González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 DOI 10.1186/s12872-016-0349-5 RESEARCH ARTICLE Open Access Ischemic aetiology, self-reported frailty, and gender with respect to cognitive impairment in chronic heart failure patients María J González-Moneo1,3,4,5, Gonzalo Sánchez-Benavides2, José M Verdu-Rotellar1,3,4,5, Mercé Cladellas3, Jordi Bruguera3, Sonia Quiñones-Ubeda2, Cristina Enjuanes3, Jordi Peña-Casanova2 and Josep Comín-Colet3* Abstract Background: Decisive information on the parameters involved in cognitive impairment in patients with chronic heart failure is as yet lacking. Our aim was to determine the functional and psychosocial variables related with cognitive impairment using the mini-mental-state examination (MMSE) with age-and education-corrected scores. Methods: A cohort study of chronic heart failure patients included in an integrated multidisciplinary hospital/ primary care program. The MMSE (corrected for age and education in the Spanish population) was administered at enrolment in the program. Analyses were performed in 525 patients. Demographic and clinical variables were collected. Comprehensive assessment included depression (Yesavage), family function (family APGAR), social network (Duke), dependence (Barthel Index), frailty (Barber), and comorbidities. Univariate and multivariate logistic regression were performed to determine the predictors of cognitive impairment. Results: Cognitive impairment affected 145 patients (27.6 %). Explanatory factors were gender (OR: 2.77 (1.75–4.39) p < 0.001), ischemic etiology (OR: 1.99 (1.25–3.17) p = 0.004), frailty (OR: 1.58 (0.99 to 2.50, p =0.050), albumin > 3.5 (OR: 0.59 (0.35–0.99) p = 0.048), and beta-blocker treatment (OR: 0.36 (0.17 to 0.76, p = 0.007)). No association was found between cognitive impairment and social support or family function. Conclusion: The observed prevalence of cognitive impairment using MMSE corrected scores was 27.6 %. A global approach in the management of these patients is needed, especially focusing on women and patients with frailty, low albumin levels, and ischemic aetiology heart failure. Keywords: Heart failure, Cognitive symptoms, Prevalence, Comorbidity Abbreviations: CHF, Chronic heart failure; CI, Cognitive impairment; LVEF, Left ventricular ejection fraction; MMSE, Mini-mental state examination; NHYA, New York heart association; Nt proBNP, pro-brain natriuretic peptide Background should be included in a more extensive neuropsycho- Cognitive impairment (CI) is particularly common in pa- logical diagnostic battery. Despite interest in the global tients with chronic heart failure (CHF) and has been assessment of CHF there is as yet no decisive informa- associated with an increased mortality rate in hospital tion on which factors play a role when deciding whether admissions and worse clinical outcomes [1]. Neverthe- routine cognition screening in clinical practice would be less, it is still uncertain when CI should be assessed in helpful for an individual patient. common clinical practice, the instruments that should The prevalence of CI depends on the test used and be employed for primary evaluation, and which patients varies widely among studies, ranging from less than 30 % to more than 80 % [2–7]. As cognitive tests are * Correspondence: josepcomin@gmail.com age-and education sensitive [8] the use of a validated 3 Grup de recerca biomedica en malalties del cor GREC (Heart Diseases screening test in our population would provide valuable Biomedical Research Group), IMIM (Hospital del Mar Medical Research information on CI prevalence. The Mini Mental State Institute), 88, Doctor Aiguader, 08003 Barcelona, Spain Full list of author information is available at the end of the article Examination (MMSE) is a widely used instrument to © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 2 of 10 evaluate cognitive function in CHF patients [9, 10]. It Recorded comorbidities were current diabetes mellitus, determines global cognitive impairment and is more history of stroke or cerebrovascular disease, peripheral ar- commonly employed in clinical practice than complex terial disease, chronic kidney disease, anemia, and chronic neuropsychological batteries [4, 11]. obstructive pulmonary disease. Once CHF diagnosis had The present study aims to assess CI prevalence with been confirmed on an outpatient basis the patient was the MMSE, and identify clinical, psychological, social, invited to be included in the heart failure registry. Clinical family, and frailty-related factors, in a large sample of and analytical data were then collected, and evalua- CHF patients. tions were performed by cardiology nurses trained in heart failure management. Methods From 2005 to 2010, 805 consecutive patients were en- rolled as a cohort of non-institutionalized patients referred Cognitive functions to hospital-primary care integrated multidisciplinary nurse All the patients who agreed to participate underwent a led heart failure program [12]. We present here the base- comprehensive psycho-social assessment. From those line assessment. Assessment was conducted at the time of invited 122 had an incomplete evaluation and were ex- patient inclusion in the heart failure program. Once diag- cluded. The MMSEwas completed at baseline in all nosis was confirmed on an ambulatory basis, the patient the 683 remaining patients, either by a neuropsycholo- was invited to be included in the heart failure registry. gist or a specialized nurse who had received specific Then, clinical and analytical data were collected, and psy- training. Of those tested, 158 had not education recorded chosocial evaluations were performed by hospital nurses and were excluded. MMSE [14] Spanish validated version specifically trained in heart failure. The evaluation was (30 items, cognitive impairment scores ≤ 24) [15]) with carried out in a single session, with the cognitive