Article Text

Original research
Hepatobiliary phenotypes of adults with alpha-1 antitrypsin deficiency
  1. Malin Fromme1,
  2. Carolin V Schneider1,
  3. Vitor Pereira2,
  4. Karim Hamesch1,
  5. Monica Pons3,4,
  6. Matthias C Reichert5,
  7. Federica Benini6,
  8. Paul Ellis7,
  9. Katrine H Thorhauge8,
  10. Mattias Mandorfer9,
  11. Barbara Burbaum1,
  12. Vivien Woditsch1,
  13. Joanna Chorostowska-Wynimko10,
  14. Jef Verbeek11,
  15. Frederik Nevens11,
  16. Joan Genesca3,4,
  17. Marc Miravitlles12,
  18. Alexa Nuñez12,
  19. Benedikt Schaefer13,
  20. Heinz Zoller13,
  21. Sabina Janciauskiene14,
  22. Nélia Abreu2,
  23. Luís Jasmins2,
  24. Rui Gaspar15,
  25. Rodrigo Liberal15,
  26. Guilherme Macedo15,
  27. Ravi Mahadeva16,
  28. Catarina Gomes17,
  29. Kai Markus Schneider1,
  30. Michael Trauner9,
  31. Aleksander Krag8,
  32. Bibek Gooptu18,19,
  33. Douglas Thorburn19,20,
  34. Aileen Marshall19,20,
  35. John R Hurst19,21,
  36. David A Lomas19,21,
  37. Frank Lammert5,22,
  38. Nadine T Gaisa23,
  39. Virginia Clark24,
  40. William Griffiths25,
  41. Christian Trautwein1,
  42. Alice M Turner7,
  43. Noel G McElvaney26,
  44. Pavel Strnad1
  1. 1 Medical Clinic III, Gastroenterology, Metabolic Diseases and Intensive Care, University Hospital RWTH Aachen, Health Care Provider of the European Reference Network on Rare Liver Disorders (ERN RARE LIVER), Aachen, Germany
  2. 2 Department of Gastroenterology, Centro Hospitalar do Funchal, Madeira, Portugal
  3. 3 Liver Unit, Hospital Universitari Vall d’Hebron, Vall d’Hebron Research Institute (VHIR), Universitat Autonoma de Barcelona, Barcelona, Catalunya, Spain
  4. 4 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Comunidad de Madrid, Spain
  5. 5 Department of Medicine II, Saarland University Medical Center, Saarland University, Homburg, Germany
  6. 6 Gastroenterology Unit, Department of Medicine, Spedali Civili and University, Brescia, Italy
  7. 7 Institute of Applied Health Research, University of Birmingham, Birmingham, UK
  8. 8 Department of Gastroenterology and Hepatology, Odense University Hospital, Odense, Denmark
  9. 9 Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Health Care Provider of the European Reference Network on Rare Liver Disorders (ERN RARE LIVER), Vienna, Austria
  10. 10 Department of Genetics and Clinical Immunology, National Tuberculosis and Lung Diseases Institute, Warszawa, Poland
  11. 11 Department of Gastroenterology & Hepatology, KU Leuven University Hospitals Leuven, Health Care Provider of the European Reference Network on Rare Liver Disorders (ERN RARE LIVER), Leuven, Flanders, Belgium
  12. 12 Pneumology Department, Hospital Universitari Vall d'Hebron, Vall d’Hebron Institut de Recerca (VHIR), Vall d’Hebron Hospital Campus, CIBER de Enfermedades Respiratorias (CIBERES), Barcelona, Spain
  13. 13 Department of Internal Medicine I, Medical University of Innsbruck, Innsbruck, Tirol, Austria
  14. 14 Department of Respiratory Medicine, Hannover Medical School, Hannover, Germany
  15. 15 Gastroenterology Department, Centro Hospitalar de São João, Faculty of Medicine of Porto University, Porto, Portugal
  16. 16 Department of Respiratory Medicine, Cambridge University Hospitals, Cambridge, UK
  17. 17 Gastroenterology Department, Centro Hospitalar de Vila Nova de Gaia Espinho EPE, Vila Nova de Gaia, Porto, Portugal
  18. 18 NIHR Leicester BRC-Respiratory and Leicester Institute of Structural & Chemical Biology, University of Leicester, Leicester, Leicestershire, UK
  19. 19 London Alpha-1 Antitrypsin Deficiency Service, Royal Free Hospital, London, UK
  20. 20 Sheila Sherlock Liver Unit and UCL Institute for Liver and Digestive Health, Royal Free Hospital, London, UK
  21. 21 UCL Respiratory, Division of Medicine, University College London, London, UK
  22. 22 Hannover Medical School (MHH), Hannover, Germany
  23. 23 Institute of Pathology, University Hospital RWTH Aachen, Aachen, Germany
  24. 24 Division of Gastroenterology, Hepatology, and Nutrition, University of Florida, Gainesville, Florida, USA
  25. 25 Department of Hepatology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, Cambridgeshire, UK
  26. 26 Irish Centre for Genetic Lung Disease, Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin, Ireland
  1. Correspondence to Dr Pavel Strnad, Medical Clinic III, Gastroenterology, Metabolic diseases and Intensive Care, RWTH Aachen University, Aachen 52074, Germany; pstrnad{at}


