Pulmonary and Hepatic Outcomes in Intermediate Alpha-1 Antitrypsin Deficiency: A TriNetX Database Analysis

1. Introduction

Alpha-1 antitrypsin deficiency (A1ATD) is a hereditary disorder, inherited in an autosomal codominant manner, and affects roughly 1 in 2,000 to 5,000 people of European ancestry [1]. The problem stems from mutations in the SERPINA1 gene, which encodes alpha-1 antitrypsin (A1AT), a serine protease inhibitor mainly produced in the liver. Most cases of deficiency involve the S and Z alleles, but the Z allele (Glu342Lys) is the primary cause, as it induces the most pronounced protein misfolding and leads to A1AT accumulation within cells [2].

A1ATD primarily causes dysfunction in two organs: the lungs and the liver. In the lungs, insufficient functional A1AT allows neutrophil elastase to act unchecked, degrading lung tissue and leading to progressive emphysema [2]. In the liver, the abnormal Z-A1AT protein misfolds and aggregates in the endoplasmic reticulum of hepatocytes. Over time, this stress leads to fibrosis, cirrhosis, and, in some individuals, hepatocellular carcinoma [3].

Serum A1AT concentrations usually correlate with a person's genotype, though there is considerable overlap, and levels can rise during acute-phase responses [4]. Individuals with the normal piMM genotype have serum concentrations between 150 and 350 mg/dL. Heterozygous piMZ and compound heterozygous piSZ individuals fall into an intermediate range. Homozygous ZZ individuals, however, have only 20-45 mg/dL, far below the protective threshold, which is approximately 57 mg/dL.

Clinically, A1ATD presents in various forms. In the lungs, it frequently leads to early-onset emphysema, often affecting the lower lobes. The NHLBI Registry has shown that people with severe deficiency have reduced survival [5]. In the liver, manifestations range from neonatal cholestasis to chronic hepatitis, cirrhosis, and hepatocellular carcinoma. Around 10% of adult PiZZ individuals develop cirrhosis, and the risk increases with age [6].

Although the PiZZ genotype is well characterized, much less is known about people whose serum A1AT levels are in the 40-80 mg/dL range, mostly those with SZ, MZ, and SS genotypes. This represents a significant gap in our knowledge, since these genotypes are more common than severe deficiency and may still contribute to clinically significant morbidity [7].

Current guidelines recommend targeted testing for individuals with early emphysema, unexplained liver disease, or a family history of A1ATD. The treatment threshold is currently set at 57 mg/dL; it is unclear if treatment thresholds should be raised due to involvement of liver and lung disease. In this study, we utilize TriNetX, a large multicenter electronic medical record network, to compare liver and lung outcomes, as well as mortality, among A1ATD patients grouped by serum A1AT levels.

2. Materials and Methods

We used the US Collaborative Network from 43 HCO(s) in the TrinetX research network. The initial search yielded 1410 patients amongst 2 cohorts, which were then adjusted to 324 patients in each cohort. The HCOs were hospitals, primary care units, or specialists, providing data on uninsured or insured patients. The TriNetX database is a global health collaborative clinical research platform that collects real-time electronic medical data from a network of HCOs. Because the data were anonymous, informed consent was waived.

The available data included information on demographics, diagnoses (based on the International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] codes), and laboratory tests (coded using Logical Observation Identifiers Names and Codes [LOINC]).

Inclusion(Table 1)-

Inclusion Criteria

Patients were included in the study if they met all of the following criteria:

• Age ≥ 18 years at the time of the index event.
• Documented diagnosis of alpha-1 antitrypsin deficiency (AATD) identified using ICD-10-CM code E88.01.
• At least one recorded serum alpha-1 antitrypsin (A1AT) level within the defined study ranges:
• Cohort 1: A1AT level between 40–60 mg/dL; Cohort 2: A1AT level between 60–80 mg/dL
• Availability of follow-up data after the index event within the TriNetX network.

Exclusion criterion

Severe AATD (e.g., serum A1AT < 40 mg/dL), to specifically focus on mild and intermediate deficiency states.

Prior diagnosis of cirrhosis or hepatic decompensation before the index date, to ensure assessment of incident outcomes.

