Published ahead of print on July 22, 2004, doi:10.1165/rcmb.2004-0027OC
© 2004 American Thoracic Society DOI: 10.1165/rcmb.2004-0027OC Alpha1-Antitrypsin as a Risk for Infant and Adult Respiratory Outcomes in a National Birth CohortMedical Research Council National Survey of Health and Development, Department of Epidemiology and Public Health, Royal Free Hospital and University College London Medical School, London; and Galton Laboratory, Department of Biology, University College London, London, United Kingdom Address correspondence to: Michael E. J. Wadsworth, Medical Research Council National Survey of Health and Development, Department of Epidemiology and Public Health, Royal Free Hospital and University College London Medical School, 1-19 Torrington Place, London WC1E 6BT, UK. E-mail: m.wadsworth{at}ucl.ac.uk
Reduced alpha1-antitrypsin (AAT) encoded by the gene SERPINA1 is a potential risk for pulmonary disease. We investigated SERPINA1 polymorphism as a risk for infant and adult pulmonary morbidity, and adult respiratory function and its change between 43 and 53 yr. We used data on a British national representative sample (n = 5,362) studied since birth in 1946 to age 53 yr (when n = 3,035), when DNA was first obtained. SERPINA1 Z and, to a lesser extent, S carriers had an increased risk of infant lower respiratory infection compared with those who were neither S nor Z carriers (Z carriers: odds ratio = 2.32, 95% confidence interval = 1.37–3.92; S but not Z carriers odds ratio = 1.58, 95% confidence interval = 1.10–2.28) after adjustment for environmental, socioeconomic, and developmental factors, and breast-feeding. There was no difference in the adult outcomes at 53 yr according to genotype, nor was there any association of genotype with change in forced expiratory volume at 1 s between 43 and 53 yr. Lower alpha1-antitrypsin, as indicated by carrier status for the Z and S alleles, was a risk for infant lower respiratory infection, but not for adult respiratory outcomes.
Abbreviations: alpha1-antitrypsin, AAT • chronic obstructive pulmonary disease, COPD • forced expiratory volume at 1 s, FEV1 • forced vital capacity, FVC • lower respiratory infection, LRI • Global Initiative for Chronic Obstructive Lung Disease, GOLD
A number of genetic, developmental, and environmental factors increase risk of childhood and adult chest disease. One of the most widely studied of the genetic factors is variation in SERPINA1, which codes for alpha1-antitrypsin (AAT), a plasma serine protease inhibitor that is synthesized predominantly in the liver and, to a lesser extent, in alveolar macrophages. The main function of AAT in the respiratory tract is to protect the lungs from proteolytic damage caused by neutrophil elastase, a serine protease produced in response to inflammation (1–3). SERPINA1 is located within a SERPIN gene cluster on chromosome 14q32.1 (2). Many SERPINA1 alleles have been described, with two variants, S and Z, being associated with reduced serum levels of AAT. If individuals with the genotype MM are considered to have normal (i.e., 100%) levels of AAT, the lower levels seen in individuals with genotypes ZZ, SZ, MZ, SS, and MS are  16, 51, 83, 93, and 97%, respectively—the Z allele causing severe AAT deficiency and the S allele causing partial deficiency (4). Serum levels of AAT increase in response to inflammation (2) and insufficient levels of AAT, as can occur in ZZ homozygotes, lead to loss of elasticity and destruction of lung tissue and the development of progressive, irreversible chronic obstructive pulmonary disease (COPD) (2, 5). Risk of COPD is increased by smoking, but particularly so for ZZ homozygotes, because smoking induces an inflammatory response, and polymerization of the AAT Z may exacerbate this response (3). In addition, respiratory function, as measured by forced expiratory volume (FEV), which normally declines with age, declines more rapidly in ZZ homozygotes, and the effect is worsened by cigarette smoking (6), although reduced lung function was not found to be associated with SERPINA1 homozygosity in young adults (7). Signs of asthma are more common in AAT-deficient adults, particularly in homozygotes (7, 8). However, although Z carrier smokers are reported to be at greater risk of decline in FEV at 1 s (FEV1) than non–Z carrier smokers (6), other studies have not found lung function to be associated with SERPINA1 heterozygosity in adulthood (9). Similarly, some studies have shown the Z allele to be overrepresented in patients with COPD (10–14), and Seersholm and colleagues found that individuals with the MZ genotype who were first-degree relatives of ZZ patients with COPD were at increased risk of hospital admission for COPD (15). In contrast, other studies found that MZ individuals carried no greater risk for developing lung disease than nondeficient individuals when corrected for age, race, sex, and smoking history (16–18). The S allele was not found to be significantly associated with predisposition to COPD or with reduced FEV1 when compared with normal individuals (11, 12, 19). It has also been reported that the A allele of a G/A single nucleotide polymorphism, situated within an enhancer element 1,237 nucleotides 3' of the SERPINA1 gene, is significantly more frequent in patients with COPD than in healthy control subjects (20, 21), although these findings have also been conflicting, with no association found in other studies (10, 22). The presence of this allele is thought to be associated with reduced levels of binding of a transcription factor and an impaired inflammatory response (1, 2). Developmental and environmental factors are also associated with risk of COPD and poor respiratory function in adulthood. Epidemiologic studies show that low birth weight and childhood pulmonary disease are important factors in the progression to these adult outcomes (23–25). Poor socioeconomic environment in childhood, high exposure to parental smoking and atmospheric pollution, and a short or nonexistent period of exclusive breast-feeding increase the risk associated with developmental sources (24, 25). Availability of DNA for the first time in a national birth cohort study allows us to assess the role of genetic polymorphism in relation to developmental and environmental factors and outcome measures of infant pulmonary disease and adult respiratory health in a prospectively studied, nationally representative population, selected only by date of birth.
