Diabetes mellitus (DM) has been described as a 21st-century epidemic in developed countries, and one of the most important factors in this regard is an aging population. In Spain, 17% of the population (>7.8 million people) are currently age 65 and older, and this figure is expected to reach 32% by 2050.[1, 2] The prevalence of DM has been estimated at 8% in adults and is thought to affect approximately one-third of elderly people (46% of them undiagnosed), which suggests a substantial burden on the individual and the healthcare system.[3, 4]
In primary prevention, the power of classic risk factors to predict the risk of cardiovascular disease accurately seems to diminish with advancing age.[5, 6] Observational studies in the oldest old (≥85) have shown that some of these risk factors, such as hypertension and dyslipidemia, become nebulous or even act in the reverse direction. Furthermore, it has been argued that, at this age, new biomarkers, namely C-reactive protein, folic acid, interleukin-6, and various lipoproteins, are better at predicting cardiovascular disease than are classic risk factors.
Although vascular complications have been well documented in adults, there remains a lack of clear data from the oldest populations. Elderly adults in studies of DM are heterogeneous, including individuals aged 60 to 80, who have a higher mortality risk, and individuals aged 80 and older, who have been shown to have a lower rate of complications. Moreover, improvements in clinical management in the general population may have reduced the incidence of vascular complications in recent decades. Nevertheless, guideline committees worldwide concur that the care of older adults with DM is complicated[4, 9] because of their clinical symptoms, functional differences, cognitive disorders, tendency to fall, and other increasingly common geriatric syndromes, all of which can affect the management of individuals with DM, their families, and the care system.[6, 7]
The present study was designed to test the hypothesis that DM is part of a multipathology that includes vascular diseases and other less frequently explored geriatric morbidities. Using a multidisciplinary approach, the main aim of the present study was to describe the rate of DM (previously diagnosed and unknown) in community-dwelling oldest-old participants. A secondary aim was to explore the effect of cardiovascular risk factors, geriatric syndromes and other morbidities on the oldest old individuals with and without DM.
- Top of page
The current data were gathered within the framework of the OCTABAIX study, a prospective population-based study of 328 community-dwelling elderly adults, all of whom were Spanish Caucasians born in 1924 (85 years old at the time of inclusion, in 2009) and who were registered with one of seven primary healthcare centers in the geographical area of Baix Llobregat (Barcelona, Spain). The combined population served includes approximately 210,000 inhabitants, and the referral hospital is the Hospital Universitari de Bellvitge. From the total of 696 potential participants, 487 were considered eligible after analysis of the data. Reasons for noninclusion were living in unregistered nursing homes (76, 10.9%), died before the study started (67, 9.6%), and had moved to another city (66, 9.4%). The participation rate was 67.5%, and there were no differences between respondents and nonresponders in terms of sex, healthcare center, or physician in charge. The sampling procedures and data collection methods have been described elsewhere. Participants were examined in their place of residence or at the primary healthcare center if they were ambulatory, in both cases by trained teams (doctors and nurses) with extensive experience in working with elderly adults. An informant interview was conducted in cases in which the participant was unable to participate fully because of compromised health or cognitive ability. The institutional ethics committee of the Jordi Gol Institute for Primary Care Research approved the study. All participants, or their caregivers in the case of cognitively impaired participants, gave written informed consent.
