim, who serves as coordinating editor of the Cochrane Heart Group, and colleague Juan Pablo-Casas, paraphrased the findings of the CTT this way: ‘Men and women, old and young, and people with and without cardiovascular disease all benefit.’

  Moreover, they suggested that medical guidelines should lower the threshold for taking statins to people whose risk was greater than 10 per cent over ten years – which includes 83 per cent of men over 50 and 56 per cent of women over 60. Indeed the title of their Lancet paper was ‘Statins for all by the age of 50?’

  Abramson and colleagues were unconvinced. They reevaluated the CTT meta-analysis – but in a curious way. They ignored the findings on reductions of heart attacks, strokes or surgery to clear blocked arteries because they considered these endpoints were subject to bias. Instead they focused on deaths from any cause. Their conclusion: for people whose risk was lower than 10 per cent statins did not significantly reduce the risk of death.

  Abramson’s decision to focus only on deaths strikes some cardiologists as strange. ‘I refuse to believe that I should only do primary prevention to prevent total mortality,’ commented Donald Lloyd-Jones at Northwestern University in Chicago, a member of the panel that developed the most recent US statin guidelines. ‘Patients are much more worried about having a stroke or a major heart attack and living with that; they’re not only focused on what’s going to kill them.’

  He adds the only reason Abramson didn’t see a ‘statistically significant’ effect in mortality is that in the lower risk group too few people died over the five-year course of these studies to observe a significant effect. Wait long enough, Lloyd-Jones says, and the ‘trend’ will reach significance.

  * * * * *

  Nevertheless, Abramson and colleagues had many supporters. By late July 2014, close to 500 people, including many doctors and researchers, had signed an open letter on the website of the non-profit Lown Institute asking the BMJ not to retract the articles. This shows that ‘in the medical community (of) practitioners and researchers, a significant number of people are sceptical about the push to spread the use of statins to lower-risk patients,’ says Vikas Saini, a cardiologist and president of the institute. ‘I certainly haven’t seen enough proof to warrant changing guidelines and prescribing statins to another billion patients on the planet.’ On 1 August, the BMJ panel decided against retracting the two papers. But that’s hardly the end of the story.

  What is one to make of it when medical titans do battle?

  One reading is that the battlelines appear to be drawn between clinicians, who see individual patients, and epidemiologists, who see data on tens or hundreds of thousands of patients.

  Clinicians believe their patients when they complain of muscle pains or memory loss after taking statins. When they go off the pills, they stop complaining. ‘I’m a simple man; that’s good enough for me,’ says GP David Bailey in the BBC TV program Trust me, I’m a doctor. Bailey refers to epidemiologists such as Rory Collins, who was also interviewed for the program, as being in an ‘ivory tower’.

  But when patients are prescribed statins, some would doubtless read the packet information with its long list of possible side effects. What if they are imagining those ailments or falsely attributing new aches and pains and memory loss – all common in middle-age – to statins? It’s known as the nocebo effect – ‘the evil twin of the placebo effect’ as Ben Goldacre, the celebrity author of Bad Science, puts it. To see it in action he suggests: ‘When sitting on a sofa with friends, suddenly ask: “does this thing have fleas in it?”.’

  Could many statin side effects be explained by nocebo?

  ‘I’ve stopped trying to work it all out, because the patients’ beliefs are the most important factor,’ admits David Sullivan, a physician and researcher at the University of Sydney. ‘I see a huge number of patients with these complaints, I don’t know to what degree it is the power of suggestion, but it’s a tragedy if they miss out on the benefits.’

  In March 2015 two studies gave more weight to the power of suggestion. One was a meta-analysis published in the European Journal of Preventive Cardiology. Goldacre, who is also an epidemiologist at the London School of Tropical Medicine and Hygiene, was a co-author. He and colleagues trawled though 62 randomised controlled trials picking out 14 of the best primary prevention trials.

  Their findings were stunning: ‘Cholesterol-lowering statins have almost no side effects,’ blared the newspaper headlines.

  Indeed, the paper found that with the exception of a slightly increased risk of diabetes, most common side effects attributed to statins were found in the placebo group at around the same rate. These included muscular weakness, muscle aches, back pain, newly diagnosed cancer, kidney problems, insomnia, gastrointestinal disturbance, nausea, indigestion, diarrhoea, constipation, fatigue, headache and a more than ten-fold increase of creatine kinase (a muscle enzyme whose levels rise after injury).

  Despite the findings, Goldacre says he has reservations about concluding that ‘statins have almost no side effects’. Most importantly, he has concerns that the data collected by randomised controlled trials, which are mostly funded by drug companies, do not necessarily reflect what happens with ordinary patients. This is because these trials routinely involve a ‘run-in’ period, up to several months, to weed out subjects who won’t take their medication reliably. That’s important for reliable data collection, but also means that some people who really can’t take statins are left out.

  Second, there are suspicions that industry-funded trials gild the lily. For instance, a 2012 Cochrane review compared trials that were funded by industry with ones from a non-commercial source. The results of the industry-funded trials were generally better. In another study in PLoS Medicine in October 2013, German researchers compared the final results that companies published to the raw data on which it was based. They found side effects were significantly under-reported. This suspicion led Goldacre and others to co-found the ALLTrials campaign in January 2013, to petition drug companies to make their raw data freely accessible.