The Sample The Medical Research Council National Survey of Health and Development is a birth cohort study of a sample (n = 5,362) of all births that occurred in England, Wales, and Scotland in 1 wk in March 1946 (26). The sample includes all single, legitimate births whose fathers were in nonmanual or agricultural occupations and a randomly selected one in four of all other single, legitimate births (26). The sampling predates the major postwar immigration into Britain, and there are no non-European sample members. The sample successfully contacted in adulthood is representative of the national population of a similar age (26). Information on health, development, education, and occupation has been collected 22 times (Table 1). Research nurses visited homes to measure a range of health outcomes and to collect data on socioeconomic circumstances of subjects in adulthood. At the most recent collection, when subjects were 53-yr-old, nurses collected information from 3,035 cohort members, 83% of the 3,673 individuals who were still alive, resident in Britain, and who had not previously refused further contact (n = 640). At age 53 yr, we did not attempt to contact 31% (1,689) of the sample selected at birth (5,362). Either because they had, by that age, died (9%), were residing abroad (10%), or already refused all further contact (12%).
In this analysis the sample was, for most purposes, restricted to those who responded to the data collection at age 53 yr, because that was when a source of DNA was first collected. The aspect of analysis concerned with mortality used the whole cohort selected at birth.
The Measures Information on prenatal growth was provided by birth weight (g) recorded by midwives and health visitors from records when the child was aged 8 wk (information on state of maturity is not available). Postnatal growth was assessed in terms of weight (g) and height (cm) measured by health visitors when the child was aged 2 yr.
Infant feeding history was collected by health visitors at home visits when the child was aged 2 yr and coded here as breast fed for 1–4 mo,
Environmental data included information on parental smoking (asked in retrospect at age 53 yr: "Did either of your parents smoke cigarettes, cigars or pipes when you lived with them as a child?", and was coded as mother smoked yes/no, father smoked yes/no), and on atmospheric pollution from coal burning, which was taken from national official sources (23), summed for the years 1946 to 1948 (ages 0 to 2 yr), and used as a continuous variable. Socioeconomic and family data were crowding (0.5 persons per room; 1 person per room; > 1 person per room), birth order (1, 2, Adult health, environment, and survival. Research nurses, trained by the research team, measured FEV1 and forced vital capacity (FVC) at home visits when subjects were 43- and 53-yr old, using the Micromedical (UK) turbine electronic spirometer. Two measurements were taken, and the maximum readings are presented here. The difference in FEV1 at 43 and 53 yr is the decline in FEV1 over that 10-yr age period. When subjects were age 53 yr, the nurses also asked "Does your chest ever sound wheezy or whistling?" (yes/no) and, if "yes," "Do you get this most days or nights?" (yes/no). At the same visit, the nurses measured height using a portable CMS (UK) stadiometer, with the subjects head in the Frankfort plane. Subject smoking history was compiled from data collected at 53 yr and coded as never smoked, ever regularly smoked > 1 cigarette/d but not now, now smoking 1–10 cigarettes/d, 11–20/d, or > 20/d. Information on death was obtained from the National Health Service Register, on which all sample members are flagged, and causes and date confirmed by death certificate. The genetic measures. At the same visit, the nurses also collected blood and buccal samples from 91 and 96% of respondents, respectively. SERPINA1 phenotypes were determined by isoelectric focusing of AAT protein present in ethylenediamine tetraacetic acid plasma, as described by Whitehouse and colleagues (27), with the following exceptions. After treatment with DTT and IAA the samples were not further treated. Solutions used for contact between the gel and the electrodes were 0.1 M sodium hydroxide for the cathode and 0.04 M glutamic acid for the anode. Samples (5 µl) were applied to the gel using a multichannel pipette and an Immobiline sample application strip (Amersham Pharmacia Biotech, Buckinghamshire, UK), which was left in place throughout the run. This allowed us to distinguish four different classes of M allele as well as three other rarer alleles and S and Z. However, for most purposes, these protein alleles were classified as either S, Z, or M (to include all M and other rare alleles). Buccal DNA was prepared using DNA extraction solution kindly provided by Ian Craig (Institute of Psychiatry, University of London, UK). Blood DNA was prepared using the Puregene DNA Isolation Kit (Flowgen, Leicestershire, UK) according to the instructions of the manufacturer. The G1237A polymorphism was analyzed by polymerase chain reaction followed by restriction digestion with TaqI, as previously described by Sandford and colleagues. (22).