Known chronic liver diseases unrelated to AATD, including:

∙

Chronic viral hepatitis (HBV or HCV)

∙

Alcohol-associated liver disease

∙

Autoimmune hepatitis

∙

Hemochromatosis

∙

Wilson disease

∙

Nicotine dependence

These criteria were designed to isolate the impact of other primary liver diseases and effect of smoking.

Outcomes-

These are defined as diagnoses, medications, procedures, or laboratory values that occurred in the time window starting after the first occurrence of the index event. It included

• Emphysema- ICD10CM:J43
• COPD exacerbation- ICD10CM: J44
• Cirrhosis- ICD10CM:K74.69 and ICD10CM:K74.60
• Hepatic decompensation- Ascites, Hepatic encephalopathy, and variceal bleeding- ICD10CM: R18, ICD10CM: I85.01, and ICD10CM: K76.82
• Death- Outcome definition in TrinetX- Deceased

Propensity Matching-

Propensity score matching was performed to minimize confounding, including variables such as age, sex, race, BMI, diabetes status, and baseline laboratory values. We used the TriNetX built-in function and matched the two groups at a 1:1 ratio. Characteristics of the cohorts before and after matching are summarized in the Table 2 below.

Statistical analyses

The Odds ratio (OR) of incident lung and liver disease was calculated for the two groups using a built-in function in the TriNetX platform. In all analyses, a 95% confidence interval (95% CI) was used to indicate statistical significance. The Kaplan-Meier method was further used. Statistical significance was defined as a P-value < 0.05.

3. Results

Patient Characteristics

Following 1:1 propensity score matching, 324 patients were included in each cohort. Baseline characteristics were well-balanced between groups, including age (53.3 ± 15.8 vs 52.7 ± 15.8 years, p = 0.654), sex distribution (49.7% vs 48.8% female, p = 0.814), and prevalence of type 2 diabetes mellitus (13.9% vs 12.0%, p = 0.483). Bronchodilator use was similar between cohorts (sympathomimetic: 42.3% vs 41.0%; anticholinergic: 21.3% vs 20.1%). Residual imbalances were noted in albumin (4.0 ± 0.8 vs 4.2 ± 0.6 g/dL, p = 0.001) and INR (1.4 ± 0.6 vs 1.2 ± 0.4, p = 0.001), suggesting more advanced liver dysfunction in the lower A1AT cohort at baseline. Outcomes are summarized in Table 3 below.

• Emphysema- Patients with serum A1AT levels of 40-60 mg/dL demonstrated a trend toward increased emphysema risk compared to those with levels of 60-80 mg/dL (10.8% vs 7.4%; OR 1.51, 95% CI 0.88-2.61, p = 0.133). Kaplan-Meier analysis revealed a 63% higher hazard of emphysema development in the lower A1AT group (HR 1.63, 95% CI 0.97-2.74), with the log-rank test approaching statistical significance (p = 0.063).
• COPD exacerbation- COPD exacerbation rates were similar between groups (14.2% vs 13.0%; OR 1.11, 95% CI 0.71-1.74, p = 0.646). Time-to-event analysis showed no significant difference in exacerbation-free survival (HR 1.21, 95% CI 0.80-1.84, log-rank p = 0.366).
• Cirrhosis- Cirrhosis developed in 10.5% of patients with A1AT levels 40-60 mg/dL compared to 7.4% with levels 60-80 mg/dL (OR 1.47, 95% CI 0.85-2.53, p = 0.169). Kaplan-Meier analysis demonstrated a 58% higher hazard of cirrhosis in the lower A1AT cohort (HR 1.58, 95% CI 0.94-2.66), with the log-rank test approaching significance (p = 0.085).
• Hepatic decompensation- Hepatic decompensation occurred in 4.6% versus 3.7% of patients (OR 1.26, 95% CI 0.58-2.74, p = 0.555). The limited number of events (n = 27) precluded detection of statistically significant differences (HR 1.33, 95% CI 0.62-2.85, log-rank p = 0.458).
• Death- No significant difference in mortality was observed between cohorts (OR 1.07, 95% CI 0.52-2.20, p = 0.854).

