Clarifying the Risk of Lung Disease in SZ Alpha-1 Antitrypsin Deficiency

Abstract

Rationale: The ZZ genotype of alpha-1 antitrypsin deficiency (AATD) is associated with chronic obstructive pulmonary disease (COPD), even among never-smokers. The SZ genotype is also considered severe; yet, its effect on lung health remains unclear.

Objectives: To determine the effect of SZ-AATD on spirometry compared with a normal-risk population and to determine the effect of smoking cessation in this genotype.

Methods: We prospectively enrolled 166 related individuals, removing lung index cases to reduce bias, and compared spirometry between 70 SZ and 46 MM/MS individuals (control subjects). The effect of AAT concentrations on outcomes was assessed in 82 SZ individuals (including lung index cases). Subsequently, we analyzed retrospective SZ registry data to determine the effect of smoking cessation on spirometry decline (n = 60) and plasma anti–neutrophil elastase capacity (n = 20).

Measurements and Main Results: No difference between SZ and control never-smokers was seen. Ever smoking was associated with a lower FEV₁% predicted (−14.3%; P = 0.0092) and a lower FEV₁/FVC ratio (−0.075; P = 0.0041) in SZ-AATD. No association was found between AAT concentration and outcomes for SZ-AATD. Longitudinal analysis of 60 SZ individuals demonstrated that COPD at baseline, but not former smoking or AAT concentrations, predicted greater spirometry decline. Finally, anti–neutrophil elastase capacity did not differ between former smokers and never-smokers (P = 0.67).

Conclusions: SZ never-smokers demonstrated no increased risk of COPD, regardless of AAT concentration. Smoking interacts with SZ-AATD to significantly increase airflow obstruction. Former smoking alone is not associated with greater spirometry decline in SZ-AATD, suggesting that cessation attenuates the obstructive process. We found no evidence that the putative protective threshold or AAT concentrations predict risk within the SZ genotype, raising further doubts over the need for intravenous AAT augmentation in this cohort.

At a Glance Commentary

Scientific Knowledge on the Subject

Although the rarer and more severe ZZ genotype of alpha-1 antitrypsin deficiency is associated with chronic obstructive pulmonary disease even among nonsmokers, the risk relating to the SZ genotype is debated. This is largely due to the fact that no previous comparison of SZ individuals with a normal-risk population has been performed.

What This Study Adds to the Field

This study provides the first comparison of an SZ cohort with normal-risk control subjects (genotypes MM and MS) and examines the utility of alpha-1 antitrypsin concentrations within the SZ genotype for determining risk.

Alpha-1 antitrypsin deficiency (AATD) remains the most common monogenic risk factor for chronic obstructive pulmonary disease (COPD). AAT is a 52-kD protease inhibitor encoded by the SERPINA1 (serpin family A, member 1) gene on chromosome 14. As the main inhibitor of neutrophil elastase (NE), AAT is crucial in maintaining protease–antiprotease homeostasis and structural integrity of the lung parenchyma.

The risk of COPD in AATD is believed to be increased in individuals with AAT concentrations below the “putative protective threshold” of 0.57 g/L (11 μM; hereinafter referred to as the “PPT”) (1), though this theory remains unproved. Intravenous augmentation therapy with exogenous AAT aims to achieve nadir concentrations above the PPT and has been shown to slow lung density loss in severe AATD (2–4). Of the most common AATD genotypes, only the SZ genotype produces an AAT concentration that straddles the PPT (1, 5). Consequently, understanding SZ-AATD is crucial to clarifying the contribution of AAT concentrations to the risk of COPD and the role of augmentation therapy in this genotype (6).

ZZ-AATD is a major risk factor for COPD (7, 8). Recent studies also demonstrate an increased risk of COPD in smokers with MZ-AATD (9, 10). The evidence in SZ-AATD remains conflicted, with meta-analysis results suggesting an increased risk for COPD (11) and cohort studies demonstrating no definitive increased risk of lung disease in SZ subjects (10, 12). Previous studies have compared SZ-AATD with ZZ-AATD and MM-COPD cohorts (13, 14), but no study has compared SZ individuals with normal-risk populations.

We report the results of a prospective, family-based study assessing the risk of COPD in SZ-AATD. We explore the area further by analyzing longitudinal registry data to assess the effect of smoking cessation on lung function decline and the activity of AAT in plasma.

