(Associate Editor: Chi Chiu Leung)
Lung function and plasma fibrinogen concentrations in the Newcastle Thousand Families birth cohort between age 49 and 51 years
Article first published online: 23 DEC 2013
© 2013 The Authors. Respirology © 2013 Asian Pacific Society of Respirology
Volume 19, Issue 1, pages 53–57, January 2014
How to Cite
Pearce, M. S., Hancox, R. J., Parker, L. and Gibson, G. J. (2014), Lung function and plasma fibrinogen concentrations in the Newcastle Thousand Families birth cohort between age 49 and 51 years. Respirology, 19: 53–57. doi: 10.1111/resp.12158
- Issue published online: 23 DEC 2013
- Article first published online: 23 DEC 2013
- Accepted manuscript online: 22 JUL 2013 05:40AM EST
- Manuscript Accepted: 10 JUN 2013
- Manuscript Revised: 6 JUN 2013
- Manuscript Revised: 29 MAY 2013
- Manuscript Received: 14 MAY 2013
- British Council
- Breathe North
Background and objective
A number of studies have suggested inverse associations between lung function and inflammatory markers, including fibrinogen. In this study, we used data from the Newcastle Thousand Families birth cohort to assess the association between contemporaneous markers of lung function and fibrinogen while adjusting for potential confounding factors throughout life.
At age 49–51 years, complete data on lung function and plasma fibrinogen were available for 380 study members. These data were analysed in relation to each other, adjusted for sex and height, with further adjustment for potential confounders within linear regression models using robust estimates.
Forced expiratory volume in 1 s was significantly inversely associated with plasma fibrinogen concentration after initial adjustments for sex and height (beta = −0.12, P = 0.011) and remained so after further adjustments for pack-years of cigarettes smoked and current smoking status. On further adjustment for standardized birthweight and duration breast-fed, the association approached statistical significance (P = 0.051). Adjusting for body mass index (BMI) resulted in a loss of significance (P = 0.09), but an unchanged regression coefficient, while, after adjustment for percent body fat, rather than BMI, the association was no longer significant (P = 0.20) and the coefficient reduced.
The association between lung function and fibrinogen remains after adjustment for potential early-life confounders and smoking. However, it is not independent of contemporaneous measures of adiposity, with evidence of confounding by percent body fat. Further studies, with measures of adiposity, are required to confirm whether associations between markers of inflammation and lung function are due to residual confounding by adiposity.
body mass index
forced expiratory volume in 1 s
forced vital capacity
Reduced lung function is associated with an increased risk of adverse cardiovascular events and risk factors.[1-5] This has led to the suggestion that impaired lung function and cardiovascular disease may share certain aetiological features. Part of the link is suggested to be the result of cigarette smoking, a primary risk factor for both outcomes. However, this explanation does not appear to be sufficient given the occurrence of both cardiovascular disease and impaired lung function among non-smokers.
Systemic inflammation has emerged as an alternative explanation. Increased concentrations of inflammatory proteins, such as C-reactive protein and fibrinogen, have been identified as a key feature of cardiovascular disease. However, these findings have also raised a number of questions about the exact role of systemic inflammation in influencing lung function, in particular whether it is purely symptomatic, for example as a result of ‘overspill’ from airway inflammation or whether there is a causative association.
A number of cross-sectional population-based studies have examined the relationship between systemic inflammation and lung function. The majority identify an inverse association between lung function and serum concentration of either C-reactive protein[9-13] or fibrinogen.[14-17] However, most of these lack information on potential confounders in early life, such as birthweight and breast-feeding that have been linked to both respiratory function and inflammation[19-21] in adulthood. Further, where adjustments have been made for body size, body mass index (BMI) has most often been used but is a measure influenced more by build than percent body fat.
The Newcastle Thousand Families birth cohort[23, 24] provided a rare opportunity to investigate the possible association between fibrinogen and lung function, while controlling for potential confounding factors acting before birth, in infancy, childhood and adulthood, including two different measures of adiposity.