The assessment included sensory status (near vision measured using the Jaeger test and hearing ability measured using the whisper test), alcohol and tobacco intake, hypertension (blood pressure ≥140/90 mmHg), dyslipidemia, and number of comorbidities (only those with a prevalence of ≥5%). These comorbidities were ischemic cardiopathy, stroke diagnosis, vasculopathy, heart failure, atrial fibrillation, dementia, cancer, anemia, and thyroid disease. Functional status was measured using the Barthel Index (BI) for basic activities of daily living (ADLs) and the Lawton Index (LI) for instrumental ADLs. The BI measures performance in seven areas: feeding, bathing, personal cleanliness, dressing, bowel and bladder control, getting on and off the toilet, and locomotion. A maximum score of 100 indicates that the individual is independent. The LI score ranges from 0 (low function, dependent) to 8 (high function, independent). Cognitive status was evaluated using the Spanish version of the Mini-Mental State Examination (MEC), which yields a score up to 35 (a score of ≤23 indicates cognitive impairment). The Braden Scale for Predicting Pressure Sore Risk was also administered. Nutritional status was assessed using the Mini Nutritional Assessment (MNA), which has a maximum score of 30; anyone scoring less than 23.5 is classified as being at risk of malnutrition. Social risk was measured using the Gijon Test, which has a maximum score of 24 points (social risk >10). Quality of life was assessed using the visual analogue scale (EQ-VAS) of perceived health from the EuroQol-5D test of quality of life (on which 0 indicates the worst state of health and 100 the best). The total number of chronic drug prescriptions and blood analysis results were recorded. Tests performed on fasting venous blood were total leukocytes, serum glucose and creatinine, estimated glomerular filtration rate (eGFR) using the Modification of Diet in Renal Disease formula (MDRD) in mL/min per 1.73 m2, total cholesterol, high-density lipoprotein cholesterol (HDL-C) with low baseline HDL-C as less than 1.0 mmol/L in men and 1.2 in women, low-density lipoprotein cholesterol (LDL-C) with high baseline as LDL-C as 2.6 mmol/L or more, vitamin B12 (usual range: 145–637 pmol/L), folic acid (usual range: 5.7–32.0 nmol/L), and thyrotropin (usual range: 0.25–5 mU/L). The dependent variable was the presence of DM. Participants were defined as having DM according to self-report, physician diagnosis, use of antidiabetic agents, or plasma glucose concentration (≥7 mm/L) at baseline. The recent recommendation to include glycosylated hemoglobin (HbA1c) in the diagnosis of DM was not considered because this study was planned, and the data compiled, before these recommendations were published. In this study, glycemia control was considered to be HbA1c less than 7%, in accordance with American Diabetes Association (ADA) treatment goals and other studies of elderly adults with DM.[4, 11] Individuals newly diagnosed with DM were defined as those with a glucose concentration of 7 mm/L or more on baseline blood tests but without any previous self-report, physician diagnosis, or prescription of antidiabetic agents. These participants were assigned to primary healthcare control, and no confirmation of their new diagnosis was obtained in this cross-sectional study.
Categorical variables are shown as frequencies and percentages, and continuous variables are presented as means and standard deviations (SDs) or medians and interquartile ranges (IQRs). In the bivariate analysis, the chi-square or Fisher exact test was used to compare categorical variables. The Student t-test or the Mann–Whitney U-test, depending on the normality of the variable, was applied to compare continuous variables. A potential association between metformin treatment and vitamin B12 deficiency was also explored. Logistic regression analysis was performed to estimate unadjusted and adjusted odds ratio (ORs) with 95% confidence intervals (CIs). The unadjusted models were developed using the presence of DM as a dependent variable and the variables with P < .20 in the bivariate analysis as independent variables one at time. The adjusted model was adjusted for sex, hypertension, ischemic cardiopathy, thyroid disease, Braden scale, Barthel index, MEC, index drugs taken, white blood cell count, total cholesterol, HDL-C, LDL-C, and folic acid. In the adjusted model, all variables were introduced at the same time. The final adjusted model was developed using the backward stepwise method. Colinerearity among the variables was checked using the generalized variance-inflation factor method. The colinearity threshold was defined at the value of 5; lower values indicate no colinearity. The results were considered statistically significant at P < .05.