Methods

Study Design and Participants

In part 1 of the study, we performed a prospective, observational, family-based study to determine the impact of SZ-AATD on lung health. Ethical approval was obtained from the institutional medical ethics board at Beaumont Hospital (Research Ethics Committee no. 16/69). Individuals diagnosed with SZ-AATD through the Irish National Alpha-1 Antitrypsin Deficiency Targeted Detection Programme were offered inclusion in the study, with those diagnosed because of lung disease classified as “lung index.” Those diagnosed for other reasons were designated “non–lung index.” All first-degree full biological relatives of index cases were invited to participate. On the basis of the statistically most likely parental genotypes of an SZ individual being MS and MZ (15), we predicted that enrollment of index cases’ parents and siblings would provide a “control” cohort of MM and MS genotypes and a study cohort of SZ individuals (see Figure 1).

Figure 1.

Study population recruitment for the family study. SZ index cases (1) are invited to take part. Parents (2) and siblings (3) of index SZ cases are enrolled and predicted to provide SZ (orange) and MS/MM (control) participants for the comparison cohorts. MZ cases are excluded from analyses.

Inclusion required age over 30 years, capacity to give informed consent, and first-degree full biological relative status for participating relatives. Individuals with known confounders to spirometry measurements were excluded (interstitial lung disease, pulmonary/lobar resection, and neuromuscular disease). Visits were scheduled at least 4 weeks after any respiratory tract infections or exacerbation of COPD.

The family study protocol included spirometry, phlebotomy for AAT typing, quantification of serum AAT concentration, blood eosinophil counts, serum IgE concentrations, liver function tests, and surrogate markers of synthetic liver function (international normalized ratio/albumin). Pedigrees and smoking histories were documented at the time of testing. Participants were invited to complete a modified version of the American Thoracic Society (ATS) Division of Lung Disease questionnaire (16), either during recruitment using a handheld tablet or by e-mail using 2019 QuestionPro survey software (https://www.questionpro.com/). All participants were recruited and tested by a single field investigator to reduce interobserver bias.

The MS genotype has previously been shown not to be associated with an increased risk of COPD (11) or to be associated with worse airflow and lung function decline among those with COPD (17) when compared with MM individuals. Therefore, to increase the power of our analysis, MM and MS genotype individuals were first compared and then pooled and designated “control subjects.” Conversely, the MZ genotype precipitates airflow obstruction by way of an interaction with smoking (9, 10, 18). As such, MZ individuals were excluded from the final analysis.

In part 2 of the study, we performed a retrospective longitudinal analysis of SZ lung function data extracted from the Irish National Alpha-1 Antitrypsin Deficiency Registry to examine the effect of smoking cessation on lung function in SZ-AATD. Inclusion required at least two spirometry measurements over a time period of at least 24 months and documented “never-smoker” or “former-smoker” status at the start of follow-up. Patients with a primary diagnosis of AATD-unrelated conditions associated with lung function decline (e.g., interstitial lung disease) were excluded.

Spirometry

In the family study (part 1), spirometry was performed before and after administration of salbutamol (albuterol) according to ATS standards (19) using the EasyOne Diagnostic spirometer (model 2001-2S; ndd Medical Technologies). Percent predicted values were calculated using the European Respiratory Society European Coal and Steel Community reference equations (20). All spirometry outcomes assessed relate to post-bronchodilator values.

Longitudinal spirometry data (part 2) comprised “real-world” data recorded by the pulmonary function laboratory of a single tertiary referral hospital at the Irish Center for Genetic Lung Disease, Beaumont Hospital, Dublin, Ireland, and captured in the Irish National Alpha-1 Antitrypsin Deficiency Registry. Dl_CO was measured in accordance with the 2017 ATS/European Respiratory Society standards for single-breath carbon monoxide uptake in the lung (21).

Smoking History

“Never-smoker” was defined as a lifetime cigarette consumption of fewer than 20 packs of cigarettes (each pack equaling 20 cigarettes), or less than 12 ounces of tobacco. Pack-years were calculated multiplying the average daily number of cigarettes consumed by the number of years smoked and dividing by 20 ([average cigarettes/d × yr smoked]/20).