Objective Alpha-1 antitrypsin deficiency (AATD) is a common, potentially lethal inborn disorder caused by mutations in alpha-1 antitrypsin (AAT). Homozygosity for the ‘Pi*Z’ variant of AAT (Pi*ZZ genotype) causes lung and liver disease, whereas heterozygous ‘Pi*Z’ carriage (Pi*MZ genotype) predisposes to gallstones and liver fibrosis. The clinical significance of the more common ‘Pi*S’ variant remains largely undefined and no robust data exist on the prevalence of liver tumours in AATD.

Design Baseline phenotypes of AATD individuals and non-carriers were analysed in 482 380 participants in the UK Biobank. 1104 participants of a multinational cohort (586 Pi*ZZ, 239 Pi*SZ, 279 non-carriers) underwent a comprehensive clinical assessment. Associations were adjusted for age, sex, body mass index, diabetes and alcohol consumption.

Results Among UK Biobank participants, Pi*ZZ individuals displayed the highest liver enzyme values, the highest occurrence of liver fibrosis/cirrhosis (adjusted OR (aOR)=21.7 (8.8–53.7)) and primary liver cancer (aOR=44.5 (10.8–183.6)). Subjects with Pi*MZ genotype had slightly elevated liver enzymes and moderately increased odds for liver fibrosis/cirrhosis (aOR=1.7 (1.2–2.2)) and cholelithiasis (aOR=1.3 (1.2–1.4)). Individuals with homozygous Pi*S mutation (Pi*SS genotype) harboured minimally elevated alanine aminotransferase values, but no other hepatobiliary abnormalities. Pi*SZ participants displayed higher liver enzymes, more frequent liver fibrosis/cirrhosis (aOR=3.1 (1.1–8.2)) and primary liver cancer (aOR=6.6 (1.6–26.9)). The higher fibrosis burden was confirmed in a multinational cohort. Male sex, age ≥50 years, obesity and the presence of diabetes were associated with significant liver fibrosis.

Conclusion Our study defines the hepatobiliary phenotype of individuals with the most relevant AATD genotypes including their predisposition to liver tumours, thereby allowing evidence-based advice and individualised hepatological surveillance.

  • fibrosis
  • liver
  • liver cirrhosis
  • cancer

Data availability statement

Data are available upon reasonable request. The deidentified participant data are available from the corresponding author upon request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Significance of this study

What is already known about this subject?

  • Pi*Z and Pi*S are the most important genetic variants causing alpha-1 antitrypsin deficiency (AATD).

  • No reliable data on hepatobiliary phenotype exist for individuals with Pi*SS and Pi*SZ genotype, despite the fact that both genotypes are seen in ~1:500 Caucasians.

  • The vast majority of subjects with AATD remain undiagnosed during their lifetime and this fact complicates AATD phenotyping.

What are the new findings?

  • In a large community-based Biobank, as well as in a multinational cohort, subjects with Pi*SZ genotype were markedly predisposed to liver fibrosis and seemed to be predisposed to primary liver cancer.

  • Compared with the characteristic severe AATD genotype Pi*ZZ, Pi*SZ genotype causes intermediate hepatobiliary phenotype, while Pi*SS does not seem to have major hepatobiliary consequences.

Significance of this study

How might it impact on clinical practice in the foreseeable future?

  • Our study defines the hepatic risks associated with the major AATD genotypes. These data, together with the individual situation/susceptibility factors, should guide the counselling and management of individuals with AATD.

  • The observed association with primary liver cancer should promote hepatological surveillance of individuals with AATD and spur longitudinal studies characterising the development of liver fibrosis and malignancy.


Alpha-1 antitrypsin deficiency (AATD) is one of the most common, potentially lethal inborn disorders, with AATD-related lung and liver disease being the major drivers of morbidity and mortality.1 Mutations in the SERPINA1 gene coding for alpha-1 antitrypsin (AAT) lead to a ‘gain of function’ proteotoxic liver injury, whereas the lack of AAT in the bloodstream facilitates the development of chronic obstructive pulmonary disease (COPD) and emphysema.1

The most common severe SERPINA1 variant is termed ‘Pi*Z’ (rs28929474).2 3 The ‘Pi*S’ variant (rs17580) is even more prevalent, but less detrimental.1 The homozygous occurrence of ‘Pi*Z’ is found in 1:2000 Caucasians2 and is termed ‘Pi*ZZ’, while heterozygous ‘Pi*Z’ carriage (termed ‘Pi*MZ’ genotype) is seen in 1:30 individuals of Northern European descent. The strong predisposition of ‘Pi*ZZ’ individuals for lung disease is supported by a large body of evidence and reflected in clinical management guidelines.1 The susceptibility for liver disease is less well documented.3 4 ‘Pi*ZZ’-related liver disease displays a biphasic pattern, with the first peak in early childhood as neonatal cholestasis and the second peak after 50 years of age.1 5 Two cross-sectional studies indicate that advanced liver fibrosis is 10 to 20 times more common in ‘Pi*ZZ’ subjects compared with individuals without a ‘Pi*Z’ mutation (non-carriers) and revealed that the non-invasive liver stiffness measurement (LSM) via transient elastography (TE) constitute a useful surrogate of liver fibrosis.4 6 ‘Pi*MZ’ individuals seem to carry moderately elevated odds for both lung and liver disease, and to be susceptible to gallstone disease.1 7

