Comprehensive Geriatric Assessment is a ... - Wiley Online Library

4 downloads 0 Views 243KB Size Report
tool for postoperative delirium after gastrointestinal surgery in old-old adults .... university hospital, almost all were elective operation cases. Furthermore, the ...
bs_bs_banner

Geriatr Gerontol Int 2016; 16: 1036–1042

ORIGINAL ARTICLE: EPIDEMIOLOGY, CLINICAL PRACTICE AND HEALTH

Comprehensive Geriatric Assessment is a useful predictive tool for postoperative delirium after gastrointestinal surgery in old-old adults Yoshihiro Maekawa,1 Ken Sugimoto,1 Makoto Yamasaki,2 Yasushi Takeya,1 Koichi Yamamoto,1 Mitsuru Ohishi,3 Toshio Ogihara,4 Ayumi Shintani,5 Yuichiro Doki,2 Masaki Mori2 and Hiromi Rakugi1 1

Geriatric Medicine, 2Gastroenterological Surgery, 5Clinical Epidemiology and Biostatistics, Osaka University Graduate School of Medicine, Suita, 3Cardiovascular Medicine and Hypertension, Graduate School of Medical and Dental Sciences, Kagoshima University, Kagoshima, and 4Graduate School of Health Sciences, Morinomiya University of Medical Sciences, Osaka, Japan

Aim: To determine whether carrying out the Comprehensive Geriatric Assessment before operations would be useful for predicting complications, particularly postoperative delirium (POD), in old-old patients. Methods: A total of 517 patients aged 75 years and older, who underwent radical surgery for gastrointestinal cancer at Osaka University Hospital, were recruited for this observational study. The Comprehensive Geriatric Assessment components and assessment of performance status were carried out before surgery, and a record of postoperative complications including POD was made prospectively until discharge from hospital. The following morphological and clinical measurements were also obtained from the medical records: age, sex, disease type, previous history, comorbid lifestyle-related diseases, POD, postoperative complications, operative method, duration of operation, hemorrhage volume, blood transfusion volume, method of anesthesia, body mass index and blood tests. Results: POD appeared in 24.0% of the 517 patients who underwent surgery. Barthel Index, Mini-Mental State Examination, instrumental activities of daily living and Geriatric Depression Scale results were associated with the incidence of POD, and the Barthel Index, Mini-Mental State Examination and Instrumental Activities of Daily Living results were extracted as independent factors associated with the development of POD after adjusting for traditional risk factors for postoperative complications and performance status. Conclusions: The Comprehensive Geriatric Assessment before gastrointestinal surgery can be a useful tool for predicting the development of POD in old-old patients. Geriatr Gerontol Int 2016; 16: 1036–1042. Keywords: Comprehensive Geriatric Assessment, gastrointestinal surgery, old-old adults, postoperative complications, postoperative delirium.

Introduction The safety of surgical operations has improved considerably as a result of dramatic breakthroughs in medicine, which has led to an increase in the number of surgical operations among older adults. However, Japan

Accepted for publication 26 June 2015. Correspondence: Dr Ken Sugimoto MD, PhD, Geriatric Medicine, Osaka University Graduate School of Medicine, 2-2, Yamada-Oka, Suita, Osaka 565-0871, Japan. Email: [email protected]

1036 | doi: 10.1111/ggi.12587

is now confronting a super-aged society. Although there is no evidence of an increase in the rate of death directly caused by surgeries among older adults, there has been an increase in reports of serious postoperative complications in these patients.1 Currently, operational risks for older patients are evaluated on a case-by-case basis, because there are no fixed standards to serve as criteria for such assessments, which would serve as the basis for assessing the operability of older patients. The most commonly used method for assessing the operability of a patient is performance status (PS); however, there have been many reports that PS alone is insufficient for a

© 2015 The Authors. Geriatrics & Gerontology International published by Wiley Publishing Asia Pty Ltd on behalf of Japan Geriatrics Society.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

CGA as a useful predictor for POD

comprehensive assessment of operability, especially in older patients.2 Furthermore, because postoperative complications include not only complications directly related to the operation itself, but also other complications, such as postoperative delirium (POD), which is considered to be frequent among older patients; the topic could require special attention.3 Indeed, there have been some reports that have focused on patients’ age and pre-existing comorbidities as the main predictors of adverse postoperative outcomes in older patients after surgery.4 In particular, POD has received substantial attention as a complication that extends the postoperative observation period, especially in older patients, because POD is associated with high mortality5 and morbidity6 after discharge. Therefore, establishing a simple and beneficial assessment that predicts POD would be useful for shortening the postoperative observation period and improving patient prognosis. The Comprehensive Geriatric Assessment (CGA) might be useful for simultaneously evaluating both cognitive function and frailty using a simple questionnaire. It is widely used as a tool for extracting problems in daily life and achieving holistic medicine in older adults. CGA makes it possible to extract problems in daily life, and because this enables specific measures and management programs for disease management and daily living assistance, CGA has become a useful tool for holistic medicine, which has also led to its increased diversity. Furthermore, the impact of CGA7,8 on mortality and morbidity has been reported in many studies.9 However, the impact of CGA as a predicator of POD has not yet been clarified, especially in the old-old patients who are candidates for cancer surgery. Therefore, we carried out an observational study to consider whether carrying out CGA before operations would be useful for predicting the development of post-

operative complications, especially POD among old-old adults who expect to undergo digestive surgical operations for gastrointestinal cancers.

