Autor:innen:
Dr. med. Norbert Ostendorf | St. Katharinen-Hospital Frechen | Germany
Birgit Meyer-Noack | St. Katharinen-Hospital Frechen | Germany
Objective: HbA1c is commonly used for diagnosis and follow-up of Diabetes mellitus. The proportion of glycated HbA1 not only depends on the average blood glucose level, but also on erythrocyte lifespan. The latter is usually assumed to be 120 days, although in pathologic conditions like bleeding or iron deficiency anemia it can be largely different. In a retrospective study, we evaluated the dependency of HbA1c results and reticulocyte count using routine laboratory data from an intermediate care hospital. To our knowledge, no study to date has investigated the potential role of reticulocyte counting to assess the reliability of HbA1c -values.
Methods: From 01.01.2009 to 31.05.2016, all hematologic samples sent to our laboratory for measurement of the reticulocyte count and/or proportion of HbA1c were identified in the laboratory information system. In 297 cases, both parameters were measured simultaneously from the same blood sample. An additional number of 176 measurements were performed from different samples within no more than 5 days. All data sets were checked for comparability of hematologic parameters and for compromising incidents before or between measurements, like bleeding, transfusion or start of iron therapy. 47 data sets were excluded, leaving a total number of 426. The data were anonymized and exported to R statistical program for further analysis.
HbA1c was measured using a competitive, immunoturbidimetric Assay (Beckman coulter Hemoglobin A1c) on a Beckman Coulter AU400 analyzer (Beckman Coulter, Krefeld, Germany). The reticulocyte count was performed on a Siemens hematology analyzer (ADVIA 120, Siemens Healthcare Diagnostics, Eschborn, Germany). In June 2016, the hematology analyzer was changed to a model from another manufacturer. Data collection was stopped at that point to avoid bias from the different assay.
Results: Using 35‰ reticulocytes as an arbitrary cut off, mean HbA1c in the samples with a reticulocyte count below this threshold was 7.0% (53 mmol/mol, 373 values, range 4.3 to 13.1% or 24 to 120 mmol/mol). In the group with reticulocyte counts equal or greater than this cut off, mean HbA1c was 6.3% (44 mmol/mol, 53 values, range 4.4 to 16.1% or 25 to 152 mmol/mol). Omitting the two highest values as outliers, the mean HbA1c was 6%.
Moreover, in 28 of the 473 measurements (5.9%), the patients received red cell transfusions before HbA1c was analyzed. This is known to produce false low results.
Conclusion: Investigation of the reticulocyte count is a useful tool to evaluate the reliability of HbA1c measurements. Low or high reticulocyte counts indicate a bias on HbA1c results.
In hospital routine, measurement of HbA1c following red cell transfusion is not infrequent and must be taken into consideration when results are implausible.