“Taller men are more likely to develop prostate cancer”, the Daily Express reported. It said that an extensive study found the risk of the disease increases by between 6 and 12% for every extra four inches in height. The newspaper continued that this suggested that a short man who took part in the research had about a 19% smaller risk of developing prostate cancer compared to a man a foot taller than him.
This study examined the heights of men with prostate cancer compared to men without it. Overall, it did not find that the risk of prostate cancer significantly increased with height, but it did find a stronger link with high-grade prostate cancer. A meta-analysis of other studies found a combined 6% increase in risk with 10cm of height but their are inherent limitations that must be considered. While increased age is the most established risk factor for prostate disease, diet and environment are also thought to contribute. Although growth and height are to some extent affected by health and nutrition during childhood, stature is largely genetically determined. If there is a true association between height and increased prostate cancer risk, then the reasons for this remain unclear and require further study.
Where did the story come from?
Luisa Zuccolo and colleagues from the University of Bristol, University of York, and the Hull York Medical School, University of Cambridge, and University of Sheffield and Royal Hallamshire Hospital, carried out the research. The randomised controlled trial was funded by the National Health Service Health Technology Assessment Programme. Other support for this research came from the National Cancer Research Institute, Cancer research UK, World Cancer Research Fund, and a fellowship from the Cancer Epidemiology Unit, University of Turin.
The study was published in the peer-reviewed medical journal: Cancer Epidemiology Biomarkers and Prevention.
What kind of scientific study was this?
The study is a nested case-control study and a systematic review with meta-analysis. The researchers were interested in the theory that prostate cancer is associated with early childhood environment and that height is a marker of this. In other words, improved diet and health is associated with growth, and this, in turn, is associated with prostate cancer.
For the case-control study, the researchers used participants in the ongoing randomised controlled trial, ProtecT, which is examining the effectiveness of treatment for localised prostate cancer. This trial invited all men aged between 50 and 69 years, registered with 400 general practices across the UK, for prostate screening. For this case-control study, the researchers allocated men with a raised prostate specific antigen (PSA) level and histologically confirmed prostate cancer in the ProtecT group to the case group. The control group was made up of all the other men in the ProtecT group without prostate cancer, i.e. men who had a PSA level below the threshold for suspected cancer, or those with a raised level but for whom histological biopsy was negative. Each patient with prostate cancer (case) was matched to six controls of the same age and from the same general practice.
All the men were asked to complete a medical questionnaire before being given their PSA results or cancer status. Questions included their height and leg measurements, weight, ethnicity, diet, lifestyle, other medical conditions and their early childhood environment. After excluding men with incomplete questionnaires, 1,357 cases (67% of the total of those with cancer) and 8,343 controls remained for analysis. For these people, the researchers explored the association between prostate cancer and height, trunk and leg length, taking into account other potential risk factors for the disease, such as family history.
The systematic review-part of the study was performed by searching nine journal databases to find cohort or case-control studies that examined the association between height and prostate cancer (database details and search terms and search dates not given). Fifty-six papers reporting on 57 studies (30 cohort and 27 case-control) were suitable for inclusion in the review and meta-analysis. This included studies of various ethnic groups from a number of countries. When combining the studies, the researchers also separately analysed the studies that had focused on the association between height and prostate cancer from those that only incidentally looked at height.
What were the results of the study?
In the ProtecT nested case-control study, of the 1,357 men with confirmed prostate cancer, 173 had advanced stage disease, and 402 had a Gleason score (which describes how abnormal cancer cells look under a microscope) of seven or over. On a scale of 2-10, a score of seven indicates that the cells look less like normal cancer cells and have the potential to spread. The most aggressive abnormal cells are scored 8 to 10). There were no obvious differences between cases and controls in age, height, leg or trunk length, BMI, birth weight, ethnicity, occupation, or number of siblings. Family history was slightly stronger, however, in the case group compared to the control group (7.4% vs. 5.2%).
Overall, there was no significant increase in risk of prostate cancer for every 10cm increase in height (odds ratio 1.06, 95% confidence interval 0.97 to 1.16). Likewise, there was no significant increase in risk per 5cm increase in leg length or trunk length. There were no significant associations for any measure when separate analyses were carried out for those with localised or advanced prostate cancer, or those with low-grade prostate cancer (Gleason score less than seven).
However, for the 402 men with high-grade cancer, each 10cm increase in height meant a significant 23% increase in risk of cancer (odds ratio 1.23, 95% confidence interval 1.06 to 1.43). A 5cm increase in leg length also had a just significant increased risk, but not an increase in trunk length.
When results of 31 cohort studies from the systematic review were combined in meta-analysis (using a statistical model that took into account the different results and methodologies of individual studies) the researchers found there to be a significant 6% increase in risk of prostate cancer per 10cm increase in height (risk ratio 1.06, 95% confidence interval 1.03 to 1.09). Looking at only the 13 studies with advanced or aggressive prostate cancers gave a slightly stronger increased risk (risk ratio 1.12, 95% confidence interval 1.05 to 1.19).
Those studies that only incidentally looked at height did not find a significant association between increased height and cancer, but the 19 studies that had considered height a primary finding did find significant associations. The individual case controls showed more widely differing results and methodologies and so combining them would not have given a reliable overall result (results from pooling a few same-population studies were not significant).
What interpretations did the researchers draw from these results?
The researchers say that based on a systematic review of 57 studies and results from ProtecT, there was evidence for an increased risk of prostate cancer associated with increased stature, but that the overall size of the effect was modest and varied by study design. They say that the findings indicate a ‘limited role’ for height, as a proxy measure of childhood environment, influencing prostate cancer risk and possibly disease progression, but that the mechanisms for this require further investigation.
What does the NHS Knowledge Service make of this study?
This is a reliable piece of research that examined the associations between height and prostate cancer in a large group of men and has followed this with supporting evidence from a systematic review. The conclusions are sensible and reflect the current level of understanding about these associations. Points to consider:
- The ProtecT study did not find an overall significant increase in risk of prostate cancer with increased height, leg length or trunk length.
- The only positive association found was for those with high-grade cancer, but this sub-analysis is of a smaller number of people and this decreases the reliability of the risk estimate. There was no association between height and prostate cancer in the 936 men with low-grade cancer, however, this group may have included a number with potentially aggressive cancers that had the potential to become advanced but were only recently diagnosed. This would affect the reliability of the risk groupings into low- and high-grade.
- Not all men with prostate cancer were included in the study (only 67%). The results may have been different if the entire eligible cohort were included.
- Combining studies in meta-analysis means including studies with different methods, study groups and reliability, and this affects the reliability of the combined result (although the researchers did take stringent measures to take these things into account).
- There may have been inaccuracies in height measures in both ProtecT and the studies in the systematic review if these were only self-reports rather than measurements taken by the researchers.
- It is not possible to say how these associations between height and prostate cancer, if true, could have any influence on the prevention or treatment of prostate cancer.
Taller men should not be overly concerned by this research. The causes for prostate cancer are not completely understood and are probably a mix of age, lifestyle, ethnic and genetic factors. In particular, it is unclear whether it is increased stature in itself that could increase risk of prostate cancer or whether it is, as the researchers speculate, a result of childhood environmental factors such as diet and nutrition. This will require much further research.
Sir Muir Gray adds...
The headline should really have been ‘links between early childhood life and the risk of prostate cancer’.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
Daily Mail, 4 September 2008
The Daily Telegraph, 4 September 2008
Daily Express, 4 September 2008
Links to the science
Cancer Epidemiology Biomarkers & Prevention 2008; 17: 2325-2336
Cochrane Database Syst Rev 2006, Issue3