test ad- scores adjusted by age and education was finally calculated ministered first. The study was conducted in accordance in 525 patients (Fig. 1). The MMSE measures overall with the Declaration of Helsinki. The study protocol was cognitive impairment, it includes brief assessments of approved by the local committee of ethics for clinical memory, language, praxis, and orientation and it is research and all patients gave written informed consent regarded as the gold standard in cognitive impairment de- after recruitment. For inclusion in the study, patients had tection [16]. It has been used in heart failure patients [5], to be in a stable condition and with a CHF diagnosis of it takes approximately 10 min to administer, and It has either reduced or preserved ejection fraction, according to been found predictor of hospital readmission [3, 15, 17]. the European Society of Cardiology diagnostic criteria Age and education corrected scores were calculated [13]. Additionally, inclusion in the study required that following Blesa MMSE validation in Spanish popula- patients were able to undergo neuropsychological testing tion. These corrections add or subtract up to 2 points and could communicate adequately in order to follow depending on age bands and education level (Table 1). the tests instructions by themselves. Exclusion criteria Other corrections have been proposed but external for the study were: significant primary valvular disease, validity of Blesa correction is particularly appropriate hemoglobin levels < 8.5 g/dL, clinical signs of fluid as it was carried out in our area. No experimental overload, pericardial disease, restrictive cardiomyop- intervention was performed. athy, hypertrophic cardiomyopathy, active malignancy and chronic liver disease. Patients with serious psychi- atric illness, unstable CHF, overt cognitive impairment which impeded psychosocial assessment; and those suffering from extra-cardiac disease with a life expect- ancy of less than 1 year were also excluded. No experi- mental intervention was performed. Clinical variables At recruitment, peripheral blood samples were taken to measure the usual biological variables needed to per- form initial evaluation in heart failure patients. Clinical and demographic information assessed included CHF etiology; New York Heart Association (NYHA) functional class, heart rate, pro-brain natriuretic peptide (Nt proBNP), blood pressure, current medical therapy and the most re- Fig. 1 Patient flow showing exclusion criteria cent left ventricular ejection fraction (LVEF) evaluation. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 3 of 10 Table 1 Correction table of MMSE by age and education. Points within specified groups were calculated and P values added/subtracted from MMSE were derived using Chi [2] tests. All the tests were two- Age sided; differences were considered significant at the p < 0.05 Education (years) ≤50 51–75 >75 level. ≤8 0 +1 +2 The following clinical variables were dichotomized be- fore entering the linear regression model: age over 65/ 9–17 −1 0 +1 below 65, higher education meaning more than 10 years >17 −2 −1 0 versus less than 10 years of formal education, mild Based on Blesa [15] NYHA class I, II versus advanced NYHA III, IV, pre- served left ventricle ejection fraction vs. impaired, heart Psycho-social evaluation rate over 70 versus HR below 70 as heart rate over 70 is Education was assessed as categorical variable: illiter- a marker of disease severity and mortality predictor [30]. ates, elementary education (4 years of education), high albumin <3,5 yes/no. NtproBNP was dichotomized over school (up to 9 years), and college (more than 10 years). the median as multiple factors influence its values We recorded marital status, living with a partner, self- beyond the predictive cut off 1000. Logistic regression administration of medication, and the need for a caregiver. models (enter) were generated to explore the relationships Dependency was measured by the validated Spanish between dichotomous and clinical variables. Clinical, func- version of the Barthel test [18] for basic daily activ- tional and biochemical values, echocardiography and co- ities, in its Spanish validation version; with scores ranging existing diseases were used as explanatory variables. from 0 to 100, those <90indicate dependency. Family func- Univariate analysis was composed of the comparison of tion was recorded following The Family APGAR question- cognitive impairment/cognitive normal according to naire in its Spanish form [19], the scale can range from 0 MMSE age-education cut-off scores. Differences be- to 10, < 7 represents family dysfunction. tween the cognitively normal/impaired groups were Depression was screened with the short form of the calculated using logistic regression model one-to one Spanish validated version [20] of the Yesavage Geriatric analyses. Depression Scale [21]. Scores > 4 indicate depression. Variables that showed statistical significance (p <0.10) Social support is positively associated with medication in the univariate analysis were included in the multivar- compliance [22–25]. We used The Duke-UNC Func- iable logistic regression (enter method) using MMSE tional Social Support Questionnaire [26], an eight-item age-and education-corrected scores as the main vari- 3Likert scale in which the higher the average score, the able. Three logistic regression models (enter) were com- greater the perceived social support. A cut-off of 32 has pleted. Model one included only the clinical factors, been proposed in its Spanish version. model two included psychosocial factors and model three Frailty was screened with the self-reported Barber both clinical and psychosocial factors. Results were pre- questionnaire [27] in its Spanish validated version [28]. sented as odds ratios (OR) with 95 % confidence intervals This questionnaire assesses functional