Ten-Year Prospective Follow-Up of Institutionalized Patients With Diabetes in a Long-Term Care Home

This was a 10-year monocentric, prospective, observational study carried out in an LTCH serving 435 residents in Saint Just Saint Rambert, France. We recorded glycemic measurements of all patients with diabetes living in the LTCH in November 2009, as well as those of newly institutionalized residents until January 2019. In total, 321 residents with diabetes were seen during the 10-year study period. A diabetes specialist trained at the Medical School of Saint Etienne in France regularly evaluated patients’ glucose records and A1C levels and adjusted treatment plans as needed. We aimed to assess the effectiveness of this approach in lowering A1C in patients with an initial A1C >8% without inducing hypoglycemia.

Lipid balance in overweight individuals and those treated for hyperlipidemia and the presence of foot ulcers were also recorded (Table 1). Geriatric end points included results of a dependence evaluation, malnutrition based on an albumin assay, cognitive disorders identified through the Mini Mental Status Evaluation (MMSE), and the results of calcium, vitamin D, liver function, and thyroid-stimulating hormone (TSH) assays.

Glycemic control monitoring was based on A1C measurements performed every 3 months, as well as monthly evaluation of capillary glucose levels for insulin-treated residents, performed by LTCH nurses either daily or three times per week for residents with stable glycemia. Capillary glucose was measured with a Dinno Lecteur S+ (Dinno Santé Laboratory). Quality control was based on regular calibration according to the manufacturer’s recommendations.

Nurses delivered insulin injections at doses prescribed by the diabetes specialist. In case of hypoglycemia, nurses spontaneously reduced the insulin dose by 4 units without seeking immediate consultation with the doctor. The reduction in insulin dose was indicated in residents’ treatment upon their entry. Assessment results were expressed as mean ± SD (range). A t test was used for statistical analysis, with P <0.05 considered statistically significant.

At LTCH entry, the average age was 79 years (range 45–98) for all patients with diabetes, 81 years (range 49–98) for women, and 79 years (range 45–93) for men. The mean overall BMI was 27.7 kg/m² (range 13.4–43.7), with mean BMI slightly higher for women and slightly lower for men (28.3 and 26.8 kg/m², respectively), and 66% of the patients were overweight (BMI >25 kg/m²).

Regarding drug treatments at entry, apart from OADs detailed separately, 287 residents (89.4%) received cardiovascular drugs, 272 (84.7%) received psychiatric or neurological treatments, 234 (72.9%) received digestive treatments, 172 (53.6%) took analgesics, 281 (87.5%) had endocrine treatments, and 14 (4.4%) took antibiotics. All residents were taking some form of medication. Thirty-three (10.3%) took one to four different drugs, 155 (48.3%) took five to nine drugs, and 133 (41.4%) took 10 or more drugs.

At entrance, renal failure was present in 135 residents (42%) with a creatinine clearance of 30–60 mL/min and in 29 residents (9%) with a creatinine clearance <30 mL/min (Figure 1). Foot ulcers were seen in 43 residents (19 men and 24 women), varicose ulcers in eight residents (four men and four women), superficial skin wounds in 14 residents (six men and eight women), bedsores in 17 residents (five men and 12 women), and four wounds on recent amputations in four men (one midthigh amputation, one subknee amputation, and two forefoot amputations, including one bilateral). Total cholesterol was >5.2 mmol/L (201.0 mg/dL) in 57 residents (17.7%), 16 residents (5%) had LDL cholesterol >4.1 mmol/L (158.5 mg/dL), 50 residents (15.6%) had HDL cholesterol >1 mmol/L (38.6 mg/dL), and 60 residents (18.7%) had triglycerides >1.7 mmol/L (150.4 mg/dL).

During follow-up, 15 residents left the LTCH to either return to their own home or to integrate into another LTCH. Over 10 years, 196 residents (61%) died with a clear difference according to age-groups: 15 residents (4.6%) <65 years of age, 27 residents (8.4%) 65–75 years of age, 77 residents (24%) 75–85 years of age, and 77 residents (24%) >85 years. The death rate was 71.6% (n = 58) in patients with an A1C >8% (64 mmol/mol) compared with 57.6% (n = 137) for patients with an A1C <8% (64 mmol/mol).