- Top of page
Eighty-five (25.9%) participants had DM. It had been previously diagnosed in 79, whereas a new diagnosis was made in the remaining six (7.1%). The median duration of DM was 5.8 years (IQR: 3.5–10.6), with 62 (72.9%) participants having had the disease for less than 10 years. Median HbA1c was 6.4% (IQR: 5.60–7.38). Sixty-eight participants used oral antidiabetic agents, and 45.9% of these took metformin and 20.0% sulfonylureas. In the group with DM, 21.2% were receiving insulin therapy, and 16.6% were untreated. Table 1 summarizes the sociodemographic and comorbidity characteristics of the two groups (participants with and without DM). There were no significant differences in relation to sociodemographic variables, but individuals with DM had a higher rate of hypertension (P = .001), ischemic cardiopathy (P = .02), and number of chronic prescriptions (P < .001). Conversely, they had poorer functional status according to the BI (P = .06) and a lower pressure sore score on the Braden scale (P = .002).
Table 1. Sociodemographic and Comorbidity Characteristics According to Presence or Absence of Diabetes Mellitus (DM) in 85-Year-Old Participants
|Characteristic||DM (n = 85)||No DM (n = 243)||P-Value|
|Female, n (%)||53 (62.4)||149 (61.3)||.97|
|Married, n (%)||30 (35.3)||104 (42.8)||.45|
|Single||5 (5.9)||15 (6.2)|
|Widowed||50 (58.8)||124 (51.0)|
|Education <6 years, n (%)||75 (88.3)||191(78.6)||.19|
|Impaired auditory acuity, n (%)||37 (43.5)||87 (35.8)||.26|
|Visual acuity, median (IQR)a||5.00 (3.00–10.0)||5.00 (3.00–10.0)||.45|
|Alcohol intake ≥1 drinks/d, n (%)||22 (25.8)||53 (21.9)||.69|
|Tobacco, n (%)||2 (2.3)||8 (3.3)||.66|
|Hypertension, n (%)||76 (89.4)||173 (71.2)||.001|
|Dyslipidemia, n (%)||43 (50.6)||125 (51.4)||.99|
|Ischemic cardiopathy, n (%)||10 (11.8)||10 (4.1)||.02|
|Previous stroke, n (%)||34 (14.0)||15 (17.6)||.52|
|Vasculopathy, n (%)||6 (7.06)||10 (4.1)||.38|
|Heart failure, n (%)||13 (15.3)||29 (11.9)||.54|
|Atrial fibrillation, n (%)||13 (15.3)||28 (11.5)||.47|
|Dementia, n (%)||7 (8.2)||24 (9.9)||.82|
|Cancer, n (%)||10 (11.8)||33 (13.6)||.81|
|Anemia, n (%)||15 (17.6)||41 (16.9)||.>99|
|Thyroid disease, n (%)||3 (3.6)||21 (8.6)||.19|
|Comorbidities, median (IQR)||1 (0.00–2.00)||1 (0.00–2.00)||.47|
|Gijon test, median (IQR)||9 (8.00–11.0)||10 (8.00–11.0)||.74|
|Braden Scale, median (IQR)||20 (19.0–21.0)||21 (20.0–23.0)||.002|
|Mini Nutritional Assessment score, median (IQR)||25 (22.5–27.0)||25 (23.0–27.5)||.49|
|Barthel Index, median (IQR)||95 (80.0–100)||95 (85.0–100)||.06|
|Lawton Index, median (IQR)||5 (3.00–8.00)||6 (4.00–8.00)||.13|
|Spanish version of the Mini-Mental State Examination, median (IQR)||27 (22.0–31.0)||29 (23.0–32.0)||.13|
|EuroQol-5D visual analogue scale, median (IQR)||60 (50.0–75.0)||60 (50.0–75.0)||.63|
Table 2 shows the differences in the analytical results and treatment data. Individuals with DM had a significantly higher white blood cell count (P < .001), folic acid level (P = .002), and number of chronically prescribed angiotensin-converting enzyme inhibitors (P = .001), angiotensin receptor antagonists (P < .001), and calcium channel blockers (P = .04). They also had lower total cholesterol (P = .01), HDL-C (P = .03) and LDL-C (P = .005) levels. A secondary analysis revealed a statistical association between DM and metformin treatment and low vitamin B12 level (P < .001), although metformin treatment was not statistically associated with folic acid level.