The Newcastle Thousand Families study began as a prospective study of all 1142 children born in May and June 1947 to mothers resident in Newcastle upon Tyne, UK.[23, 24] The health, growth and development of the cohort were followed in great detail, initially up to age 15 years. The cohort underwent a major follow-up at age 49–51 years. Participants in the current investigation were members of the cohort who were either traced through the National Health Service Central Register or contacted the study team in response to media publicity. Between October 1996 and December 1998, health and lifestyle questionnaires were sent out for completion and return, and study members invited to attend for clinical examination, which took place over the same time period. The study received ethical approval from the appropriate local research ethics committees, and all participants gave their written consent.
Forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) were measured using a portable spirometer (Vitalograph Compact II, Vitalograph, Buckingham, UK) by one of two trained research nurses. The highest value from a minimum of three attempts was recorded for each individual, and these values were used for calculation of the FEV1/FVC ratio. Fasted plasma fibrinogen concentration was derived, on the day of the clinical examination, from prothrombin time using an automatic coagulator. Height, weight, and waist and hip circumferences were measured according to the protocol of the World Health Organisation MONICA project, and BMI was derived. Per cent body fat was estimated from bioelectrical impedance measured using a Holtain body composition analyser (Holtain Ltd, Crymych, Wales, UK).
Information on early life was recorded prospectively for all study members.[23, 24] Birthweights, as recorded by the midwife at the time of the child's birth, were standardized for gestational age and sex. Duration breast-fed was defined as the length of time a study member was at least partly breast-fed, as recorded prospectively by the health visitors.
The number of pack-years of cigarettes smoked (1 pack-year = 1 pack of cigarettes smoked per day for 1 year) was estimated from the study members' smoking habits at ages 15, 25, 35 and 50, as ascertained at age 49–51 years. Current smoking status (at the time of questionnaire completion at age 49–51 years) was also derived (never, ex-smoker, current smoker).
Twins were excluded from all analyses due to potential lack of independence in the statistical models. Current smoking status was treated as a categorical variable. All other variables were treated as continuous. Associations between plasma fibrinogen levels and FEV1, both measured at age 49–51 years, were analysed using multiple linear regression, with FEV1 as the dependent variable, with all models adjusted for sex and height. Further adjusted models included potential confounders in a stepwise fashion to allow the potential extent of a confounding relationship to be shown. Huber/White estimators of variance were used to provide robust P-values and confidence intervals (CI) due to the presence of heteroskedasticity, as determined using the Cook–Weisberg test. Regression coefficients (b) and corresponding 95% CI were estimated within the regression models. Beta coefficients, estimated in regression models when all variables are standardized to have mean 0 and standard deviation 1, were also calculated. Analyses were repeated for both FVC and the FEV1/FVC ratio. All statistical analyses were done using the statistical software package Stata, version 12 (StataCorp, College Station, TX, USA).
Of the original 1142 study members, 832 were traced out of the 932 children whose families remained in Newcastle for at least the first year of life. Of these, 574 completed the health and lifestyle questionnaire, and 412 attended for clinical examination. The sample of the cohort included in this investigation has previously been shown to be representative of the original cohort except for sex, with women more likely to attend the clinical examination than men (P < 0.001). After the exclusion of 11 twins, 4 individuals on anti-thrombin medication (warfarin) and a further 17 for whom fibrinogen or usable respiratory data were not available, a complete dataset was available for 380 singleton study members (166 men and 214 women).
Descriptive statistics for all continuous variables included in this study are given in Table 1, with medians and interquartile ranges given for those variables significantly non-normally distributed, and means and standard deviations otherwise. Of the 378 individuals in this study with recorded cigarette smoking data, 156 (41%) had never smoked, 120 (32%) were former smokers, and 102 (27%) were current smokers. BMI and per cent body fat were highly correlated (r = 0.65) so were not included together in adjusted models due to possible colinearity.