Many residents had cognitive impairment at LTCH entry, with MMSE scores <28/30 in 298 residents (92.8%). Severe dementia, with MMSE scores <13/30 or infeasible, was found in 137 residents (42.7%). The most severe dementia scores occurred in 16 subjects <65 years of age (47%) and 49 residents 75–85 years of age (42%). Protection measures, whether guardianship or curatorship, were taken for 154 residents (49%), with 98 measures taken in women and 56 in men.

Dependence was evaluated with the Autonomie Gérontologique Groupes Iso Ressources (AGGIR; Gerontological Autonomy Iso Groups Resources) grid, which is a tool used in France to assess the degree of autonomy or loss of autonomy mainly in elderly people, whether they live in an institution or at home. AGGIR scores 1 and 2 correspond to totally dependence. AGGIR scores 3 and 4 indicate a loss of autonomy, such that the person may have reduced mobility or impaired mental abilities and may need occasional help with daily activities. Elderly people with AGGIR scores 5 and 6 are considered autonomous. No residents of the LTCH were rated AGGIR 5 or 6. An AGGIR score of 4 was found for 73 residents (22.7%), 74 residents (23%) were rated AGGIR 3, 128 residents (39.9%) were rated AGGIR 2, and 46 residents (14%) were rated AGGIR 1 (Table 2).

Treatment of diabetes at entry was an OAD alone in 75 patients (23%), combined OAD and insulin therapy in 10 patients (3%), or insulin therapy alone in 132 patients (41%) (Figure 2). No resident at entry received a glinide, glucagon-like peptide 1 receptor agonist, or dipeptidyl peptidase 4 inhibitor. Diabetes was diagnosed on initial assessment in eight patients at LTCH entry. FBG levels were <3.3 mmol/L (60 mg/dL) on initial assessment in three residents; 59 residents had an FBG level between 3.3 and 5.5 mmol/L (60 and 100 mg/dL); and 28 residents had an FBG level >11 mmol/L (200 mg/dL). Diabetes developed during follow-up in 104 residents (32.4%) and was detected through screening at annual biological assessments in 33 men and 71 women. Conversely, OAD treatments had to be interrupted in 31 residents because of recurrent hypoglycemia and/or an A1C <6% (42 mmol/mol).

The average A1C at entry was 7.4% (57 mmol/mol), with no significant difference between men and women, but with a difference among different age-groups (Table 3). Initial A1C was <7.0% (53 mmol/mol) in 138 residents (43.0%), between 7.0 and 8.0% (53 and 64 mmol/mol) in 100 residents (31.1%), between 8.0 and 9.0% (64 and 75 mmol/mol) in 50 residents (15.6%), between 9.0 and 10.0% (75 and 86 mmol/mol) in 18 residents (5.6%), and >10.0% (86 mmol/mol) in 13 residents (4.0%) (Figure 3). Monitoring of A1C values showed a significant difference (P <0.0001) between initial values and values at 1-year follow-up but no significant differences thereafter in the total group or among only women or only men (Figure 4). There was no significant difference in A1C between men and women throughout follow-up (Table 3).

When considered by age-group, among people <65 years of age, there was a significant difference between initial and 3-month A1C values but no difference in subsequent A1C values except at the 2-year assessment (P = 0.042). Among those 65–75 years of age, there was a persistent significant difference during the first year of follow-up and at the 4-year evaluation. Identical results were obtained for those who were 75–85 years of age. For those >85 years of age, there was a significant difference only between initial A1C and A1C values at the 3- and 9-month assessments (Table 4 and Figure 5).

Insulin therapy was prescribed either before LTCH entry or thereafter when diabetes was diagnosed if an OAD was not possible, in very old subjects, or during follow-up as patients’ diabetes progressed and they became insulin- requiring. Daily insulin doses were <20 units/day in 57 residents (33%), 20–60 units/day in 93 residents (54%), 60–100 units/day in 13 residents (8%), and >100 units/day in 9 residents (5%). A total of 172 residents (53.6%) received insulin therapy during follow-up.

Adjustments in treatment plans were made based on A1C results for residents treated with OADs and on A1C and capillary glucose values for residents treated with insulin. The therapeutic objective was to obtain an A1C level <7.0% (53 mmol/mol) for patients who were still autonomous (as in the younger population) or an A1C <8.0% (64 mmol/mol) for patients who were considered undertreated.

