The role of the OGTT in the diagnosis of diabetes in the general population has been the subject of debate for decades due to the test’s poor reproducibility, which is mainly caused by random variants in glucose metabolism. Thus, it is necessary to take the reproducibility of the OGTT into consideration when evaluating the use of the test in ACS. To the best of our knowledge, this is the first meta-analysis designed to evaluate the accuracy of the OGTT in ACS and to compare it with that in non-ACS cases. In patients with ACS, OGTT has a slightly lower SPE in diagnosing diabetes compared with those without ACS, while the SEN values are comparable. A lower SPE (true negative rate) indicates that performing an OGTT in patients with ACS will result in a higher proportion of false positive results. According to the result of the subgroup analysis, less than one in every 10 ACS patients diagnosed with diabetes using an OGTT before discharge will have a different result at the follow-up OGTT resulting in a change in diagnosis. After adjusting the interval between repeated tests and age in the multiple meta-regression model, performing an OGTT in ACS patients was not associated with lower diagnostic accuracy compared with non-ACS patients. This means that it is reasonable to perform an OGTT to screen for diabetes in patients with ACS before discharge.
It is believed that aging is associated with function of the pancreatic β cells, which may affect the accuracy of OGTT. In the subgroup analysis, advanced age (≥60 years) was associated with lower SPE, while the SEN is similar when compared with the younger age (< 60 years) group. The younger patients, with well preserved function of the pancreatic β cells, maintained their plasma glucose in normal range in the setting of stress and only those with previous unknown DM could be identified during ACS. On the other hand, due to compromised function of the pancreatic β cells, the patients of advanced age were unable to maintain the plasma glucose level during the stress and the plasma glucose returned to the normal range only after the stress was eliminated, which made the OGTT less accurate in patients of advanced age enduring ACS. However, the effect of age on the accuracy of OGTT during ACS is not observed when using the meta-regression model.
Due to the poor reproducibility of the OGTT, it may be reasonable to repeat the test after discharge, which is also recommended in the WHO 1999 criteria. Although the optimal time to repeat an OGTT is still unknown, OGTT could be repeated in 3 months after hospital discharge if necessary, given most of the studies repeated the OGTT 3 months after discharge.
Hyperglycemia during ACS was thought to be “stress hyperglycemia”, which develops due to a highly complex interplay between hormones (such as catecholamines, growth hormones, and cortisol) and cytokines, ultimately leading to excessive hepatic glucose production and insulin resistance. A study by Choi et al. showed that acute myocardial infarction (MI) patients with IGT or diabetes exhibited higher levels of high-sensitivity C-reactive protein and interleukin-6 levels compared with acute MI patients with normal glucose tolerance or well-controlled diabetes, indicating that glycol metabolism in acute MI is associated with acute stress and inflammation. However, this finding was not supported by the other studies included in this meta-analysis. Neither C-reactive protein nor the extent of myocardial damage was found to be related to hyperglycemia in ACS patients[21, 23, 24]. Thus, the evidence of stress hyperglycemia in ACS is still equivocal.
It is well-known that in-hospital hyperglycemia is associated with both the short-term and long-term prognosis in AMI patients[33, 34]. However, either hyperglycemia on admission, fasting plasma glucose or HbA1c, in non-diabetic patients with AMI are not sensitive enough to uncover previously undiagnosed abnormal glucose tolerance or diabetes[35–37]. Furthermore, previous studies have indicated that the 2-h post challenge plasma glucose level was a significant predictor of cardiovascular events in patients with previous MI and the prognosis of AMI patients with a new diagnosis of diabetes or impaired glucose tolerance defined by an OGTT during hospitalization was significantly worse than that of patients with impaired fasting glucose and normal glucose regulation[39–41]. Individuals who had normoglycemia, but whose 2-hour plasma glucose concentrations did not return to the FPG levels during an OGTT have been shown to have increased risk of cardiovascular diseases. Thus, early performance of OGTT during the AMI hospitalization may provide an opportunity to detect the high risk population and establish the undiagnosed diabetes or impaired glucose tolerance. A recent study has shown that discontinuation of anti-hyperglycemic therapy during AMI hospitalization is common and associated with higher mortality rates after discharge in older patients. Early diagnosis of diabetes during AMI hospitalization may be vital to improve compliance with life style changes or anti-hyperglycemic therapy.
Studies of repeated OGTT performed only in patients with abnormal or normal OGTT results in the first test were excluded from this meta-analysis, since these studies may lead to underestimation or overestimation of the accuracy of OGTT. Studies with intended medical intervention between the two OGTTs were also excluded, because several interventions such as exercise and medication have been shown to be associated with improved glucose tolerance.
The use of the hemoglobin A1C test to diagnose DM, with a threshold of 6.5%, had been recommended by ADA 2010 criteria due to its stability and convenience. Recently, Ramachandran et al. divided newly diagnosed DM patients with ACS into two groups based on A1C level: undiagnosed preexisting diabetes was considered possible if the hemoglobin A1C values were ≥6.0%, and stress hyperglycemia was considered if the hemoglobin A1C values were <6.0%. Surprisingly, after 3 months, all of the undiagnosed preexisting diabetes patients remained diabetic, while only 16.7% of the stress hyperglycemia patients remained diabetic. Thus, new consensus statements on the care of the hyperglycemia in ACS patients have recommended its use during hospitalization with an OGTT.
The studies have several limitations. First, even though the analysis identified statistically insignificant publication bias, all of the articles that fulfilled the inclusion criteria were not included because some authors were unwilling to offer the original data and there may be data relevant to this topic which have never been published. Second, these findings are based on an indirect comparison of two groups of participants, and no comparison of the reproducibility of the OGTT between ACS and non-ACS patients was available in a single study. Finally, this is a meta-analysis of study level, instead of individual level, which makes it impossible for us to determine the optimal timing for OGTT during hospitalization.