Trends in HbA_1c and LDL Cholesterol in Patients With Type 2 Diabetes Receiving First-Time Treatment in Northern Denmark, 2000–2017: Population-Based Sequential Cross-Sectional Analysis

The evolving evidence base for early diabetes detection and intensive treatment over the recent decades has likely led to a change in the profile of patients initiating type 2 diabetes treatment (1–3). Knowing in which direction the clinical profile of patients with diabetes is moving is important because treatment guidelines rely on the generalizability of clinical trials to contemporary populations (3). Many of these trials were conducted when both the diagnostic criteria and treatment targets were quite different than today (1–3). Population-based data investigating long-term trends in diabetes management and treatment targets are scarce. We aimed to examine 18-year changes in HbA_1c and lipid testing and control among people initiating glucose-lowering drugs (GLDs) for type 2 diabetes. We performed one of the first population-based studies, using laboratory and health care databases covering the entire population of Northern Denmark (∼1.8 million people) (4).

We identified people redeeming their first-ever GLD prescription at age ≥30 years. For all 94,175 GLD initiators 2000–2017, we examined pretreatment HbA_1c and lipid levels and proportions of testing. For initiators 2000–2016, we assessed 12 months posttreatment lipid-lowering therapy and achievement of glycemic and LDL cholesterol targets. Mean HbA_1c and LDL levels and reductions were plotted with 95% CIs.

Median age at first GLD treatment fell from 64 years in 2000 to 61 in 2017; 56% of the study population were men. The proportion of patients with at least one HbA_1c test within 12 months after GLD initiation increased from 53% (2000) to 95% (2016) (Fig. 1A). Concurrently, mean pretreatment HbA_1c decreased from 9.2% (77 mmol/mol) (2000) to 7.9% (63 mmol/mol) (2017), with a nadir occurring in 2011 (7.3% [56 mmol/mol]). For mean posttreatment HbA_1c, a smaller decline was seen from 7.1% (54 mmol/mol) (2000) to 6.6% (49 mmol/mol) (2016). (Fig. 1B). The proportion of patients achieving posttreatment HbA_1c target <7% (53 mmol/mol) increased from 54% to 81% during 2000–2016 and for target <6.5% (48 mmol/mol) increased from 37% to 56% (Fig. 1C). The proportion with a pretreatment HbA_1c below 6.5% (42 mmol/mol) increased from 7% in 2000 to 31% immediately preceding the 2012 introduction of HbA_1c as a diagnostic criterion. After the change in diagnostic criteria, the group with HbA_1c <6.5% before treatment dropped substantially, to only 12% in 2017 (Fig. 1D). As shown in Fig. 1E, GLD initiators below the current 6.5% diagnostic threshold did not experience any posttreatment HbA_1c reduction. In contrast, patients in successively higher pretreatment HbA_1c categories had increasingly large posttreatment reductions.

The proportion of patients who had at least one blood lipid test within 12 months following their first-ever GLD treatment increased from 82% (2000) to 99% (2016). The proportion receiving lipid-lowering therapy within 12 months quintupled from 12% (2000) to 61% (2016) but declined after peaking at 68% in 2011 (Fig. 1F). Mean pretreatment LDL cholesterol declined from 3.5 mmol/L (2000) to 2.8 mmol/L (2017), while the mean posttreatment value declined more, from 3.3 mmol/L (2000) to 2.3 mmol/L (2016) (Fig. 1G). The proportion achieving LDL cholesterol target <2.6 mmol/L increased from 23% (2000) to 65% (2016) and for target <1.8 mmol/L from 5% (2000) to 29% (2016) (Fig. 1H).

We found evidence that real-life patients with first-treated type 2 diabetes have changed markedly in the past 18 years. A large decline in HbA_1c levels before first-time GLD therapy is probably a main driver of the improvement in glycemic target achievement. Newer GLDs with a decreased risk of hypoglycemia may be another potential contributor. Of note, our investigated HbA_1c targets may not apply to all patients (3). Improvements in LDL cholesterol over time may relate to more intensive lipid-lowering therapy. Overall, recent developments likely reflect a combination of evolving clinical practices (earlier and more complete diabetes detection and coding practices), secular demographic changes, and true improvements in treatment. It is difficult to pinpoint one key driver of the observed changes. The main factors driving these changes—changing diagnostic and treatment guidelines, demography, and increasing treatment options—are seen in other Western countries (3). One study limitation is that HbA_1c measurements in early years may have been restricted predominantly to patients with anticipated glycemic control problems, which could lead to overestimation of HbA_1c improvements over time.

Although monitoring and treatment of glucose and cholesterol has improved considerably, there is room for further improvement, especially in proportions initiating lipid-lowering therapy. Finally, the introduction of HbA_1c for diagnosis of diabetes will have led to the exclusion of patients with blood glucose but not HbA_1c in the diabetic range from this study of GLD initiators.