Methods Participants A total of 517 patients aged 75 years and older, who underwent radical surgery for gastrointestinal cancer between December 2005 and December 2013 at Osaka University Hospital and satisfied the criteria shown in Figure 1 were recruited for the present study. This study was approved by the ethics committee of Osaka University Hospital as the protocol entitled “the observational study for discovery of the beneficial factors influencing on postoperative course in the old subjects undergoing surgery,” and informed consent for gastrointestinal surgery and oral consent for the CGA were obtained from each patient. The recruitment period of this observational study is until March 2016, and the follow-up period is until March 2017.

Measurements Based on a request from the Department of Gastrointestinal Surgery at Osaka University Graduate School of Medicine, before admission or within 7 days of hospitalization, CGA was carried out in each patient. Two assistants who were previously trained by an experienced research investigator according to the criteria set in the CGA and PS manuals carried out the following assessments: Mini-Mental State Examination (MMSE), Geriatric Depression Score (GDS), Vitality Index, Barthel Index, instrumental activities of daily living (IADL; calculated IADL: the score divided by 5 in

Figure 1 study. © 2015 The Authors. Geriatrics & Gerontology International published by Wiley Publishing Asia Pty Ltd on behalf of Japan Geriatrics Society

Flow of patients in the

|

1037

Y Maekawa et al.

males, 8 in females, respectively) was used for analyses in the current study.10 In addition, the PS (Eastern Cooperative Oncology Group) parameters were as follows: fully active with no restrictions on activities: 0; unable to do strenuous activities, but able to carry out light housework and sedentary activities, 1; able to walk and manage self-care but unable to work and out of bed more than 50% of waking hours, 2; confined to bed or a chair more than 50% of waking hours and capable of limited self-care, 3; completely disabled and totally confined to a bed or chair and unable to do any self-care, 4; and death, 5. The following morphological and clinical measurements were also obtained from the medical records: age, sex, disease type (esophagus, stomach, liver, colon), previous history (cerebrovascular or cardiovascular diseases), comorbid lifestyle-related diseases, POD, postoperative complications, operative method (celiotomy or laparoscopy), duration of operation, hemorrhage volume, blood transfusion volume, method of anesthesia, body mass index (BMI) and blood tests (albumin [Alb], serum sodium, hemoglobin [Hb], serum creatinine [Cr], C-reactive protein).

Diagnosis of postoperative complications and POD The outcomes of the present observational study were postoperative complications including delirium, and short- and long-term prognosis; however, we focused only on postoperative outcomes in the current study. The assessment of postoperative complications and duration of hospitalization were extracted from the hospital medical records in accordance with previous reports. Delirium was diagnosed according to the Confusion Assessment Method (CAM) algorithm: acute onset and fluctuation, inattention, disorganized thinking, and/or altered level of consciousness.11 Delirium was evaluated by two independent doctors who were previously trained according to the criteria set in the CAM manual. Patients who scored positive on at least one CAM assessment were diagnosed with delirium after discussion between the two doctors. POD was defined as delirium observed from the day just after the operation to the day before discharge.

Statistical analysis We used proportional odds logistic regression to assess the independent effect of CGA components on POD, with the incidence of POD as the dependent variable. Proportional odds logistic regression, also known as ordinal logistic regression, is a popular model for ordinal categorical outcome variables. In addition to the CGA components (Barthel Index [the scores divided by 5], Vitality Index, MMSE, calculated IADL and GDS), 11 other covariates were 1038

|

included in the multivariable model: age (continuous), sex (dichotomous), type of surgery (dichotomous; celiotomy or laparoscopic surgery), operative duration (continuous), anesthesia type (dichotomous), transfusion (dichotomous), BMI, serum albumin, Hb concentration and serum Cr level (continuous), and performance status (PS; continuous). These covariates were selected based on clinical judgment; age, type of surgery, operative duration, anesthesia type, transfusion, BMI, serum albumin, Hb concentration and serum Cr level are known risk factors for the development of postoperative complications including delirium. PS was specifically selected to evaluate the advantage of the CGA components, because it is a popular and widely accepted preoperative assessment. All analyses were carried out using JMP Pro (version 11; SAS Institutes, Cary, NC, USA), and are presented as the mean ± SD. All statistical inferences were made at two-sided 5% significance level.