Computed Tomography Results

When available, chest computed tomography (CT) reports by hospital radiologists were reviewed for documented evidence of emphysema and confounding diagnoses such as interstitial lung disease.

AAT Typing and Serum Quantification

Concomitant C-reactive protein concentrations were measured to identify spuriously elevated AAT concentrations attributable to an acute-phase response (5, 22). AAT phenotype was determined by immunofixation of serum glycoforms via isoelectric focusing, performed using the Hydrasys electrophoresis platform (Sebia) and the Hydragel 18 A1AT Isofocusing kit (Sebia) (23). DNA was collected for genotyping in the event of inconclusive phenotyping. AAT concentrations were measured by turbidimetry, and internal quality control of assay accuracy was performed by comparing serial dilutions of the World Health Organization standard (WHO International Standard, 1st International Standard For Alpha-1-Antitrypsin, Plasma-Derived; National Institute for Biological Standards and Control code 05/162) (24) with measured values on the Olympus AU5800 platform (Olympus Corp.).

Anti-NE Capacity

Exogenous NE (Elastin Products Co.) was incubated with a range of concentrations of SZ plasma using NE alone as a control. Samples were mixed with fluorescence resonance energy transfer substrate, and fluorescence was recorded by spectrophotometry. NE activity was quantified by comparison with an NE standard curve of known NE activity. Anti-NE capacity was calculated via the percentage inhibition of NE from the plot of the percentage of remaining activity versus the plasma concentrations (see full methods in the online supplement).

Statistical Analysis

Statistical analysis was performed using RStudio version 1.1.463 software running R version 3.5.2 (www.cran.r-project.com). Continuous data were validated for normality of distribution using the Shapiro-Wilk test. Normally and nonnormally distributed data were analyzed using Student’s t test and Mann-Whitney U test, respectively. Categorical variables were analyzed using the chi-square test. Models with percent predicted spirometry results were adjusted for smoking history (ever-smoking and pack-years). Absolute spirometry results (in L), FEV₁/FVC ratio, forced expiratory times (FETs), and categorical COPD (defined as FEV₁/FVC ratio <0.7 in all analyses) were also adjusted for age, height, weight, and sex (and, for FET, the measured FVC). Analyses were adjusted for multiplicity using the Bonferroni correction.

In the family study, lung index cases were excluded from the final case–control analyses to remove referral bias. In SZ-only analyses, lung index cases were included and index status was modeled as a fixed effect. The effect of the SZ genotype on spirometry was defined in never-smokers and ever-smokers using mixed model analyses examining the covariates of age, genotype, smoking, and the PPT to assess differences between categories. We fit a linear mixed-effects model with predictors and confounders as fixed effects and a kinship coefficient matrix to model relatedness as a random effect (25). The association of the SZ genotype with COPD in all persons and in ever-smokers was performed with the GMMAT (generalized linear mixed model association test) R package (26) using a logistic mixed model adjusted for age, sex, pack-years of smoking, and familial relatedness (27).

In the longitudinal study of persons with SZ-AATD, a linear mixed model regression analysis modeling subject identifier and time as random effects (each subject with independent slope and intercept) was used to calculate age-, sex-, height-, and weight-adjusted slopes for each spirometric parameter against time (using lmer in R). Individual slopes were extracted using the random effects output of lmer and assessed for association with categorical fixed effects (never-smoker or former smoker, COPD at baseline, AAT concentrations above/below the PPT, and baseline presence of emphysema) in multivariable linear mixed models (glm in R).

Results

Between November 1, 2016, and January 30, 2019, we enrolled 166 participants from 44 SZ genotype–containing families comprising 82 SZ, 27 MS, 32 MZ, 19 MM, 4 ZZ, and 2 MI individuals. Lung index cases (12 SZ, 4 MZ, and 1 ZZ) were excluded from the final case–control analyses to remove referral bias. No significant difference in AAT concentrations was seen between lung index and non–lung index SZ individuals (median, 0.60 g/L vs. 0.61 g/L; P = 0.476) (Table E1 in the online supplement).

The characteristics of MM, MS, and MZ participants are included in Tables E2 and E3. Other than AAT concentrations (P < 0.001), no significant differences in baseline demographic, physiologic, or biochemical data were found. Furthermore, only the SZ cohort demonstrated an AAT range inclusive of concentrations both above and below the PPT (Figure E1), with 40.6% of non–lung index SZ subjects having AAT concentrations below the PPT.