|FEV1 (L) at age 49–51 years||380||2.9||2.5–3.6|
|FVC (L) at age 49–51 years||380||3.9||3.4–4.7|
|FEV1/FVC ratio at age 49–51 years||380||0.75||0.71–0.78|
|Plasma fibrinogen (g/L) at age 49–51 years||380||2.9||2.5–3.3|
|Height (cm) at age 49–51 years||379||167||147–190|
|Duration breast-fed (weeks)||372||9.3||3.8–30.1|
|Pack-years of cigarettes smokeda||222||20.5||7–30|
|Percent body fat at age 49–51 years||377||39.3||8.7|
FEV1 was significantly inversely associated with plasma fibrinogen concentration after initial adjustments for sex and height (beta = −0.12, P = 0.011) and remained so after further adjustments for pack-years of cigarettes smoked and current smoking status (Table 2). A beta value of −0.12 means that a standard deviation change in fibrinogen (0.69 g/L) is associated with around a 10th of a standard deviation change in FEV1 in the opposite direction. As the standard deviation for FEV1 was 0.73 L, this means that a one standard deviation reduction in fibrinogen is associated with a 0.09 L increase in FEV1. On further adjustment for the early-life variables (standardized birthweight and duration breast-fed), the association approached statistical significance (P = 0.051). Adjusting for BMI resulted in a loss of significance (P = 0.09), but an unchanged regression coefficient, while after adjustment for per cent body fat rather than BMI, the association was no longer significant (P = 0.20) and the coefficient was reduced (beta = −0.07). On limiting the study to the 156 individuals who had never smoked, no association was seen between FEV1 and fibrinogen (b = −0.12, 95% CI = −0.30 to 0.07, beta = −0.10, P = 0.21), although the regression coefficient was the unchanged, while in those who had ever smoked, the association remained significant (b = −0.24, 95% CI = −0.37 to −0.10, beta = −0.23, P = 0.001), including after adjustment for height and sex (b = −0.13, 95% CI = −0.23 to −0.04, beta = −0.13, P = 0.006).
|Beta||Coefficient (95% CI)||P||Beta||Coefficient (95% CI)||P|
|Sex and height||379||−0.12||−0.12 (−0.21 to −0.03)||0.011||−0.09||−0.13 (−0.23 to −0.03)||0.013|
|Sex, height and pack-years smoked||377||−0.10||−0.11 (−0.20 to −0.01)||0.024||−0.08||−0.12 (−0.22 to −0.02)||0.020|
|Sex, height, pack-years and current smoking status||377||−0.09||−0.10 (−0.19 to −0.004)||0.041||−0.08||−0.12 (−0.22 to −0.01)||0.028|
|Sex, height, pack-years, current smoking status, standardized birthweight and duration breast-fed||369||−0.09||−0.09 (−0.19 to 0.00)||0.051||−0.08||−0.12 (−0.23 to −0.02)||0.025|
|Sex, height, pack-years, current smoking status, standardized birth weight, duration breast-fed and BMI||369||−0.08||−0.09 (−0.19 to −0.01)||0.093||−0.07||−0.11 (−0.22 to −0.01)||0.072|
|Sex, height, pack-years, current smoking status, standardized birth weight, duration breast-fed and per cent body fat||367||−0.07||−0.07 (−0.18 to 0.04)||0.195||−0.06||−0.09 (−0.21 to 0.03)||0.159|
Similar results were seen for FVC (Table 2). There was no significant association between plasma fibrinogen and the FEV1/FVC ratio (coefficient adjusted for sex and height = −0.005, 95% CI = −0.02 to 0.004, P = 0.28).
There were no significant interactions between sex and plasma fibrinogen concentration on any of the respiratory outcomes (P > 0.3). BMI and per cent body fat were highly correlated (r = 0.65) and hence were not included together in adjusted models due to possible colinearity. All models were valid in terms of not violating the assumptions surrounding linear regression. Using a log-transformed form of duration breast-fed made very little difference to the findings (results not presented).
This study showed significant inverse associations between both FEV1 and FVC and plasma fibrinogen concentration at age 49–51 years when adjusting for the standard adjustment variables, sex and height. These associations remained significant after adjusting for a number of potential confounders, including those operating in early life and of total pack-years of cigarettes smoked and contemporaneous BMI. However, for both FEV1 and FVC, the association was not statistically significant after adjusting for contemporary per cent body fat, although regression coefficients were similar throughout all adjustments, even in the case where statistical significance was lost.
The main strength of this study is the ability to analyse prospectively collected data from different stages of life simultaneously and allow cross-sectional associations at age 49–51 years to be adjusted for potential confounders, including those from much earlier in life. The data collected at age 49–51 years included a direct estimate of per cent body fat based on bioelectrical impedance, which, in contrast with BMI, is largely independent of build and may represent a more accurate measure of potential confounding by level of adipose tissue than do more build-based measures.
