select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
An individual participant data meta-analysis published in The Lancet looks at blood pressure lowering and risk of new-onset type 2 diabetes.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This is an impressive study. Though there is good evidence that lowering people’s blood pressure, if it is too high, can have important health benefits in reducing the risk of heart attacks and strokes, it hasn’t been clear whether lowering blood pressure can reduce the chance of developing type 2 diabetes in the future. There have been several previous studies that have investigated, in various ways, whether there is an association between blood pressure and diabetes risk, but for a range of different reasons the findings from them haven’t been definitive. For example, there has been evidence from observational studies that having higher blood pressure is associated with an increased risk of developing type 2 diabetes, but in that kind of study it’s difficult to know whether the raised blood pressure actually causes the increased diabetes risk. In these studies, there would typically be many other differences between people with higher and lower blood pressure, apart from the blood pressure difference, and it could be some of those differences that really cause the increased diabetes risk. (Statistical adjustments can be, and are, made to allow for these differences, where data are available, but those adjustments can’t take account of all possible differences.) Also, in observational studies like this, it could be that some health condition that is an early stage of diabetes caused the high blood pressure in people who would later develop full-blown type 2 diabetes, so that it’s essentially the diabetes causing the high blood pressure rather than the other way round.
“The new research gets round these complications of cause and effect in two different ways. Its main analysis takes data from several randomised clinical trials which had the intention of lowering blood pressure, but also recorded whether the trial participants developed type 2 diabetes. Those trials allow the assessment of whether lowering blood pressure could reduce type 2 diabetes risk. In a typical trial of this sort, people would be chosen at random either to receive the treatment intended to lower blood pressure (such as a drug), while others would receive no active treatment (so, for example in a drug trial, a placebo, that is, a treatment that looks just like the active drug but contains no active ingredients). Because the treatments are allocated at random, there will (on average) be no difference between the people who have the active treatment and the others, so if there is any difference (in terms of blood pressure reduction, or in terms of diabetes risk), it can really only be caused by the treatment being studied.
“The researchers in this new study used data from 19 different randomised clinical trials, with in total almost 146,000 participants. They put together the data from all these trials in a procedure called individual participant meta-analysis, a statistically appropriate method of dealing with data of this kind. The headline finding was that reducing systolic blood pressure by 5 mm Hg reduced the risk of becoming a new case of type 2 diabetes, in a given period of time, by 11%. There’s some inevitable statistical variability in that estimate and the data are compatible with a reduction in risk of between 5% and 16%. Systolic blood pressure is the higher of the two figures that you’ll have been given if your blood pressure has been measured – it’s the measurement of the pressure in the blood vessels when the heart contracts to pump the blood round the body.
“Given that type 2 diabetes is a pretty common condition, with many possible adverse consequences for health, the researchers clearly feel that a reduction in risk of somewhere between 5% and 16% is very worthwhile, and they point out that a 5 mm Hg reduction in systolic blood pressure is not all that large, and can be achieved in several different ways. To give an idea of the impact of such a 5 mm Hg reduction on diabetes risk, think of a group of 1000 people, who are on average like those who participated in the clinical trials that were analysed. In 1000 trial participants who did not receive active treatment to lower their blood pressure, about 82 would develop a new case of type 2 diabetes in a period of five years, according to data summarised in the new study. (The analysis did not include people who already had type 2 diabetes when they entered the trials.) On the basis of the study results, if all those 1000 people lowered their systolic blood pressure by 5 mm Hg, then the number who would be newly diagnosed with type 2 diabetes in five years would fall to somewhere between about 78 and 69, with the best estimate being about 73. In other words, this lowering of their blood pressure would save between 4 and 13 of our thousand people from developing type 2 diabetes in the next five years. That’s not an immense reduction in risk, but sounds worthwhile the type 2 diabetes is both common and potentially serious.
