Global life expectancy has increased over the last 150 years, and it has more than doubled since 1900 [1
]. People born at the beginning of the 1800s lived an average of 30 years [1
], whereas those born in 2019 have a life expectancy of 73 years [2
]. Many developed countries have life expectancies above 80 years, with Japan currently ranked highest [2
]; thus, longevity has reached epidemic proportions [3
]. Several studies have focused on explanatory variables for longevity, and genetic factors have been found to explain approximately 25% [4
]. Thus, modifiable
risk factors (i.e., lifestyle factors) may be the most important determinants for reaching old age.
Early population-based cohort studies offer a unique possibility to study the effect of multiple risk factors in midlife on survival to advanced ages. Some of the first to conduct such a study were Terry et al., who examined the association between cardiovascular risk factors measured at the ages of 40–50 years and survival to the age of 85 in the Framingham Heart Study (FHS) [5
]. The results showed the predictive power of a simple risk score, formed by counting the number of risk factors, e.g., women with all risk factors (five in that study) had a 14% chance of reaching the age of 85 compared to 65% among those with no risk factors [5
]. The importance of midlife risk factors on longevity was later confirmed in studies including men only [6
]. However, to the best of our knowledge, research on the association between number of risk factors in midlife and survival to old age in women is scarce.
The aim of this study is to examine the association between single risk factors and a number of risk factors in midlife and older age (up to 64 years) and survival to the age of 85 years in women, using data from a large population-based cohort in Norway.
We found that daily smoking, physical inactivity, being unmarried, and obesity in midlife (age 45–49 years) were significant single risk factors for death before age 85. The chance of reaching the age of 85 decreased gradually with increasing number of risk factors: from 67% for those with no risk factors to 28% for those with four or five risk factors. Those who quit smoking and those who became physically active prior to Tromsø4 (age 60–64) had higher survival than those who continued to smoke and those who remained physically inactive, respectively.
The ability of the number of risk factors (risk score) to predict survival to age 85 is in accordance with other studies that included men only [6
] and with the FHS, which included both men and women [5
]. However, the variables considered differed between this study and the FHS, which may be partly explained by different methodologies. In the FHS, stepwise model selection was used to determine which variables to include in the risk score; these ultimately included sex, systolic blood pressure, serum cholesterol, glucose intolerance, smoking history, and education. In contrast, we chose six risk factors a priori and included total cholesterol, but not glucose intolerance. This choice was partly based on what was found to predict survival to age 90 in men in The Tromsø Study [6
]. Also, except for marital status, these risk factors are some of the established risk factors for death identified in women in the Global Burden of Disease (GBD) project, though the GBD moved from estimating total cholesterol in 2016 to estimating LDL cholesterol in 2017 [19
A physical activity index was considered in the FHS but was not included in the final (most parsimonious) model. In contrast, we found that being physically active was a significant predictor of survival and longevity, consistent with other studies [20
]. Another difference between our results and those of the FHS is that we did not observe any effect of high systolic blood pressure on survival, and we saw a non-significant association with total cholesterol. In contrast to our findings, the GDB project ranked systolic blood pressure as the number one risk factor for death in women [19
]. However, further analyses of our data (presented in the Appendix A
) showed an association between a stricter definition of high blood pressure and survival.
Though women who were daily smokers in midlife had a lower chance of reaching age 85, the proportion who did so was still 45%. This relatively high number of survivors can be explained by the fact that many quit daily smoking later in life (Figure 3
). This higher survival among quitters is consistent with previous findings from the Netherland Cohort Study [21
]. However, that study had the limitation of using baseline information to define quitters due to lack of follow-up surveys, whereas we were able to use information from later surveys. We observed no improvement in survival time among daily smokers who did and did not quit smoking between Tromsø2 and Tromsø3, which is similar to what was observed previously for men [6
]. This phenomenon may be difficult to comprehend and we have no explanation for these null findings; however, we did observe that quitters had significantly higher survival than persistent smokers when we compared Tromsø2 and Tromsø4 and later surveys. We interpret this as strong evidence that quitting smoking increases the probability of survival.
As previously found for men in the Tromsø Study [6
], we observed that women who were married in midlife had higher survival than those who were unmarried, a result that is in accordance with other studies [22
]. Unlike other studies [5
], we did not find a significant effect of education. However, the risk factor of obesity became insignificant when education was included in the model. A similar finding was reported for BMI in Terry et al. [5
All six of the investigated risk factors are modifiable, thus individuals can make lifestyle changes in midlife or later in life that may improve their longevity. We consider this to be an important message to communicate to the public. At the same time, lasting lifestyle changes can be difficult to achieve without the support of community-based interventions. For instance, since 1975, various legislative measures were implemented in Norway, which is a likely cause for the large decline in cigarette smoking [15
Strengths of this study include the prospective study design with a follow-up time of 40 years, anthropometric measurements that were performed by trained personnel, and objective measurements obtained from blood samples. Another strength is the population-based sample and the high attendance rate among women aged 45–49 years (90.6%) in Tromsø2 and subsequent surveys (Figure 3
We did not include morbidity-free survival in this analysis, which may be considered a limitation, as increased attention has been placed on healthy aging. However, previous knowledge suggests that factors that promote longevity are highly correlated with delayed morbidity [23
]. Another limitation is the observational study design, which means that we cannot claim causality, only associations. Finally, there are other risk factors that may be of importance to reach longevity. Thus, our conclusion comes with the limitation that only six risk factors were considered.