Table 2. Analytical and Chronic Treatment Parameters According to Presence or Absence of Diabetes Mellitus (DM) in 85-Year-Old Participants
|Parameter||DM (n = 85)||No DM (n = 243)||P-Value|
|Estimated glomerular filtration rate, mL/min per 1.73m2 median (IQR)|| 61.0 (53.1–61.0)|| 61.0 (52.0–61.0)||.46|
|Total cholesterol, mmol/L, mean ± SD|| 4.8 ± 0.91|| 5.1 ± 0.95||.01|
|High-density lipoprotein cholesterol, mmol/L, median (IQR)|| 1.3 (1.16–1.64)|| 1.4 (1.19–1.73)||.03|
|Low-density lipoprotein cholesterol, mmol/L, mean ± SD|| 2.8 ± 0.78|| 3.0 ± 0.81||.005|
|White-cell count, 1,000/mL, mean ± SD|| 7.4 ± 1.93|| 6.6 ± 1.79||<.001|
|Folic acid, median (IQR; usual range: 5.7–32.0 nmol/L)|| 22.4 (18.2–29.2)|| 19.1 (14.3–26.8)||.002|
|Vitamin B12, median (IQR; normal range: 145–637 pmol/L)||269 (229–383)||278 (219–370)||.95|
|Thyrotropin, median (IQR)|| 2.2 (1.33–3.68)|| 2.1 (1.37–3.08)||.59|
|Drugs taken, median (IQR)|| 8.0 (6.00–9.00)|| 5.0 (3.00–8.00)||<.001|
|Angiotensin-converting enzyme inhibitor, n (%)|| 45 (52.9)|| 79 (32.5)||.001|
|Angiotensin receptor antagonist, n (%)|| 28 (32.9)|| 35 (14.4)||<.001|
|Beta-blocker, n (%)|| 16 (18.8)|| 25 (10.3)||.06|
|Calcium channel blocker, n (%)|| 28 (32.9)|| 51 (21.0)||.04|
|Statin, n (%)|| 37 (43.5)|| 78 (32.1)||.08|
|Antiplatelet, n (%)|| 38 (44.7)|| 84 (34.6)||.12|
|Diuretic, n (%)|| 40 (47.1)||103 (42.4)||.53|
|Metformin, n (%)|| 39 (45.9)|| 0 (0.0)||<.001|
|Sulfonylurea, n (%)|| 17 (20.0)|| 0 (0.0)||<.001|
|Insulin, n (%)|| 18 (21.2)|| 0 (0.0)||<.001|
Finally, Table 3 shows the results of the logistic regression. This analysis indicated that the factors significantly associated with DM were BI (OR = 1.03, 95% CI = 1.01–1.05, P = .007), number of chronic drug prescriptions (OR = 1.28, 95% CI = 1.15–1.42, P < .001), white blood cell count (OR = 1.34, 95% CI = 1.15–1.56, P < .001), folic acid level (OR = 1.04, 95% CI = 1.01–1.07, P = .02), Braden risk score (OR = 0.87, 95% CI = 0.79–0.97, P = .01), thyroid disease (OR = 0.23, 95% CI = 0.06–0.92, P = .04) and LDL-C (OR = 0.63, 95% CI = 0.43–0.92, P = .02). No colinearity was found in the final model because all variables had a generalized variance-inflation factor lower than 5.
