Invisible Women: Exposing Data Bias in a World Designed for Men
Page 21
Finally, the authors didn’t consider the number of drug treatments that might be beneficial to women but never even reach human testing because they were ruled out at the cell and animal trial stage. And this number could be substantial. Sex differences in animals have been consistently reported for nearly fifty years, and yet a 2007 paper found that 90% of pharmacological articles described male-only studies.62 In 2014 another paper found that 22% of animal studies did not specify sex, and of those that did, 80% included only males.63
Perhaps most galling from a gender-data-gap perspective was the finding that females aren’t even included in animal studies on female-prevalent diseases. Women are 70% more likely to suffer depression than men, for instance, but animal studies on brain disorders are five times as likely to be done on male animals.64 A 2014 paper found that of studies on female-prevalent diseases that specified sex (44%), only 12% studied female animals.65 Even when both sexes are included there is no guarantee the data will be sex-analysed: one paper reported that in studies where two sexes were included, two-thirds of the time the results were not analysed by sex.66 Does this matter? Well, in the 2007 analysis of animal studies, of the few studies that did involve rats or mice of both sexes, 54% revealed sex-dependent drug effects.67
These sex-dependent effects can be extreme. Dr Tami Martino researches the impact of circadian rhythms on heart disease, and during a 2016 lecture to the Physiology Society she recounted a recent shock discovery. Together with her team, she conducted a study which found that the time of day you have a heart attack affects your chances of survival. A heart attack that hits during the day triggers, among other things, a greater immune response. In particular, it triggers a greater neutrophil response (neutrophils are a type of white blood cell that are usually first on the scene in response to any injury), and this response correlates with a better chance of survival. This finding has been replicated many times over many years with many different animals, becoming, explained Martino, the ‘gold standard for survivorship in the literature’.
So Martino and her team were ‘quite surprised’ when in 2016 another group of researchers released a paper which also found that daytime heart attacks triggered a greater neutrophil response – but that this correlated with a worse chance of survival. After a substantial amount of head-scratching, they realised there was one basic difference between the historic studies and this one new study: the old studies had all used male mice, while this new paper had used female mice. Different sex: totally opposite result.
As for cell studies, a 2011 review of ten cardiovascular journals found that when sex was specified 69% of cell studies reported using only male cells.68 And ‘when sex was specified’ is an important caveat: a 2007 analysis of 645 cardiovascular clinical trials (all published in prominent journals) found that only 24% provided sex-specific results.69 A 2014 analysis of five leading surgical journals found that 76% of cell studies did not specify sex and of those that did, 71% included only male cells and only 7% reported sex-based results.70 And again, even for diseases that are more prevalent in women, researchers can be found ‘exclusively’ studying XY cells.71
As in animal and human studies, when sex has been analysed in cell studies, dramatic differences have been found. For years researchers were puzzled by the unpredictability of transplanted muscle-derived stem cells (sometimes they regenerated diseased muscle, sometimes they didn’t do anything) until they realised that the cells weren’t unpredictable at all – it’s just that female cells promote regeneration and male cells don’t. Perhaps of more urgent concern for women’s health is the 2016 discovery of a sex difference in how male and female cells respond to oestrogen. When researchers72 exposed male and female cells to this hormone and then infected them with a virus, only the female cells responded to the oestrogen and fought off the virus. It’s a tantalising finding that inevitably leads to the following question: how many treatments have women missed out on because they had no effect on the male cells on which they were exclusively tested?
In light of all this evidence, it’s hard to see how researchers can continue to argue in good faith that sex doesn’t matter. Rather, it seems clear that McGill University neuroscientist Jeffrey Mogil was right when he told the Organisation for the Study of Sex Differences that failing to include both sexes ‘right at the very beginning’ of your research ‘is not only scientifically idiotic and a waste of money, it is an ethical issue as well’.73 Nevertheless, women continue to be routinely under-represented in medical research, and you can’t even expect sex-specific trials to adequately represent women. When the ‘female Viagra’74 that was released with much fanfare in 2015 was found to potentially interact negatively with alcohol (as most readers will know, the absorption of alcohol differs between men and women75), its manufacturer, Sprout Pharmaceuticals, quite rightly decided to run a trial – for which they recruited twenty-three men and two women.76 They did not sex-disaggregate the data.
