Variant haemoglobin may affect measurements

BMJ 1995;310:740 (18 March)

Letters

Variant haemoglobin may affect measurements

EDITOR,--We agree that Eric S Kilpatrick and colleagues' studysuggests the need for greater standardisation of methods usedfor estimating glycated haemoglobin concentrations.¹ The estimationof glycated haemoglobin in patients with variant haemoglobinsalso urgently needs standardisation. An external quality assessmentof the measurement of glycated haemoglobin in such patientsby laboratories in the North West Thames region showed alarmingdiscrepancies in the results obtained (C E Andrew and P Harrison,national meeting of Association of Clinical Biochemists, 1991).For some samples the variability was so large as to alter completelythe interpretation from one of very poor diabetic control toone of overcontrol likely to lead to hypoglycaemic episodes.An additional worrying finding was the varying methods usedto calculate results, which partly accounted for the discrepancies.While some of the laboratories that used electrophoresis orion exchange methods expressed haemoglobin A1 and haemoglobinA1c as a percentage of total haemoglobin--that is, haemoglobinA plus variant--others expressed their results as a percentageof the non-glycated fraction, haemoglobin A0, or of haemoglobinA0 plus haemoglobin A. In these circumstances comparison ofresults obtained with different methods from any one samplebecomes meaningless.

To add to these difficulties, most ion exchange and electrophoreticmethods have the problem of coelution of haemoglobin F (HbF)with glycated haemoglobin, which leads to overestimation ofthe latter.² Small increases in HbF are not uncommon in diabeticpatients and are often unrecognised.³ Hence methods that cannotresolve HbF from glycated haemoglobin adequately are unsuitablefor estimating glycated haemoglobin and should be abandonedas soon as possible.

Until we have greater standardisation and guidelines about estimatingglycated haemoglobin concentrations in diabetic patients withvariant haemoglobins, all patients with newly diagnosed diabetesshould undergo a haemoglobinopathy screen. This should includemeasurement of HbF at the time of initial presentation to enablecorrect interpretation of results.

All laboratories should be willing to provide information aboutthe method used to measure glycated haemoglobin, the mode ofcalculations of results, and interference from variant haemoglobinsincluding HbF. Clinicians would also be well advised to ascertainnecessary information about the methods used from their laboratorycolleagues.

Consultant clinical pathologist Newham General Hospital, LondonE13 8RU

S Bulusu