The case–control population characteristics are summarized in Table 1. No significant age-related difference in spirometry between the control and SZ cohorts was noted in univariate analysis (Figure E2). Among ever-smokers, a significant difference in lung function outcomes was demonstrated between control and SZ cohorts in univariate regression analysis, with FEV₁% predicted (FEV₁pp) and FEV₁/FVC ratio both being more negatively correlated with pack-years (Figure 2).

Table 1.

Family Study Cohort Characteristics

Characteristics	Control Cohort	SZ Cohort	P Value
Subjects, n	46	70
Age, yr	53.39 ± 13.59	52.93 ± 10.93	0.84
AAT concentration, g/L	1.24 (1.13–1.40)	0.60 (0.50–0.68)	<0.001
CRP, mg/L	1.40 (1.05–1.75)	1.75 (0.8–3.0)	0.30
Sex, M	21 (45.7)	31 (44.3)	1
Ever-smoker	24 (52.2)	33 (47.1)	0.73
Pack-years smoked (ever-smokers)	18 (8.25–28.95)	14 (6–22)	0.41
BMI, kg/m2	27.30 ± 3.94	28.21 ± 4.93	0.29
Height, cm	167.91 ± 10.16	169.39 ± 10.88	0.47
Weight, kg	77.04 ± 16.20	81.36 ± 16.50	0.17
FEV1pp	99 (91–112)	103.5 (92.25–114)	0.51
FVCpp	113.27 ± 15.18	112.20 ± 19.7	0.76
FEV1/FVC ratio	0.74 (0.71–0.79)	0.77 (0.71–0.81)	0.20
MMEF25–75pp	65.87 ± 27.09	71.60 ± 31.66	0.32
Positive post-BD response	10 (21.7)	16 (23.5)	1
Hb, g/dl	14.30 ± 1.18	14.20 ± 1.15	0.67
Serum eosinophils, per 109/L	0.19 (0.11–0.31)	0.16 (0.11–0.26)	0.61
Bilirubin, μmol/L	9 (7–11)	9 (7–10)	0.67
ALT, U/L	24 (18–34.75)	30 (23–39)	0.025
ALP, U/L	80 (69–100.25)	89.00 (74–107)	0.07
GGT, U/L	28 (24–44.75)	29.5 (21–48.5)	0.96
INR	1.08 (1.02–1.14)	1.04 (1.00–1.06)	0.005
Albumin, g/L	45 (42.75–46)	44.50 (43–47)	0.8
IgE, IU/L	35 (7.25–114.25)	29 (15–78)	0.84

Definition of abbreviations: AAT = alpha-1 antitrypsin; ALP = alkaline phosphatase; ALT = alanine aminotransferase; BD = bronchodilator; BMI = body mass index; CRP = C-reactive protein; FEV₁pp = FEV₁ percent predicted; FVCpp = FVC percent predicted; GGT = γ-glutamyltransferase; INR = international normalized ratio; MMEF_25–75pp = maximal midexpiratory flow percent predicted.

Data are presented as mean ± SE for parametric data, median (interquartile range) for nonparametric data, and n (%) for categorical data unless otherwise specified.

Figure 2.

Univariate linear regression analyses of effect of pack-years smoked on spirometry. A significant difference in effect on (A) FEV₁ percent predicted (FEV₁pp) (P = 0.025) and (C) FEV₁/FVC ratio (P = 0.026) is seen. No significant difference was observed for (B) FVC percent predicted (FVCpp) or (D) forced expiratory time (FET).

In the final linear mixed model analyses, no significant difference was noted in spirometry between the never-smoker control and SZ cohorts (Table 2). Conversely, smoking exerted a significantly greater effect on FEV₁pp (−14.23% [95% confidence interval (CI), −24.94 to −3.52]; P = 0.009), FEV₁/FVC ratio (−0.075 [95% CI, −0.13 to −0.02]; P = 0.0041), and FET (+2.83 s [95% CI, +0.75 to +4.91 s]; P = 0.008) in ever-smoker SZ individuals versus control subjects. These findings were validated in a subanalysis that included SZ lung index smokers (modeling lung index status as a covariate, SZ FEV₁pp, −13.59% [95% CI, −24.33 to −2.86]; P = 0.013) and accentuated in a subanalysis of >20–pack-year smokers (SZ FEV₁pp, −23.14% [95% CI, −39.61 to −6.68]; P = 0.006).

