Beyond Genotype in Alpha-1 Antitrypsin Deficiency

Evidence

Alpha-1 antitrypsin deficiency, or AATD, is inherited, but its pulmonary presentation is not determined by genotype alone. Two independent studies approach that heterogeneity from different directions. A fresh European registry analysis groups people by their broader clinical profiles. A separate German imaging study measures where emphysema appears within the lungs of patients spanning reduced-to-normal, moderate, and severe alpha-1 antitrypsin levels. The first identifies patient-level clusters; the second shows what finer anatomical measurement can add when familiar global measures look similar.

The registry study used baseline data from the European Alpha-1 Research Collaboration, or EARCO. Investigators applied K-prototypes, a clustering method that can combine categorical and numerical variables, and identified six clinical phenotype clusters. The reported cross-validated weighted F1 score was 0.926, a measure of how consistently the analysis assigned the predefined cluster labels during validation. A separate random-forest classifier ranked age at diagnosis, lung function, and tobacco consumption as the most important distinguishing features beyond the AATD-associated genotype. Because the ingested source is an abstract, the cluster sizes, variable definitions, missing-data procedures, and stability across countries cannot be assessed here.

The quantitative-CT study retrospectively analyzed 75 patients seen at one specialist center from March 2012 through February 2024. Researchers grouped 13 patients with reduced-to-normal alpha-1 antitrypsin levels, 16 with moderate deficiency, and 46 with severe deficiency. The groups contained multiple genotypes, including rare variants. Smoking exposure was uneven: the moderate-deficiency group averaged 39.4 pack-years, versus 14.9 and 15.1 in the other groups. Mean predicted FEV1, a standard airflow measure, was 54.1% across the cohort and did not differ significantly among the three groups.

Automated CT analysis nevertheless detected structural differences. Mean lung density and the 15th-percentile density decreased as alpha-1 antitrypsin levels fell, with group-comparison p values of 0.007 and 0.043. The whole-lung emphysema index rose from 14.6% in the reduced-to-normal group to 25.9% in the severe group, but that global comparison was not statistically significant. Regional analysis was more discriminating. The reduced-to-normal and moderate groups tended toward upper-lobe-predominant disease, whereas the severe group had its highest mean emphysema index in the middle lobe at 35.5%, followed by the lingula at 30.5% and the lower lobes.

Across all 75 patients, 38 had homogeneous emphysema, 22 had middle-lobe- or lingula-predominant disease, eight had upper-lobe-predominant disease, and seven had lower-lobe-predominant disease. Twenty-one of the 46 patients with severe deficiency fell into the middle-lobe and lingula category. In a stepwise logistic-regression analysis, older age and lower serum alpha-1 antitrypsin levels were associated with that regional pattern after gender and body-mass index were retained for adjustment. The paper reports odds ratios of 1.07 per year of age and 0.95 per milligram per deciliter of serum alpha-1 antitrypsin, with 95% confidence intervals of 1.01–1.12 and 0.92–0.98.

Analysis — Phenotypes Depend on Measurement Scale

The cross-study inference is that AATD heterogeneity becomes clearer when researchers move beyond a single genetic label or a single whole-lung summary. This is analysis, not a conclusion tested jointly by the two studies. The EARCO analysis finds separable clinical profiles using genotype together with age at diagnosis, tobacco exposure, and lung function. The CT cohort then shows why measurement scale matters: groups with no significant difference in predicted FEV1 still differed in lung-density measures and regional emphysema distribution. That convergence supports a research model in which inherited deficiency sets part of the biological context while accumulated exposure, age, serum level, and anatomical pattern shape the observed phenotype. It does not prove that the six registry clusters correspond to the imaging categories, because the studies did not analyze the same participants or variables. A useful next test would prospectively apply locked EARCO cluster definitions and standardized lobe-level CT measurements in an external cohort, then ask whether either layer predicts progression beyond genotype and smoking history. Until that happens, the clusters and CT patterns are promising descriptions, not validated disease trajectories or treatment-selection tools.

Limitations

The fresh registry study was available only at abstract depth. Its participant count, country distribution, cluster composition, input-variable handling, and sensitivity analyses are therefore not available in the ingested evidence. A high cross-validated F1 score describes classification performance within that analysis; it does not show that the clusters will reproduce in a new registry or forecast future outcomes. Baseline clustering also cannot establish which features cause a phenotype.

The CT study was retrospective, single-center, and small, particularly after division into three serum-level groups and rare genotypes. Patients received CT scans for different clinical indications, creating possible referral and selection bias. Smoking exposure differed sharply by group, three participants lacked genotyping, and augmentation therapy was much more common in severe deficiency. The regional regression used a limited sample and stepwise variable selection, which can produce unstable estimates. CT density and emphysema distribution are structural measurements, not direct measures of symptoms, progression, or clinical benefit.

Finally, the studies did not use the same cohort. One clusters broad registry variables; the other groups patients by serum level and quantifies lung anatomy. Their agreement is conceptual rather than a replication of specific phenotypes. Longitudinal, multicenter validation would be needed to determine whether the reported groupings remain stable, predict change, or add information beyond established clinical measures.