It is important to note that there were very few symptomatic hypoglycemia episodes. Most instances of hypoglycemia were detected during routine daily measurements of capillary glucose performed by nurses, most often before breakfast for residents with only one daily insulin injection and in the morning and evening for those taking two daily insulin injections. For capillary glucose levels <5.5 mmol/L (100 mg/dL), nurses reduced insulin dose by 4 units to prevent a recurrence of hypoglycemia and to reassure the health care team (e.g., nurses fearing repeated episodes of particularly harmful hypoglycemia in these elderly patients).

Many residents had cognitive impairments that likely prevented them from alerting nurses when they experienced symptoms of hypoglycemia. Some residents were highly dependent physically and unable to use their room bell to call for a nurse. Moreover, hypoglycemia symptoms are less typical in older patients, making hypoglycemia more difficult to identify by the health care team. In contrast, the development of diabetes discovery revealed by an episode of ketoacidosis occurred in only one resident of the 104 diagnosed during follow-up.

The number of residents in each age-group decreased significantly over time, and few residents exceeded >5 years of follow-up.

Comparisons between men and women showed significant differences in average age (79 vs. 81 years, respectively, P = 0.0001), weight (74.7 vs. 68.5 kg, respectively, P = 0.002), creatinine level (99 vs. 79 µmol/L [1.11 vs. 0.88 mg/dL], respectively, P = 0.001), calcium level (2.30 vs. 2.33 mmol/L [9.21 vs. 9.33 mg/dL], respectively, P = 0.04), and albumin level (33.9 vs. 32.6 g/L [3.39 vs. 3.26 g/dL, P = 0.03]). There were no significant differences in values for urea, vitamin D, TSH, and lipid parameters (Table 1).

By age-groups, there was a significant difference in weight between younger residents <65 years of age and those who were 65–75 years of age (73.5 vs. 78.2 kg, respectively, P = 0.04) and between those who were 75–85 years of age and those >85 years of age (70.7 vs. 66.6 kg, respectively, P = 0.03). There was also a significant difference in albumin values between residents <65 years of age and those aged 65–75 years (36.6 vs. 34.1 g/L [3.66 vs. 3.41 g/dL], respectively, P = 0.01) and between those 65–75 years of age and those 75–85 years of age (34.1 vs. 32.5 g/L [3.40 vs. 3.25 g/dL], respectively, P = 0.05). There were no other statistically significant differences in other parameters (Table 2).

The prevalence of type 2 diabetes is increasing. Recent projections estimate that there will be >500 million people with diabetes in 2030 and 700 million in 2045 (2). A subgroup of this population will experience a loss of autonomy, possibly associated with cognitive impairment, that will necessitate LTCH entry (3). Elderly patients have well-known risk factors associated with aging, which increases their fragility. Available OADs are the same as for younger patients (12); however, their use requires consideration of common comorbidities and treatment side effects, with the potential increased risk of hypoglycemia particularly deleterious in this population (20,21). In addition, cognitive disorders and geriatric syndromes, including malnutrition, sarcopenia, falls and fractures, incontinence, and depression are frequently present. Treatment plans must consider all of these factors to successfully improve patients’ quality of life. These considerations apply not only to pharmacologic therapy choices, but also to the number of daily capillary glucose checks, which often can be excessive among LTCH residents (22–24). Individuals in this population face one of two main risks: either excessive treatment with an increased risk of hypoglycemia (14) or inadequate treatment leading to the progression of degenerative complications and excess mortality (11,25). The risks of hypoglycemia and microvascular complications increase significantly with diabetes duration. In addition, the risk of dementia increases with the number of hypoglycemia episodes in elderly individuals (26).

Are residents of institutional settings with diabetes treated excessively? A literature review highlights great disparities in findings of studies depending on the definitions used. In LTCHs, 16–74% of residents may be overtreated, while 1.4–14.8% of residents could be undertreated (27). In one-third of these patients, A1C is <6.5% (48 mmol/mol), increasing the risk of hypoglycemia. OADs must be reduced for individuals in this population who have an A1C <7.5% (58 mmol/mol) (14). This fragile population also has prevalence of comorbidities, with an average of four comorbid conditions (15) (mainly high blood pressure, cardiovascular pathologies, and depression [28] in addition to diabetes).