status, psycho- (CI). The Hosmer-Lemeshow chi-square test was used to social functioning, neurosensory deficits, self-reported assess the models’ goodness-of-fit. Various models were health, social support, and previous year hospital admis- developed, including several combinations of adjustment sions. The three major domains in frailty [29]-physical, variables. Co-linearity problems were not observed as the psychological and social-are represented. It is a 9-item change of the standard coefficient errors was not relevant screening tool, easy to understand and self-administer, in terms of loss of statistical significance. which identifies individuals who may be at risk of All models were carried out in the final patient sample dependency. Any positive item indicates frailty. A cutoff (n = 525). SPSS® version 13.0 (IBM, Armonk, NY, USA) of 2 was applied to increase the positive predictive value was employed for statistical analyses. of the test since most patients had a recent admission in the previous year. Results Cognitive impairment and clinical markers Statistical methods Of the total sample size (n = 525), 145 participants Demographic and clinical data were summarized with (27.6 %) were affected by CI as determined by MMSE basic descriptive statistics in the total cohort. For age-and education-corrected scores. There were no quantitative variables arithmetic mean (± standard differences between participants and non-participants deviation) or median (interquartile range) were calcu- with respect to gender, education level, marital status, lated as appropriate, and P values derived from a two- CHF etiology, natriuretic peptide levels, ventricular sample t-test (U-Mann–Whitney tests were used for function, cardiovascular risk factors, reactive depres- skewed data). For qualitative variables, percentages sion, comorbidity, or need of a caregiver. The former González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 4 of 10 had hypercholesterolemia less frequently (45, 38.5 % vs. female. CI was less common in patients with higher edu- 360, 54.1 %, p = 0.001) and were more often included in cation (defined as more than 9 years of formal education) the advanced NYHA functional class (59, 58.4 % vs. 295, (Table 2). More men had completed higher education 44.3 % p = 0.005). than women [153 (47.8 %) vs. 46 (22.4 %) p < 0.001]. CI The clinically relevant affected MMSE areas in CHF was more common in women after adjusting by education patients with CI were orientation, attention, recall, lan- and NHYA class although the relationship could be mod- guage, and copying. The worst scores were found in erated by ischemic etiology: CI was more likely in women NHYA advanced stages, and only registration was stable with ischemic etiology heart failure [OR: 2.303, (95 % IC in the functional class (Fig. 2). Markers of disease sever- 1.204–4.405, p < 0.001)]. ity such as higher levels of Nt pro-BNP, heart rate > 70, and advanced NHYA functional class were related to CI Co-morbidities (Table 2). Mean albumin levels were lower in patients We observed a relationship between CI and mean glom- with CI, with a 3.5 cutoff being the most significant; al- erular flow. CI was also more frequent in patients with bumin levels > 3.5 were protective, in contrast, albumin chronic kidney disease, diabetes, and peripheral vascular levels < 3.5 were related to CI (Fig. 3). Ischemic and disease. Cerebrovascular disease almost reached signifi- hypertensive patients were more impaired than those cance (Table 2) and was included in the final model. with heart failure from other causes. Beta-blocker pre- Anemia was more common in the CI group but not scription provided a significant reduction in CI (RR: 0.35 significant (p = 0 .078). Pulmonary obstructive disease (IC95%, 0.18–0.66)) whilst non-use showed an inverse was not related with CI in our sample. relationship (RR 1.94 (IC 95 %, 1.37–2.74)). No associ- ation was found between CI and other drugs commonly Psychosocial factors used such as angiotensin-converting-enzyme inhibitor Cognitively impaired patients were more likely in need and angiotensin receptor blockers of a caregiver, and not able to administer treatment by themselves. Decreased CI was related to greater patient Gender and education autonomy for the management of treatment (Table 3). Patients with CHF and CI impairment were older, had less Depressive symptoms were more frequent in cogni- than 9 years of formal education, and were predominantly tively impaired patients, almost reaching significance Fig. 2 Cognitive impairment in heart failure patients according to Mini-Mental-State Examination sub-items median scores in initial and advanced NHYA functional class. Statistical significance: (Mann-Whinney U): Orientation: p <0.001, registration: p =0.126, Attention: p =0.004, Recall: p =0.168, Language and copying: p = 0.001. NHYA: New York Heart Association González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 5 of 10 Table 2 Demographics and baseline characteristics of Chronic Heart Failure patients according to cognitive function (CI) n = 525 Overall Normal Cognitive impairment n p value Demographics, n% 525 380 (72) 145 (28) 525 Age, years, mean (SD) 71 (11) 70 (11) 75 (9) 525 <0.001 Female gender n (%) 205 (39) 126 (33) 79 (55) 525 <0.001 NYHA class n% (I, II / III, IV) 300 (59) / 211 (41) 229 (62) / 139 (38) 71 (50) / 72 (50) 511 0.007 Heart rate > =70, bpm, n (%) 304 (59) 210 (56) 94 (66) 515 0.034 SBP, mean (SD) 126 (22) 126 (23) 124 (21) 517 0.324 Preserved LVEF %, mean (SD) 180 (34) 131 (34) 49 (34) 525 0.484 Aetiology of CHF, n% 382 (73) 268 (71) 114 (79) 520 0.042 Hypertension n% 412 (79) 295 (79) 114 (79) 520 0.459 Diabetes n% 239 (47) 162 (44) 77 (54) 511 0.023 Peripheral artery disease n% 99 (19) 63 (17) 36 (25) 521 0.020 CKD 286 (55) 195 (52) 91 (63) 523 0.014 Anaemia n% 176 (33) 120 (32) 56 (39) 525 0.078 Cerebrovascular