select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
A pre-print study by the University of Bristol demonstrates there may be a protective benefit against breast cancer in women who wake earlier.
Dr Sowmiya Moorthie, Senior Policy Analyst (Epidemiology), PHG Foundation, said:
“This is a robust paper that clearly covers the strengths and limitations of the study, the major strength being the use of multiple approaches to examine the links between sleep traits and breast cancer, which allows the researchers to demonstrate consistency in their findings. They acknowledge the limitations of the work, notably the generalisability of findings to the wider female population, since the UK Biobank population examined was mainly women of European ancestry.
“In terms of the implications of the research, it supports existing evidence that sleep patterns influence cancer risk, but it remains unclear how individual preferences for early or late rising interact with actual sleep behaviours. Whether attempts to modify sleep behaviours can help people manage their health and reduce their risk of cancer therefore needs further investigation.”
Dr Dipender Gill, Clinical Research Training Fellow, Imperial College London (and has published many studies that use Mendelian randomization), said:
“Although informative and interesting, this study alone does not warrant any action other than further investigation – people should not be changing their sleep patterns based on the evidence presented here. The statistical method used in this study, called Mendelian randomisation, does not always allow causality to be inferred. In this instance there may be a shared genetic cause for being a ‘lark’ and risk factors related to breast cancer. For example, the genetic determinants of sleep may also affect other neuronal mechanisms that affect breast cancer risk independently of sleep patterns. In such a scenario, sleep patterns may be associated with risk of breast cancer, but not directly cause it.
“This paper is a pre-print, which is a way of distributing research before it has been published in a peer-reviewed journal. The standard is generally close to that of a peer-reviewed publication, however the findings are typically disseminated faster. The main downside is that the work has not gone through any formal peer-review process by other academics expert in the field.”
Dr Stephen Burgess, Sir Henry Welcome Post Doctoral Fellow, MRC Biostatistics Unit, University of Cambridge, said:
Q: Does the press release accurately reflect the science?
“More or less. The face-value interpretation is that people with a genetic predisposition to be an evening person are more likely to suffer from breast cancer compared to those with a genetic predisposition to be a morning person. The implication is that changing one’s lifestyle to be more of a morning person would have a similar effect on breast cancer risk. This is clearly evidenced by the paper.
“I have some reservations about the choice of numbers – the numbers quoted in the press release ‘a preference for mornings reduced the risk of breast cancer by 40% compared with being an evening type’ and “morning preference reduced the risk of breast cancer by 48%” do not appear to come from the manuscript itself. A detailed investigation reveals that the 40% and 48% decreases in risk are for people answering that they are ‘definitely an evening person” versus ‘definitely a morning person’. The headline comparisons given in the paper include people answering that they are ‘more an evening person than a morning person’ or ‘more a morning person than an evening person’. Hence the numbers don’t quite tally between the paper and the press release. This is unfortunate, as it means (for example) that confidence intervals for these values are not available.”
Q: Have the authors accounted for confounders? Are there important limitations to be aware of?
“Mendelian randomization works by considering the associations of genetic variants with disease outcomes for genetic variants that are linked with modifiable risk factors. The premise is that because people don’t choose their genes, and because the genetic code is fixed at conception, genetic associations are less likely to be subject to confounding and reverse causation than associations from typical observational studies. This is because genetic variants can’t be influenced by confounders acting after the person is born, and the genetic variants come first, so can only ever be the cause of a phenomenon, and not the effect. Mendelian randomization studies are generally recognized to make conclusions that reflect causal relationships, although care is needed to ensure that the genetic variants are specific predictors of the risk factor of interest, and not simply markers of poor health in general.
“The key limitation is lack of mechanism. The authors do not show any biological mechanism by which sleep timing preference could influence breast cancer risk. Another limitation is that sleep timing preference (chronotype) is self-reported, and the investigation did not specifically recruit individuals with different sleep patterns, such as night-shift workers.”
Q: What it means that this is a pre-print – is there sufficient information to assess the evidence?
“Being a pre-print means that the manuscript has not yet passed peer review. It is rare that peer review in this area exposes a fatal flaw in a manuscript, as the methods used here are fairly routine. However, it is not uncommon for a paper to change substantially through the peer review process, particularly relating to how data are presented.”
Q: Is this good quality research? Are the conclusions backed up by solid data?
“In general, yes.”
Q: How does this work fit with the existing evidence?
“This manuscript adds to an increasing body of evidence from a variety of different study designs indicating that sleep patterns can have a detrimental effect on health outcomes. Despite the focus here on breast cancer, the evidence suggests poor sleep patterns are likely to adversely influence many disease outcomes.”
Q: What are the implications in the real world? Is there any over-speculation?
“In isolation, this manuscript only provides one strand of evidence – linking genetic predisposition to sleep timing and sleep duration patterns to one particular outcome. However, this paper is one in a succession of papers using this technique to link genetic predisposition to sleep timing and sleep duration patterns to various disease outcomes. And this technique is one of a number of techniques for investigating relationships between modifiable risk factors and disease outcomes.”
Q: Does this mean that people should be changing their sleep patterns, or worried about their sleep patterns?
“It is increasingly clear that healthful sleep patterns are an important component of a healthful lifestyle.”
‘Investigating causal relationships between sleep traits and risk of breast cancer: a Mendelian randomization study’ by Richmond et al. is a pre-print study available at https://doi.org/10.1101/457572 and is being presented as Abstract no: Poster 1822
Declared interests
Dr Sowmiya Moorthie: No declarations received
Dr Gill: “No conflicts to declare”
Dr Stephen Burgess: “No conflicts to declare”