Humans carrying both the ‘Pi*Z’ and the ‘Pi*S’ variants (termed ‘Pi*SZ’) are as frequent as 1:500 in certain geographical regions,8 while the occurrence of homozygous carriage of the ‘Pi*S’ variant (termed ‘Pi*SS’) might be even higher.9 Two studies demonstrated that ‘Pi*SZ’ individuals display a less severe lung phenotype than Pi*ZZ subjects,10 11 whereas the extent of their liver disease was not systematically studied. Children with ‘Pi*SZ’ genotype develop a clinically relevant liver disease markedly less often than Pi*ZZ individuals.12 13 Similar findings have been reported in adults,14 but multiple case reports have described ‘idiopathic’ liver cirrhosis in ‘Pi*SZ’ subjects.15 Finally, while the ‘Pi*SS’ genotype is considered to confer minimal if any risks, little clinical data are available to support this directly.16

Probably the greatest limitation when studying the AATD phenotype is the fact that the majority of AATD cases remain undiagnosed and the proportion is even higher in individuals with less severe genotypes.1 A Swedish birth cohort-based study partially addressed this issue, but this study examined the individuals only up to 45 years of age, that is, before the peak of AATD-related adult liver disease and focused on Pi*ZZ individuals.17 To provide unbiased information about the hepatobiliary phenotype of individuals with major AATD genotypes, we used the UK Biobank, a community sample from the UK totalling nearly 500 000 individuals with available ‘Pi*Z’ and ‘Pi*S’ genotyping. To corroborate our findings, we prospectively recruited the largest, multinational cohort of Pi*SZ subjects without previously known chronic liver disease and compared their lung-related and liver-related parameters to those of Pi*ZZ participants and non-carriers. The goal of our study was to provide data for evidence-based management and counselling of these individuals.


Population-based UK Biobank participants (cohort 1)

The ‘UK Biobank’ (UKB) is a population-based cohort study conducted in the UK, which recruited 502 511 volunteers aged 37–73 years at baseline. All participants underwent an initial examination, which was the basis for our study and gave informed consent for genotyping and data linkage to medical reports. Ongoing inpatient hospital records beginning in 1996 were used to identify diagnoses according to ICD10 codes (International Classification of Diseases, 10th revision). Genotyping for both the Pi*Z (rs28929474) and Pi*S (rs17580) mutations of SERPINA1 was available in 487 503 subjects. Follow-up measurement of liver enzymes was conducted in 16 010 participants.

We excluded participants with viral hepatitis (ICD10: B16-B19: 713 MM, 20 MZ, 2 SZ) or risky alcohol consumption (>60 g alcohol/day for men, >40 g alcohol/day for women: 3775 MM, 153 MZ, 9 SZ, 3 ZZ, 4 SS). We compared SERPINA1 variants to well-known genes that modulate the risk of liver disease, that is, PNPLA3 p.I148M (rs738409), HSD17B13:T (rs72613567) and TM6SF2 p.E167K (rs5854926); homozygous carriers were compared with non-carriers.

The presence of the following primary ICD10 codes was evaluated: fibrosis and cirrhosis (K74.0-2+K74.6), primary liver cancer (C22.0), non-alcoholic fatty liver disease (NAFLD, K76.0), non-alcoholic steatohepatitis (NASH, K75.81), cholelithiasis (K80), emphysema (J43.1+J43.2+J43.8+J43.9) and chronic bronchitis (J44). The presence of metabolic syndrome was based on the IDF (International Diabetes Federation) definition, which consists of central obesity (defined as waist circumference with gender and ethnicity-specific values) plus any two of the following four factors: (1) raised serum triglycerides ≥150 mg/dL (1.7 mmol/L) or specific treatment for this lipid abnormality; (2) reduced serum HDL cholesterol <40 mg/dL (1.03 mmol/L) in male or <50 mg/dL (1.29 mmol/L) in female or specific treatment for this lipid abnormality; (3) raised systolic blood pressure (BP) ≥130 mm Hg or diastolic BP ≥85 mm Hg or treatment of previously diagnosed hypertension; (4) raised fasting plasma glucose ≥100 mg/dL (5.6 mmol/L), or previously diagnosed type 2 diabetes.