In this latter failure, they are not alone. Several reviews of papers published in major journals over the past ten years have all identified a routine failure to either present results by sex, or to explain why the influence of sex has been ignored.77 A 2001 US Government Accounting Office (GAO) audit of FDA records found that about a third of documents didn’t sex-disaggregate their outcomes and 40% didn’t even specify the sex of the participants. The auditors concluded that the FDA had ‘not effectively overseen the presentation and analysis of data related to sex differences in drug development’,78 a finding that was confirmed in a 2007 analysis of new drug applications submitted to the FDA which found a failure to establish standards for data analysis of applications.79 In 2015 the GAO criticised the US National Institutes of Health (NIH) for failing to routinely track whether researchers had actually evaluated any differences between the sexes.80 Things are often even worse in non-government-funded trials – which represent the majority of studies. A 2014 investigation into sex analysis in cardiovascular trials found that thirty-one of sixty-one NIH-sponsored trials analysed outcomes by sex compared with only 125 of 567 non-NIH-sponsored clinical trials.81
The lack of sex-disaggregated data affects our ability to give women sound medical advice. In 2011 the World Cancer Research Fund complained that only 50% of studies into the impact of diet on cancer that included both men and women disaggregated their data by sex, making it hard to establish dietary guidelines for cancer prevention that are valid for both sexes.82 Women, for example, should probably eat more protein than men as they age (because of muscle mass loss), but ‘the optimal dose per meal to support muscle protein synthesis in older women has not been determined’.83
The failure to sex-disaggregate when you’ve actually gone to the effort of including both sexes is baffling, not to mention, as Londa Schiebinger at Stanford University puts it, ‘money wasted [and] research that is lost to future meta-analysis’.84 And when female representation in trials is so low, the ability to conduct meta-analysis can mean the difference between life and death.
In 2014 a review of the FDA database of a cardiac resynchronisation therapy device (CRT-D – essentially a more complicated kind of pacemaker) trials found that women made up about 20% of participants.85 The number of women included in each individual study was so low that separating out the data for men and women didn’t reveal anything statistically significant. But when the review authors combined all the trial results and sex-disaggregated that data, they found something alarming.
A CRT-D is used to correct a delay in your heart’s electrical signals. They are implanted for established heart failure and the D stands for defibrillator. This defibrillator (a larger version of which most of us will have seen in one hospital drama or other) performs something like a hard reset on the heart, shocking it out of its irregular rhythm so that it can restart in its correct rhythm. A doctor I spoke to described CRT-Ds as ‘symptom control’. They aren’t a cure, but they prevent many early deaths, and if your heart takes 150 milliseconds or longer to complete
a full electrical wave, you should have one implanted. If your heart completes a full circuit in under that time, you wouldn’t benefit from one.
Unless, the meta-analysis found, you happened to be female. While the 150 milliseconds threshold worked for men, it was twenty milliseconds too high for women. This may not sound like much, but the meta-analysis found that women with an electrical wave of between 130-49 milliseconds had a 76% reduction in heart failure or death and a 76% reduction in death alone from having the advanced pacemaker implanted. But these women would not be given the device under the guidelines. And so because the trials treated male bodies as the default, and women as a side-show, they had condemned hundreds of women to avoidable heart failure and death.
The CRT-D is far from the only piece of medical tech that doesn’t work for women – which is unsurprising given a 2014 analysis which found that only 14% of post-approval medical-device studies included sex as a key outcome measure and only 4% included a subgroup analysis for female participants.86 A 2010 paper found that ‘the female gender is associated with an increased risk of acute complications during primary pacemaker implantation, being independent from age or type of device implanted’.87 In 2013, a supposedly revolutionary artificial heart was developed that was too big for women.88 Its designers are working on a smaller version, which is great, but it’s striking that, like other artificial hearts,89 the female version comes years after the default male one.
Even something as basic as advice on how to exercise to keep disease at bay is based on male-biased research. If you run a general search for whether resistance training is good for reducing heart disease, you’ll come across a series of papers warning against resistance training if you have high blood pressure.90 This is in large part because of the concerns that it doesn’t have as beneficial an effect on lowering blood pressure as aerobic exercise, and also because it causes an increase in artery stiffness.
Which is all true. In men. Who, as ever, form the majority of research participants. The research that has been done on women suggests that this advice is not gender-neutral. A 2008 paper, for example, found that not only does resistance training lower blood pressure to a greater extent in women, women don’t suffer from the same increase in artery stiffness.91 And this matters, because as women get older, their blood pressure gets higher compared to men of the same age, and elevated blood pressure is more directly linked to cardiovascular mortality in women than in men. In fact, the risk of death from coronary artery disease for women is twice that for men for every 20 mm Hg increase in blood pressure above normal levels. It also matters because commonly used antihypertensive drugs have been shown to be less beneficial in lowering blood pressure in women than in men.92
So to sum up: for women, the blood-pressure drugs (developed using male subjects) don’t work as effectively, but resistance training just might do the trick. Except we haven’t known that because all the studies have been done on men. And this is before we account for the benefits to women in doing resistance training to counteract osteopenia and osteoporosis, both of which they are at high risk for post-menopause.
Other male-biased advice includes the recommendation for diabetics to do high-intensity interval training; it doesn’t really help female diabetics93 (we don’t really know why, but this is possibly because women burn fat more than carbs during exercise94). We know very little about how women respond to concussions,95 ‘even though women suffer from concussions at higher rates than men and take longer to recover in comparable sports’.96 Isometric exercises fatigue women less (which is relevant for post-injury rehabilitation) because men and women have different ratios of types of muscle fibre, but we have ‘a limited understanding of the differences’ because there are ‘an inadequate number of published studies’.97
When even something as simple as ice-pack application is sex-sensitive, it’s clear that women should be included in sports-medicine research at the same rates as men.98 But they aren’t.99 And researchers continue to research men and act as if their findings apply to women. In 2017, a Loughborough University study100 was hailed around the UK news media as proving that a hot bath has anti-inflammatory and blood-sugar response benefits similar to exercise.101 Published in the journal Temperature with the sub heading ‘A possible treatment for metabolic diseases?’ the study included no women at all.
