Summary
Background
In the elderly, cognitive impairment is associated with loss of independence and may be predictive of mortality.
Aims
Our aim was to determine if cognitive impairment correlated to poor prognosis in an elderly (> 70 years) hospitalized population with cardiovascular diseases. Our other goal was to explore other factors that might influence mortality risk. Better understanding of these factors should help practitioners select tools to assess these patients and prevent the occurrence of adverse outcomes.
Methods
During 4 years of follow-up, medical events and all-cause mortality were reported in 331 patients aged above 70 years, as well as clinical and biological variables and Mini Mental State Examination scores.
Results
Patients with cognitive impairment were older and had a lower body mass index than patients without cognitive impairment ( P = 0.023). When all factors were forced into the Cox model, cognitive impairment remained an independent predictor of mortality ( P < 0.001). High plasma glucose, low body mass index and low plasma albumin were associated with overall mortality, independent of cognitive impairment.
Conclusion
In elderly inpatients aged above 70 years with cardiovascular diseases, cognitive impairment and malnutrition are associated, and both are predictors of all-cause mortality. Early nutrition programmes may help to delay mortality, as well as screening the impairment of neuropsychological functioning using the total Mini Mental State Examination score.
Résumé
Contexte
Chez les patients âgés, les troubles cognitifs sont accompagnés d’une perte d’indépendance qui pourrait conduire à la mortalité.
Objectifs
L’objectif de cette étude était de déterminer, de façon globale, si les troubles cognitifs sont corrélés à un pronostic plus mauvais chez les sujets âgés (> 70 ans) hospitalisés, ayant des maladies cardiovasculaires. L’autre objectif était d’explorer les différents facteurs qui pourraient modifier le risque de mortalité. Une meilleure compréhension de ces facteurs pourrait aider les praticiens dans leurs pratiques de prévention.
Méthodes
Pendant quatre ans de suivi, les événements intercurrents et les causes de mortalité ont été rapportés chez 331 patients âgés plus de 70 ans.
Résultats
Les analyses statistiques ont montré que les troubles cognitifs représentent un facteur de risque indépendant de la mortalité. Les patients présentant des troubles cognitifs étaient les plus âgés, avaient un index de masse corporelle plus bas que les autres ( p = 0,023). Quand les autres facteurs de risque étaient introduits dans le modèle statistique, les troubles cognitifs demeuraient un facteur de risque indépendant prédictif de la mortalité ( p < 0,001). L’hyperglycémie, le faible IMC et l’hypoalbuminémie étaient significativement associés à la mortalité indépendamment des troubles cognitifs.
Conclusion
Chez les sujets âgés (> 70 ans) hospitalisés ayant des maladies cardiovasculaires, les troubles cognitifs et la dénutrition sont associés et tous les deux sont prédictifs de la mortalité. Des programmes de nutrition adaptés pourraient contribuer à un meilleur pronostic.
Introduction
Impairment of cognitive functions is a common disorder in older people. Cognitive impairment (CI) is associated with decreased memory function and/or other neuropsychological deficits, such as impairment of executive functions, apraxia, agnosia or aphasia. CI represents a major risk factor associated with loss of personal independence and, finally, with the development of dementia, especially Alzheimer’s disease, which is the most common form of neurodegenerative dementia in the elderly . Recent studies on CI, particularly those involving Mini Mental State Examination (MMSE) determinations, have shown the increasing prevalence of dementia with advancing age, affecting around 7% of individuals aged above 65 years and 30% of those aged above 80 years . Thus, patients with CI could exhibit difficulties in completing instrumental activities of daily living, such as managing finances, organizing medications and food preparation . Therefore, a better understanding of the role of multiple factors (as clinical markers) associated with the onset and progression of cognitive decline with advancing age would be useful in the development of new prevention methods. Each year, as an increase in the average life span is observed, the need to perform such studies in the elderly becomes urgent, not only in the ‘young elderly’, but also in those aged above 70 years .
Neuropsychological evaluations and mostly novel neuroimaging methodologies suggest that major modifications within the nervous tissue itself (mainly the cerebral white matter) may be at the origin of CI in the elderly . However, there are many arguments suggesting that vascular damage might make a substantial additional contribution . In addition, metabolic and nutritional modifications occur in older people and may interact with vascular factors; both may contribute to overall risk. Finally, increased arterial stiffness may be deleterious for the cardiovascular system, especially in the presence of inflammation or malnutrition . On the other hand, in the elderly population, only type 2 diabetes mellitus remains an important risk factor, as other cardiovascular risk factors do not have the same effect in adults or older people .