Real-life cohort without previously known liver disease (cohort 2)

Study population

One thousand one hundred and four individuals were recruited as part of our global Alpha-1 Liver initiative, a multicentre registry effort for AATD-related liver disease carrying out baseline assessment, which are shown herein, as well as a prospective follow-up. The majority of non-carriers and Pi*ZZ participants was previously published.6 7 The inclusion criteria were (1) age ≥18 years, (2) the ability to provide written informed consent and (3) no known pregnancy. To prevent an enrichment with individuals with liver involvement, the following exclusion criteria were used: (1) the presence of known liver disease or a liver comorbidity identified in clinical and laboratory workup, (2) at least two independent visits with elevated liver enzymes in medical records prior to baseline.

All participants underwent clinical and laboratory workup including standardised questionnaires with demographic parameters, previously known chronic diseases and relevant comorbidities, as well as the evaluation of alcohol and cigarette consumption. The need for augmentation therapy and long-term oxygen therapy and the COPD assessment test (CAT) values constituted determinants of lung phenotype. AAT serum level was measured and genotyping was conducted by the responsible national AAT reference laboratories, which performed PCR analysis and/or isoelectric focusing as described.6

The Pi*SZ population consisted of 239 subjects from 10 European countries (Germany, UK, Portugal, Spain, Italy, Austria, Belgium, Ireland, Denmark, Poland) and the USA (online supplemental table 1). Among the Pi*ZZ participants, 413 (70.5%) were from Germany. Non-carriers (n=279) were defined as individuals with normal serum AAT levels (ie, >110 mg/dL), in whom the presence of the ‘Pi*Z’ and ‘Pi*S’ variants was excluded as described.6 Of them, 219 (78.5%) were recruited in Germany, and 60 (21.5%) in Austria.

Supplemental material

Statistical analysis

All continuous variables were analysed by unpaired, two-tailed t-tests or Mann-Whitney U test, and by a multivariable model to account for confounders (age, sex, body mass index (BMI), diabetes mellitus and alcohol consumption) and shown as mean (SD) (normal distribution) or median (IQR) (non-normal distribution). Categorical variables were displayed as relative (%) frequencies and analysed using the χ2 test. ORs were presented with their corresponding 95% CI. Multivariable logistic regression tested for independent associations. Correlations were assessed by Spearman correlation coefficients, where appropriate. Differences were indicated as statistically significant when p<0.05. The data were analysed using SPSS Statistics V.26 (IBM) and Prism V.8 (GraphPad, LaJolla, California, USA).


Lung-related and liver-related parameters of AATD subjects in UK Biobank (cohort 1)

The 482 380 eligible UK Biobank participants comprised 138 Pi*ZZ (frequency 1:3496), 864 Pi*SZ (1:558), 1014 Pi*SS (1:476) and 17 006 Pi*MZ individuals (1:28; figure 1A). All subgroups displayed a similar age and sex distribution as well as a comparable—mostly low or moderate—alcohol consumption. While diabetes mellitus was infrequent, it was less common in Pi*ZZ subjects compared with non-carriers (5% vs 2%, p<0.0001; table 1). As expected, Pi*ZZ individuals showed a significantly lower FEV1/FVC ratio compared with all other genotypes despite the lowest cigarette consumption (table 1). The percentage of individuals with FEV1/FVC<70% was the highest among Pi*ZZ participants but was also significantly higher in Pi*SZ individuals compared with non-carriers. Regarding liver-related blood parameters, mean alanine aminotransferase (ALT) values were significantly higher in all analysed AATD genotypes compared with non-carriers (table 1, figure 2A). Pi*MZ and Pi*SZ subjects presented with higher AST values than non-carriers; however, Pi*ZZ individuals had significantly higher AST values than any other assessed AATD subgroup (table 1, figure 2B). Gamma-glutamyl transferase (GGT) values were comparable in non-carriers, Pi*MZ, Pi*SS and Pi*SZ individuals, while Pi*ZZ subjects significantly more often displayed elevated GGT levels (table 1). Alkaline phosphatase (ALP) was significantly elevated in Pi*MZ and Pi*SZ participants when compared with non-carriers, however, comparable to subjects without AATD mutation in Pi*ZZ and Pi*SS individuals (table 1, figure 2C). The OR of having elevated AST was the highest in Pi*ZZ individuals (adjusted OR=4.5 (2.8–7.3), p<0.0001; figure 3) and surpassed the odds seen in established genetic liver disease modifiers such as homozygous PNPLA3 or TM6SF2 mutation (figure 3; online supplemental tables 2-4). Pi*ZZ participants also had a moderately increased risk for elevated ALT values (adjusted OR=2.1 (1.2–3.6), p<0.0001; figure 3) with odds comparable to the ones seen in subjects with a homozygous PNPLA3 or TM6SF2 mutation (figure 3; online supplemental tables 2 and 3). Individuals with Pi*MZ, Pi*SS and Pi*SZ genotype had all significantly increased risk of presenting with elevated ALT values (OR=1.2–1.5; figure 3). To determine, whether AATD predisposes to elevated liver enzymes even in metabolically inconspicuous individuals, we reperformed the liver enzyme analysis after exclusion of individuals with the diagnosis NAFLD (online supplemental figure 1) and after exclusion of individuals with metabolic syndrome (online supplemental figure 2). Both analyses yielded largely identical results, thereby establishing the effect of AATD mutations even in metabolically inconspicuous individuals. In line with this, an additional adjustment for the PNPLA3 allele as the second strongest—and due to the high frequency of the risk allele—most relevant genetic risk factor for metabolic liver disease did not affect the results (data not shown).