Statistical Analysis The population size for each protein allele of SERPINA1 and for G1237A is given in Table 2. The first analysis examined whether the genotype distributions of the G1237A polymorphism and the M, S, and Z protein alleles deviated from the Hardy-Weinberg equilibrium. Then we examined the allelic association of G1237A with the protein alleles of SERPINA1, using the expectation maximization (EM) algorithm from the Arlequin program (30).
Next, the association of SERPINA1 with the outcomes was examined using  2 and t tests. Then the association of infant LRI with the geographic, social, developmental, and environmental factors was examined, and the geographic distribution of carrier status was compared with that of infant LRI using univariate logistic regression and 2 tests. Logistic regression analysis was then used to test whether the association of carrier status with infant LRI could be accounted for by the developmental, social, and environmental factors. First, an unadjusted model was used; then separate models, adjusting for social, developmental, nutritional, and environmental factors; and finally a model adjusting for all factors. Interactions of SERPINA1 with atmospheric pollution and with parental smoking were tested, because of their inflammatory action, and because other studies (6) report greater FEV1 decline in SERPINA1 Z carrier smokers than in SERPINA1 Z carrier nonsmokers. The associations of SERPINA1 with FEV1, the GOLD criteria, FEV1/FVC < 70%, and reported wheeze, all at 53 yr, were examined using logistic and normal regression analysis as appropriate, adjusting for height, sex, and smoking history. These analyses were repeated with the G1237A polymorphism. Finally, the difference between FEV1 at 43 and 53 yr was examined in relation to SERPINA1 carrier status, and a conditional regression model was generated with FEV1 at 53 yr as the outcome, adjusting for FEV1 at 43 yr, sex, adult height, and smoking history. The interaction of FEV1 decline with smoking history was tested. Because loss to follow-up is an inevitable problem in long-term longitudinal studies, we looked at how that might have affected the results. First, we used a Cox proportional hazard model to see whether loss through death in the population of subjects who had had infant LRI differed from that in the population of those who did not have that illness. Second, we investigated whether the population for which we have SERPINA1 data at 53 yr (90% of those providing information at that age) differed from others not seen at that age in terms of known respiratory risk factors and experience of infant LRI. The level of statistical significance was taken as P = 0.05 throughout. Analyses were performed using SPSS, Inc. (Chicago, IL). The n values differ between analyses because data are taken from several data collections with missing values in each, with consequent reduction in n in analyses using data from > 1 collection.
The distribution of demographic characteristics of the sample is given in the APPENDIX Table A1.
Allelic Distribution in the Sample
Allelic Association
Infant LRI and SERPINA1 SERPINA1 Z and S carriers each had a raised risk of infant LRI compared with noncarriers (Table 4). The risk was highest in Z carriers, lowest in noncarriers, and intermediate in S carriers. The G1237A polymorphism was not associated with infant LRI (data not shown).
The Association of SERPINA1 and Environmental Factors with Infant LRI Infant LRI was associated with all the social factors (P  0.01), but was not associated with sex. The developmental factors were not associated with infant LRI, but breast-feeding was a significant protective factor (P = 0.02). Atmospheric pollution and parental smoking were risks (P 0.001 and P 0.04, respectively). Table 5 shows that in all but two regions, risk of infant LRI was higher in S and Z carriers. However, this was a statistically significant risk in only two regions, as the number of Z carriers was too small for this analysis, as the wide confidence intervals show. The geographic distribution of carrier status and of infant LRI is given in APPENDIX Table A2.
In an unadjusted logistic regression model, SERPINA1 was associated with infant LRI (Table 6), and that association remained significant in models adjusting for each group of risk factors and for all factors (Table 6). The unadjusted analysis shown in Table 6 was repeated (not shown) excluding the ZZ homozygotes and SZ heterozygotes (n = 6) in order to determine whether the effect of those high-risk genotypes accounted for the findings, but they did not; infant LRI remained strongly associated with the three groups (noncarriers odds ratio [OR] = 1.00, S carriers OR = 1.35 [95% confidence interval {CI} = 0.96–1.88], Z carriers OR=2.34 [95% CI=1.46–3.77] P = 0.001). Rerunning each of the models shown in Table 6, allowing inclusion of all possible sample members in each (not shown), yielded similar results to those presented in Table 6.