Table 2.

Family Study Main Results

	Control Cohort (n = 46)	SZ Cohort (n = 70)	P Value
Never-smokers
FEV1pp	Ref	+4.85 ± 3.57	0.17
FEV1, L	Ref	+0.14 ± 0.11	0.19
FVCpp	Ref	−0.82 ± 5.15	0.87
FVC, L	Ref	+0.14 ± 0.14	0.3
FEV1/FVC	Ref	−0.02 ± 0.02	0.29
MMEF25–75pp	Ref	+9.86 ± 7.14	0.17
MMEF25–75, L	Ref	+0.37 ± 0.25	0.13
FET, s	Ref	−1.29 ± 0.93	0.17
Ever-smokers
FEV1pp	Ref	−14.23 ± 5.47	0.009
FEV1, L	Ref	−0.367 ± 0.18	0.045
FVCpp	Ref	−8.830 ± 5.29	0.09
FVC, L	Ref	−0.231 ± 0.20	0.26
FEV1/FVC	Ref	−0.075 ± 0.03	0.004
MMEF25–75pp	Ref	−8.502 ± 6.95	0.22
MMEF25–75, L	Ref	−0.399 ± 0.27	0.13
FET, s	Ref	+2.830 ± 1.06	0.008

Definition of abbreviations: FET = forced expiratory time; FEV₁pp = FEV₁ percent predicted; FVCpp = FVC percent predicted; MMEF_25–75pp = maximal midexpiratory flow percent predicted; Ref = reference population.

Data are presented as mean ± SE. Method: lmekin mixed model.

When further exploring the relationship of smoking to lung function, a significant interaction between pack-years of smoking and SZ-AATD (compared with control) was found in ever-smokers. The effect of pack-years of smoking on lung function in SZ versus control subjects was measured by FEV₁pp (−1.24 vs. −0.0615; P < 0.001), FVC% predicted (−0.657 vs. −0.126; P = 0.032), and FEV₁/FVC ratio (−0.007 vs. −0.003; P = 0.002) (Table E4). No significant increased risk of categorical COPD was seen for SZ individuals compared with family-based control subjects (odds ratio [OR], 3.1 [95% CI, 0.7–13.66]; P = 0.14). The effect of the SZ genotype tended to be greater, though this finding was not statistically significant (OR, 4.65 [95% CI, 0.51–42.3]; P = 0.17), when we restricted the analysis to ever-smokers.

Predictors of Outcome for the SZ Genotype

In the SZ-only analyses, smoking (pack-years) was associated with a decrease in all spirometry outcomes, greatest in FEV₁pp (−1.261 per pack-year [95% CI, −1.61 to −0.915]; P < 0.001), whereas lung index status also predicted lower FEV1pp (−22.18% [95% CI, −33.75 to −10.63]; P < 0.001). AAT concentrations were not significantly associated with any spirometry outcome when modeled either as a dichotomous outcome using the PPT or as numerical values in g/L (Tables 3 and E5).

Table 3.

Multivariate Analysis of Estimated Effect of Predictors on Spirometry Outcomes in SZ Cohort

Outcome (n = 82)	Covariate Effect
	Ever-Smoker*	Pack-Years†	Lung Index	Below PPT
FEV1pp	−15.63‡	−1.261§	−22.182§	0.830
FVCpp	−5.12	−0.775§	−4.248	−1.496
FEV1/FVC‖	−0.07‡	−0.006§	−0.125§	0.003
MMEF25–75pp	−18.53‡	−0.920§	−31.268§	−4.884
FET, s‖¶	+2.54**	0.164§	1.269	1.129

Definition of abbreviations: FET = forced expiratory time; FEV₁pp = FEV₁ percent predicted; FVCpp = FVC percent predicted; MMEF_25–75pp = maximal midexpiratory flow percent predicted; PPT = putative protective threshold (0.57 g/L; 11 μM).

All analyses include random variable (unique patient identifier) and adjustment for kinship matrix. Statistical model: lmekin.

^*When modeled without adjustment for pack-years.

^†When modeled with adjustment for pack-years + ever-smoker status.