Figure 1

Overview of analysed cohorts. (A) Cohort 1: Population-based study analysing UK Biobank participants aged 37–73 years at baseline. (B) Cohort 2: Prospectively recruited individuals available in a multinational, cross-sectional Alpha-1 Liver initiative. AAT, alpha-1 antitrypsin.

Table 1

Comparison of lung and liver phenotype in individuals with Pi*SS and Pi*SZ genotype compared with Pi*ZZ, Pi*MZ and non-carriers (cohort 1)

Figure 2

Liver-related parameters in individuals heterozygous for the Pi*Z variant (Pi*MZ), homozygous for the Pi*S variant (Pi*SS), heterozygous for both Pi*S and Pi*Z (Pi*SZ), and homozygous for the Pi*Z variant (Pi*ZZ) compared with non-carriers (cohort 1). A total of 422 506 non-carriers, 17 006 Pi*MZ subjects, 1014 Pi*SS individuals, 864 Pi*SZ subjects and 138 Pi*ZZ individuals underwent laboratory analysis. P values were adjusted for age, sex, body mass index, alcohol consumption and presence of diabetes mellitus. Scatter plots of serum level of alanine aminotransferase (ALT; A), aspartate aminotransferase (AST; B) and alkaline phosphatase (ALP; C), all normalised to the sex-specific upper limit of normal (ULN).

Figure 3

Risk of Pi*SS and Pi*SZ subjects to show elevated AST or ALT compared with heterozygous (Pi*MZ) and homozygous (Pi*ZZ) Pi*Z carriers as well as homozygous carriers of PNPLA3 p.I148M (rs738409), HSD17B13:T (rs72613567) and TM6SF2 p.E167K (rs5854926) (cohort 1). Adjusted OR (aOR) with their corresponding 95% CI are shown for aspartate aminotransferase (AST; A) and alanine aminotransferase (ALT; B). The risk to display levels higher than the corresponding sex-dependent upper limit of normal (ULN) was compared with the respective non-carriers. ORs were adjusted for age, sex, body mass index, alcohol consumption and diabetes mellitus.

Next, we assessed whether the analysed liver enzymes remain stable or fluctuate over time. Here, we took advantage of follow-up measurements that were available in a subset of UK Biobank individuals and correlated them with baseline values. Serum levels of GGT and ALP showed a strong correlation (r=0.7–0.85; online supplemental table 5), whereas the correlation between baseline and follow-up transaminases/bilirubin levels was somewhat weaker (online supplemental table 5), which is in line with previous reports.18

In the least studied genotypes Pi*SS and Pi*SZ, male sex, age ≥50 years and smoking were associated with higher rates of decreased %FEV1/FVC (figure 4). With regard to transaminases, the presence of BMI≥30 kg/m2 or diabetes mellitus conferred an increased chance of displaying elevated values. Age≥50 years was associated with increased AST, but not ALT values (figure 4).

Figure 4

Rate of Pi*SS and Pi*SZ subjects with decreased Tiffeneau index, elevated AST or elevated ALT in different subpopulations (cohort 1). Relative frequencies (%) are shown and visualised by a colour coding (right panel). Decreased Tiffeneau index is defined as FEV1/VC<70%. Smokers are defined as ‘ever-smokers’ and non-smokers are defined as ‘never-smokers’. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index (kg/m²); DM, diabetes mellitus; FEV1, forced expiratory volume in 1 s; VC, vital capacity.

Lung-related and liver-related diagnoses of AATD subjects in UK Biobank (cohort 1)

With regard to lung-related ICD codes, the diagnosis of COPD and emphysema was >20 times enriched in Pi*ZZ participants compared with non-carriers. It was not more common in Pi*SS and Pi*SZ subjects, while the much more frequent Pi*MZ individuals displayed a moderately increased OR for emphysema (adjusted OR=1.6 (1.3–1.9), p<0.0001; online supplemental table 1; figure 5). Consistent with a previous publication,19 gallstone disease was enriched in Pi*MZ carriers versus non-carriers (adjusted OR=1.3 (1.2–1.4), p<0.0001), but in none of the other AATD genotypes. The diagnosis liver fibrosis/cirrhosis was 20 times more common in Pi*ZZ individuals compared with non-carriers (adjusted OR=21.7 (8.8–53.7), p<0.0001), but also markedly enriched in Pi*SZ subjects (adjusted OR=3.1 (1.1–8.2), p=0.027) and moderately in Pi*MZ participants (1.7 (1.2–2.2); p=0.001; online supplemental table 6; figure 5). Both Pi*SZ and Pi*ZZ individuals harboured a numerically higher risk of liver fibrosis/cirrhosis than any of the previously described genetic liver disease modifiers (figure 6A; online supplemental tables 2-4, 6). Similarly, both Pi*SZ and Pi*ZZ subjects, but none of the other AATD genotypes, possessed a markedly increased risk for the diagnosis of primary liver cancer. Again, the risk of Pi*ZZ individuals for primary liver cancer surpassed the odds seen in individuals with other genetic modifiers, while Pi*SZ was comparable to known risk factors TM6SF2 and PNPLA3 (figure 6, online supplemental tables 2-4, 6).

