bs_bs_banner

Geriatr Gerontol Int 2015; 15: 72–79

ORIGINAL ARTICLE: EPIDEMIOLOGY, CLINICAL PRACTICE AND HEALTH

Association between body mass index and cause-specific mortality as well as hospitalization in frail Chinese older adults Tuen-Ching Chan,1,2 James Ka Hay Luk,1 Leung-Wing Chu2 and Felix Hon Wai Chan1 1

Department of Medicine and Geriatrics, Fung Yiu King Hospital, and 2Division of Geriatric Medicine, Department of Medicine, Queen Mary Hospital, LKS Faculty of Medicine, The University of Hong Kong, Hong Kong

Aim: A U-shaped relationship between body mass index (BMI) and all-cause mortality has been reported, but there are few studies examining the association between BMI and cause-specific mortality and hospitalization. We carried out a longitudinal study to examine these associations in Chinese older adults with multiple comorbidities, which could provide a reference for the recommended BMI in this population. Methods: From 2004 to 2013, a retrospective cohort of Chinese older adults was selected from a geriatric day hospital in Hong Kong. They were divided into groups according to their BMI: BMI 30. Other assessments included medical, functional, cognitive, social and nutritional assessment. Results: A total of 1747 older adults (mean age 80.8 ± 7.1 years, 44.1% male, 46.1% living in nursing homes, Charlson Comorbidity Index 2.0 ± 1.6) with a median follow up of 3.5 years were included. Older adults with BMI 24–28 had the lowest all-cause, infection-related and cardiovascular mortality (P < 0.001). Multivariate analysis showed that there was an inverted J-shaped association between BMI and hazard ratio for all-cause and infectionrelated mortality in both nursing home and community-dwelling older adults. The rate of all-cause hospitalization was lower in older adults with BMI 22–28 (P = 0.002). Multivariate analysis showed that there was an inverted J-shaped association between the odds ratio of recurrent hospitalization and BMI. Conclusion: Chinese older adults with BMI 24–28 had lower all-cause mortality, infection-related mortality, cardiovascular-related mortality and all-cause hospitalization. This study provides a reference for the recommended BMI in this population. Geriatr Gerontol Int 2015; 15: 72–79. Keywords: body mass index, Chinese older adults, hospitalization, mortality, obesity, overweight.

Background The prevalence of obesity in nursing home residents in the USA has increased from 16.9% to 25.8% over the past decade.1,2 Epidemiological studies have documented that obesity is an independent risk factor for diabetes, coronary heart disease, stroke and mortality.3,4 In contrast, being underweight is also common in the

Accepted for publication 25 November 2013. Correspondence: Dr Tuen-Ching Chan MBBS FHKAM, The University of Hong Kong Division of Geriatrics, Department of Medicine, Queen Mary Hospital, The University of Hong Kong, Pokfulam Road, Hong Kong SAR, China. Email: [email protected]

72

|

doi: 10.1111/ggi.12230

geriatric population, with 8% of community-dwelling older adults and 19.7% of nursing home older adults having a body mass index (BMI) less than 18.5.5,6 Being underweight has been associated with increased mortality in different studies, and the association is even greater than with obesity.7 These observations have contributed to the U-shaped association between BMI and all-cause mortality reported in different studies.8,9 However, most of these studies were carried out in healthy adults, and there are few studies examining the association between BMI and cause-specific mortality as well as hospitalization. Furthermore, most studies were carried out in Caucasian older adults, with few examining Chinese older adults. The association could provide a reference for the recommended BMI in this population, and might help with implementing different © 2014 Japan Geriatrics Society

BMI, mortality and hospitalization

nutritional interventions to reduce mortality as well as hospitalization. Hence, we carried out a longitudinal study to examine these associations in Chinese older adults with multiple comorbidities.

Methodology We carried out a retrospective cohort study examining older adults from the Geriatric Day Hospital (GDH) of Fung Yiu King Hospital (FYKH) in the Hong Kong West Cluster, one of seven major health districts in Hong Kong under the Hospital Authority. The Hospital Authority provides a public hospital service for all Hong Kong citizens. The GDH of FYKH is a 22-place day hospital that receives one to two new case referrals each day. Older adults are referred to the GDH for multidisciplinary rehabilitation.10 Each patient attends two times per week, and receives a total of 10–16 sessions of rehabilitation. Each session involves 1.5 h of physiotherapy in the morning and 1.5 h of occupational therapy in the afternoon, with 2 h of rest in between. Measurements of bodyweight and body height, and a multidisciplinary assessment are made on the older adult’s first visit and at discharge. These assessments include medical, functional, cognitive, social and nutritional assessment. The inclusion criteria for the study were all Chinese older adults ≥65 years of age attending the GDH of FYKH from 2004 to 2011. They were divided into groups according to their BMI at discharge from the GDH: BMI 30. Data retrieved from the medical records of the older adults at discharge from the GDH included demographic data, medical status, functional status, cognitive status, social status and serum albumin. Demographic data included age and sex. For medical status, the number and the type of comorbidities were quantified using the Charlson Comorbidity Index, which is a valid and reliable method to measure comorbidity.11–14 In the Charlson Comorbidity Index, each comorbidity is assigned a score of 1, 2, 3 or 6 depending on the risk of death associated with this condition. As anemia and chronic renal disease are common and important prognostic indicators for older adults, serum hemoglobin level and estimated glomerular filtration rate (eGFR; estimated by Modified Diet for Renal Disease equation adjusted for the Chinese population) were also recorded.15–19 Functional status was assessed using the functional independence measure, which is a valid and reliable method of measuring functional ability in older adults.20–22 The functional independence measure has a total of 18 items divided into six major subgroups: self-care ability (eating, grooming, bathing, dressing upper body, dressing lower body, toileting), sphincter control (bladder control and bowel control), transfer (bed/chair transfer, © 2014 Japan Geriatrics Society