Of 1142 men and women recruited at birth in 1947, 380 (33%) were included in this study. As with all studies of this nature, a concern should be that mortality and morbidity in study members may introduce a bias, where those with severe disease may have either not survived or been well enough to take part in the study. Other than sex, with proportionally more women than men from the original cohort being included, the included study members were representative of the original cohort in terms of the early-life variables considered. What is more, inclusion of cohort members who had moved out of the study region (18% of those who participated in the health check were resident outside the Northern Region) increased the representativeness of the population studied. The main limitation of this study is statistical power. It is possible that with a larger sample size, the associations between lung function and plasma fibrinogen would have remained significant after adjustment for per cent body fat. Body fat is known to impair lung function because of the physical effects of obesity on the thorax. It is also known that adiposity is positively associated with inflammation.[28-30] In this cohort, per cent body fat has been shown to be significantly associated with FEV1 (P < 0.001), with no such association for BMI (P = 0.91). Further, a previous analysis of the fibrinogen data used in this study showing significant associations between fibrinogen and both BMI and per cent body fat at the univariate level, but only for per cent body fat in the adjusted model. Thus, a confounding role is possible, and the difference in associations with FEV1 between per cent body fat and BMI in this cohort may explain why per cent body fat is a more likely confounding factor than BMI. However, it may also be possible that adiposity may impair lung function through an inflammatory pathway. Other inflammatory measures, such as high-sensitivity C-reactive protein, were not available in this investigation.
A number of studies have shown associations between markers of systemic inflammation and lung function in healthy individuals.[9-11, 13-17, 31] This group of studies includes a number that have identified inverse associations between fibrinogen and measures of lung function.[14-17] Consistent with these previous studies, inverse associations, present after adjustment for smoking, were found in this study between fibrinogen and both FEV1 and FVC. The slight attenuation of the coefficients and P-values on adjustment for smoking history and current smoking status may mean that fibrinogen is a predictor of lung function but on the causal pathway between cigarette smoking and lung function. Our previous analysis of the fibrinogen data as an outcome showed significantly reduced fibrinogen levels among current non-smokers. No association was seen between fibrinogen and time since stopping smoking among the former smokers. With further data collection, we hope to address the temporal issues of the current study in the future. The associations also remained after adjustment for both standardized birthweight and duration breast-fed, both of which are suggested to be related to both lung function and inflammation[19-21] in adulthood. To our knowledge, this is the first time such an adjustment has made when assessing these associations.
It was not possible to assess the temporal nature of the associations investigated in this study, so we cannot rule out that a decrease in lung function may result in an increase in fibrinogen levels rather than the other way around. However, a previous study showed that measurements of fibrinogen in young adults were inversely related to both FEV1 and FVC in middle age, although the authors noted limitations in the adjustment for confounding by obesity. A study of individuals born in Dunedin, New Zealand found an association between serum C-reactive protein and lung function in young adults, independently of BMI, suggesting that the association exists before either disease is present in a clinical form. If the results of this study are confirmed by others, longitudinal studies of changes in all three of inflammation, lung function and levels of adipose tissue will be the most helpful. This will allow the exploration of potential issues such as reverse causality.
In conclusion, the inverse association between lung function and plasma fibrinogen remains after adjustment for potential early-life confounders and lifetime smoking history. However, in this study, it is not independent of percent body fat. While this may be a chance finding, or down to a lack of statistical power, it may also suggest that studies that have adjusted for BMI may have overadjusted for build and underadjusted for adipose tissue. Further studies, with measures of both build and adiposity, are required to confirm whether associations between markers of inflammation and lung function are due to residual confounding by adipose tissue, and if so, by what mechanism.
We thank all the Newcastle Thousand Family Study members for taking part in this study and the study teams past and present, particularly Professors Alan Craft and George Alberti, who along with Professor Louise Parker, led the age 49–51 year follow-up, and Emma Thompson who provides the administrative support for the study. We thank the previous funding bodies who have contributed to this study since it began. The preliminary data analyses for this paper were done during a Research Exchange visit by MSP to the University of Otago, funded by the British Council with further funding contributed by Breathe North.
- 6Atherosclerosis, COPD and chronic inflammation. Respir. Med. COPD Update 2008; 4: 60–65..
- 25World Health Organisation. MONICA Manual Part III, Section I. World Health Organisation Cardiovascular Diseases Unit, Geneva, 1990.