Background Screening for A1AT-deficiency at our institution involves serum protein assay, then phenotyping where A1AT≤1.1g/L. RT-PCR genotyping for S/Z variants is available for clarification. We recently established a multi-disciplinary respiratory/hepatology London A1AT Deficiency Service and re-examined the cost-effectiveness of our diagnostic algorithm.

Method We studied all patients who had a serum A1AT request over the three years 1/1/2010–31/12/2012. We went on to examine all requests for phenotyping and/or genotyping over the ten years 1/1/2003–31/12/2012.

Results From 4460 requests over 2010–2012, 240(5.4%) had serum A1AT≤1.1g/l. Of these, phenotyping was not available in 33 and of the remainder the following phenotype-prevalence data were observed (n,%): MM (75,36%), MZ (89,43%), MS (29,14%), ZZ (6,3%), SZ (6,3%), SS (1,1%), GM (1,1%). In error, 28 patients had phenotyping with A1AT>1.1g/l: 27 were PiMM and 1 PiMS.

Over 2003–2012, 525 phenotyping tests were performed. In 41 the result was unclear or there was evidence of in vivo degradation. The prevalence of the respective phenotypes and mean (SD) A1AT concentrations are reported in the Table. Over the same period, 24 equivocal samples were sent for genotyping. This confirmed 13 = PiMM, 4 = PiZZ, 3 = PiMZ, 2 = PiMS and 2 = PiSZ.

The NHS costs of A1AT serum assay, phenotyping and genotyping are £2.45, £35.50 and £87.50. Using the 2010–2012 data, our mean annualised spend on A1AT serum, phenotyping and genotyping was £3,642 £3,170 and £376 respectively. The cost per ZZ/SZ case diagnosed was £1198.

The highest A1AT serum concentration recorded in the ZZ/SZ patients were 0.4 and 1.0g/l respectively suggesting that the ≤1.1g/l cut-off is appropriate. The specificity, sensitivity and positive predictive values are 26.5%, 100% and 6.5%.

Conclusion 5.4% of serum A1AT assays yielded results ≤1.1g/l. Of these, 6% were found to represent SZ/ZZ variants at risk of clinically relevant disease (0.27%, 1/372 requests). The ≤1.1g/l cut-off was 100% sensitive but only 26.5% specific for clinically relevant deficiency alleles. The ≤1.1g/l cut-off could be lowered with improved specificity and cost-saving if the desire was to detect only ZZ patients.