Table 3. Logistic Regression Results According to the Presence or Absence of Diabetes Mellitus in 85-Year-Old Participants
|Factor||Odds Ratio (95% Confidence Interval) P-Value|
|Unadjusted Analysis||Adjusted Analysisa|
|Female||1.04 (0.63–1.74) .87||–|
|Hypertension||0.29 (0.14–0.62) .001||–|
|Ischemic cardiopathy||3.11 (1.25–7.75) .01||–|
|Thyroid disease||2.59 (0.75–8.90) .13||0.23 (0.06–0.92) .04|
|Braden Scale||0.92 (0.85–1.00) .047||0.87 (0.79–0.97) .01|
|Barthel Index||1.00 (0.99–1.01) .79||1.03 (1.01–1.05) .007|
|Spanish version of the Mini-Mental State Examination||0.98 (0.94–1.01) .18||–|
|Drugs taken||1.22 (1.12–1.32) <.001||1.28 (1.15–1.42) <.001|
|White-cell count||1.28 (1.11–1.46) <.001||1.34 (1.15–1.56) <.001|
|Total cholesterol||0.70 (0.53–0.93) .01||–|
|High-density lipoprotein cholesterol||0.48 (0.24–0.95) .03||–|
|Low-density lipoprotein cholesterol||0.64 (0.46–0.88) .006||0.63 (0.43–0.92) .02|
|Folic acid||1.03 (1.01–1.06) .007||1.04 (1.01–1.07) .005|
- Top of page
This study has documented a higher DM prevalence (26%) than reported in other studies, such as the Newcastle study group (14%) and the Leiden octogenarian group (16%), but similar to the rate found in another study (23%) in institutionalized 85-year-olds. There are a number of potential reasons for this. One is the extensive diagnostic criteria used in the present study (self-report, fasting blood glucose levels, medical index diagnoses, and antidiabetic prescriptions). Another is the absence of exclusion criteria based on health or morbidities, the purpose of which was to avoid the systematic exclusion of elderly people with poor health status (dementia, cancer, heart failure). Alternatively, the rate obtained may reflect a real increase in the incidence of DM in this population group. The proportion (7%) of participants newly diagnosed with DM (those with a plasma glucose concentration ≥7 mm/L in the baseline blood test but without any previous self-report or physician diagnosis of DM or antidiabetic prescription) was lower than might be expected. Previous studies have reported rates of newly diagnosed DM of 11% in octogenarians and of 45% in people aged 65 and older, which is probably because of the broad diagnostic criteria used to avoid underrepresentation. Moreover, cardiovascular risk factors such as hypertension and dyslipidemia increase steadily in most elderly people.[7, 14] Nevertheless, it was hypothesized that, in spite of the increase in absolute terms with age, the incidence of DM might stabilize around the age of 85 and may even decline. For instance, one study found that the incidence fell from 24% in 75-year-olds to 19% in 85-year-olds; similarly, a study conducted in the same area as the present one found that DM incidence fell from 28% in octogenarians to 14% in community-dwelling nonagenarians.
The characteristics of participants were similar to those of other populations described recently.[3, 12] Although the group with DM maintained high ADL scores, they were more likely to report difficulties in performing functional tasks. However, only BI score (as a measure of initial functional disability) remained an independent factor associated with DM, which is consistent with previous studies. With regard to cognitive impairment, this was greater in participants with DM, although the results were inconsistent. Although there is an association between cognitive impairment and DM, in this study, the high comorbidity and other vascular risk factors in the group without DM might have attenuated the differences between the two groups in part.[7, 16] Similarly, Braden Pressure Sore Scale score appeared to be inversely associated with DM, and other morbidity factors are again likely to have influenced this relationship. In relation to health-related quality of life, the results show a high index in both groups. This is probably because of the low social risk (measured using the Gijon test) and low-risk lifestyle components (no smoking and alcohol habits and relatively high levels of physical activity) found in the individuals with DM.[4, 17] These findings suggest that, in older adults, DM should be monitored to prevent deterioration in geriatric conditions, especially functional capacity, and that a multidisciplinary approach is required to manage its complexity.[18, 19]
The overall rate of hypertension (78%) in this group was higher than that reported for other populations of the same age (57%) and was even higher (89%) in the group with DM.[7, 12] Possible explanations for this include the different definition of hypertension used (140/90 mmHg vs 160/95 mmHg) and the different published guidelines (130/80 mmHg for the ADA vs 140/80 mmHg for the National Institute for Health and Clinical Excellence in DM without kidney, eye or cerebrovascular disease), suggesting, as mentioned by the European Society of Hypertension, that the blood pressure level below which vital organ perfusion is impaired is likely to differ according to participant characteristics.