toilet transfer, bathtub transfer), locomotion (walking on level ground, stair use), communication (comprehension and expression) and social/cognition (social interaction, problem-solving, memory). Each item receives a score from 1 (completely dependent) to 7 (completely independent). Cognitive status was assessed using the Chinese version of the Mini-Mental State Examination (C-MMSE) and the abbreviated mental test (AMT), which are both validated screening tests for cognitive impairment in Chinese older people.23–25 With regard to the social status, we documented whether the older adults were living at home or in a nursing home, as this might affect mortality and hospitalization.26,27 Serum albumin, which could provide evidence about nutritional status, was also recorded.28 The clinical status of each older adult after discharge from the GDH up until May 2013 was retrieved from the computer management system (CMS), which has complete hospitalization and mortality records for each patient. Data was collected on the number of nonscheduled hospitalizations that occurred 1 year before discharge from the GDH, as this is a predictor for mortality.29 The main outcome measures were mortality and hospitalization. Cases of mortality were reviewed with cause of mortality confirmed according to the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM). They were broadly divided into all-cause mortality, infection-related mortality and cardiovascular-related mortality. In order to avoid the effect of misclassification of BMI on outcomes, we followed suggestions from Manson et al. and Lin et al., whereby mortality in the early phase of follow up (6 months) was excluded from analysis.30,31 Each hospitalization was reviewed and recurrent hospitalization was defined, according to Stott et al., as two or more hospitalizations in 1 year.32 The longitudinal study was approved by the institutional review boards of the University of Hong Kong and Hospital Authority Hong Kong West Cluster. Statistical Package for Social Science (Windows version 18; SPSS, Chicago, IL, USA) was used for statistical analysis. Continuous variables were expressed as either mean ± standard deviation of the mean (mean ± SD) or median with interquartile range as appropriate. The Kruskal–Wallis test was used to compare the rate of hospitalization between groups. The association between different factors (including BMI, age, sex, Charlson Comorbidity Index, functional independence measure, MMSE, residence status [nursing home vs home), serum albumin, serum haemoglobin and eGFR) and mortality was estimated using univariate Cox regression. Significant factors identified from univariate analysis were further analyzed using multivariate stepwise Cox regression. The association between different factors and hospitalization was |

73

T-C Chan et al.

estimated using univariate logistic regression. Significant factors identified from univariate analysis were further analyzed using multivariate stepwise logistic regression. The World Health Organization has suggested that the normal BMI range for Chinese adults is BMI 20–22, and this was used as the reference group during multivariate analysis.33 Statistical significance was inferred by a two-tailed P-value of 0.05 or less.

Results Baseline characteristics A total of 1747 older adults were included (BMI 30, n = 63 [3.6%]). The mean age was 80.8 ± 7.1 years, and 770 (44.1%) participants were male. There were 805 (46.1%) participants living in nursing homes. Other clinical characteristics are summarized in Tables 1 and 2.

Mortality After a median follow-up period of 3.5 years (range 1.5–8 years), 581 (33.3%) of the 1747 older adults had died. The leading causes of mortality were infection (n = 336, 57.8%), followed by cardiovascular disease (n = 182, 31.3%) and malignancy (n = 64, 11.0%). For BMI 30, 50.8%, 48.3%, 38.3%, 34.7%, 23.7%, 29.1%, 20.6%, 22.4%, and 28.6% of older adults had died, respectively (P < 0.001). After multivariate analysis using Cox regression, when compared with BMI 20–22, there was an inverted J-shaped association between hazard ratio and BMI for all-cause mortality (Fig. 1). A similar inverted-J shaped association was seen in subgroup analysis for residence status (nursing home older adults and community-dwelling older adults) (Fig. 2). For infection-related mortality, there was also an inverted-J shaped association between hazard ratio and BMI (Fig. 3a). For cardiovascular-related mortality, the hazard ratio was generally lower in older adults with BMI more than 22 (Fig. 3b).