These findings suggest that there are no statistically significant correlations between the exposure and any of the five outcomes (death, cirrhosis, hepatic decompensation, COPD exacerbation, or emphysema). The results are consistent with no effect because all p-values are greater than 0.05 and all 95% confidence intervals for the risk and odds ratios encompass 1.0. The p-values range from 0.133 to 0.854, all well above the conventional 0.05 threshold for statistical significance. However, Patients with A1AT 40-60 mg/dL demonstrated trends toward increased emphysema risk (10.8% vs 7.4%) and cirrhosis (10.5% vs 7.4%). No significant differences were observed for COPD exacerbation (14.2% vs 13.0%, p=0.646), hepatic decompensation (4.6% vs 3.7%, p=0.555), or mortality (OR 1.07, p=0.854). While none of the associations reached statistical significance, the point estimates suggest potential trends with lower A1AT levels associated with worse outcomes. We cannot conclude a true absence of effect, only that we observed no significant associations.

4. Discussion

Current guidelines focus on individuals with severe A1ATD, particularly those with the ZZ genotype, given their well-established risk for progressive liver and pulmonary disease. However, the clinical implications of mild-to-intermediate deficiency states, such as piMZ, piSZ, and piSS, remain less well defined. While traditionally considered carriers, growing evidence suggests that these individuals may be at increased risk of liver disease, especially in the presence of additional insults such as MASLD or alcohol use. This has led to ongoing interest regarding whether heterozygous individuals represent a truly benign group or a population with modifiable risk.

Our study uses a large, real-world dataset to evaluate outcomes in these understudied populations, thereby addressing an important gap not fully captured in prior registry-based or single-center studies. Based on the study by Ferrarotti et al. of 6,057 individuals, serum A1AT levels correlate with genotype. In their study, the MM genotype was associated with 105–164 mg/dL, whereas the MS genotype was associated with 88–137 mg/dL. In contrast, genotypes associated with intermediate deficiency were SS (73–106 mg/dL) and MZ (66–100 mg/dL), with considerable overlap between these two groups. SZ individuals had the lowest intermediate levels at 49–66 mg/dL, placing them in an intermediate-severe category. Notably, there was very little overlap in serum A1AT levels between the genotypes not associated with increased risk (MM and MS) and those with intermediate deficiency (SS, MZ), suggesting that serum A1AT concentration can help distinguish risk categories. However, genotyping remains essential for definitive classification [8].

In an institutional study by Danto et al., 80 mg/dL was used as the cutoff to identify at-risk patients. Only 1.6% of adult piMS samples fell below this cutoff, whereas in the homozygous piZZ population, nearly all patients fell in the high-risk range with levels always falling below 50 mg/dL [9]. The piSZ population (85.7%) has high-risk levels, with nearly 10% falling within a similar range to the piZZ genotype. A quarter of the MZ populations had a high-risk range, with none falling below 50 mg/dL whereas 20% of piSS genotypes had values under 80 mg/dL. The findings were loosely correlated with the protein levels estimated in our study. Based on these established ranges, our 40-60 mg/dL cohort likely comprised predominantly piSZ individuals, while our 60-80 mg/dL cohort likely captured the lower end of piMZ and piSS genotypes.

The SZ genotype has traditionally been considered less severe than piZZ, largely because of the Z variant protein's lower intracellular accumulation [10]. The Z allele (Glu342Lys) results in misfolded A1AT that is retained within hepatocytes, forming periodic acid–Schiff–positive inclusions and inducing endoplasmic reticulum stress and oxidative injury. In contrast, the S variant (Glu264Val) leads to less pronounced misfolding and does not form the same pathogenic polymers, thereby reducing hepatocellular toxicity.

Data from large population cohorts, including the UK Biobank, demonstrate that individuals with the piZZ genotype have the highest risk of liver-related complications, including transaminase elevation, cirrhosis, and hepatocellular carcinoma. Heterozygous and intermediate genotypes, such as piMZ and piSS, are generally associated with milder biochemical abnormalities, although piMZ individuals may still exhibit increased hepatobiliary risk. Notably, piSZ individuals appear to represent an intermediate risk group, with higher liver enzyme levels and an increased prevalence of fibrosis and cirrhosis compared to other non-ZZ genotypes.