Although the population profiles of industrialized countries show dramatic increases in the number of elderly people, few studies have focused on CI in very old age . Recently, Strandberg et al. reported that a low MMSE score was predictive of mortality in home-dwelling patients with cardiovascular disease (CVD) aged 75 years or above . However, this study did not explore hospitalized patients, despite them being the most vulnerable and dependent patients and also the most likely to have multiple risk factors for mortality.
Our first working hypothesis was that, in elderly hospitalized subjects aged above 70 years and presenting CVD, CI studied prospectively should also be a good predictor of overall and cardiovascular mortality risk. In addition, we investigated the factors significantly correlated to CI and all-cause mortality, to suggest new therapeutic and prevention methods, particularly in patients presenting CVDs.
Methods
Study cohort
Inclusion criteria were as follows: age above 70 years; history of CVD (defined in this study as presence or history of arterial hypertension, coronary heart disease, cerebrovascular disease, history of clinically assessed heart failure or any vascular event of the arteries of the upper or lower limbs, the thoracic or abdominal aorta or the renal arteries); MMSE greater than 15 out of 30 (to make sure the patient could give informed consent to participate in the study), with no severe visual or hearing impairment, absence of fatal disease and life expectancy less than 1 month. Patients with cachexia (body mass index [BMI] < 17 kg/m 2 ) and/or advanced cancer and/or renal failure (plasma creatinine > 250 μmol/L) were not included.
From May 2000 to November 2001, 331 consecutive patients (86 men and 245 women) with a mean age ± standard deviation of 87 ± 7 years (range: 72–104 years), hospitalized in the geriatric departments of two Île-de-France (Paris suburbs) hospitals (Charles Foix and Emile Roux), were included in the PRonostic cardiovasculaire et Optimisation Thérapeutique En GERiatrie (PROTEGER) study. All patients had CVD; 17% had a history of dyslipidaemia, 21% diabetes, 75% arterial hypertension, 78% coronary heart disease, 22% heart failure, 28% cerebral infarction and 28% peripheral artery disease.
The PROTEGER study was approved by the Committee for the Protection of Human Subjects in Biomedical Research of the Saint-Germain Hospital in Île-de-France. Written informed consent was obtained from all participants after relevant information was given to them and/or to their relatives. Variables that are important for the present analysis are presented here.
Social and clinical characteristics of patients
Information compiled from the questionnaire filled out at inclusion comprised: sex; age; weight; height; personal history of cardiovascular event; presence of diabetes mellitus, dyslipidaemia or arterial hypertension; smoking habits; and previous diseases. The reason for hospitalization and total MMSE score were registered once, at inclusion, in haemodynamically-stabilized conditions. The education level (1 indicates primary school; 2, college degree; 3, bachelor degree; and 4, university degree) was also registered. In some subjects, such information was provided by relatives and/or recorded from the most recent previous hospitalization.
Follow-up procedures
Follow-up started from the baseline examination and lasted until April 2004. Among the 331 participants in the present study, three (1%) were lost to follow-up. Information was obtained from the patients themselves, from their relatives or from general practitioners. Interim telephone and clinic contacts were used to assess all the hospitalizations, outpatient cardiovascular diagnoses and overall mortality. In the case of hospitalization, discharge reports from medical specialists were obtained. Fatal and non-fatal cardiovascular events and all-cause mortality were reported. Follow-up time was defined as the time from the baseline visit to April 2004 for those who survived, to the last contact date for those who were lost to follow-up or to death for the others.
Mini Mental State Examination evaluation
The MMSE, the most commonly administered psychometric screening assessment of cognitive functioning, is a simple and rapid test that may be done by any clinician to evaluate global cognitive functions, especially in primary care . The MMSE has been reported to be a reliable and reproducible method, measuring five areas of cognition labelled as ‘orientation’, ‘registration’, ‘attention and calculation’, ‘recall’, and ‘language’ . A lower MMSE score has been previously reported to predict a higher mortality in subjects with neurological diseases and with CVD . Scores of less than 23–24 points out of 30 have been considered to indicate CI. MMSE scores have been previously reported to be affected by age, education and cultural background, but less by sex . There are no standard cut-offs for the MMSE with regard to education, as prior studies have suggested different cut-offs . In a French population, MMSE was adjusted to educational level and Kalafat et al. defined the pathological cut-off scores as 20–21 in low educational groups and 24 in high educational groups . In this study, MMSE was adjusted to educational level and significant CI was defined as MMSE less than 20 in subjects with primary education (level 1) and less than 24 in subjects with higher education (levels 2, 3 and 4) .
