In This Issue of Diabetes Care

The relationship between C-peptide and glucose levels during progression to type 1 diabetes might be used to identify individuals at high risk, according to Ismail et al. (p. 2264). Specifically, it seems an inflection point exists where levels of C-peptide transition from an increase to a decrease at ∼1.5 years prior to diagnosis. According to the authors, the inflection point may reflect changes in immunologic and environmental factors and thus potential research options. Perhaps more pertinently, it might also help identify high-risk individuals for participation in future prevention studies for type 1 diabetes. The findings come from further analyses of oral glucose tolerance test data from the Diabetes Prevention Trial—Type 1 (DPT-1) (n = 72) and the TrialNet Pathway to Prevention (n = 82) trials. Specifically, participants had complete oral glucose tolerance test data at baseline and then every 6 months up to 2 years prior to diagnosis. In both trials, the authors found that glucose levels increased from baseline to about 0.5 years prior to diagnosis. In contrast, C-peptide levels increased from baseline until ∼1.5 years prior to diagnosis and then decreased over the following year. This would contrast with the view that sudden C-peptide loss precedes a diagnosis of type 1 diabetes. Rather, it seems to follow a transition phase. Using individual and combined values of glucose and C-peptide at 1.5 years prior to diagnosis as a threshold, the authors also examined risks in the wider trial populations (n = 639 in DPT-1 and n = 6,485 in TrialNet). For those above the thresholds, 5-year risks were elevated, although the level of risk depended on the measure used. Commenting further, author Heba Ismail said: “Evidence from this study supports the concept that the metabolic progression to type 1 diabetes is better understood when glucose and C-peptide are assessed in the context of each other. Also, the findings underscore the importance of using C-peptide, in addition to glucose, for staging the progression to disease.”

Continuous glucose monitoring (CGM) is safe and effective in guiding insulin therapy for individuals with diabetes in inpatient hospital settings, according to Spanakis et al. (p. 2369). Compared with point-of-care (POC)–guided insulin adjustment, CGM resulted in a similar improvement in glycemic control and a reduction of recurrent hypoglycemic events. In the latter case, the reduction was mainly seen overnight at a time when POC-measured insulin levels are rarely checked. Author Guillermo Umpierrez said this is a period when POC-measured insulin levels are rarely checked. “Bedside POC glucose testing for glycemic control has been the standard of care in hospitalized patients for over 50 years,” he said. “And yet, it leaves extended intervals of time when glucose is not monitored. This leads to missing information that is important for glycemic control.” The findings come from a randomized open-label trial involving 185 patients with diabetes requiring insulin who were admitted for general medical or surgical care. All patients were treated with a standard basal-bolus insulin regimen and received POC capillary glucose testing before meals and bedtime. The standard-of-care group wore a blinded CGM device but had insulin adjusted according to POC measurements. In contrast, the CGM group had insulin adjusted according to CGM data. There were largely no differences between the groups in terms of a range of glycemia metrics, but there was a significant reduction in hypoglycemia reoccurrence and lower percentage of time below range in the CGM group, particularly during nocturnal hours. Historically, there have been concerns around the use of CGM in such settings due to lack of knowledge around the technology and perceived hypoglycemia risks. However, as Dr. Umpierrez points out, the balance of evidence may be changing. “Recent observational and prospective studies have reported high accuracy between POC and CGM, as well as increased detection of both hypoglycemic and hyperglycemic events with the use of CGM,” he added. “Our prospective randomized study indicates that real-time CGM devices are safe to adjust insulin therapy in hospital settings.”

A reanalysis of data from the Diabetes Control and Complications Trial (DCCT) suggests that any association between estimated continuous glucose monitor (CGM) time-in-range (eTIR) glucose and microvascular complications is due to the correlation of eTIR with HbA1c, according to Lachin et al. (p. 2445). Such a finding would invalidate the view that CGM time-in-range glucose measurements are “an acceptable end point for clinical trials,” as proposed originally by Beck et al. (doi.org/10.2337/dc18-1444). Repeating the original analysis that assessed the relationship between eTIR and retinopathy/microalbuminuria, the current authors (who are from the DCCT/EDIC Research Group) also adjusted for HbA1c and other covariates, something that the original analysis by Beck et al. did not include. In the original unadjusted analyses, eTIR was significantly associated with all microvascular outcomes (i.e., full, primary, and secondary cohorts for two microvascular outcomes). When the DCCT authors then adjusted for HbA1c alone, the hazard ratios narrowed much closer to 1. Only the association between eTIR and nephropathy reached significance. When fully adjusted, there were barely any significant relationships. Reversing the associations, HbA1c had statistically significant relationships with the outcomes when unadjusted and (with a few exceptions) after adjusting for eTIR alone and with other covariates. Based on the new analysis, the authors suggest that there is no evidence that times in range derived from CGM measurements would have an association with eye and kidney microvascular complications of diabetes. Rather, HbA1c should remain the primary measure for determining microvascular risks associated with (type 1) diabetes—in contrast to the original conclusions by Beck et al. Commenting further, author John Lachin said: “There is no doubt that TIR and other CGM metrics provide critically useful information to guide insulin therapy. However, the association of these metrics with outcomes is largely mediated by their association with HbA1c, and therapies aimed at reducing the risk of complications would best be evaluated by their effects on HbA1c.”

Less aggressive use of supplemental insulin in hospitalized patients with type 2 diabetes appears to give glycemic control equivalent to that of a more aggressive approach, according to Vellanki et al. (p. 2217). They add that, potentially, supplemental insulin initiated at higher glucose levels with basal-bolus insulin regimens are “expected to lead to easier implementation of insulin treatment without compromising glycemic control.” More intensive supplemental insulin use for non–intensive care unit (ICU) patients with type 2 diabetes is recommended in guidelines, as it can improve glycemic control. However, it comes at a cost of increased treatment burden and risks for hypoglycemia, according to the authors. Based on previous work, they hypothesized that there would be no difference in glycemic control between intensive and nonintensive supplemental insulin use at night in such patients. Using an open-label study design, the authors recruited patients with type 2 diabetes who were hospitalized either for acute/chronic medical illnesses or for elective/emergency surgery. Patients were then assigned to intensive (blood glucose >140 mg/dL) or nonintensive (blood glucose >260 mg/dL) administration of rapid-acting supplemental insulin before meals and bedtime to lower blood glucose in addition to basal-bolus insulin regimens. They found that nonintensive treatment resulted in mean daily blood glucose of 173 ± 43 mg/dL, while intensive treatment led to mean glucose levels of 172 ± 38 mg/dL. According to the authors, this meant nonintensive treatment was noninferior to the intensive approach. Indeed, fewer patients received sliding-scale insulin in the nonintensive (34%) versus the intensive (91%) treatment groups. Commenting further on the findings, author Priyathama Vellanki said: “The use of supplemental insulin starting at glucose >140 mg/dL added on to a basal-bolus insulin regimen is complicated and leads to hypoglycemia. Our study shows that, among hospitalized patients with mild to moderate hyperglycemia, daily titration of a basal-bolus insulin regimen is sufficient to achieve glycemic control without the need for a supplemental sliding scale.”