This was an epidemiological, observational, cross-sectional, multicenter national study conducted between October 2014 and April 2015. Eighty-two investigators, who were primary care physicians of the various autonomous communities, participated in the study, and their distribution was proportional to the population of each autonomous community in Spain.

The study population consisted of all patients 65 years of age or older with known or newly diagnosed DM2 according to the 2011 ADA criteria16 (cardinal symptoms of diabetes and random glucose level ≥200mg/dL on one occasion; fasting blood glucose ≥126mg/dL, HbA1c≥6.5% or blood glucose≥200mg/dL 2h after an oral overload with 75g of glucose). If the presence of hyperglycemia was not unequivocal, the results were confirmed by repeating the analysis.

The patients were required to sign an informed consent document (themselves or their legal representative), and they could not be participating in any other epidemiological study.

The study excluded individuals younger than 65 years, those who did not meet the 2011 ADA diagnostic criteria for DM2, those in terminal condition and those whose medical history provided insufficient data for the study, in the assessment committee's judgment.

Each medical researcher was asked to select 12 patients with DM2 who meet the inclusion criteria. To this end, a sequence of random numbers was provided from which these patients could be chosen from the medical researcher's list of all patients with DM2 who were 65 years or older.

The study included 939 patients, whose mean age was 76.4±6.7 years. The values for the anthropometric variables, functional capacity and comorbidity distributed by age group are shown in Table 1. The results of the degree of glycemic control, baseline fasting glucose and the rate of hypoglycemia according to the degree of dependence and comorbidity are shown in Table 2. The HbA1c level showed statistically significant differences based on the degree of disability between the elderly patients with DM2 and severe dependence and those who were independent (p=.01) or had slight dependence (p=.03). The patients with severe dependence had a higher mean HbA1c level (7.9±1.1%). There was also a significant difference between the mean HbA1c level of the elderly patients with diabetes and moderate dependence versus those who were independent (7.43% vs. 7, p=.047). A high percentage of the patients with total dependence (62.5%) were undergoing insulin treatment.

According to the Charlson index, 57.2% of the elderly with DM2 had medium comorbidity (score 1–2), 28.4% had high comorbidity (3–4), and 14.4% had very high comorbidity (≥5).

The HbA1c levels increased as the comorbidity increased, with statistically significant differences between the HbA1c readings of the elderly patients with DM2 and medium comorbidity versus those with very high comorbidity (p=.012) (Table 2). As the comorbidity increased, the percentage of patients treated with insulin alone or combined with secretagogues increased, while the percentage of patients treated with oral diabetes drugs decreased.

Table 3 shows the data for the glycemic control and prevalence of cardiovascular risk factors. Forty percent of the patients had a body mass index>30, 43.7% had systolic blood pressure readings>140mmHg and/or diastolic readings>90mmHg. Only 3.2% of the patients had total cholesterol readings>250mg/dL. Some 84.3% of the patients without cardiovascular disease had LDL cholesterol levels<130mg/dL. In 69.1% of those who had prior cardiovascular disease, the LDL cholesterol levels were<100mg/dL. The HDL cholesterol levels were appropriate in 60.7% of the patients (HDL>50mg/dL in women and >40mg/dL in men), and 66.6% of the patients maintained triglycerides levels <150mg/dL. LDL cholesterol levels were <70mg/dL in 34% of the patients with macrovascular disease and in 24.1% of the patients with advanced renal failure.

Some 58.4% of the population had macrovascular and/or microvascular complications (Table 4). Some 28.6% of the patients had at least one macrovascular disease, and 45.7% had some microvascular complication.

Ninety-one percent of the patients followed a drug therapy: 25.5% with insulin and 82% with noninsulin therapy. Seventeen percent of the patients used insulin along with oral diabetes drugs. Sixteen percent of the patients took sulfonylureas, the most common of which was gliclazide (54% of the sulfonylureas). Twenty-eight percent of the patients were treated with dipeptidyl-peptidase-4 inhibitors (DPP4i). This group of inhibitors included sitagliptin (46.5%), vildagliptin (30%) and linagliptin (19%). The most widely used oral diabetes drugs (or combinations) were metformin (47%), metformin plus DPP-4i (19%), metformin plus sulfonylureas (9), DPP4i (6%) and DPP4i plus metformin plus sulfonylureas (5%).

Some 12.5% of the patients had hypoglycemic episodes recorded in their medical history. Eighteen percent of these episodes were considered severe, because they required the assistance of another individual. Eighty-six percent of the severe hypoglycemic events occurred in patients treated with insulin versus 14% that occurred in patients treated with secretagogues (sulfonylureas or repaglinide). There was a higher percentage of insulin and secretagogue use among patients who had greater comorbidity and dependency. These drugs were more often used by the patients with poorer metabolic control (67% of the patients with an HbA1c level >7% were treated with insulin and/or secretagogues vs. 25% of the patients with an HbA1c level <7%) (Table 5).