  But it seems unwarranted to tarnish the CTT trialists with this broad brush; they are a collaboration of academics and they do have access to the raw data. It is true that they do not share that data outside their collaboration and are criticised by other researchers who would like to be able to check their calculations. But the trialists fear mischief, especially from drug companies seeking to discredit the data of their rivals or from other people with vested interests. Explains Keech, ‘the problem with ad hoc analyses are that they can use methods to produce a particular result. The most reliable analyses are the ones done using the methods we published in 1995. The rules were set out before we started.’ And he points out these analyses are cross-checked by the academic collaborators: ‘Everything is replicated.’

  Nevertheless, Ebrahim agrees randomised clinical trials have their problems and that the people they include may not represent the majority of patients. So he co-authored a paper looking at 90 studies of the side effects observed in people being prescribed statins in the general population. The findings, published in March 2014 in BMC Medicine, turned out quite similar to those reported in Goldacre’s study: that is, that statins are safe. Interestingly, they foun
d that statin takers appeared to be slightly protected against memory loss.

  So how will the statins war be resolved?

  More data is on the way from the CTT. Their past analyses focused on major benefits and major risks such as cancer or diabetes. Now Keech says they are collecting the data on the common, milder side effects. This data, widely agreed to be the best that science can deliver, should finally rule on whether side effects such as muscle pain or memory loss have anything to do with statins. Could the protective effect on memory be real? Many doctors are anxiously waiting to know. ‘It’s a mess out there. Doctors are pulling their patients off statins because of concerns over dementia. But mini-strokes are a major cause; I fear this (stopping statins) risks more harm than good,’ says Bryce Vissel, who studies brain diseases at Sydney’s Garvan Institute.

  So perhaps it’s just a matter of more time? Or not. As Sullivan points out ‘people have been arguing about statins for an awfully long time. Old habits die hard.’

  How I rescued my brain

  Honest placebos

  An uneasy alliance:

  Our debt to tools and their persistence in our nightmares

  Elizabeth Bryer

  A friend and I leapt aboard a combi that accelerated and slowed around Trujillo, a coastal city in northern Peru, moving farther out into the fringing districts until we had left the bustle behind. It turned down a narrow track; when it ended we clambered out and hiked to a distant ticket booth, beyond which, in the still desert, rose two hills. Between them were the weathered, slumped pyramids of the Moche, one a sun temple, the other a moon temple, but both, from where I was standing, seemingly nothing more than formless mounds of sand. The air tasted of clay.

  On a whim my friend asked the volunteer guide, an archaeology student, if we could see a mural he had heard whisperings about, one that depicted machines killing people. I turned to him in surprise. Maybe I hadn’t heard right; maybe ‘máquina’ didn’t mean machine. But the guide proceeded to respond.

  What struck me wasn’t her regretful tone but the fact that she was unperturbed by what, to my mind, had to be a crackpot conspiracy. She explained that the mural was high up in the Temple of the Moon and badly deteriorated, and not even she had been allowed to see it. Instead, she showed us around the temple’s lower reaches, past other murals, their red and white paint still visible, and past walls built with bricks marked by personal signatures, perhaps because the bricks had been exhorted as a kind of tax.

  Two years later I was in my university library reading an English translation of the Huarochirí Manuscript, a testimony of ancient memories of the Andeans compiled in Kechwa circa 1608 in Peru’s central highlands. The tome spoke of a great deluge, of a mountain deity who expelled the fire deities of old, of ceremonies devoted to landscape. And, in one brief paragraph, it described a time when the sun died. The darkness lasted for five days, during which time ‘mortars and grinding stones began to eat people’.

  I had a tingling sensation of I-think-I’ve-read-this-before. But I hadn’t, had I? Slowly I realised, after a laborious dredging of foggy, incidental memories, what the declaration reminded me of: the mural in the Moche’s Temple of the Moon that we hadn’t been able to see.

  I found a representation of the Moche mural in a scholarly journal, the camphorous rosin that lifted from the page a world away from the clean mineral tang of desert where the real-life mural crumbled out of anyone’s sight. I saw that my friend’s ‘machines’ were in fact weapons and military regalia, and that they attacked warriors, their owners. Weapons weren’t exactly cooking utensils, but still, both were important tools, and in both imaginings they had come to life to rise against humans. How could the imaginations of the Andean and Moche civilisations, distant across time and space, be so connected?

  Connections have also been drawn even further afield, with the Mayas and their Popol Vuh. There it is written that early humans were spoken to by their maize grinders, cooking griddles, plates, pots and grinding stones, which, in revenge for the way humans had treated them, crushed their owners’ faces, ground up their flesh, landed on their heads and flattened their bodies. The correspondences across time and space suggest that this Revolt of the Objects, as it is usually called, is an ancient American myth.