Tests for interactions of SERPINA1 with maternal and paternal smoking and with atmospheric pollution were not statistically significant (P = 0.99, P = 0.98, and P = 0.22, respectively, data not shown).
SERPINA1 and Adult Respiratory Health Although mean decline in FEV1 between ages 43 and 53 yr was greatest in the Z carrier group, differences in mean decline across the three categories of SERPINA1 carrier status were not significant (P = 0.49, data not shown), nor was a regression model of carrier status in relation to conditional change in FEV1 (P = 0.65). There was no interaction between FEV1 decline and smoking history (P = 0.89).
Mortality in Relation to Infant LRI
Missing Sample Members
To our knowledge, this is the first study to examine the relationship of SERPINA1 with infant and adult respiratory outcomes in the same sample. We have the advantage of measurements and data collection on risk exposures made throughout the life of sample members in this longitudinal study (birth to 53 yr), with the exception of data on parental smoking, which were collected in retrospect. The DNA resource from which SERPINA1 information was derived was collected at age 53 yr. We show that SERPINA1 played a significant and independent role in the risk of infant LRI, but there was no interaction with atmospheric pollution, as might be expected from the findings of von Ehrenstein and colleagues (31). The results suggest that the level of AAT may be critical during this developmental period (1). Although we found no evidence of interactions between parental smoking and SERPINA1 in relation to infant LRI, this may not be a reliable result, as parental smoking was recollected. The absence of an interaction with atmospheric pollution may be the result of the small number of Z carriers. The absence of association of SERPINA1 with adult respiratory outcomes was similar to the findings of others, who also failed to show an association with heterozygosity for SERPINA1 S and Z (9, 16–18). We also failed to find an association with reported wheeze, which others have found in those severely deficient in AAT (7, 8). Our findings may reflect the nature of our cohort in which, although relatively large, only 6 individuals would be expected to be severely deficient, and the fact that, unlike other large studies, our sample was not taken from a population of patients. We have no information on AAT deficiency in relatives, which Seersholm and colleagues (15) showed to be a risk. Our study is thus similar to the sample used by Silva and colleagues (9), who also reported no association, but differs from that of several hospital-based studies, in which associations were found (10–15, 19). It may be that in this sample at later ages, as the risk of COPD and the slope of functional decline each increase, their association with SERPINA1 will change (10). Those for whom we have no DNA sample at 53 yr, but who had provided data 10 yr earlier, had lower mean FEV1 at that age (P = 0.01), suggesting that they would also have had a lower FEV1 at 53 yr. This, together with the significantly higher exposure to atmospheric pollution in infancy among those not providing data at 53 yr (P = 0.05), and a greater extent of contact loss among those who were smokers at 43 yr (P < 0.001), indicates that the effects of smoking and infant exposure to atmospheric pollution may have been underestimated in our analyses. The significantly elevated risk of death before age 53 yr (when the source of DNA was first collected) among those who had infant LRI indicates that the observed associations between SERPINA1 alleles and the disease outcomes in infancy and adulthood are likely to be weaker than the true effect of SERPINA1 on respiratory function. However, death rates by any cause in this population are representative of those in the national population of a similar age (26).
SERPINA1 was shown to be a risk for infant LRI in this community sample, independently of environmental factors. However, unlike infant LRI, SERPINA1 was shown not to be a significant risk for adult respiratory outcomes by age 53 yr. There was no significant association of the G1237A polymorphism with either infant LRI or adult respiratory outcomes.
Carrier status by region
Region by infant lower respiratory infection
Region by infant lower respiratory infection (S carriers only)
Region by infant lower respiratory infection (Z carriers only)
The authors are grateful to Prof. Sue Povey of the Department of Biology at University College London for comments and encouragement; to the Medical Research Council for funding the MRC National Survey of Health and Development, and the collection of data and material for making DNA; to Prof. Ian Craig and Mr. Bernard Freeman at the Molecular Genetics Group, Institute of Psychiatry, London for their advice and help with buccal DNA preparation; and to the National Centre for Social Research, whose research nurses collected the data and biological samples.
Conflict of Interest Statement: M.E.J.W. has no declared conflicts of interest; L.E.V. has no declared conflicts of interest; A.L.J. has no declared conflicts of interest; R.J.H. has no declared conflicts of interest; D.B.W. has no declared conflicts of interest; S.L.B. has no declared conflicts of interest; W.S.H. has no declared conflicts of interest; J.U.L. has no declared conflicts of interest; and D.M.S. has no declared conflicts of interest. Received in original form January 27, 2004 Received in final form July 19, 2004
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
5 mo, or bottle fed. 
1-antityrpsin heterozygotes of phenotype PiMZ. Am J Respir Crit Care Med 161:81–84.
