One Z allele is not a clean bill of health for the liver

Design

Retrospective cohort with multivariable Fine-Gray competing-risk models; genotypes ascertained by validated natural language processing

Population

22,537 US veterans with alpha-1 antitrypsin genotyping (19,665 Pi*MM, 1,656 Pi*MZ, 281 Pi*SZ, 935 Pi*ZZ), median follow-up 15.9 years

Primary outcome

Major adverse liver outcomes: decompensation, hepatocellular carcinoma, liver transplantation or liver-related death

Effect

Adjusted HR vs Pi*MM: 1.25 (1.11-1.40) Pi*MZ, 1.51 (1.17-1.94) Pi*SZ, 1.80 (1.57-2.07) Pi*ZZ; five-year risk 3.5%, 5.5%, 5.3%, 8.1%

Alpha-1 antitrypsin deficiency is usually treated as a liver problem only at Pi*ZZ, with Pi*MZ filed as carrier status. A retrospective cohort of 22,537 US veterans who had genotype testing, followed over 352,612 person-years with a median of 15.9 years, argues against that.

Major adverse liver outcomes rose stepwise with Z-allele burden. Against Pi*MM, adjusted hazard ratios were 1.25 (1.11-1.40) for Pi*MZ, 1.51 (1.17-1.94) for Pi*SZ and 1.80 (1.57-2.07) for Pi*ZZ. Five-year probability of an adverse liver outcome went 3.5%, 5.5%, 5.3%, 8.1% across the four genotypes. Pi*ZZ carried more of every component - decompensation, hepatocellular carcinoma, transplantation and liver-related death. Pi*MZ and Pi*SZ carried more decompensation, transplantation and liver-related death, but not more hepatocellular carcinoma. A sensitivity analysis limited to patients with metabolic dysfunction-associated steatotic liver disease found the same pattern, so the Z allele appears to add risk on top of metabolic liver disease rather than merely marking it.

This is retrospective, genotypes were identified by natural language processing rather than a testing protocol, and a veteran cohort is older and overwhelmingly male. The absolute numbers are modest - Pi*MZ raises five-year risk from about 3.5% to 5.5%, two extra events per hundred patients over five years, not a transformed prognosis.

What changes is what you do with an MZ result already sitting in the notes. It is a reason to keep the patient in fibrosis surveillance and to press harder on alcohol and weight, not a reason to reassure and discharge. Alpha-1 genotyping is not routine in most Indian centres, so this mainly matters for patients tested abroad or as part of a family workup.

• Do not file a Pi*MZ result as 'carrier, no action' - keep the patient in fibrosis surveillance
• Look for a second hit: alcohol, obesity and metabolic liver disease are where the Z allele does its damage
• Hepatocellular carcinoma surveillance is still driven by cirrhosis, not by MZ or SZ status alone
• Test first-degree relatives when a Pi*ZZ is found, and explain that heterozygotes are not risk-free
• Record the genotype in the problem list so the next clinician does not have to find the old report

The statistics, in plain English

An adjusted hazard ratio of 1.25 is a relative figure; the absolute translation matters more, and it is a five-year risk of 5.5% instead of 3.5% - about two extra events per hundred patients. All three intervals sit above 1.0, so the association is unlikely to be chance, but this is observational: genotype was not randomly assigned, and confounding by alcohol, viral hepatitis or metabolic disease cannot be excluded even after adjustment. The absence of a cancer signal in Pi*MZ may simply reflect too few events rather than absent risk.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.