We know that men and women have different metabolic systems. We know that diabetes, one of the diseases particularly singled out as being relevant to this discovery, also affects men and women differently,102 and that it is a greater risk factor for cardiovascular disease in women than in men.103 But despite all this, the paper’s authors consistently failed to acknowledge any relevance of sex differences to their research. They cited animal studies that had similarly been conducted in all male populations, and perhaps most shockingly of all, in a section specifically looking at ‘limitations with the present investigation’ they completely failed to mention the fact that the study was all-male as a potential drawback, only referring to their ‘relatively small sample size’.
There have been some attempts to force researchers to properly represent females in medical research. Since 1993, when the US passed the National Institute of Health Revitalization Act, it has been illegal not to include women in federally funded clinical trials. Australia’s main funding body made similar rules for the research it funds,104 as has the EU, which in fact went even further, also requiring both sexes to be studied in pre-clinical animal studies. This requirement did not come into effect in the US until January 2016,105 which is also when the NIH introduced the requirement that the data in trials it funded be disaggregated and analysed by sex (unless there is a compelling reason not to).106
Other positive developments include the German Society of Epidemiology which has for more than a decade required researchers to justify including only one sex in any study where the results could potentially affect both sexes;107 and the introduction of the same by the Canadian Institutes of Health in 2012, as well as mandatory questions about the consideration of sex and gender in the study design. Some academic journals also now insist that papers submitted for publication should provide information about the gender of participants in clinical trials, for example.108
Trailing behind everyone is the UK, whose main funders ‘make no substantive reference to, or requirements regarding, the consideration of gender in research design and analysis’,109 and despite the at-risk population of women suffering more morbidity and mortality,110 UK research funding for coronary artery disease in men is far greater than for women. Indeed, such is the dearth of gender-based clinical research from within the UK, that Anita Holdcroft, emeritus professor at Imperial College London, has written that for cardiovascular treatment, ‘it is pertinent to use studies from North America and Europe where these issues have been investigated’.111
Still, while the situation in the UK is dire, significant problems remain elsewhere. For a start, the evidence we’ve just seen on the representation of women in trials suggests that these policies are not being rigorously enforced. And, indeed, this is what analyses of the NIH have found. Four years after the NIH announced their first policy calling for the inclusion of women in medical trials, a report was released by the GAO which criticised the NIH for having ‘no readily accessible source of data on the demographics of NIH study populations’, making it impossible to determine if the NIH was enforcing its own recommendations.112 By 2015 the GAO was still reporting that the NIH ‘does a poor job of enforcing rules requiring that clinical trials include both sexes’.113
There also remain plenty of loopholes for US drug manufacturers who don’t want the cost and complication of including unharmonious females with their messy hormones in their neat clinical trials, because the rules only apply to NIH-funded trials; independent drug manufacturers can do whatever they want. And the evidence suggests that many of them do: a 2016 paper found that ‘a quarter of the drug manufacturers in an industry survey did not deliberately recruit
representative numbers of women as participants in drug trials.’114 When it comes to generic drugs, the FDA only specifies ‘guidelines’ rather than rules and, as we’ve seen, these guidelines are being roundly ignored. And the NIH policy on including female subjects in clinical trials doesn’t apply to cell studies.
Then of course there’s the issue of legacy drugs. Two million women per year take Valium for conditions ranging from anxiety to epilepsy, and it was aggressively marketed towards women for decades.115 And yet, a 2003 paper points out,116 this ‘mother’s little helper’ was never tested in randomised clinical trials with female subjects. A 1992 survey by the US General Accounting Office (the Congressional watchdog) found that less than half of publicly available prescription drugs had been analysed for sex differences.117 A 2015 Dutch paper baldly states that ‘The specific effect on women of a huge number of existing medications is simply unknown.’118
There is clearly a long way to go, and we must begin to address these gaps as a matter of urgency, because while they remain open, women (who ingest approximately 80% of pharmaceuticals in the US119) are dying. Some drugs used to break up blood clots immediately after a heart attack can cause ‘significant bleeding problems in women.’120 Other drugs that are commonly prescribed to treat high blood pressure have been found to lower men’s mortality from heart attack – but to increase cardiac-related deaths among women.’121 Statins, which are regularly prescribed around the world as a preventative measure for heart disease have mainly been tested in men and recent research from Australia suggests that women taking statins at higher dosages may face an increased diabetes risk122 – which in turn is a higher risk factor for cardiovascular disease in women than in men.123 In 2000 the FDA forced drug manufacturers to remove phenylpropanolamine, a component of many over-the-counter medications, from all products because of a reported increased risk of bleeding into the brain or into tissue around the brain in women, but not in men.124 Drug-induced acute liver failure has also been reported more often in women,125 and certain HIV medications are six to eight times more likely to cause an adverse drug reaction (ADR) in women.126