Haemoglobin A1c is besides glucose the most requested analysis for diagnosing diabetes due to its ability to predict diseases that occur due a high blood sugar. The analysis has some limitations when a haemoglobin variant is present as they may cause falsely low or high results. In this study, we used whole blood from routine requested haemoglobin A1c samples for screening for haemoglobin variants with Capillarys 3 TERA. The samples with haemoglobin variants was then analysed with Cobas c501, Roche and Afinion, Alere for comparison of the methods. The aim of this study was to investigate the presence of haemoglobin variants in the county of Uppsala by screening over 3000 haemoglobin A1c samples. Samples with known haemoglobin variant was analysed as above to compare the results with those found in our county to study the possibility to identify the haemoglobin variants with Capillarys 3 TERA. The presence of haemoglobin variants in the county of Uppsala was approximately 0.5%. Correlation between Cobas c501 and Capillarys 3 TERA was strong for samples with haemoglobin F (r = 0.988) and for the samples with haemoglobin variants the correlations coefficient was 0.94. Correlation coefficients between Afinion and Capillarys 3 TERA for samples with haemoglobin F was 0.987 and for the haemoglobin variant 0.937. The presence of haemoglobin variants in the county of Uppsala is still low. Capillarys 3 TERA is capable to separate the most common variants and despite the good correlation every haemoglobin variant needs to be handled separately and the results should be answered with a comment that there is a possibility that the haemoglobin variant is giving an incorrect result.