Hospitalization There were 6390 all-cause hospitalizations during follow up. The median number of hospitalizations per year was lower in older adults with BMI 22–28 and higher in other groups (BMI ≤16–22, median 1, interquartile range 0–2; BMI 22.1–28, median 0, interquartile range 0–1; BMI ≥28, median 1, 74

|

Table 1 Baseline characteristics of 1747 older adults† Factor Body mass index 30 Age Sex (male) Functional independence measure Mini-Mental State Examination Abbreviated mental test Serum creatinine (μmol/L) Estimated glomerular filtration rate (mL/min/1.73 m2) Serum hemoglobin (g/dL) Serum albumin (g/L) Nursing home residence Charlson Comorbidity Index Hypertension Diabetes Stroke Ischemic heart disease Congestive heart failure Atrial fibrillation Dementia Cirrhosis Chronic obstructive pulmonary disease Renal impairment No. hospitalizations in the preceding year †

120 (6.9) 180 (10.3) 274 (15.7) 366 (21.0) 316 (18.1) 220 (12.6) 141 (8.1) 67 (3.8) 63 (3.6) 80.8 ± 67.1 770 (44.1) 88.2 ± 18.0 17.7 ± 6.3 6.1 ± 2.1 97.3 ± 56.0 79.6 ± 32.4 12.0 ± 1.7 36.5 ± 5.2 805 (46.1) 2.0 ± 1.6 1135 (65.0) 557 (31.9) 556 (31.8) 257 (14.7) 163 (9.3) 220 (12.6) 335 (19.2) 10 (0.5) 184 (10.5) 288 (16.5) 1 (0–1)

Mean ± standard deviation or n (%).

interquartile range 0–1.5; P = 0.002). After multivariate analysis using logistic regression, there was an inverted J-shaped association between odds ratio of recurrent hospitalization and BMI (Fig. 4).

Discussion The present study showed that, in both Chinese community-dwelling and Chinese nursing home older adults with multiple comorbidities, those with BMI 24–28 had the lowest mortality. According to the World Health Organization BMI classification, BMI 24–28 is regarded as overweight in Asians.33 Our findings concurred with those from studies focusing on communitydwelling healthy older adults. In the longitudinal study carried out by Lin et al., for people aged over 65 years, © 2014 Japan Geriatrics Society

© 2014 Japan Geriatrics Society

Hazard ratio

2

1

0 30

BMI

Figure 1 Association between body mass index (BMI) and all-cause mortality among 1747 older adults. *Multivariate analysis by Cox regression. Hazard ratio adjusted for age, sex, Charlson Comorbidity Index, functional independence measure, residence status, serum albumin, serum haemoglobin and estimated glomerular filtration rate.

2 Home nursing home

Hazard ratio

Albumin unit: g/L. BMI, body mass index; CCI, Charlson Comorbidity Index; FIM, Functional independence measure; MMSE, Mini-Mental State Examination. †

78.9 ± 7.2 29.4 (19) 87.7 ± 19.7 18.0 ± 6.9 2.3 ± 2.0 37.0 ± 4.9 23.3 (15) 78.9 ± 6.3 33.3 (47) 89.3 ± 17.6 18.6 ± 5.3 2.0 ± 1.8 37.7 ± 4.9 27.8 (39) 80.9 ± 7.1 46.7 (128) 88.3 ± 17.9 17.9 ± 6.3 2.0 ± 1.9 36.8 ± 4.9 42.5 (116) 82.8 ± 7.0 37.6 (68) 87.3 ± 19.4 17.2 ± 6.3 1.7 ± 1.5 34.8 ± 5.2 54.1 (97) Age Sex (male) FIM MMSE CCI Albumin† Institutionalized

82.7 ± 6.2 32.5 (39) 88.4 ± 18.8 17.4 ± 6.2 1.6 ± 1.4 34.2 ± 5.1 42.5 (94)

80.5 ± 6.4 49.7 (182) 88.3 ± 17.4 17.6 ± 6.4 2.0 ± 1.7 36.9 ± 5.1 38.6 (141)

80.0 ± 6.3 38.3 (120) 89.2 ± 16.8 18.2 ± 6.6 2.0 ± 1.8 37.6 ± 4.8 34.4 (109)

79.7 ± 7.5 45.8 (101) 91.2 ± 18.2 18.5 ± 6.4 2.1 ± 1.9 37.4 ± 5.0 27.3 (60)

79.9 ± 6.0 22.6 (15) 89.2 ± 16.3 18.1 ± 5.6 2.3 ± 1.6 36.8 ± 4.8 22.6 (15)

>30 (n = 63) 28–30 (n = 67) 26–28 (n = 141) 24–26 (n = 220) 22–24 (n = 316) 20–22 (n = 366) 18–20 (n = 274) 16–18 (n = 180) BMI

Association between body mass index and cause-specific mortality as well as hospitalization in frail Chinese older adults.

A U-shaped relationship between body mass index (BMI) and all-cause mortality has been reported, but there are few studies examining the association b...
226KB Sizes 0 Downloads 0 Views