In our study, lower serum A1AT levels were associated with a trend toward worse hepatic outcomes, including cirrhosis and hepatic decompensation. However, these associations did not reach statistical significance, likely reflecting limited power rather than the absence of a true effect.

A 2022 review by Franciosi et al. found no clear evidence supporting an increased risk of chronic obstructive pulmonary disease (COPD) based solely on the previously proposed protective threshold of 57 mg/dL, suggesting that genotype may be a more reliable predictor of pulmonary risk than serum A1AT concentration alone [11].

The pathogenesis of AATD-related lung disease is driven by a protease–antiprotease imbalance, in which insufficient circulating A1AT permits unopposed neutrophil elastase activity, leading to degradation of elastin and other structural components of the extracellular matrix and ultimately resulting in emphysema.

However, environmental factors, particularly cigarette smoking, play a critical role in modulating disease expression. Multiple studies have demonstrated that piMZ and piSZ heterozygotes who do not smoke have little to no increased risk of clinically significant lung disease, underscoring the importance of environmental interactions rather than A1AT levels or genotype alone.

In our study, lower A1AT levels were associated with a trend toward higher rates of emphysema and COPD exacerbations; however, these associations did not reach statistical significance. This finding is consistent with prior literature and further supports the idea that environmental exposures, particularly smoking, may exert a stronger influence on pulmonary outcomes than modest differences in serum A1AT levels within the mild to intermediate deficiency range.

The National Heart, Lung, and Blood Institute Registry identified older age, lower FEV₁% predicted, and lung transplantation status as key risk factors for mortality in patients with AATD, with respiratory failure being the leading cause of death, followed by complications of liver cirrhosis [12]. While increased mortality has been well established in individuals with the piZZ genotype, outcomes in intermediate genotypes remain less clearly defined.

Green et al. (2015) compared survival across genotypes in a cohort of 699 ZZ, 126 piSZ, and 316 piMM individuals, demonstrating that piSZ patients had significantly better survival than piZZ patients, but slightly worse survival compared to piMM individuals [13]. Similarly, Dahl et al. (2002) reported a modest increase in mortality among piMZ heterozygotes. However, direct comparisons among piSS, piSZ, and piMZ genotypes remain limited, and existing data suggest that mortality risk in these groups is relatively low compared to piZZ individuals [14].

In our study, no statistically significant difference in mortality was observed between groups stratified by serum A1AT levels. This finding may reflect the relatively low baseline mortality risk in individuals with mild to intermediate deficiency, as well as limited statistical power, rather than the absence of clinically meaningful differences.

This study has several important limitations. First, the relatively small sample size following propensity score matching may have limited statistical power to detect modest but clinically meaningful differences between groups. As a result, the absence of statistically significant associations should be interpreted with caution, as it may reflect type II error rather than a true lack of effect. Second, inherent limitations of the TriNetX platform must be acknowledged. Genotype data were not available, precluding direct correlation between specific A1ATD genotypes and clinical outcomes. Although serum A1AT levels were used as a surrogate, there is known overlap between genotypes, which may have resulted in misclassification and attenuation of observed association.

Third, the use of ICD-10 coding to define diagnoses introduces the possibility of misclassification bias. Prior studies, including those by Greulich et al., have highlighted the potential for underreporting, coding inaccuracies, and variability in diagnostic practices across institutions, which may impact the reliability of outcome ascertainment [15]. Finally, as a retrospective observational study based on electronic health record data, this analysis is subject to selection bias and cannot establish causality. Additionally, variability in follow-up duration and incomplete capture of longitudinal data may further limit the interpretation of outcomes.

5. Conclusions

A1AT deficiency, lower serum A1AT levels (40-60 mg/dL) were associated with numerically higher rates of emphysema and cirrhosis compared to levels of 60-80 mg/dL, though none of the associations reached statistical significance. COPD exacerbation, hepatic decompensation, and mortality were similar between groups.

The study was underpowered to detect modest but clinically meaningful differences. The absence of genotyping data, smoking history, and alcohol use limits the ability to draw definitive conclusions. Additionally, residual baseline differences in albumin and INR between cohorts suggest that the lower A1AT group may have had more advanced subclinical liver disease at the time of enrollment, which could not be fully accounted for through propensity matching.