Interpretation
OGTT remains the gold standard test for diagnosing GDM. With the pressure on health care resources, anxiety associated with the diagnosis, and the need to start management in timely manner, providing accurate, precise and rapid results for OGTT is beneficial. The advancing technology of POC devices can help achieve these goals.
Several studies have investigated the use of a variety of variable POC glucometers using different diagnosis criteria. Summary of these studies are included in table 2. The studies used a variety of VPG reference measurements: from sending samples to the hospital laboratory without special arrangements as we have done here, which has the advantage being a comparison to usual clinical practice; through sending to the hospital laboratory paying strict attention to The National Association of Clinical Biochemist (NACB) for the diagnosis of GDM guidelines to minimise impact of pre-analytical factors; through to use of Isotope Dilution Gas Chromatography Mass Spectrometry. Most studies showed good agreement between POC testing and VPG with acceptable diagnostic performance, but no study has advocated for completely replacing laboratory testing with POC-testing for GDM screening. Other studies comparing POC with Lab-VPG methods in 50g OGGT also reported satisfactory performance for POC in GDM diagnosis.(15-17) However, the case for comparing POC testing to laboratory testing is far from straightforward due to multiple factors including clinical and analytical factors.
Clinically, the diagnostic performance is partly depending on the diagnostic thresholds. The HAPO study showed there is a linear relationship between increasing glucose concentrations at OGTT and adverse pregnancy outcomes, such as macrosomia, neonatal hypoglycaemia and caesarean birth, with no threshold effect.(1)Health care systems impose diagnostic thresholds: those at or above threshold are managed as GDM and those below are labelled as no GDM. This is relatively easy to administer and allows healthcare resources to be directed for those at higher risk of adverse pregnancy outcomes. However, in this situation, when a linear parameter is converted into a binary outcome, when different systems are used to measure the linear parameter there will be diagnostic disagreement, particularly when close to the diagnostic threshold.
Analytically, POC glucometer performance is subjected to analytical interferences from variation in haematocrit, pH and oxygen and sample matrix effect.(9) However, the laboratory methods, against which POC devices are compared, have inherent analytical and pre-analytical errors, of which the effect of in-vitro glycolysis is particularly significant. Uninhibited in-vitro glycolysis can result in 5-10% reduction in VPG(18) and GDM misclassification. To prevent this, NACB guidelines recommends collecting samples in sodium fluoride additive tubes, transferring them on slurry ice to be centrifuged with 30 minutes of collection. Alternatively, citrate tubes can immediately inhibit the glycolysis.(19) In routine practice adherence to these guidelines is suboptimal.(20-22) So the discrepancy in diagnostic performance might be partly attributed to the negative bias with lab methods rather than positive bias with POC. Generally, studies that compared POC without strict measures to control in-vitro glycolysis, like our study, have reported positive analytical bias,(23, 24) high sensitivity and NPV(25-29) for POC with potential for over-diagnosis. While some studies that religiously applied NACB guidelines have reported negative analytical bias and potential for misdiagnosis.(30, 31)