^‡P < 0.01.

^§P < 0.001.

^‖Adjusted for age, sex, height, weight, and pack-years.

^¶Also adjusted for FVC (L).

^**P < 0.05.

Respiratory Symptoms

Respiratory symptom data revealed no significant differences between non–lung index SZ individuals and control subjects (Table 4). The overall response rate for respiratory symptom questions was 80% (133 of 166), including 12 of 12 SZ lung index (100%), 51 of 70 SZ non–lung index (72%), and 29 of 47 (62%) control cases.

Table 4.

Self-reported Symptom and Intervention Requirements among Family Study American Thoracic Society Division of Lung Disease Respondents

	Control Cohort	SZ Cohort	P Value	Method
Subjects, n	29	51
Symptoms
Cough	5 (17.2)	8 (15.7)	1	Chi-square test
Phlegm	10 (34.5)	6 (11.8)	0.031	Chi-square test
Wheeze	14 (50.0)	25 (49.0)	1	Chi-square test
mMRC dyspnea scale score			0.54	Chi-square test
N/A	0 (0.0)	2 (3.9)
1	16 (55.2)	33 (64.7)
2	11 (37.9)	12 (23.5)
3	1 (3.4)	1 (2.0)
4	1 (3.4)	3 (5.9)	0.67
Flare of chest symptoms in past yr	9 (31.0)	17 (34.0)	0.98	Chi-square test
Intervention requirement
Antibiotics for chest in past yr	8 (27.6)	8 (16.0)	0.35	Chi-square test
Steroids for chest in past yr	2 (6.9)	2 (4.0)	0.97	Chi-square test

Definition of abbreviations: mMRC = modified Medical Research Council; N/A = not applicable.

Data are presented as n (%) for categorical data unless otherwise specified.

Chest CT Results

By study completion, CT results from usual clinical follow-up were available for 50 of 82 SZ (60.09%) participants (Table E6). Emphysema was not reported in any never-smokers (0 of 17; mean age, 51.35 ± 12.69 yr) compared with 33% (11 of 33) of ever-smokers (mean age, 53.03 ± 10.63 yr), with this prevalence rising to 60% (9 of 15) among >20–pack-year smokers. Among those with CT-evident emphysema, 32 of 33 (97%) demonstrated an upper zone–predominant distribution, in contrast to the lower zone–predominant distribution typical of ZZ-AATD.

Longitudinal Lung Function Outcomes and Anti-NE Capacity in the SZ Genotype

Data from SZ individuals enrolled in the Irish National Alpha-1 Antitrypsin Deficiency Registry was filtered according to the inclusion criteria, yielding 60 individuals (Table 5). No difference in AAT concentrations was noted between never-smokers and former smokers (0.61 g/L for both; P = 0.89).

Table 5.

Longitudinal Assessment Cohort Baseline Characteristics

Characteristics	SZ Never-Smoker Cohort	SZ Former-Smoker Cohort	P Value
Subjects, n	27	33
Age at baseline, yr	46.78 ± 16.43	48.15 ± 15.22	0.74
AATD diagnosis age, yr	45.89 ± 16.33	48.73 ± 15.65	0.49
Sex, M	9 (33.3)	17 (51.5)	0.25
Lung index	2 (7.4)	17 (51.5)	0.001
Follow-up time, mo	63 (47–79.5)	58 (36–75)	0.29
Number of measurements	5.3 (2.81–7.8)	5.62 (4.10–9.01)	0.21
Pack-years smoked	0	20 (7.5–40)	—
AAT concentration, no CRP validation, g/L	0.56 (0.50–0.71)	0.58 (0.50–0.69)	0.96
AAT concentration below PPT	15 (55.6)	15 (45.5)	0.60
BMI, kg/m2	25.96 (22.11–27.34)	27.57 (23.23–30.48)	0.32
Baseline FEV1/FVC <0.7	5 (18.5)	16 (48.5)	0.032
Baseline FEV1pp	100.00 (92.50–108.00)	90.00 (67.80–100.00)	0.007
Baseline FVCpp	108.00 (101.00–115.50)	104.00 (93.00–123.00)	0.32
Baseline FEV1/FVC ratio	77.00 (71.50–82.00)	70.00 (57.00–78.00)	0.004
Baseline DlCOpp	90.00 (78.50–101.50)	83.00 (65.30–93.25)	0.047
CT-defined emphysema	0 (0.0)	11 (33.3)	0.002

Definition of abbreviation: AAT = alpha-1 antitrypsin; AATD = alpha-1 antitrypsin deficiency; BMI = body mass index; CRP = C-reactive protein; CT = computed tomography; Dl_COpp = Dl_CO percent predicted; FEV₁pp = FEV₁ percent predicted; FVCpp = FVC percent predicted; PPT = putative protective threshold (0.57 g/L; 11 μM).

