Postprandial or Fasting Hyperglycemia: Time to Move On?

In this issue of Diabetes Technology & Therapeutics, Kang et al. ¹ report on the relative contributions of fasting plasma glucose (FPG) and postprandial glucose (PPG) to overall glycemia as a function of different hemoglobin A1c (HbA1c) ranges in untreated, newly diagnosed patients with type 2 diabetes. Using continuous glucose monitoring, they found that PPG accounted for 77% of overall hyperglycemia in those with an HbA1c level below 7%, but only 23% in those with an HbA1c level above 9%. Many earlier studies have tried to disentangle the relative contributions of FPG and PPG across different HbA1c ranges, so far almost always in patients treated with oral agents. ^{2 –7} How did this field start? Which methodological issues have to be taken into account when interpreting the findings? And what, if any, message should we take home when treating our patients?

From epidemiological data on people undergoing a glucose tolerance test, we know that PPG is more strongly related to mortality than FPG. ^8,9 However, it goes for many endocrine systems that a stimulated value tells you more than an unstimulated value. Fasting cortisol will only give the diagnosis of Addison's disease in case of deep hypocortisolemia, but usually an ACTH stimulation test is needed to obtain the diagnosis. Similarly, it would only seem natural that a glucose value stimulated by a meal is more strongly related to poor outcomes than an unstimulated, fasting value. Nevertheless, it was postulated that PPG should perhaps become a treatment target as such because of these epidemiological analyses. ⁸

In 2003 Monnier et al. ² took this a step further, trying to quantify the contributions of basal glucose and PPG to overall hyperglycemia over the day in 290 patients. Their assessment was that postprandial excursions were relatively stable across HbA1c strata, but that the relative contribution of PPG was largest in the lowest HbA1c stratum, decreasing linearly from around 70% in the HbA1c <7.3% stratum to around 30% in the HbA1c >10.2% stratum. Two medium-sized studies confirmed these findings, in 52 Caucasian ³ and 66 Japanese ⁵ patients both on oral agents, as well as one larger analysis drawn from a pharmacy-initiated study in 973 patients on metformin. ⁷ The medium-sized study by Kang et al. ¹ adds to the database as it is the first study in treatment-naive patients.

Other studies, however, showed different findings. One study in 121 patients showed a higher contribution of PPG in those with an HbA1c level ranging from 5.7% to 7%, but a roughly 50:50 distribution for FPG and PPG in all four strata with an HbA1c level above 7%. ⁴ By far the largest study on this topic reported on 1,699 patients on oral drugs about to start insulin treatment in various randomized controlled trials. ⁶ Over all five HbA1c strata, they found the postprandial contribution to be around 22%, with perhaps slightly higher at 24% in those in the lowest HbA1c stratum with an HbA1c level of <8%.

In summary, the finding that those with an HbA1c level of around 7% and below have a relatively high contribution of PPG seems consistent across studies. The picture becomes more blurred in the higher strata where study results are contradictory.

Can the different results of these studies perhaps be explained by methodological issues? Some studies use self-monitored blood glucose, others continuous glucose monitoring. The baseline from which basal hyperglycemia is defined differs between 5.5 and 6.1 mmol/L. The study reported here ¹ brings up the question whether this baseline should be a single cutoff at 5.5 or 6.1 mmol/L or a varying baseline quantified by studying a group of people with normal glucose tolerance, as the authors also did. When taking the mean derived from a group of people with normal glucose tolerance as the baseline, the numbers changed considerably, but the overall trend—that is, a high PPG contribution in those with a lower HbA1c level—remained unchanged. Taken together, these methodological issues will have an impact but do not seem to give a fully satisfactory explanation for the different findings of the studies on those with somewhat higher HbA1c values. A patient-level meta-analysis could be a useful step forward.

Can we translate these findings into clinical practice? Most oral agents do not target PPG specifically, with the exception of acarbose and glinides, which are not used widely. One major trial compared insulin initiation with rapid-acting analogs to basal insulin. ¹⁰ As expected, a lower FPG was reached with basal insulin, whereas the rapid-acting analog approach controlled PPG better, but there was no difference in HbA1c level. The basal insulin approach resulted in fewer hypoglycemic events, fewer injections, less self-monitoring of blood glucose, and greater patient satisfaction. Another trial used a similar comparison to assess cardiovascular outcomes in a population of 1,115 patients with type 2 diabetes after myocardial infarction. ¹¹ Again, a lower FPG was reached with basal insulin, whereas the rapid-acting analog approach controlled PPG better, but no difference in the primary outcome was seen. Of course, this is only one trial assessing hard outcomes, and it would be interesting to see similar trials in patients with a lower cardiovascular risk.

In summary, the relative contribution of PPG as compared with FPG is high in those with the lowest HbA1c values, both in people on oral agents and as now reported by Kang et al. ¹ also in newly diagnosed patients. But taking all the available evidence together, it is not surprising that treatment guidelines do not recommend specific targeting of PPG values in those with lower HbA1c values.