disease n% 62 (12) 40 (11) 22 (15) 521 0.089 COPD n% 114 (22) 79 (21) 35 (24) 517 0.239 ACEI or ARBs n% 426 (82) 311 (83) 115 (80) 519 0.243 Betablockers n% 482 (92) 358 (95) 124 (86) 522 0.001 MRA n% 228 (44) 167 (44) 61 (42) 521 0.383 Nt-pro BNP n% 248 (48) 169 (45) 79 (56) 515 0.014 Albumin mean (SD) 4 (0) 4 (0) 3.6 (0) 501 <0.001 Albumin < 3.5 97 (19) 57 (16) 40 (28) 501 0.002 eGFR mean (SD) 59 (22) 60 (23) 55 (21) 523 <0.017 Cognitive impairment defined by age and education corrected MMSE scores. Data are presented as means ± standard deviation, medians (with interquartil range) or numbers (with percentages) where appropriate. Percentages are rounded. NYHA class: New York Heart Association functional class: LVEF: left ventricle ejection fraction. SBP: systolic blood pressure. CHF: chronic heart failure. Aetiology of CHF: patients with ischemic and hypertensive aetiology versus other causes. CKD: Chronic kidney Disease: eGFR < 60 mL/min/1.73 m2. Anaemia was defined as haemoglobin level <12 g/dL in women and <13 gr/dL in men. Cerebrovascular disease defined as. COPD: chronic obstructive pulmonary disease defined following GOLD (Global Initiative of Chronic Obstructive Lung disease) Guidelines. ACEI:angiotensin-converting-enzyme inhibitor. ARBs: Angiotensin II receptor blockers. MRA: Mineral corticoid receptor blockers. Nt-proBNP: N-terminal pro Brain Natriuretic Peptide over 1512 (median scores). eGFR: estimated glomerular filtration rate (p = 0.064). Lack of social support measured by Duke’s included (p = 0.050) (model 3). Gender and ischemic questionnaire was minimal in our sample and non- etiology prevailed in the explanatory model when all significant. Less than 15 % of the patients lived in a dys- variables were added. In addition, albumin levels > 3.5 functional family, and this factor was un-related to CI. and beta-blocker treatment emerged as potentially pro- tective CI factors. Multivariate analyses Adjusting for the significant factors identified in the un- Discussion varied analyses, we performed three different multivari- Cognitive impairment ate binary logistic regression models to evaluate (i) In this single center study we have found that cognitive clinical factors, (ii) functional ones, and (iii) both. Age- impairment determined by age-and education-corrected and education-corrected MMSE scores were employed MMSE scores affected 145 of 525 patients (27.6 %), as the main variable in all 525 patients (Table 4). In a which is consistent with other studies that used this test clinical model (model 1), patients with overt CI were in patients with heart failure [31]. Nevertheless, other significantly more likely to be women, and have ischemic screening methods employed in this type of patients etiology and low albumin levels. They were less probably have obtained different figures, as CI prevalence varies receiving beta-blocker treatment. Dependency and according to the test used. Lately, MOCA has emerged frailty presented a significant association with CI in as a possible screening tool although it does not take the functional model (model 2). The presumed association into account the potential false positives that can pro- between CI and self-reported frailty almost reached vide prevalence figures up to 80 % [1]. In a comparative statistical significance after clinical variables were study of the MOCA and MMSE in a small sample of González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 6 of 10 Fig. 3 Parameters associated with overt cognitive impairment defined by age and education corrected MMSE scores in univariate analysis with a P value < .10. n = 525. Risk estimate and confidence limits (CL). Higher education meaning more than 10 years versus less than 10 years of formal education. NYHA: New York Heart Association functional Class. LVEF: left ventricle ejection fraction. DM: diabetes mellitus. CKD: Chronic kidney Disease: eGFR < 60 mL/min/1.73 m2. Anemia defined by hemoglobin < 12. COPD: chronic obstructive pulmonary disease. Nt-proBNP: N-terminal pro Brain Natriuretic Peptide hormone. Albumin levels >3.5. Depressive symptoms: measured by Yesavage scores above 5. Family dysfunction: Family Apgar below 7. Social support meaning lack of social support by DUKE-UNC questionnaire < 32. Frailty when Barber questionnaire above 2. Dependency defined by Barthel test <90 patients, Athilingam et al. (Heart & Lung 2011) found making necessary extreme adjustments (up to 4 points in no relationship between clinical parameters of cardiac Spanish-speaking individuals; Zhou 2015) [33]. In this function and the high rate of CI detected by the regard, we think that MMSE is still useful for its use in MOCA (54 %). low-educated subjects. As a consequence, we employed a Howkins et al [4] compared the MOCA and MMSE validated Spanish MMSE adjusted for age and education against a neuropsychological battery in 106 patients. They which had not only good internal validity but also excel- reported adequate sensitivity for both tests (64–70 %) lent external validity as it was validated in our area. although in terms of specificity, the MMSE was slightly Other methods use brief screening tools evaluating only better than the MOCA (70 % vs. 64 %, respectively). This memory, such as the Memory Impairment Screening test difference in specificity had already been highlighted by (MIS) which in the EFICARE study resulted in 46 % preva- Lees et al. [32] in patients with vascular CI. lence [34], whilst the MMSE implements more varied sub- Both the MOCA and MMSE are clearly influenced domains [5–7, 35, 36]. The fact that the prevalence found by age and education, the former is considered more depends on the test used [5, 37, 38] underlines the import- appropriate in populations with more than 12 years ance of using tests validated for the population in focus, of schooling. The characteristics of the MoCA subtests, with appropriately age-and-education corrected versions. which include several complex executive demanding tasks, Although there are various tests for mild CI, only MMSE make it more prone to false positives when it comes to and neuropsychological batteries are recommended for its detect cognitive impairment in low educated subjects, early detection in Neurology Practice Parameters [39]. