<h3>Background</h3> In 2007, it was estimated that nearly 1.2 million people in Malaysia had diabetes mellitus (DM). The Diabetes Control and Complications Trial (DCCT) and UK Prospective Diabetes Study (UKPDS) demonstrated conclusively that risk for complications of DM are directly related to glycaemic control as measured by HbA1c. Many commercial methods for HbA1c available are based on different analytical principles, giving rise to different results, thus harmonisation of this is important to provide comparable results. <h3>Aims</h3> To evaluate the analytical performance of HbA1c assay on TINA-QUANT Roche COBAS 6000 against the HBA1c goals and correlation to the high-performance liquid chromatography (HPLC) ADAMS<sup>TM</sup> A1c HA-8160 (ADAMS) analyser. Interference study was performed on these two analysers and Primus PDQ analyser. <h3>Methods</h3> Cobas 6000 employs automated immunoturbidimetry. ADAMS uses cation exchange HPLC and Primus PDQ uses boronic affinity HPLC method. Imprecision studies were performed using patients' samples and two commercial control sera levels. Each sample was assayed in duplicate for 20 runs and 5 consecutive days for within-run and between-run, respectively. Recovery experiment was performed using 20 multilevel National Glycohaemoglobin Standardization Program (NGSP) samples. Method comparison was performed using 54 patients' samples on Cobas 6000 with ADAMS as the reference system. Interference study was performed using 10 patients' samples in each group of patients with thalassaemia, uraemia and hyperbilirubinaemia. <h3>Results</h3> The within-run coefficients of variation (CV) were 0.7%, 2.5% and 1.40% at 5.7%, 7.7% and 10.0% HbA1c, respectively. The total imprecisions were 0.9% and 2.0% for 5.7% and 9.8% HbA1c, respectively. The bias at the decision level of 7% HbA1c was 4.2% and the calculation of the sigma matric based on CAP PT 2007 quality requirement was 5.4. The mean recoveries for Cobas 6000 and ADAMS were 101% and 97%, respectively. The Cobas 6000 correlates well with the reference system. Cobas 6000 and ADAMS showed higher HbA1c value for uraemia and hyperbilirubinaemia samples. <h3>Conclusion</h3> The automated immunoturbidimetry (TINA-QUANT Roche COBAS 6000) demonstrated acceptable precision, good recovery and is comparable with the HPLC method.