In this issue of the 2 part series about the biomarker HbA1c, the Lab Insights team interviewed Prof Aw Tar Choon, Director of Chemical Pathology at Changi General Hospital, Singapore. He is on the editorial boards of several journals including Proceedings of Singapore Healthcare, The Malaysian Journal of Medical Science and Journal of Hormonal Disorders and Endocrine Research. Prof Aw’s research interests include endocrine disorders (thyroid disease, diabetes), immunoassays, cancer and cardiac biomarkers.

In part 2 of this Q&A, we see how HbA1c is implemented in the lab.

What is the HbA1c sampling workflow like in your lab?

We process about 4000 samples for HbA1c testing per month, with about three-quarters of that workload coming from outpatient clinics. During the day, all samples are loaded and analysed via a cobas c513 (high throughput analyser), being conscious of speed and the volume of samples. Conversely, when there is a lower number of tests being run after normal lab hours or during the weekends, we can use the cobas c502 (the lower throughput analyser).

The cobas c513 analyser gives us a quick turnaround time (mean of 17 minutes), which we can quickly relay to our colleagues in the outpatient department (e.g. endocrine or diabetic clinics).

However, for locations with lower sample volumes, a POCT device takes about 8-10 mins analytical time  and samples are processed sequentially. Hence, about 6-7 patients samples are processed per hour. There are a number of NGSP certified POCT devices in the market. Additionally, the same tube (EDTA) is used at hematology for blood cell counting.

What factors do you consider when selecting a new HbA1c assay/solution?

Speed. The reason is that the uncertainty surrounding a patient’s condition will diminish if results are available and decisions can be made more accurately, so we consider that.

Assay type. It will be a NGSP certified method. If HPLC is desired, you can only diagnose variants with certainty for S, C, D and E. If other variants are present, the chromatograms may show abnormal peaks or large A1 and H0 peaks if they co-elute in those regions. Attempts to shorten the HPLC run time to improve TAT is counterproductive as it compromises the separation of the peaks and the chromatogram quality.

For immunoassays, the antibodies recognises amino acids 4 to 10 on the beta chain and variants can affect the first 4/5/6 aa of the N-terminal site of the beta chain; the best estimates (about the uncommon variants) that I’ve gotten from experts is a prevalence of one in 50 thousand. The S/C/D/E variants are more prevalent than this.

Price. Excluding POCT, most immunoassays on the market are cheaper than HPLC.

Hb variants are often “hidden” and having them in the HPLC chromatograms can lead to dangerous over and under-treatment of patients. What are your thoughts on this?

We should not be overly concerned with variants as they are only of concern when they are homozygous. Homozygosity is also uncommon, and as seen in the NUH study (as a referral centre), where the prevalence was roughly <1%. When do we suspect the presence  of variants? A rough rule of thumb I follow is when the A1c is <4-4.5% and >14%. Just because variants are present, doesn’t necessarily mean there will be over or under-treated. Patients themselves can be their own controls-higher or lower A1c reflects worse or better glycemic control.

Some schools of thought might call for additional testing in the presence of variants, for example using glycated albumin. But these are not well evidenced in the context of HbA1c literature at the moment as they lack large longitudinal studies like DCCT or UKPDS and the methods are not standardised.

How do laboratories handle and manage samples with Hb variants present?

This is straightforward. Suspicion must be confirmed with globin chain synthesis and specialised methods, if required. Homozygous variants only matter in patients who are women of reproductive age because when they conceive  the baby might have thalassemia. In addition, homozygous variants are quite uncommon.

For heterozygous variants, the impact is minimal and we can suspect that if A1c <4-4.5%/>14%, as stated previously. We would add a coded comment to our lab report  and suggest additional testing if required.

What would be your recommendation for laboratories, similar to yours, that are considering a new HbA1c solution?

For laboratories that are set up similarly to us, I highly recommend having and using a rapid and high throughout analyser that can cater to the volume of the samples and deliver a quick turnaround time. The real bottleneck is not the analytical portion, but the phlebotomy. If you organise your phlebotomy stations in an outpatient clinic well, this process can be streamlined. Once this has been sorted, it is also crucial to have in place a rapid sample delivery system from the clinic to the lab, such as pneumatic tubes, to complete the whole workflow from patient to sample analysis.

To learn more about how your lab could benefit from an HbA1c test and its operational implementations, please reach out to Yingli Huang, Business Manager (Core Lab) at Roche Diagnostics Asia Pacific. ([email protected]).