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 7 of 10 Table 3 Psychosocial, functional and cognitive scores according to cognitive function: n = 525 Overall: 525 Normal n = 380 Cognitive impairment n = 145 n p value Higher education 199 (38) 164 (43) 35 (24) 525 <0.001 Living with someone n % 287 (57) 217 (59) 70 (52) 504 0.098 Depressive symptoms n % 35 (7) 21 (6) 14 (10) 513 0.064 Dependency n % 206 (39) 133 (35) 73 (50) 525 0.001 Need caregiver, n (%) 127 (34) 83 (30) 44 (45) 370 0.008 Patient administer treatment n (%) 201 (60) 157 (61) 44 (47) 353 0.014 Lack of social support n % 41 (8) 27 (7) 14 (10) 515 0.282 Family dysfunction n (%) 72 (14) 50 (13) 22 (15) 520 0.325 Frailty n (%) 279 (55) 186 (50) 93 (67) 509 0.001 MMSE total age and education corrected scores mean (SD) 25 (4) 27 (2) 20 (3) 525 <0.001 MMSE orientation 9 (1) 10 (1) 8 (2) <0.001 n % task not completed 239 (45) 117 (31) 128 (83) MMSE registration 3 (0) 3 (0) 3 (0) <0.001 n % task not completed 27 (5) 11 (3) 15 (10) MMSE attention 3 (2) 4 (1) 2 (1) <0.001 n % task not completed 277 (53) 137 (36) 139 (96) MMSE recall 2 (1) 2 (1) 1 (1) <0.001 n % task not completed 378 (72) 251 (66) 126 (87) MMSE language and copying 8 (2) 8 (1) 6 (2) <0.001 n % task not completed 289 (55) 157 (41) 131 (90) Percentages are rounded. Data are presented as numbers (with percentages) or means (standard deviation, SD), where appropriate. n represents the number of patients with variable recorded which were analyzed for each item. Higher education defined as more than 9 years of formal education. Depressive symptoms measured by Yesavage test >5. Dependency in daily living activities measured by Barthel test (total <20/ severe 21–60/ moderate 61–90/slight 91–99 /independence: 100. Social support by The Duke-UNC Functional Social Support Questionnaire (FSSQ), weak social support scores < 32. Family dysfunction by APGAR index < 7; frailty when Barber self-administer questionnaire above 2 (see text) MMSE: mini-mental state examination (Folstein [14]) corrected by age and education (Blesa [15]): cognitive impairment scores ≤24 The MMSE measures general cognitive functioning, in as NHYA advanced class, and Nt proBNP [43]. The our study all the sub-items were lower in patients in the relationships were not, however, sustained in a multi- advanced functional class, with the exception of registra- variate model. LVEF was not related to CI. tion, which is the ability to acquire the information in Hypoalbuminemia, previously considered a mortality the first place, and refers to short-term memory, more predictor following myocardial infarction [44], has also influenced by unstable organic conditions [40], and are been reported to be predictive of 1 year mortality preserved in many cognitively impaired subjects. among heart failure patients [45]. Classically linked to frailty, in our sample, albumin levels appeared to be re- Clinical markers lated to CI. We found a cutoff of 3.5 meaningful, as We observed that in heart failure patients, ischemic levels < 3.5 were more likely associated to CI. This ef- aetiology was linked to CI. In a recent study, lower fect persisted after controlling for renal function and recall scores in cognitive tests, and cerebral grey mat- was sustained in a regression model. ter loss in magnetic resonance images, were reported In our sample, patients under beta blocker treatment in patients with ischemic heart disease compared to showed less CI after adjusting for functional class, and non-ischemic ones [41]. Mild cerebral ischemia/hypoxia the effect was maintained after adjusting for multiple resulting from chronic heart failure could augment the factors in a regression model. In the 1990s beta- ischemic effect and cause synaptic dysfunction, as recently blocker treatment was associated to cognitive impair- reported [42]. Hypoperfusion could be an explanation ment [46, 47]. Nevertheless, recent experimental stud- although further research is needed to confirm this ies have suggested a protective effect [48]. Such an possibility. effect was also observed in a retrospective study of In concurrence with previous research we found hypertensive patients treated with beta blockers [49] other factors directly linked to heart failure, as such versus other drugs. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 8 of 10 Table 4 Multivariable regression model considering all significative variables (enter) Clinical variables Model 1 Psychosocial variables Model 2 Clinical and functional Model 3 OR (95 % CI) p value OR (95 % CI) p value OR (95 % CI) p value Gender (women) 2.98 (1.85–4.82) <0.001 2.77 (1.75–4.39) <0.001 Ischemic aetiology 1.90 (1.16–3.08) 0.010 1.99 (1.25–3.17) 0.004 Cerebrovascular disease 1.76 (0.94–3.30) 0.079 Peripheral Arterial disease 1.57 (0.91–2.73) 0.105 DM 1.15 (0.74–1.79) 0.545 Heart rate >70 1.44 (0.91–2.27) 0.118 Albumin > 3.5 0.56 (0.33–0.95) 0.030 0.59 (0.351–0.99) 0.048 eGFR 1.01 (0.64–1.58 0.965 Nt-ProBNP 1.00 (0.63–1.59) 0.990 Betablockers 0.33 (0.16–0.69) 0.003 0.36 (0.17–0.76) 0.007 Social support 0.83 (0.36–1.94) 0.668 Frailty 1.58 (1.02–2.46) 0.040 1.58 (0.99–2.50) 0.050 Family dysfunction 1.02 (0.52–1.99) 0.964 Depressive symptoms 1.72 (0.82–3.63) 0.156 Dependency 1.65 (1.08–0.53) 0.020 1.30 (0.83–2.05) 0.255 Cognitive impairment defined by MMSE age and education corrected scores DM: Diabetis Mellitus. eGFR: estimated glomerular filtration rate. Nt-proBNP: N-terminal pro Brain Natriuretic Peptide. Social support measured by The Duke-UNC Functional Social Support Questionnaire (FSSQ); lack of social support scores < 32. Frailty when Barber questionnaire above 2 (see text). Family dysfunction when APGAR index < 7. Depressive symptoms measured by Yesavage test >5. Dependency in daily living activities measured by Barthel test <90. MMSE age and education corrected scores. OR: odds ratio. CI: confidence interval. Hosmer-Lemeshow chi-square test: Model 1 p = 0.411; Model 2p = 0.766; Model 3:p = 0.713. Nagelkerke R Square: Model 1:R2 = 0.154; Model 2: R2 = 0.052; Model 3: R2 = 0.150 Gender and education (dependency level, self-reported frailty, and need of a We observed that CI was more frequent in women, who caregiver) which was associated with less CI. In addition, also were less educated. In the Women’s Health Initia- social support is positively related to medication compli- tive Study [50], self-reported cardiovascular disease was ance [22–24]. In our sample, as previously described in well documented as a factor that incremented the risk of Mediterranean countries [55], almost all the patients had cognitive decline in postmenopausal women, particularly a self-perception of having a healthy social network and in those with myocardial infarction and vascular disease satisfactory family function. (HR 2.10; 95 % CI:1.40–3.15), whereas no association Depressive symptoms almost reached signification. was found with self-reported heart failure. Heart failure Depression was associated with poorer performance of ischemic origin was associated to cognitive impair- on multiple cognitive domains in heart failure in a re- ment in women in our study. Further research in women cent study [53]. is needed to explore this relationship. Functional limitation has emerged as a predictive value in patients with mild CHF. [56] An elevated rela- Co-morbidities tionship between frailty and mortality at 12 months Co-morbidities, including cerebrovascular and periph- follow-up has been found using the Barthel test, the eral arterial disease [51], must be taken in account in Yesavage depressive symptoms tests and a geriatric CHF as they have been associated with CI and influence exam [56]. We found a link between self-reported avoidable hospitalization and mortality [52]. A global frailty and cognitive function. Frailty and heart failure assessment in CHF patients with co-morbidities illness would share a consistent correlation with some inflam- has been suggested [12]. We found diabetes and periph- matory biomarkers such as interleukin-6 and C- eral vascular disease to be associated with CI, as previ- reactive protein [57]. While there is a lack of consen- ously described [53, 54]. sus on the definition of frailty, assessing it could help tailoring treatment in selected patients. Short screen- Psychosocial factors ing instruments such as Barber tests could be easy to Our study is aimed at both clinical and psychosocial use in clinical practice. We observed that self-reported factors and we identified some that require a global frailty remained a stable predictor of cognitive impair- evaluation focused on practice. The identified psycho- ment in CHF patients after adjusting for clinical social factors included increased patient autonomy variables. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 9 of 10 Strengths and limitations Ethical approval and consent to participate The strengths of this study include the wide sample used, The study protocol was approved by the local committee of ethics for clinical research (Comité ético de Investigación clínica del Instituto Municipal the detailed characterization of the profiled patients, and de Asistencia Sanitaria (CEIC-IMAS-IMIM Hospital del Mar Medical Research the multi-dimensional assessment including both clinical Institute N°: 2004/1788/l) and all patients gave written informed consent and psychosocial factors. Our study has some limitations. after recruitment. Its cross-sectional design implies that conclusions should Author details be confirmed through relevant prospective cohort studies. 1 Department of Medicine, Universitat Autònoma de Barcelona, Barcelona, In addition, from those invited to participate, 122 patients Spain. 2Neurofunctionality of Brain and Language Group, Neurosciences Research Programme, IMIM (Hospital del Mar Medical Research Institute), had an incomplete evaluation and were excluded. The Barcelona, Spain. 3Grup de recerca biomedica en malalties del cor GREC excluded participants may well have had worse outcomes (Heart Diseases Biomedical Research Group), IMIM (Hospital del Mar Medical or have been more impaired as they were in more ad- Research Institute), 88, Doctor Aiguader, 08003 Barcelona, Spain. 4San Martin Primary Care Center, Barcelona, Spain. 5Jordi Gol University Institute for vanced functional class, resulting in higher figures in our Research Primary Healthcare, Barcelona, Spain. population than reported. Moreover, education was only recorded in 525 patients in a way that could be later Received: 21 January 2016 Accepted: 19 August 2016 analyzed. The unregistered patients lacking education could be also have been more impaired or have had worse CHF markers. It might not be possible, therefore, to ex- References 1. Cameron J, Ski CF, Thompson DR. Cognitive impairment in chronic heart trapolate our results to patients with other education failure and the need for screening. Am J Cardiol. 2011;107:1547–8. levels. Furthermore, our sample only included Spanish 2. Athilingam P, D’Aoust RF, Miller L, Chen L. Cognitive profile in persons with patients so the results may not be applicable to popula- systolic and diastolic heart failure. Congest Heart Fail. 2013;19:44–50. tions from other regions and ethnic groups. 