Dyslipidemia was the most important cardiovascular risk factor in the United Kingdom Prospective Diabetes Study, and the rate observed in the present group was similar to that reported for other similar populations, although paradoxically, an inverse association was found between DM and LDL-C, perhaps because individuals with DM were under treatment and were thus receiving better medical care. Some studies have reported an association, especially in elderly adults, between the use of statins and a slightly higher incidence of DM. Therefore, the fact that dyslipidemia treatment was more common in the DM group could reduce the effect of this cardiovascular risk factor. The group with DM did not reach the objective for LDL-C control, corroborating the results of other studies in which fewer than 50% of participants attained the recommended levels for LDL-C (2.6 mmol/L).[23, 24]
The present study failed to identify any differences in the overall rate of comorbidities, although an association was observed for some specific comorbidities. Thyroid disease was more common in individuals without DM, although the sample size means that this result should be treated with caution. By contrast, the group with DM was more likely to have macrovascular complications, such as vasculopathy, although this did not emerge as an independent factor. The results also show that older adults with DM were more likely to have other cardiovascular conditions such as heart failure and atrial fibrillation and noncardiovascular morbidities such as hearing impairment and anemia, although in the current study, no significant differences were found from the group without DM.[6, 8, 18] The number of chronic drug prescriptions was similar to that reported by other groups and was high in the group with DM.[6, 16] In most of the individuals with DM, this included oral medications prescribed alone, without insulin, with metformin being the most common oral antidiabetic agent (45%). This is consistent with current ADA recommendations regarding first-line therapy for DM in the absence of renal insufficiency or heart failure.
The analytical data showed that DM was associated with high folic acid levels. This may be due to the supplements prescribed to individuals with DM or to other side effects of their medical treatment in general, although only two of the six participants with DM (with vitamin therapies) received the supplements, and the levels reached by these two participants (16 and 53 pg/mL) remained within the normal range, so it would not seem that prescription bias regarding the supplements explains this effect. A trend toward an association with vitamin B12 deficiency was also found in the group with DM, this being consistent with previous reports. Vitamin B12 deficiency could result from digestive absorption problems in the context of diabetic microangiopathy or because of the effect of metformin treatment on vitamin B12, which could also produce poor absorption of vitamin B12. The present data therefore highlight the effect of the association between metformin treatment and low levels of vitamin B12, which is a clinically important and treatable condition.
Another important analytical finding was the high white blood cell count and low HDL-C level in the DM group, which appear to be indicators of inflammation. These results support the theory that inflammation plays a role in cardiovascular disease or in producing insulin resistance, although because the results for other acute-phase reactants were unclear, their pathogenic role also remains to be elucidated, with further research being required in this regard.[27, 28]
The strengths of the present study include the sample size, the multicenter design, and the use of a community-based sample with similar characteristics, although this last strength might also constitute a weakness by limiting the generalizability of the results. The main limitation of the study is that the cross-sectional design prevents any causal relationship from being postulated. Another limitation concerns the relatively small number of centers from which the study group was drawn; further multicenter studies or meta-analyses are required to corroborate the results. Another limitation was that microalbuminuria was not determined, and the MDRD equation, although it is indicated in individuals with DM,[4, 9] systematically underestimates normal and high glomerular filtration rate. Nonetheless, this equation was chosen over other probably more-reliable tools for calculating estimated glomerular filtration rate in octogenarians such as the Chronic Kidney Disease Epidemiology Collaboration equation because it was the only formula available at the primary healthcare centers where the study was performed, and this approach probably facilitated the generalizability of the results for later assessment of this group. Finally, this study included the new group as people with DM with no further validation because of the cross-sectional study design. Therefore, despite their clear differences, they were accepted as people with DM because this group is more representative of the real situation in health care, where not at all people are correctly diagnosed or receive appropriate treatment.
In conclusion, a high prevalence of DM and comorbidities was found at age 85. Individuals with DM are likely to have greater disability, more chronic drug prescriptions, and higher white blood cell count and folic acid levels than their counterparts without DM. Further study of the effect of morbidities on DM is required to manage the increasing complexity of clinical conditions in elderly adults.