Recommendations regarding diabetes management are the same as for younger subjects (12). However, glycemic goals must be individualized in patients with cognitive impairments, reduced autonomy, lower life expectancy, and an increased number of comorbidities. The risk of hypoglycemia also increases in treated patients, so therapy must be intensified in a stepwise manner, starting with a single OAD, then OAD combinations, and then insulin. It is also possible to use new OAD classes with lower associated risks of hypoglycemia and a possibly beneficial effects on weight (23).

Sinclair et al., in a 2012 position statement of the International Association of Gerontology and Geriatrics, the European Diabetes Working Party for Older People, and the International Task Force of Experts in Diabetes, made recommendations for diabetes management in the elderly that were not specific to those in institutional settings (29). They focused on quality-of-life management and confirmed the value of hyperlipidemia treatment with statins. For glycemic management, they recommended an A1C target of 7.5–8.0% (58–64 mmol/mol) for elderly individuals with few comorbidities and a target of 8.0–9.0% (64–75 mmol/mol) for those with multiple comorbidities or short life expectancy. A1C targets also have been proposed to guide prescribers in considering criteria of fragility (30,31).

Metformin is the first-line treatment for diabetes except where contraindicated, and especially in cases of renal insufficiency with a creatinine clearance <30 mL/min. For the current study, MDRD (Modification of Diet in Renal Disease) was not measured in all patients, which would have been a more reliable assessment of renal insufficiency. Finally, there is a low risk of lactic acidosis with metformin use. Insulin therapy is used next instead of other OADs, which are often poorly tolerated in elderly patients with severe renal insufficiency.

Among institutionalized patients, those with diabetes have more comorbidities than those without diabetes (16). They also have higher use of prescription medications than their counterparts without diabetes, including ACE inhibitors (46 vs. 24%), diuretics (44 vs. 34%), statins (40 vs. 18%), and antiplatelet drugs (37 vs. 18%), according to a U.S. study (7). This fragile population often experiences polypharmacy, with most patients receiving more than five drugs (17), and this increases the risk of errors and iatrogenic complications (18). Treatment objectives must be individualized based on life expectancy and cognitive status to avoid both hypoglycemia and hyperglycemia-related complications (32). Recent recommendations suggest appropriate minimum A1C targets for this fragile population (31). Reducing OAD treatments and lipid-lowering medications also may be appropriate (12).

Institutionalized patients with diabetes comprise a fragile population with many comorbidities. Treatment with OADs must be individualized as much as possible. It seems possible to improve glycemia in patients with an A1C >8% (64 mmol/mol) without causing an increase in hypoglycemia with the presence of and monitoring by nurses in LTCHs. Nurses can check daily capillary glucose levels in case of unusual behavior. Nurses’ training is also essential to detect hypoglycemia and adjust insulin doses by spontaneously lowering doses without first consulting with the doctor. This level of autonomy for LTCH nurses can reduce the stress induced by hypoglycemia episodes and enhance nurses’ sense of being valued team members. Moreover, nurses’ autonomy can improve the quality of life of these fragile patients who have many comorbidities and reduce hospitalizations (33).

In our study, we were surprised by the low number of hypoglycemic events. It is essential to verify that no increase in hypoglycemia occurs with glycemic treatment in this population. In older people, hypoglycemia symptoms are not typical, and LTCH residents may be too dependent to alert nurses when hypoglycemia occurs. Increasing the number of daily capillary glucose checks or implementing continuous glucose monitoring could be considered to improve monitoring in at-risk patients. It also could be interesting to carry out this type of study in several LTCHs with different specialists to verify whether our results are reproducible without increasing hypoglycemia in fragile patients. If so, it could be possible to reduce the number of daily blood glucose checks and thereby improve residents’ quality of life.

These results were obtained through monthly review of capillary glucose levels for insulin-treated residents and regular A1C measurements for all residents with diabetes. Routine monitoring, coupled with close medical follow-up by a diabetes specialist, optimized patients’ outcomes in this LTCH.

No potential conflicts of interest relevant to this article were reported.

As the sole researcher and author, N.V. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.