Lifestyle and exercise

Male habits knock decade off life

“Cigarettes and alcohol will take 10 years off your life,” said The Independent. The newspaper reported that for the first time doctors have quantified the effects of smoking, high blood pressure, and high cholesterol, described as “the three major killers of middle-aged men”. Failing to give up smoking or to control blood pressure and cholesterol were reported to reduce life expectancy by 10 to 15 years.

The Whitehall study that provided the data set for this new publication is a large cohort study that started in 1967. It provides over 30 years of follow-up data for cause-specific mortality in a large population of civil servants. This study found that there has been a clear improvement in rates of cardiovascular deaths across the decades, and also an improvement in the control of high blood pressure and high cholesterol, in addition to a reduction in smoking rates. However, despite this, those with the combined risk factors of smoking, high cholesterol and high blood pressure when aged 50 were found to live an average of 10 years less than those without.

The study has only been conducted in a specific population of men but the findings concur with the numerous other health studies demonstrating the impact of smoking, blood pressure and cholesterol upon health and mortality. The study did not specifically assess alcohol use.

Where did the story come from?

This research was conducted by Robert Clarke and colleagues of University of Oxford, University College London Medical School and the London School of Hygiene and Tropical Medicine. The study was funded by the British Heart Foundation and Medical Research Council, and was published in the British Medical Journal.

What kind of scientific study was this?

The Whitehall study is a large cohort study collecting many types of data from workers in the civil service. The researchers of this new study used data from the Whitehall study to assess life expectancy in relation to the three major cardiovascular risk factors in middle-age: smoking, high cholesterol and high blood pressure. It used data collected from 19,019 men aged from 40 to 69 years.

At each participant's entry into the study (baseline) initial health assessments took place (full medical history, examination and investigations, including blood, heart and lung tests). Subjects were entered into the study from 1967 to 1970. Researchers used Office for National Statistics procedures to trace the records of 18,863 of the men (99%) up to 2005. Of these men, 13,501 had died during this period, with a cause of death recorded for 84% of them using standard coding systems. In only 43 cases, cause of death was classified as unknown.

In 1997-98, all 8448 surviving participants were re-invited to take part in follow-up assessments. A total of 7044 (83%) responded, having their blood pressure, height and weight measured. They were also requested to take blood tests, which 77% of them provided. For 4811 men (57% of the surviving cohort), baseline and follow-up data was available on blood pressure, blood cholesterol and body mass index (BMI).

To compare their data with UK mortality trends the researchers obtained annual cause-specific death rates between 1950 and 2005 from the World Health Organization (WHO) and looked at the death rates in middle-age (35 to 69 years) and old age (70 to 79 years).

The researchers used the data to compare trends of cardiovascular and non-cardiovascular mortality in the Whitehall study with those of the general UK population. This included data on how life expectancy beyond age 50 related to the three risk factors individually and combined. They also used the data to estimate life expectancy in relation to more accurate characterisations of a range of cardiovascular risks, such as diabetes and BMI in addition to the main risk factors.

What were the results of the study?

The WHO data showed that from 1950 to 2005, standardised rates of UK mortality due to cardiovascular causes in middle-aged men were about twice of those for middle-aged women. However, for both men and women there has been an improvement in mortality rates over time, with a decline of about 2% a year from 1970 onwards. There has also been a decline in the proportion of deaths attributed to cardiovascular disease. The Whitehall study demonstrated a similar pattern to this UK data.

At the start of the Whitehall study, 42% of the men currently smoked, 39% had high blood pressure, and 51% had high cholesterol. At the re-examination in 1997, 13% were current smokers and 58% were ex-smokers, with an average quitting age of 52 years). Only one-third of those smoking at study start were still smoking. 

For those classed as having low and high blood pressure, the average (mean) differences between high and low readings had declined by two-thirds across the study period (30.6mmHg difference in 1967 compared with 8.3mmHg difference in 1997). A similar pattern was seen for low and high levels of cholesterol (1.86mmol/l difference compared with 0.49mmol/l).

These reductions imply that both high blood pressure and high cholesterol were better controlled 30 years on. However, there was less of a decline in BMI difference between obese and non-obese individuals between the start and the end of the study.

About a quarter of all deaths in the cohort occurred before the age of 70. More deaths during middle-age were attributed to cardiovascular causes than during older age. Compared with men without any baseline risk factors, the presence of all three risk factors (current smoking, high cholesterol and high blood pressure) at the start of the study was associated with a life expectancy 10 years shorter (23.7 extra years from age 50 compared with 33.3 extra years).

The researchers assigned participants a score based on smoking, diabetes, blood pressure, cholesterol, BMI and employment grade. Compared with men in the lowest 5% of risk, men in the highest 5% had a 15-year shorter life expectancy from age 50 (20.2 compared with 35.4 years). The single factor of a participant being a smoker upon entry to the study was associated with an average life expectancy 6.3 years shorter than those of non-smokers.

What interpretations did the researchers draw from these results?

The researchers conclude that, despite substantial changes in the risk factors of smoking, cholesterol and blood pressure over the 30 years, baseline differences in risk factors were associated with 10 to 15 year shorter life expectancy from age 50 onwards.

What does the NHS Knowledge Service make of this study?

The Whitehall study provides over 30 years of follow-up data for cause-specific mortality in a large population of male civil servants, and offers an opportunity to assess the impact that cardiovascular risk factors can have on middle-age on life expectancy.

The study found that there has been a clear improvement in cardiovascular mortality rates across the decades, plus improvements in the control of high blood pressure and high cholesterol, in addition to a decline in smoking rates. However, despite these improvements, those with the combined risk factors of smoking, high cholesterol and high blood pressure at the age of 50 had life expectancy around 10 years below that of men without these risk factors.

This valuable study assessed a large number of men and followed them over 30 years from middle to old age. Also, although a large number of these men died over this period, the researchers had an almost complete set of mortality information available for use in their analysis. However, there some limitations:

  • The study has examined men only, and being a population of civil servants, the participants may have slightly different health behaviour and lifestyles than the general population. However, there are demonstrated similarities between mortality patterns in this group and these seen in the general population, so the issue is unlikely to affect applicability to other groups too widely.
  • The effect of medical management and drug therapies in mitigating these risk factors over time, although assumed, cannot be directly assessed.
  • The effects of smoking frequency and duration cannot be clearly assessed, only very broad definitions of currently smoking or not smoking are given.
  • Measures of smoking, blood pressure and cholesterol were only taken at baseline and again many years later for a smaller group of survivors. The study therefore does not capture changes in exposure status during the 30-year follow-up period and the effect that this may have on outcomes. For example, some people may have stopped smoking and some people may have started smoking).
  • Despite newspaper headlines, this study does not seem to have assessed the impact of alcohol upon life expectancy.

The findings concur with the numerous other health studies that demonstrate the impact that smoking, blood pressure and cholesterol can have upon health and mortality. While this will be unsurprising to most people, the study has strengths in providing an estimation of the size of this risk. As the authors conclude, continued public health strategies to lower these risk factors could result in further improvements in life expectancy.

NHS Attribution