3. Dodson JA, Truong TT, Towle VR, Kerins G, Chaudhry SI. Cognitive impairment in older adults with heart failure: prevalence, documentation, and impact on outcomes. Am J Med. 2013;126:120. Conclusions 4. Hawkins MA, Gathright EC, Gunstad J, et al. The MoCA and MMSE as screeners for cognitive impairment in a heart failure population: a study The observed prevalence of cognitive impairment using with comprehensive neuropsychological testing. Heart Lung. 2014;43:462–8. MMSE corrected scores was 27.6 %. Screening should 5. Jurgens CY, Faulkner KM, Lee CS. Phenotypic profiling of cognitive impairment focus specially on patients with low albumin levels, risk among patients with heart failure: a literature review of the usefulness of cardiac-related variables. Eur J Cardiovasc Nurs. 2013;12:109–31. frailty, or heart failure of ischemic aetiology, and 6. Triposkiadis FK, Skoularigis J. Prevalence and importance of comorbidities in women in particular. We believe that integrated care patients with heart failure. Curr Heart Fail Rep. 2012;9:354–62. improves results in heart failure management and facili- 7. van den Hurk K, Reijmer YD, van den Berg E, et al. Heart failure and cognitive function in the general population: the Hoorn Study. Eur J Heart tates patient monitoring focused on enhancing quality Fail. 2011;13:1362–9. of life. Almost one in three patients resulted impaired, 8. Gonzalez Moneo MJ, Escriva FR, Vinyoles BE, Espel MC, Davins MJ, Borrell that justifies the screening in CHF patients as MMSE is MM. Estimation of the prevalence of cognitive impairment according to the a feasible screening instrument, easy to administer on test used. Aten Primaria. 1997;20:173–9. 9. Bauer L, Pozehl B, Hertzog M, Johnson J, Zimmerman L, Filipi M. A brief routine settings. Identifying related factors facilitates neuropsychological battery for use in the chronic heart failure population. the selection of patients in which screening would be Eur J Cardiovasc Nurs. 2012;11:223–30. especially helpful. 10. Vogels RL, Oosterman JM, Van HB, et al. Profile of cognitive impairment in chronic heart failure. J Am Geriatr Soc. 2007;55:1764–70. 11. Tombaugh TN, McIntyre NJ. The mini-mental state examination: a Acknowledgments comprehensive review. J Am Geriatr Soc. 1992;40:922–35. We appreciate the help of the Primary Healthcare Division of Barcelona 12. Comin-Colet J, Verdu-Rotellar JM, Vela E, et al. Efficacy of an integrated (Institut Català de la Salut) in the development of the study. We thank to hospital-primary care program for heart failure: a population-based analysis Aida Cuenca Royo and Paula Martinez Bel who collected part of the sample. of 56,742 patients. Rev Esp Cardiol. 2014;67:283–93. 13. McMurray JJ, Adamopoulos S, Anker SD, et al. ESC guidelines for the Availability of data and materials diagnosis and treatment of acute and chronic heart failure 2012: The The data is available upon request from the authors. Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in Authors’ contributions collaboration with the Heart Failure Association (HFA) of the ESC. Eur J JC, MJG, GS, JV participated in the study conception and design. JC, MJG, JV, Heart Fail. 2012;14:803–69. GS, CE, SQ, MC, JPC, and JB participated in the analysis and interpretation of 14. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical data. MJG and JC participated in the drafting of the manuscript. Critically method for grading the cognitive state of patients for the clinician. revising the manuscript for important intellectual content and providing final J Psychiatr Res. 1975;12:189–98. approval for the manuscript to be submitted was a collective task 15. Blesa R, Pujol M, Aguilar M, et al. Clinical validity of the ‘mini-mental state’ participating JC, MJG, JV, GS, CE, SQ, MC, JPC, and JB. All authors read and for Spanish speaking communities. Neuropsychologia. 2001;39:1150–7. approved the final manuscript. 16. Cameron J, Worrall-Carter L, Page K, Stewart S, Ski CF. Screening for mild cognitive impairment in patients with heart failure: Montreal Competing interests cognitive assessment versus mini mental state exam. Eur J Cardiovasc The authors declare that they have no competing interests in this work. Nurs. 2013;12:252–60. 17. National Collaborating Centre for Mental Health (UK). Dementia: a NICE-SCIE Consent for publication Guideline on Supporting People With Dementia and Their Carers in Health Not applicable. and Social Care. Leicester: NICE Guidelines; 2011. González-Moneo et al. BMC Cardiovascular Disorders (2016) 16:163 Page 10 of 10 18. Cabanero-Martinez MJ, Cabrero-Garcia J, Richart-Martinez M, Munoz- 42. Hofmeijer J, Mulder AT, Farinha AC, Van Putten MJ, Le FJ. Mild hypoxia Mendoza CL. The Spanish versions of the Barthel index (BI) and the Katz affects synaptic connectivity in cultured neuronal networks. Brain Res. index (KI) of activities of daily living (ADL): a structured review. Arch 2014;1557:180–9. Gerontol Geriatr. 2009;49:e77–84. 43. Kerola T, Nieminen T, Hartikainen S, Sulkava R, Vuolteenaho O, Kettunen R. 19. Delgado ALJ. Validez y fiabilidad del cuestionario de función familiar B-type natriuretic peptide as a predictor of declining cognitive function and Apgar-familiar. Aten Primaria. 1996;18:289–95. dementia–a cohort study of an elderly general population with a 5-year 20. Ramos Brieva JA, Montejo Iglesias ML, Lafuente LR, Ponce De Leon HC, follow-up. Ann Med. 2010;42:207–15. Moreno SA. Validation of the Geriatric Depression Screening Scale. Actas 44. Azab B, Bibawy J, Harris K, et al. Value of albumin-globulin ratio as a Luso Esp Neurol Psiquiatr Cienc Afines. 1991;19:174–7. predictor of all-cause mortality after non-ST elevation myocardial infarction. 21. Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a Angiology. 2013;64:137–45. geriatric depression screening scale: a preliminary report. J Psychiatr Res. 45. Uthamalingam S, Kandala J, Daley M, et al. Serum albumin and mortality in 1982;17:37–49. acutely decompensated heart failure. Am Heart J. 2010;160:1149–55. 