Figure 5

Odds ratios (ORs) of ICD10 diagnoses in individuals heterozygous or homozygous for the Pi*Z variant (Pi*MZ/Pi*ZZ), homozygous for Pi*S variant (Pi*SS) and heterozygous for Pi*S and Pi*Z (Pi*SZ) (cohort 1). Adjusted OR (aOR) with their corresponding 95% CI are shown for Pi*MZ, Pi*SS, Pi*SZ and Pi*ZZ subjects compared with non-carriers. ORs were adjusted for age, sex, body mass index, alcohol consumption and diabetes mellitus. If in one group no cases are available, the corresponding aOR was set as 1 [1;1].

Figure 6

Odds ratios (ORs) of ICD10 diagnoses in individuals heterozygous (or homozygous) for Pi*Z (Pi*MZ/Pi*ZZ), heterozygous for Pi*S and Pi*Z (Pi*SZ) or homozygous for displayed genetic liver fibrosis modifiers. Genetic liver fibrosis modifiers include PNPLA3 p.I148M (rs738409), HSD17B13:T (rs72613567) and TM6SF2 p.E167K (rs5854926)). Adjusted OR (aOR) with their 95% CI compared with the corresponding non-carriers. ORs were adjusted for age, sex, body mass index, alcohol consumption and diabetes mellitus.

A sensitivity analysis revealed that in the Pi*SZ and Pi*MZ populations, the ‘fibrosis/cirrhosis’ phenotype is markedly enriched in males, obese individuals and subjects ≥50 years old (online supplemental figure 3).

Lung and liver phenotype in a multinational AATD cohort (cohort 2)

Our multinational cohort consisted of 586 Pi*ZZ subjects, 239 Pi*SZ individuals and 279 non-carriers, all without previously known or coexisting liver disease (table 2; figure 1B). Pi*SZ subjects were under-represented when compared with the community-based UK Biobank cohort. All three subgroups showed similar rates of diabetes mellitus and alcohol consumption, while differences in other demographic factors were seen (table 2). When only participants without augmentation therapy were considered, Pi*SZ individuals showed intermediate AAT serum levels (63.8±19.8 mg/dL vs 139.5±25.1 mg/dL in non-carriers vs 28.3±16.0 mg/dL in Pi*ZZ, all p<0.0001, table 2; online supplemental figure 4A). The cut-off AAT level of 99.5 mg/dL differentiated well between Pi*SZ individuals and non-carriers (sensitivity 97.9%, specificity 98.4%, table 2; online supplemental figure 4A). Pi*SZ individuals had an intermediate lung phenotype as reflected by their CAT scores and need for long-term oxygen treatment, that is, the levels/frequencies were higher than in non-carriers, but significantly lower/less frequent compared with Pi*ZZ individuals (table 2). With regard to liver enzymes, Pi*SZ individuals had lower AST and ALT than Pi*ZZ subjects, while GGT was higher in Pi*SZ subjects than non-carriers. Mean ALP levels were the highest in Pi*SZ individuals (online supplemental figure 4C-F, online supplemental table 7), whereas GLDH and bilirubin levels did not show obvious differences among the subgroups (online supplemental table 7).

Table 2

Characteristics of Pi*SZ individuals in comparison to Pi*ZZ subjects and non-carriers in a multicentre registry cohort (cohort 2)

In TE, Pi*SZ individuals had intermediate LSM values, that is, LSMs were higher than in non-carriers (5.2±2.5 kPa vs 4.6±1.6 kPa, p=0.002), but lower than in Pi*ZZ subjects (5.2±2.5 kPa vs 6.6±5.2 kPa; p=0.022, table 2; online supplemental figure 4B) and similar results were seen when only non-obese individuals were assessed (online supplemental table 8). In the entire cohort, 13% of Pi*SZ individuals showed LSM values ≥7.1 kPa, suggesting liver fibrosis stage of at least 2 on a 0–4 scale20 compared with 5% of non-carriers (adjusted OR=2.6 (1.1–6.1), p=0.024; table 2) and 24% of Pi*ZZ subjects (adjusted OR=0.5 (0.2–0.8), p=0.013; table 2). Pi*SZ individuals with LSM ≥7.1 kPa had significantly higher BMI values and were more frequently diabetic (online supplemental table 9). Vice versa, individuals with diabetes more frequently displayed elevated AST and ALT values than subjects without diabetes (online supplemental figure 5).