The management of patients with DM2 requires an appropriate individual assessment to establish the control objectives and the best therapeutic strategy. Age, life expectancy, functional capacity, comorbidity, disease progression and patient collaboration are the pillars for basing the clinical decisions and establishing the objectives of glycemic control.19

For elderly patients with DM2, it is especially important to prevent hypoglycemic episodes and adverse drug effects, which can worsen the quality of life and cause falls, accidents and cognitive impairment. The strict control of metabolic parameters in these patients is less likely to provide long-term benefits,5,6,8 especially in the most elderly. Our study highlights the low HbA1c concentrations of elderly patients with DM2.

The mean HbA1c was 7%, despite the fact that these patients had a mean disease progression time of more than 10 years (11.6±7.9). Seventy-five percent of the patients had HbA1c levels <7.5%. These readings are indicative of the therapeutic intensity and could explain the reasons behind the hypoglycemic events recorded in the medical history (one of every 8 patients had experienced at least 1 episode of hypoglycemia). Of these, 18% required external assistance for their recovery.

The data on glycemic control in this study are similar to those of other studies conducted in our setting20–22 or other countries. An analysis of 1288 patients older than 65 years23 from a United States health survey (NHANES) indicated the possible overtreatment of elderly patients, given that 61.5% of the study sample had HbA1c levels <7%, and 41.9% had levels <6.5%. Another health survey performed in Germany found that 65% of patients with diabetes had HbA1c levels <7%.24

The results showed poorer glycemic control when the elderly patient with diabetes had a higher degree of comorbidity, although the difference in HbA1c levels between the groups with greater and lesser comorbidity was only 0.36%, a difference of little clinical relevance. Poorer blood glycemic control was also observed as the degree of dependence increased, except in cases of total dependence, which presented mean levels similar to the elderly patients with no degree of dependence. Strict glycemic control for completely dependent elderly patients with diabetes could be associated with the intensification of insulin therapy.

A study performed in France25 on patients older than 70 years with diabetes investigated the relationship between metabolic control and morbidity and mortality. The study evaluated 3 cohorts according to the treatment employed: insulin alone, insulin plus oral diabetes drugs and oral diabetes drugs alone and/or glucagon-like peptide-1 analogs. A high rate of hypoglycemia (53%) was observed in the first group, who also had poorer glycemic control and higher rates of retinopathy, kidney disease, cardiovascular impairment and geriatric complications. The patients treated exclusively with oral diabetes drugs or glucagon-like peptide-1 analogs had a very similar disease progression, characteristics and comorbidities to those of our sample and had similar mean HbA1c levels (7.1%), with 75% of patients with HbA1c levels <7.5%.

As was expected, the elderly patients in our sample had a high number of cardiovascular risk factors and vascular complications. Arterial hypertension, present in 78% of patients, remained the most difficult-to-control factor. Approximately one third of the elderly patients with DM2 had macrovascular disease. Like other Spanish studies,20,21,26 the main complication was ischemic heart disease (13.9%), followed by cerebrovascular disease (11.9%) and peripheral arterial disease (9%). However, the prevalence of macrovascular disease was much lower compared with that of other Western countries such as the United States,23,27 Germany,28,29 Scotland30 and France,25 even in younger cohorts.

Microvascular complications affected 45.7% of the patients, the most prevalent of which was chronic kidney disease. The high prevalence of chronic kidney disease should affect the drug therapies employed and be the subject of more studies.

In terms of drug therapy, metformin was the most frequently used drug, followed by DPP4i. The use of insulin was not residual, given that it was used in 25% of cases. Understanding how to manage its use is therefore important. Insulin and secretagogues had the closest relationship with hypoglycemia, and their use was more frequent among the patients with poorer metabolic control. It is also striking that glibenclamide still represented 7% of the sulfonylureas employed, despite its long half-life and capacity for causing hypoglycemia.

This study has limitations that are inherent in cross-sectional studies. The data were extracted from the medical history, and some of the results, such as the prevalence of hypoglycemia and geriatric syndromes, are probably underestimated. However, the size of the study sample and patient selection method (with researchers distributed throughout Spain, using a series of random numbers to select patients) leads us to believe that the results are representative of the current situation in Spain.

Mean HbA1c levels in elderly patients with DM2 analyzed in Spain are below those recommended by the main clinical practice guidelines. These values are higher in the patients with a higher level of comorbidity and were influenced by the patient's degree of disability.

This study was made possible thanks to financial support by Boehringer-Ingelheim. The financial sources had no role in designing the study, collecting the data or interpreting the data. The financial sources also did not participate in the preparation, review or approval of the manuscript.

The authors declare that they have no conflicts of interest.