In This Issue of Diabetes Care

Further analysis of the Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) by Garvey et al. (p. 562) reveals several baseline factors that were associated with the glycemic outcomes in the trial. In a series of statistical analyses, the authors found that only younger age (i.e., <58 years) and baseline HbA_1c ≥7.4% increased the risk of reaching the glycemic outcome (defined as treatment failure at HbA_1c ≥7.0%) by year 4. In a multivariate model, there was also evidence of associations with the glycemic outcome for younger age, higher baseline HbA_1c and fasting glucose, and treatment group (i.e., sitagliptin, liraglutide, glimepiride, or glargine treatment group). A much wider range of factors was identified in univariate analysis. For example, younger age, Hispanic ethnicity, high values of fasting glucose and HbA_1c, lower insulin secretion, and higher insulin resistance were among several factors associated with the glycemia outcome at years 1 and 4. However, several other factors surprisingly were not associated with the outcomes, including BMI, waist-to-hip ratio, and diabetes duration. A limited number of factors were also selectively associated with a differential response to the medications at year 1. For example, liraglutide was more effective in women and younger White individuals at year 1, while sitagliptin was associated with greater risk of the glycemic outcome in White but not in Black people. By year 4, no difference in effectiveness between the four medications was seen. The authors note that GRADE did not include sodium–glucose cotransporter 2 inhibitors or the newer (and more effective) glucagon-like peptide 1 receptor agonists and related therapies, which may well further alter decision-making around add-on therapies. “These results from the GRADE study argue for more aggressive treatment and follow-up in patients who present at a younger age of onset and/or with high HbA_1c values,” said author W. Timothy Garvey. “These patients should be followed closely for treatment failure with intensification of therapy without delay as HbA_1c values rise. In addition, the results point to the need for research examining whether initial combination therapy might be advantageous in the subset of patients.”

A head-to-head comparison by Kirkman et al. (p. 594) of four widely used diabetes medications added to metformin suggests that about three-quarters of participants in the Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) experienced a composite of an increase in HbA_1c, weight gain, and hypoglycemia. Specifically, however, the risk for the composite outcome differed among treatment groups, with those randomized to liraglutide having the lowest risk, while those randomized to sitagliptin, glargine, or glimepiride had increasingly higher risk. Analyzed individually, the risks for weight gain and hypoglycemia were in the same order, while risk for glycemic deterioration was lowest for glargine followed by liraglutide, glimepiride, and then sitagliptin. As previously reported, all treatments resulted in reductions in HbA_1c, with glargine and liraglutide being the most effective. Liraglutide also had the lowest risk for weight gain. Notably, however, when insulin was added as a rescue therapy when HbA_1c increased, any protection against weight gain was lost, even when treatments with proven weight reduction characteristics were continued. Clear differences were also evident between the treatments for risk of hypoglycemia, with higher risk associated with glimepiride and glargine compared with liraglutide and sitagliptin. Finally, the authors report that treatment satisfaction was generally high and similar among the groups, other than slightly lower satisfaction with glargine. They note that the study design could have contributed to this (free drugs and follow-up), and there was modest evidence that those who reached the composite outcome had lower satisfaction levels. According to author M. Sue Kirkman, “People with diabetes and their providers choose second-line therapies based not just on glycemic effects but on their potential for adverse effects such as weight gain and hypoglycemia. These analyses support shared decision-making about the many options available for treating type 2 diabetes.” Commenting further, author Rodica Pop-Busui added, “These findings from the GRADE study support informing all individuals with diabetes about the entire spectrum of effects of a given diabetes medication beyond strictly glucose lowering.”

The addition of basal insulin to metformin monotherapy did not increase emotional distress in adults with early type 2 diabetes relative to other treatments in the Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE), according to Gonzalez et al. (p. 610). Contrary to expectations, the basal insulin used in the study (insulin glargine) resulted in lower levels of the primary diabetes distress outcome compared with the other add-on treatments after 1 year. All the treatments resulted in decreases in both measures of emotional distress over 1 year and that diabetes distress was lower with insulin glargine than the other treatments. Liraglutide also lowered diabetes distress more than glimepiride and sitagliptin. Over 3 years, there were no differences in diabetes distress between the treatments. A component of diabetes distress, interpersonal diabetes distress, was also lower with liraglutide compared with the other treatments. The authors note there may be limitations with the findings, particularly with generalizability. Of note, treatment was provided at a level of care that might not be representative of typical settings in the U.S. In particular, treatments and associated care were provided free of charge, removing concerns about insurance or the typical financial co-pay costs associated with diabetes drugs, including insulin. The authors suggest the findings should be informative for both providers and patients, who are often hesitant about commencing insulin therapy. “We were excited to use GRADE’s unique design to provide for a strong test of the widely held belief that basal insulin is more burdensome for individuals with type 2 diabetes as compared with other alternatives,” said author Jeffrey Gonzalez. “We expected that those randomly assigned to add this medication to metformin would experience more emotional distress than those assigned to other glucose-lowering medications. The findings clearly failed to support that expectation, and we hope that our results will encourage acceptance of basal insulin when clinically indicated.”

Type 1 diabetes and diabetic ketoacidosis (DKA) were diagnosed less often in children on weekends and public holidays and during school vacations in Germany in the period 2013–2022, according to Kamrath et al. (p. 649). Rates of diagnosis were also higher at the beginning of the week (i.e., Monday and Tuesday) than the end of the week. The authors suggest the differing rates of diagnosis are due to a combination of family and primary care factors. In particular, the differences may be attributable to the way health care is provided in Germany, with pediatric care mostly available on weekdays and only general care available on weekends. The authors found that the vast majority (87%) of individuals with type 1 diabetes were diagnosed on weekdays, with the highest proportion diagnosed on Mondays and Tuesdays (41%). They found adjusted odds ratios for being diagnosed with type 1 diabetes were all much lower on weekends and public holidays and during school vacations. Cases of DKA followed a very similar pattern, as did cases of severe DKA. Of note, the study covers the period of the coronavirus disease 2019 pandemic, and previous studies have suggested there was an uptick in DKA during that period. However, in this study, odds of developing DKA during the pandemic were not significantly different from the odds in the period before the pandemic. “The manifestation of type 1 diabetes is often misdiagnosed, especially in times when there is less expertise in the care of children,” said author Clemens Kamrath. “This is reflected in an accumulation of diabetes diagnoses as well as in ketoacidotic derailments at the beginning of the week. These data indicate a need for training on the typical signs of diabetes manifestations in children and adolescents for physicians providing acute care outside regular office hours.”