Data are presented as mean ± SE for parametric data, median (interquartile range) for nonparametric data, and n (%) for categorical data unless otherwise specified.

Smoking status (ever vs. never), baseline COPD, and AAT concentrations (above vs. below the PPT) were assessed as predictors of slope of lung function (e.g., FEV₁ in ml/yr). In this model, only COPD at baseline was associated with greater decline in FEV₁ (−12.87 ml/yr [95% CI, −24.92 to −0.83]; P = 0.041) (Figure 3). Former smoking (n = 33) was not associated with greater decline in FEV₁ (−3.81 ml/yr vs. never-smokers; P = 0.51) (Table 6). Furthermore, below-PPT concentrations (Table E7), quantitative AAT concentration, or >20–pack-year smoking status was not associated with significant decline.

Figure 3.

Rate of change of FEV₁ (ml/yr) per individual categorized by (A) smoking status and (B) baseline presence of chronic obstructive pulmonary disease (COPD). Significant variability is demonstrated in both never-smokers and former smokers. Greater association with decline is seen in individuals with COPD at baseline (−12.87 ml/yr [95% confidence interval, −24.92 to −0.83]; P = 0.041 by mixed model analysis). ns = not significant.

Table 6.

Fixed Effect of Former Smoker Status on Lung Function Trends in Longitudinal Registry Cohort

	SZ Never-Smoker Cohort	SZ Former Smoker Cohort	P Value	Method
Subjects, n	27	33
ΔFEV1, ml/yr	Ref	−3.81 ± 5.71	0.51	glm
ΔFVC, ml/yr	Ref	−4.47 ± 5.53	0.42	glm
ΔFEV1/FVC, /yr	Ref	−0.01 ± 0.02	0.69	glm
ΔDlCO, ml/min/mm Hg/yr	Ref	+0.012 ± 0.07	0.86	glm

Definition of abbreviations: glm = generalized linear model; Ref = reference population.

Data are presented as mean ± SE unless otherwise specified.

Anti-NE Capacity Compared by Smoking Status

Plasma from 10 SZ never-smokers and 10 SZ former-smokers, with former smokers required to have quit for more than 6 months, was collected to compare anti-NE capacity after smoking cessation. The demographics of the anti-NE comparison cohort are included in Table E8. No significant difference was demonstrated between the anti-NE capacity of never-smokers and former smokers (Figure 4).

Figure 4.

Comparison of plasma anti–neutrophil elastase (anti-NE) activity of never-smoker (n = 10) and former smoker (n = 10) plasma (see full methods in the online supplement).

Discussion

Our findings suggest that SZ never-smokers are not at increased risk of lung disease and that neither the PPT nor AAT concentrations are useful for predicting risk in this genotype. Conversely, currently smoking SZ individuals are at a significantly increased risk of airflow obstruction compared with control smokers. Former smoking alone was not associated with accelerated decline in our registry cohort, suggesting that smoking cessation attenuates the interaction of smoking with SZ-AATD. A greater decline in FEV₁pp was seen in SZ former smokers with COPD, a finding also described in non-AATD populations (28). To adequately assess whether former smoker lung function decline differs between individuals with SZ-COPD and those with MM-COPD, a study directly comparing the two would be required.

In our study, we have accounted for the impact of referral bias to more precisely assess the risk for airflow obstruction for the SZ genotype compared with normal-risk genotypes. We present the largest number of non–lung index SZ individuals to date, and, for the first time, to our knowledge, we compare them with a well-matched control population. The adjusted OR for COPD in SZ smokers did not achieve statistical significance in our data (OR, 4.65; P = 0.17). Nonetheless, a significant difference in both the FEV₁/FVC ratio and the interaction effect of pack-years with SZ-AATD on FEV₁/FVC ratio was demonstrated. Furthermore, SZ smokers who were found to have emphysema by CT of the chest were noted to have an upper lobe–predominant distribution of disease, as opposed to the lower zone–predominant emphysema classically described in ZZ-AATD. These findings support those previously published by other investigators in larger CT-specific studies (29), suggesting that the pathophysiology of COPD in ZZ-AATD and SZ-AATD may differ significantly.