The simultaneously assessed CAP as a surrogate of hepatic steatosis did not show major differences between Pi*SZ and non-carriers, nor between Pi*SZ individuals and Pi*ZZ subjects (table 2). However, increased liver enzyme levels were seen primarily in Pi*SZ individuals with liver steatosis (as revealed by an analysis of individuals with CAP ≥248 dB/m) who displayed higher AST, GGT and ALP levels than ‘steatotic’ non-carriers (online supplemental figure 6).


We analysed the hepatobiliary phenotype of individuals with the most common AATD genotypes using the UK Biobank as a unique, openly available resource with deep genetic, physical and health data.21 It does not constitute an entirely representative population sample since 94% of subjects are classed as white British and 6% within ethnic minority groups, compared with 80.5% and 19.5%, respectively, in UK census data, and it is skewed towards higher income classes.21 Nevertheless, it represents the best available approximation of such a cohort in that it recruited and systematically genotyped participants independently of their known SERPINA1 genotype. This approach is crucial since the vast majority of AATD individuals remain undetected and were therefore not considered in previous studies. The AAT genotyping used in our study was extracted from the UK Biobank Axiom array and the results remained unknown to the study subjects. The frequencies of analysed genotypes agreed well with their published occurrence in Caucasian population8—an observation that further validates our approach.

An important limitation of our study is the difficulty to reliably identify all individuals with NAFLD and NASH since these disorders were not systematically assessed in the UK Biobank baseline visits and may remain underdiagnosed in the clinical routine. To offset this limitation, we repeated the analyses after excluding individuals with the ICD code for NAFLD as well as with presence of metabolic syndrome and demonstrated that the differences persisted in this subgroup. Moreover, liver transaminase levels significantly fluctuated over time and therefore a single measurement is not sufficient to comprehensively evaluate the liver phenotype of AATD individuals. Notably, the limited usefulness of single ALT measurements for evaluation of AATD individuals was reported previously.22 However, our manuscript aimed to provide ‘typical liver enzyme levels’ seen in subjects with different AATD genotypes.

In the UK Biobank cohort, Pi*ZZ participants suffered a >20 times higher risk of liver fibrosis and cirrhosis as well as ~45 times higher risk of primary liver cancer. The former finding is in line with previous reports demonstrating that signs of advanced fibrosis are ninefold to 20-fold more common in Pi*ZZ individuals compared with people without AAT mutations, as well as the observation that Pi*ZZ individuals are 20 times more likely to require liver transplantation than the general population.6 23 The odds of Pi*ZZ subjects to develop advanced liver fibrosis/cirrhosis are substantially higher than the ones reported for other established genetic conditions such as mutations in PNPLA3, TM6SF2 or HSD17B13 gene.23 24 While the predisposition to liver fibrosis is now supported by a solid body of evidence, reports on liver cancer in Pi*ZZ individuals are very limited1 and further analyses are needed. Pi*ZZ participants had the highest AST/ALT values, but their ALP levels were similar to the ones seen in non-carriers and they did not present with an increased risk of cholelithiasis. Since gallstones consist mainly of lipids such as cholesterol, the alterations in lipid metabolism that were observed in Pi*ZZ individuals (ie, lower serum levels of triglycerides, very low-density lipoproteins and low-density lipoproteins compared with controls) indicate that an impaired hepatic secretion of lipids might play a role.6 Collectively, our data revealing a marked susceptibility of Pi*ZZ individuals to end-stage liver disease should prompt their thorough hepatological monitoring.

The availability of genetic information allowed us to systematically study AATD genotypes that are not assessed in clinical routine such as Pi*MZ and Pi*SS. With regard to Pi*MZ, our data confirmed previous findings of a mild increase in transaminases as well as a moderately increased risk of liver fibrosis/cirrhosis and cholelithiasis.7 19 24 25 On the other hand, the increased occurrence of emphysema was seen in some, but not all population-based studies.1 With regard to the Pi*SS genotype, our data are novel and support the current opinion that these individuals display no or only minimal predisposition to both lung and liver disease. It provides an important guidance for physicians and a relief for the carriers of this genotype.

A focus of our work was on the Pi*SZ phenotype that is under-represented in clinical routine compared with Pi*ZZ subjects. This might be due to their less conspicuous AAT serum levels as well as their less pronounced disease phenotype.10 11 26 Consistent with published data, the Pi*SZ individuals available in the UK Biobank displayed no or only minimal lung phenotype,10 11 while our multicentre cohort was skewed towards more lung-diseased individuals, likely due to the fact that it often prompted the diagnosis of AATD. Although Pi*SZ individuals display normal or only minimally elevated transaminases, both analysed cohorts revealed a marked predisposition to liver fibrosis. The UK Biobank cohort also suggested an increased susceptibility to primary liver cancer that was not assessed in the second cohort. The more pronounced association with liver fibrosis compared with lung emphysema might be attributable to the fact that the liver phenotype constitutes a ‘gain-of-function’ toxicity, while lung injury seems to arise due to a loss-of-function situation. Accordingly, the intermediate AAT serum levels seen in Pi*SZ individuals might be sufficient to protect the lung from proteolytic damage, while misfolding and polymerisation of AAT may generate biologically relevant proteotoxic stress in the liver. The identified hetero-polymerisation between Pi*S and Pi*Z27 might be responsible for the greater liver fibrosis burden than that of the Pi*MZ state despite the absence of any Pi*SS signal (suggesting no clinically significant challenge with Pi*S misfolding alone). While Pi*SZ subjects display clear predisposition to liver fibrosis and primary liver cancer, their susceptibility is markedly lower than the one seen in Pi*ZZ individuals, which is consistent with the observed lower levels of intracellular polymers and a less pronounced lung phenotype.10 11