  I contemplated the reproduction of the Moche mural: headdresses, clubs and weapons sprouted arms and legs and chased after people, attacked them and dragged them by their hair. Chaos reigned on the walls as humans were domesticated by the tools they had created: one man grimaced in pain, another stumbled as a shield attacked him. The militaristic Moche were thrust to the bottom of the food chain by the technologies responsible for their regional dominance – their most important tools – just as the Huarochirí informants and the Maya had imagined their ancestors were.

  I closed the book and leaned back in my chair. The neon lights were casting a clinical sheen across the desk and there were murmurs drifting from the stacks. Sometimes a piece of the past flies through time at just the right angle to cast new light on the present. Images from a steady diet of pop-culture rose in my mind: cool, terrifying robots, humans in farms and the planet a wasteland. It seemed that the modern, Western imagination to which I was subject wasn’t so different from this ancient American one.

  * * * * *

  Tools: they have long been recognised as our evolutionary ally, yet we humans are not actually alone in our fashioning and making use of them. To name just a few other animals that do the same, chimpanzees have a toolkit for termite hunting that includes a short stick to penetrate above-ground mounds, a large stick to drill holes into subterranean nests and a probe that they first pull through their teeth to craft a frayed end, all the better for collecting the insects. Bottle-nosed dolphins, to protect from abrasions as they search for food on the seafloor, tear off pieces of sponge and manoeuvre these around their snouts. And elephants, once they have dug a waterhole, rip bark from a tree, chew it into a circular shape and then place it over the hole to limit evaporation. The difference between tool use among humans and these other animals is, of course, in the tools’ comparative level of complexity.

  While recent discoveries in Kenya indicate that stone tools were fashioned 3.3 million years ago by hominids that may not have belonged to our Homo genus, it was around 2.6 million years ago that Homo hominids first reached the level of tool use of modern-day chimpanzees, elephants and dolphins. In their case, they used hunks of rock for pounding and sharp flakes for cutting through hide. It probably wasn’t until 1.8 million years ago and the appearance of Homo erectus that our ancestors, through their fashioning of axes and cleavers, became more technologically advanced than other species in any time period. More tools followed. Four hundred thousand years ago came spears; 200 000 years ago came glue; 28 000 years ago came rope. In the fourth millennium BCE appeared the wheel; in the second century BCE, paper; in the ninth century CE, gunpowder.

  If I were to keep on in this way, listing chronologically our most significant tools and the dates on which they were invented, the diversity of these would increase and the distance between dates would diminish until we arrived at the explosion of tools that is the present. It is perhaps not surprising, then, that we Homo sapiens sapiens have not only evolved as a result of tools – we have been shaped by them. We have lost our sharp fingernails and our heavy jaw musculature. Our eyesight is worse nowadays than it was 10 000 years ago. Over the past 30 000 years our brains have decreased in size. All of these apparent signs of devolution point to one indisputable fact: we have become a species conditioned to use tools. In other words, children told to get off the computer and go play outside are, contrary to what well-meaning parents believe, actually doing the opposite of what has become natural.

  This brings me back to those ancient American imaginings that depict a confused ancestral past in which the debt owed to tools for human success was acknowledged and mulled over. Surely we can view such myths as expressing an acknowledgement of the power of tools, of h
ow, while we may have created them, we have come to rely on them – our bodies have even devolved because of them. If we rely so much on tools, the American myth encourages us to ask, then who, exactly, rules whom?

  In more recent centuries in the West, we can see a traversal of the same imaginative territory, which shows that this preoccupation has persisted in time as a powerful allegory. In 1651 English philosopher Thomas Hobbes, in his book Leviathan, wrote that humans, by virtue of their machines, would create a new intelligence. In 1921, the play by Czech writer Karel Čapek, R.U.R.: Rossum’s Universal Robots, imagined a future populated with artificial people called robots that rebel and cause the extinction of the human race. And in 1968, Stanley Kubrick’s film 2001: A Space Odyssey posited a world in which humankind, having reached the pinnacle of evolution on Earth, sets out into space. There, the ultimate tool, the HAL 9000 computer, controls the spacecraft. When it erroneously reports a problem and the astronauts decide to switch off its higher-mind functions, it kills four of them and won’t open the hatch for another floating outside. But the remaining astronaut manages to re-enter the spacecraft and subsequently shut down HAL, forever ending humanity’s alliance with tools.

  While these scenarios take place in the imagined future, rather than in the ancestral past, the effect is the same. Both settings point to unease: in ancient America, the tools’ revolt had already happened at the time of the telling, which, given the belief in the cyclical nature of space-time, suggested that it would happen again. In the Western case, the future setting evokes a similar sense of real possibility in that it suggests not just that this scenario could, but that one day it might, occur. By framing these imaginings as possibilities, the cultures that produced them reveal that the imaginings spring from, and perhaps are designed to evoke in the audience, a sense of anxiety. This is especially visible in the Western case when the science fiction depicting Artificial Intelligence is in cautionary-tale mode: in addition to the Čapek and Kubrick examples, think the Wachowski brothers’ The Matrix, in which humans are no longer born but grown for AI consumption, or James Cameron’s The Terminator, in which an AI defence network becomes self-aware and initiates a nuclear holocaust of the human race.