Assessment of haemodynamic and biological variables
Haemodynamic measurements were performed in the morning (8.00–10.00 a.m.). Brachial blood pressure (BP) was measured after 15 minutes’ rest, using the semiautomatic oscillometric device (Dynamap-Kontron, Paris, France). Aortic pulse wave velocity was determined using the foot-to-foot method in only 283 patients. Venous blood samples were obtained in patients after an overnight fast. Plasma was separated without delay at 4 °C in a refrigerated centrifuge and stored at 4 °C until analysis. Total cholesterol, triglycerides, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, plasma creatinine, plasma albumin and orosomucoid concentrations were measured as previously described .
Statistical analysis
Social, clinical and biological variables, as well as previous cardiovascular events and cardiovascular mortality and all-cause mortality, were compared between subjects with or without CI, using Student’s t test for quantitative variables and the Chi 2 test for qualitative variables. Logistic regression models were used to investigate determinants of CI with the stepwise method. The independent association of all-cause mortality with CI was studied by stepwise Cox regression analysis in all subjects and subjects without CI, after adjustment for confounders. Statistical analysis was performed using SAS software, version 9.1 (SAS institute, Cary, NC, USA). A P value < 0.05 was considered statistically significant. In Table 1 , only significant results have been mentioned in model B.
| β | Hazard ratio a | 95% CI | P | |
|---|---|---|---|---|
| Model A b : Chi 2 = 21.2; P < 0.001 | ||||
| Cognitive impairment (1, 0) | 0.71 | 2.03 | 1.38–2.98 | < 0.001 |
| Model B c : Chi 2 = 54.1; P < 0.001 | ||||
| Cognitive impairment (1, 0) | 0.71 | 2.04 | 1.32–3.15 | 0.001 |
| Albumin (g/L) | –0.47 | 0.63 | 0.49–0.80 | < 0.001 |
| Glucose (mmol/L) | 0.29 | 1.33 | 1.11–1.61 | 0.003 |
| BMI (kg/m 2 ) | –0.38 | 0.68 | 0.52–0.89 | 0.004 |
| HDL cholesterol (mmol/L) | –0.32 | 0.72 | 0.56–0.93 | 0.011 |
a Hazard ratios were calculated for a 1 standard deviation increase in all quantitative variables and the presence of qualitative variables against the absence.
b Model A = age, male sex, hospital location and cognitive impairment
c Model B = model A + education level, residence condition, systolic and diastolic blood pressures, heart rate, smoking, antidepressive therapy, total cholesterol and triglycerides. Age, male sex and hospital location were forced in both models. Only significant results are mentioned.
Results
General characteristics of the population
The population was composed of 331 patients (86 men, 245 women) with a mean age of 87 ± 7 years. All patients had a past history of CVD but the reasons for their hospitalization were as follows: a cardiovascular event (19% of all patients); various physical injuries mostly due to falls (36%); unspecified neuropsychological problems (14%); social reasons (mainly a bridge to a nursing home) (10%); various infections (10%); and other miscellaneous causes (11%). A total of 138 patients (42%) had CI and 191 (58%) had no CI. The mean total MMSE score was 22.2 ± 5.2. After a mean follow-up of 380 ± 196 days (extremes: 50–848 days), 110 subjects (33%) had died.
Table 2 compares the clinical, biological and cardiovascular haemodynamic variables, as well as the medication of CI versus non-CI subjects at the end of follow-up. CI patients were older ( P = 0.017), although the prevalence of CI with aging was significant in men ( n = 86; P = 0.04) but not in women ( n = 244; P = 0.29). CI patients had a lower BMI ( P = 0.023) than non-CI patients, an indicator of malnutrition. In contrast with non-CI patients, CI subjects usually lived in an assisted-living community (nursing home) before hospitalization ( P = 0.050) and tended to have a low mood as they consumed more antidepressive drugs ( P = 0.092).