In addition to the characterisation of the hepatobiliary phenotype of individuals with AATD, we demonstrated that male sex, obesity, diabetes and higher age are associated with increased risk of liver fibrosis/cirrhosis as well as primary liver cancer. Notably, the same factors were previously implicated in liver fibrosis development in Pi*MZ and Pi*ZZ individuals.4 6 7 25 28 Among them, obesity and diabetes are potentially modifiable and their effects as drivers of NAFLD extend beyond AATD.29 30 They are associated with increased oxidative stress and lipolysis and may aggravate the endoplasmic reticulum stress occurring in AATD.31 32 Male sex is another parameter linked to AATD since the production of AAT is stimulated by testosterone and males therefore produce higher amounts of the potentially toxic protein.33

In conclusion, our data characterise the hepatobiliary phenotype of adults with major AATD genotypes with a focus on Pi*SZ and should help in patient management and counselling. While Pi*ZZ individuals need a closer monitoring, the surveillance of Pi*MZ and Pi*SZ subjects needs to be adjusted to the overall clinical context that includes the presence of hepatic comorbidities/metabolic risk factors, other genetic factors as well as the presence/absence of baseline liver fibrosis as evaluated by non-invasive methods. The association with primary liver cancer should spur hepatological surveillance of both Pi*ZZ and Pi*SZ individuals. However, further studies are warranted to determine whether screening is needed for all Pi*SZ/Pi*ZZ individuals or only those with advanced liver fibrosis/cirrhosis. Longitudinal assessment is needed to define the rate of disease development and tumour occurrence in the individuals with different AATD genotypes.

Data availability statement

Data are available upon reasonable request. The deidentified participant data are available from the corresponding author upon request.

Ethics statements

Patient consent for publication

Ethics approval

The study has been approved by the UKB Access Committee (Project Number 47527).


We thank the national patient organisations (ie, Germany: Marion Wilkens and Gabi Niethammer; Austria: Ella Geiblinger; Switzerland: Gottfried Grünig; Belgium: Frank Willersinn; Denmark: Gunhil Norhave as well as Spain, Italy, Poland and Portugal for their help with the execution of our study. We also thank all patients for their participation in our study. This research has been conducted using the UK Biobank Resource.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • MF and CVS contributed equally.

  • Correction notice This article has been corrected since it published Online First. The authors list has been updated.

  • Contributors Study concept and design: MF, CVS and PS. Acquisition of data: MF, CVS, VP, KH, MP, MCR, FB, PE, KT, MM, BB, VW, JC-W, JV, FN, JG, MM, AN, BS, HZ, SJ, NA, LJ, RG, CG, KMS, MT, AK, BG, DT, AM, JRH, DAL, FL, NTG, VC, WG, CT, AMT, NGM and PS. Analysis and interpretation of data: MF, CVS and PS. Drafting of the manuscript: MF, CVS and PS. Critical revision of the manuscript for important intellectual content: all authors. Figures and tables: MF, CVS and PS. Statistical analysis: MF and CVS. Obtained funding: CVS, KH, PE, MM, DAL, CT, AMT and PS. Study supervision: MF, CVS and PS. All authors had full access to all of the data and approved the final version of this manuscript. All authors take responsibility for the integrity of the data and the accuracy of the data analysis.

  • Funding This work was supported by the EASL registry grant on alpha-1 antitrypsin-related liver disease, the Deutsche Forschungsgemeinschaft (DFG) consortium SFB/TRR57 “Liver fibrosis” (both to PS and CT), the DFG grant STR1095/6-1 (to PS), unrestricted research grants from CSL Behring and Arrowhead Pharmaceuticals (to PS), the Peter-Scriba-MD-Scholarship (to CVS), the START program within the medical faculty at RWTH Aachen University (to KH), the German Liver Foundation (to KH), the Joseph-Skoda Award of the Austrian Society of Internal Medicine (to MM). DAL is supported by the Medical Research Council (UK), the Alpha-1 Foundation (USA), Alpha-1 Association (UK) and the NIHR UCLH Biomedical Research Centre. He is an NIHR Senior Investigator. BG is supported by the Alpha-1 Foundation (USA), Medical Research Council (UK) and the British Lung Foundation. AMT has current research grants from the Alpha 1 Foundation, ATS Foundation, NIHR, CSL Behring, AstraZeneca, Chiesi and Health Foundation. PE is supported by the Alpha 1 Foundation and ATS Foundation.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.