“In addition, the study looked specifically at the effects of various different drugs, intended to reduce blood pressure, on diabetes risk, and found that some (ACEs and ARBs) did have a protective effect against diabetes, while others did not, even though they do lower blood pressure. But it’s worth pointing out that there are other potential ways to reduce blood pressure than taking drugs, such as various lifestyle changes.
“I should point out that this study did not look at any effects of lowering blood pressure on people who already have type 2 diabetes.
“An interesting and important feature is that, as well as putting together the results of the randomised clinical trials, the study also analysed some data from observational studies using a method called mendelian randomisation. This approach uses associations between a person’s genetic make-up and their chance of having high blood pressure. Someone’s genetics are determined essentially at random, at the time of conception, so it’s a bit like a randomised trial in which people are randomly assigned to different chances of having high blood pressure. One can then analyse the chance of a type 2 diabetes diagnosis in groups with different genetic make-up. The fact that the genes are allocated essentially at random means that it is possible to draw conclusions about cause and effect from observational data, using this method, rather as if the data really came from a randomised trial. The procedure involves some assumptions, in particular that the genes being studied cannot affect diabetes risk except by affecting blood pressure, but some checks on the validity of these assumptions can be made (and were made here). The upshot, in this new research, was that the mendelian randomisation estimate of the effect of a 5 mm Hg reduction in systolic blood pressure on the risk of type 2 diabetes, obtained by a different statistical approach on different data from a different kind of study, was almost the same as the estimate from the meta-analysis of the clinical trials. The estimate of the risk reduction found using mendelian randomisation was 12%, compared to 11% from the meta-analysis of data from randomised trials, with a margin of statistical error in the mendelian randomisation running from an 8% to 16% reduction in risk, compared to a margin of error of 5% to 16% from the meta-analysis. The close comparability of the two sets of results is very reassuring, and indicates that there are unlikely to be any major biases in either of the sets of results.”
Dr Marc George, Blood Pressure clinical lead for University College London Hospital (UCLH), said:
“Lowering blood pressure (BP) prevents heart attacks, strokes and kidney failure and this new large and comprehensive study published in The Lancet also shows that it lowers the risk of developing diabetes. Until now this effect was not clear. The study used data from multiple clinical trials and a genetic analysis to confirm it findings. A modest reduction in systolic BP of 5 mmHg reduced the risk of developing diabetes by 11%. Since diabetes is a common condition which causes significant harm this is an important finding. Also important was that not all BP-lowering drugs had the same effect. First-line BP drugs Angiotensin-Converting-Enzyme (ACE) inhibitors and Angiotensin-Receptor Blockers (ARBs) reduced the risk of diabetes whereas the study confirmed previous findings that second-line thiazide diuretics and third-line beta blockers actually increase the risk of diabetes in an affect unrelated to their BP lowering. However, it is vital to note that these drugs are still of net-benefit in lowering BP and the risk of cardiovascular disease as summarised in this previous, similar study in The Lancet (https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(15)01225-8/fulltext). Nonetheless, this new information re-enforces the status of ACE inhibitors and ARBs as a first-line treatments for hypertension and suggests they should be considered preferentially where a patient has diabetes or pre-diabetes.”
Dr Dipender Gill, NIHR Clinical Lecturer in Clinical Pharmacology and Therapeutics at St George’s, University of London, said:
“This is a comprehensive study triangulating clinical trial and genetic data to find support for effects of blood pressure reduction through particular pharmacological targets on glycaemic control and risk of type 2 diabetes. While the genetic evidence alone may be limited by lack of insight into the mechanisms by which the variants affect the pharmacological targets, the robust nature of the trial evidence provides convincing overall support.”
‘Blood pressure lowering and risk of new-onset type 2 diabetes: an individual participant data meta-analysis’ by Milad Nazarzadeh et al. was published in The Lancet at 23:31 UK time on Thursday 11 November.
Declared interests
Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee. My quote above is in my capacity as an independent professional statistician.”
Dr Marc George: “No conflicts.”
Dr Dipender Gill: “I am employed part-time by Novo Nordisk, unrelated to the work under consideration.”