“Short people are more likely to develop heart disease,” BBC News reported. It said that men under 5ft 4in (163cm) and women under 5ft (152cm) were 1.5 times more likely to develop and die from heart disease than tall adults.
The news story is based on a large review of 52 studies in more than 3 million people. The review does suggest that there is a link between shorter height and risk of cardiovascular events. However, it is not clear why this is the case, or exactly how strong the link is.
The press release for this study makes the important point that height is only one factor shown to be associated with heart disease risk, and height cannot be controlled, while other factors, such as weight and lifestyle habits, can. Individuals of all heights can aim to reduce their risk of cardiovascular disease by reducing their exposure to modifiable risk factors.
Where did the story come from?
The study was carried out by researchers from Tampere University Hospital. It was funded by the Finnish Cultural Foundation, Tampere University Hospital, the Aarno Koskelo Foundation and the Finnish Foundation for Cardiovascular Research. The study was published in the peer-reviewed European Heart Journal.
What kind of research was this?
This was a systematic review and meta-analysis of studies looking at whether short stature is associated with coronary heart disease. The researchers say that the first report of a link between shorter stature and increased coronary heart disease risk was published in 1951, and that almost 2,000 studies have dealt with this question since then. They say that, although there have been several reviews of this issue, none have systematically assessed and pooled the results of the research to date.
A systematic review is the best way of identifying and summarising the best quality research evidence about a particular question. Meta-analyses pool the results from multiple studies, and can increase the detection of differences between observed groups when compared with individual studies. They can also increase the precision of the results obtained. When carrying out a meta-analysis researchers need to ensure that the studies are similar enough for their pooling to make sense. Statistical tests can be used to determine whether the studies appear to be similar enough to justify pooling them.
What did the research involve?
The researchers carried out searches of established scientific literature databases (MEDLINE, PreMEDLINE and All EBM Reviews) to identify studies looking at the relationship between height and coronary heart disease. Their final search was carried out in December 2007. They then selected relevant studies that met their inclusion criteria and used reference lists of these studies to identify more relevant studies. The results of the included studies were then pooled to determine how short stature affected the risk of cardiovascular outcomes.
The researchers only included systematic reviews, meta-analyses, randomised controlled trials, clinical trials, cohort or case-control studies. To be eligible, studies also had to:
- include more than 200 participants
- be in healthy people or people with symptomatic coronary heart disease at the start of the study
- look at the effect of height as a continuous variable or compare different height groups
- look at important outcomes, including death from any cause, death from cardiovascular disease, death from coronary heart disease or other cardiovascular outcomes
- cohort studies had to follow up individuals for at least two years to assess these outcomes
Studies that only looked at height as a confounding factor were excluded, as were studies only looking at birth height, and non-English language studies. Two researchers independently assessed whether studies met inclusion criteria and a third reviewer resolved any differences of opinion. The researchers graded the quality of these studies using set criteria (with the maximum score being 15). One researcher extracted data from the included studies and two researchers checked these data.
The identified studies compared differing height categories. The researchers decided to compare the shortest group in each study with the tallest group, rather than specifying in advance what would be considered ‘short’ or ‘tall’. The researchers were interested in obtaining the relative risks [RRs] of each outcome: the proportion of people having an outcome in the shorter group divided by the proportion having the outcome in the taller group. The RR was either taken from the included papers or calculated using the available data where possible. Where an odds ratio (OR, which is a related but not identical measure) was provided, the researchers used a formula to calculate the RR from this number.
The researchers used accepted statistical methods to investigate whether the included studies had significantly different results. This analysis did show that there were differences between the studies, which suggests that the results of the pooling should be cautiously interpreted. This led the researchers to use methods that take into account differences between studies.
What were the basic results?
The researchers’ initial searches identified 1,902 articles, and 52 of the studies described in these articles met their inclusion criteria. These studies included more than 3 million people in total (3,012,747 individuals). Twenty-two of these studies could be included in the statistical pooling of results as they either presented RRs or had enough data to enable the calculation of RRs. The studies were given quality scores ranging from 7 to 14 (with the highest score possible being 15).
Across the studies, on average, short individuals were less than 160.5cm tall (about 5ft 3in) and tall individuals were over 173.9cm (about 5ft 8in). Short individuals were about 35% more likely to die from any cause during follow-up than tall individuals (relative risk [RR] 1.35, 95% CI 1.25 to 1.44).
Shorter individuals were also about 50% more likely to die from cardiovascular disease (CVD), to have or die from coronary heart disease (CHD) or to have a heart attack than taller individuals (RR for CVD death 1.55, 95% CI 1.37 to 1.74; RR for CHD 1.49, 95% CI 1.33 to 1.67; RR for heart attack 1.52, 95% CI 1.28 to 1.81).
Overall, the shortest adults were 46% more likely to have one of the cardiovascular outcomes assessed than the tallest (54 results pooled from 22 studies; RR 1.46, 95% CI 1.37 to 1.55).
Being short was associated with increased risk of these outcomes in both men and women.
How did the researchers interpret the results?
The researchers concluded that the “relationship between short stature and CVD appears to be a real one”. Adults within the shortest height category had about a 50% higher risk of coronary heart disease and death from this cause than the tallest individuals.
This study used the most appropriate design for summarising the available high quality research evidence on a question. It includes a large number of individuals and its findings are likely to be reasonably robust. The explanation for this link between height and cardiovascular risk is unclear, but it seems unlikely that short stature itself ‘causes’ this increase in risk, and more likely that it is another linked factor. The authors suggest that short individuals may have smaller coronary arteries and that this possibility needs looking into. As individuals cannot be randomised to being different heights, studies looking at this link can only observe what happens in the general population, and as such they are affected by the possibility of confounding.
Individuals who are shorter may differ from taller individuals in a number of ways, for example, in their socio-economic status, nutrition, general health and ethnicity. These other factors may themselves be contributing to the difference in cardiovascular risk seen between the shorter and taller groups. This is known as confounding.
The study’s strengths include:
- the fact that it systematically searched for and assessed studies, and used set inclusion/exclusion criteria to decide whether studies were eligible
- the large number of studies and individuals involved
There are a few limitations:
The researchers pooled the relative risks across a range of different cardiovascular outcomes and gave a summary RR of 1.46 for the combined outcome. This also involved pooling multiple different outcomes from individual studies. It is not clear how appropriate this technique is. Although it tells us that overall risks of the outcomes assessed appear to be increased, it cannot tell us which outcomes are increased. This is because, for example, non-fatal heart attacks were counted on their own in some studies but only deaths from heart attack in others. The inclusion of multiple results from individual studies may inadvertently increase the strength of the link seen.
- Individual studies will have taken into account different potential confounding factors. These adjustments are likely to lead to differing strengths of association between height and cardiovascular risk. This means that it is hard to judge the degree of confounding remaining in the pooled result.
- The researchers did find statistical evidence that the pooled studies had differing results, suggesting that the pooled results should be interpreted cautiously. Although they did use appropriate analysis methods, ideally the researchers would have investigated why the study results differed (for example, whether the differences were due to differing study designs, populations or outcomes assessed).
- The exact number of individuals included in each meta-analysis was not reported, nor were the absolute risks of the events in the individual studies.
Overall, the results of this large review suggest a link, but why this link exists is not clear. It is not possible to say how strong the link would be if all known cardiovascular factors were taken into account. Importantly, the findings do not mean that tall people are protected from heart disease, and they should pay attention to the same modifiable risk factors as shorter people: stopping smoking, improving diet and increasing physical activity.