22. Graven LJ, Grant JS. Social support and self-care behaviors in individuals 46. Rogers TK, Bowman CE. Cognitive impairment associated with beta- with heart failure: an integrative review. Int J Nurs Stud. 2014;51:320–33. blockade in the elderly. Postgrad Med J. 1990;66:1050–2. 23. Hajduk AM, Lemon SC, McManus DD, et al. Cognitive impairment and 47. Solomon S, Hotchkiss E, Saravay SM, Bayer C, Ramsey P, Blum RS. self-care in heart failure. Clin Epidemiol. 2013;5:407–16. Impairment of memory function by antihypertensive medication. Arch Gen 24. Maeda U, Shen BJ, Schwarz ER, Farrell KA, Mallon S. Self-efficacy mediates Psychiatry. 1983;40:1109–12. the associations of social support and depression with treatment adherence 48. Kumar A, Dogra S. Neuroprotective effect of carvedilol, an adrenergic in heart failure patients. Int J Behav Med. 2013;20:88–96. antagonist against colchicine induced cognitive impairment and oxidative 25. McCullough PA, Philbin EF, Spertus JA, Kaatz S, Sandberg KR, Weaver WD. damage in rat. Pharmacol Biochem Behav. 2009;92:25–31. Confirmation of a heart failure epidemic: findings from the Resource 49. Wagner G, Icks A, Abholz HH, Schroder-Bernhardi D, Rathmann W, Kostev K. Utilization Among Congestive Heart Failure (REACH) study. J Am Coll Antihypertensive treatment and risk of dementia: a retrospective database Cardiol. 2002;39:60–9. study. Int J Clin Pharmacol Ther. 2012;50:195–201. 26. Broadhead WE, Gehlbach SH, De Gruy FV, Kaplan BH. The Duke-UNC 50. Haring B, Leng X, Robinson J, et al. Cardiovascular disease and cognitive Functional Social Support Questionnaire. Measurement of social support in decline in postmenopausal women: results from the Women’s Health family medicine patients. Med Care. 1988;26:709–23. Initiative Memory Study. J Am Heart Assoc. 2013;2, e000369. 27. Barber JH, Wallis JB, McKeating E. A postal screening questionnaire in 51. Zuccala G, Marzetti E, Cesari M, et al. Correlates of cognitive impairment preventive geriatric care. J R Coll Gen Pract. 1980;30:49–51. among patients with heart failure: results of a multicenter survey. Am J 28. Larizgoitia A, Larizgoitia I. Adaptación en nuestro medio de una encuesta Med. 2005;118:496–502. para la detección de ancianos con riesgo de dependencia. Rev Gerontol. 52. Ottenbacher KJ, Karmarkar A, Graham JE, et al. Thirty-day hospital 1996;6:224–31. readmission following discharge from postacute rehabilitation in fee-for- 29. De Vries NM, Staal JB, Van Ravensberg CD, Hobbelen JS, Olde Rikkert MG, service Medicare patients. JAMA. 2014;311:604–14. Nijhuis-van der Sanden MW. Outcome instruments to measure frailty: a 53. Hawkins MA, Dolansky MA, Schaefer JT, et al. Cognitive Function in Heart systematic review. Ageing Res Rev. 2011;10:104–14. Failure Is Associated With Nonsomatic Symptoms of Depression But Not 30. Swedberg K, Komajda M, Bohm M, et al. Effects on outcomes of heart rate Somatic Symptoms. J Cardiovasc Nurs. 2015;30:E9–E17. reduction by ivabradine in patients with congestive heart failure: is there an 54. Newman AB, Fitzpatrick AL, Lopez O, et al. Dementia and Alzheimer’s influence of beta-blocker dose?: findings from the SHIFT (Systolic Heart disease incidence in relationship to cardiovascular disease in the failure treatment with the I (f) inhibitor ivabradine Trial) study. J Am Coll Cardiovascular Health Study cohort. J Am Geriatr Soc. 2005;53:1101–7. Cardiol. 2012;59:1938–45. 55. Carpenter I, Gambassi G, Topinkova E, et al. Community care in Europe. The 31. Graham S, Ye S, Qian M, et al. Cognitive function in ambulatory patients Aged in Home Care project (AdHOC). Aging Clin Exp Res. 2004;16:259–69. with systolic heart failure: insights from the warfarin versus aspirin in 56. Kono Y, Yamada S, Iwatsu K, et al. Predictive value of functional limitation reduced cardiac ejection fraction (WARCEF) trial. PLoS ONE. 2014;9, e113447. for disease severity in patients with mild chronic heart failure. J Cardiol. 32. Lees R, Selvarajah J, Fenton C, et al. Test accuracy of cognitive screening 2012;60:411–5. tests for diagnosis of dementia and multidomain cognitive impairment in 57. Afilalo J, Karunananthan S, Eisenberg MJ, Alexander KP, Bergman H. Role of stroke. Stroke. 2014;45:3008–18. frailty in patients with cardiovascular disease. Am J Cardiol. 2009;103:1616–21. 33. Zhou Y, Ortiz F, Nunez C, et al. Use of the MoCA in Detecting Early Alzheimer’s Disease in a Spanish-Speaking Population with Varied Levels of Education. Dement Geriatr Cogn Dis Extra. 2015;5:85–95. 34. Hanon O, Vidal JS, De GP, et al. Prevalence of memory disorders in ambulatory patients aged >/=70 years with chronic heart failure (from the EFICARE study). Am J Cardiol. 2014;113:1205–10. 35. Chaudhry SI, McAvay G, Chen S, et al. Risk factors for hospital admission among older persons with newly diagnosed heart failure: findings from the Cardiovascular Health Study. J Am Coll Cardiol. 2013;61:635–42. 36. Ganguli M, Fu B, Snitz BE, Hughes TF, Chang CC. Mild cognitive impairment: incidence and vascular risk factors in a population-based cohort. Neurology. 2013;80:2112–20. 37. Davis KK, Allen JK. Identifying cognitive impairment in heart failure: a review of screening measures. Heart Lung. 2013;42:92–7. 38. Hajduk AM, Kiefe CI, Person SD, Gore JG, Saczynski JS. Cognitive change in heart failure: a systematic review. Circ Cardiovasc Qual Outcomes. 2013;6:451–60. 39. Petersen RC, Stevens JC, Ganguli M, Tangalos EG, Cummings JL, DeKosky ST. Practice parameter: early detection of dementia: mild cognitive impairment (an evidence-based review). Report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology. 2001;56:1133–42. 40. Arslanian-Engoren C, Giordani BJ, Algase D, Schuh A, Lee C, Moser DK. Cognitive dysfunction in older adults hospitalized for acute heart failure. J Card Fail. 2014;20:669–78. 41. Almeida OP, Garrido GJ, Beer C, Lautenschlager NT, Arnolda L, Flicker L. Cognitive and brain changes associated with ischaemic heart disease and heart failure. Eur Heart J. 2012;33:1769–76.