The hypothesis that individuals with SZ-AATD, and particularly those with concentrations below the PPT, are at increased risk of COPD is based on previously reported concentrations of AAT in this genotype (1), coupled with evidence of physiological anomalies in asymptomatic individuals with SZ-AATD (30), though the outcomes examined were not typical markers of COPD.

Whether genotype or AAT concentration is the greater predictor of lung disease in AATD is an ongoing point of discussion; however, the point may be moot. An overlap of AAT concentrations between SZ and ZZ cohorts is not seen outside of the acute phase, with ZZ concentrations not exceeding the PPT (1, 5, 22). Moreover, the S and Z isomers differ in their biochemical properties and interaction with NE (31, 32). Comparing different genotypes on the basis of AAT concentrations alone is therefore fundamentally flawed. Consequently, the true protective capacity of a given genotype is likely to be a composite effect of both the anti-NE/antiinflammatory capacity and the inherent ability of that genotype to mount an increase in AAT concentrations during the acute phase that is sufficient to meet the challenge of inflammatory insults. The evidence to date suggests that the ZZ genotype is insufficient on both these fronts while also promoting exaggerated protease activity (33–35). Other phenotypes, such as SZ or MZ, may have sufficient anti-NE/antiinflammatory capacity in general but less ability than the MM phenotype to meet the chronic inflammatory challenges of cigarette smoking.

If a protective threshold exists within the SZ range, it is essential that it be assessed within a population of SZ individuals alone to remove confounding by other genotypes. In previous studies, when the PPT has been examined as a categorical variable of the SZ population, no association with worse clinical status has been demonstrated, with CT findings being unrelated to AAT concentration and indeed both physiology and symptoms often worse among those above the PPT (14, 29). Furthermore, previous studies have reported a relative minority of SZ participants (∼10–20%) to have AAT concentrations below the PPT (1, 13, 14, 36). In these studies, the acute-phase response was not quantified, and consequently the mean AAT concentrations may have been transiently elevated. Our data suggest that a significantly higher proportion of SZ individuals (40.6%) than previously described have resting AAT concentrations below the PPT, with a range of 0.4–0.74 g/L when measured in the absence of acute-phase response and using a highly purified AAT standard (24). In our analyses, we have found no predictive value for the PPT or AAT concentrations in the SZ cohort.

To date, the only randomized placebo-controlled trial demonstrating the clinical efficacy of intravenous AAT (2) required participants to have AAT concentrations below the PPT. Of 180 participants, only 2 were SZ-AATD. As such, no clinical study has adequately assessed the benefits of intravenous AAT in SZ-AATD. Recommendations for the use of intravenous AAT previously specified concentrations below the PPT as an indication for treatment (37, 38), although more recently recommendations have been made against its use in those actively smoking or in those with the MZ genotype (39).

Despite these recommendations, intravenous AAT is prescribed to at least 1,000 MZ and SZ individuals in the United States alone (40, 41) at an estimated annual cost of over $80,000 per patient (42), or approximately $80 million each year. Moreover, these cohorts report active smoking rates of 7–11%, despite evidence that cigarette smoking may directly reduce the effectiveness of the therapy itself (43). Even more concerning is the fact that that some physicians prescribe intravenous augmentation for even the mild MS form of AATD (44). Our results suggest that there is no increased risk of reduced lung function in never-smoking individuals with SZ-AATD and no evidence of CT-defined emphysema among never-smoking SZ individuals.

Identification of SZ individuals who smoke and subsequent smoking cessation should be a focus of care with the aim of preventing the onset of COPD. Furthermore, we found no evidence that AAT concentrations or the PPT are useful predictors of risk in this genotype. Therefore, this study raises significant questions regarding the validity of the current PPT of 0.57 g/L as an indication for therapy or as a target for the efficacy of such therapy. Consequently, there remains an absence of clinical evidence to support the need for augmentation therapy in SZ-AATD.