Results Patients in the present study Although there were initially 736 entries, some were excluded as shown in Figure 1, leaving 580 cases of gastrointestinal surgery carried out on patients aged 75 years and older as eligible participants of the present study. The mean age of the 517 participants that were finally recruited for the study was 79.3 ± 3.6 years, and the male-to-female ratio was 378:182. There were 215 cases of upper gastrointestinal cancer (stomach/ duodenum 142, esophagus 73), 174 cases of lower gastrointestinal cancer, and 128 cases of liver, gallbladder, bile duct and pancreas disease (liver 68, gallbladder/bile duct 20, pancreas 40). Because the cases were from a university hospital, almost all were elective operation cases. Furthermore, the attending surgeons had diagnosed them to be operable during the patients’ outpatient visits.

Association between factors including PS, CGA and POD A total of 124 patients (24.0%) were observed with POD, and 26 patients (5.0%) with anastomotic leakage, 25 (4.8%) with pneumonia, 62 (12.0%) with surgical cite infections, 33 (6.4%) with cardiovascular complications, 11 (2.1%) with bleedings and 20 (3.9%) with Ileus. Mean ± SD of days of the onset of POD was 2.95 ± 3.09 days after surgery. The average of duration of POD was 3.81 ± 4.29 days. As shown in Table 1, patients who developed POD showed higher values on the PS and GDS, and lower values on the Barthel Index, Vitality Index, MMSE and calculated IADL compared with patients who did not develop POD.

© 2015 The Authors. Geriatrics & Gerontology International published by Wiley Publishing Asia Pty Ltd on behalf of Japan Geriatrics Society

CGA as a useful predictor for POD

Table 1 Demographics, surgical history, laboratory findings and Comprehensive Geriatric Assessment characteristics of the study population

Factors Age (years) Sex, male (%) BMI (kg/m2) Operative duration (min) Operative procedure Celiotomy, n (%) Laparoscopic surgery, n (%) Hemorrhage volume (mL) Blood transfusion, n (%) Amount of blood transfusion (mL) Anesthesia Epidural anesthesia, n (%) TAP block, n (%) Previous history Cerebrovascular disease, n (%) Cardiovascular disease, n (%) Lifestyle-related disease Hypertension, n (%) Dyslipidemia, n (%) Diabetes mellitus, n (%) Blood test Alb (mg/dL) Na (mEq/L) Hb (g/dL) Cr (mg/dL) CRP (mg/dL) Preoperative assessments PS CGA Barthel Index Vitality Index MMSE Calculated IADL† GDS

All

No delirium

Delirium

(n = 517) 79.3 ± 3.6 351 (67.9) 22.1 ± 3.1 283 ± 138

(n = 393) 79.1 ± 3.5 270 (68.7) 22.3 ± 3.1 278 ± 135

(n = 124) 79.9 ± 3.7 81 (65.3) 21.5 ± 3.3 296 ± 147

282 (54.5) 235 (45.5) 486 ± 857 132 (25.5) 193 ± 490

207 (52.7) 186 (47.3) 481 ± 888 98 (24.9) 186 ± 488

75 (60.5) 49 (39.5) 501 ± 749 34 (27.4) 213 ± 497

215 (41.6) 36 (7.0)

160 (40.7) 28 (7.1)

55 (44.4) 8 (6.5)

37 (7.2) 96 (18.6)

21 (5.3) 67 (17.1)

16 (12.9) 29 (23.4)

316 (61.1) 128 (24.8) 114 (22.1)

236 (60.1) 92 (23.4) 83 (21.1)

80 (64.5) 36 (29.0) 31 (25.0)

3.64 ± 0.47 138.5 ± 6.4 11.8 ± 1.8 0.90 ± 0.39 0.64 ± 1.60

3.69 ± 0.45 138.7 ± 7.0 11.9 ± 1.8 0.88 ± 0.31 0.56 ± 1.52

3.48 ± 0.52 138.2 ± 3.8 11.4 ± 1.9 0.96 ± 0.57 0.87 ± 1.85

0.35 ± 0.67

0.25 ± 0.55

0.68 ± 0.96

97.6 ± 8.6 9.88 ± 0.52 26.0 ± 3.4 0.92 ± 0.16 3.02 ± 2.91

98.8 ± 4.9 9.92 ± 0.38 26.9 ± 2.58 0.95 ± 0.11 2.64 ± 2.72

93.7 ± 15.1 9.74 ± 0.83 23.3 ± 5.28 0.83 ± 0.26 4.21 ± 3.42

Date presented as mean ± SD. †Calculated instrumental activities of daily living (IADL); the IADL score divided by 5 in males and 8 in females. Alb, albumin; BMI, body mass index; CGA, Comprehensive Geriatric Assessment; Cr, creatinine; CRP, C-reactive protein; GDS, Geriatric Depression Score; Hb, hemoglobin; IADL, Instrumental Activities of Daily Living; MMSE, Mini-Mental State Examination; Na, serum sodium; PS, performance status; TAP, transversus abdominis plane.

Regarding associations with items other than CGA, patients who developed POD often had low BMI, a history of cerebrovascular disease and cardiovascular disease, and had low levels of Alb and Hb, and high levels of Cr, which were most commonly observed in celiotomy cases. To clarify the impact of CGA as a predictor of POD in patients with gastrointestinal cancers, multiple logistic regression analyses were carried out. ADL (Barthel Index), Vitality Index, MMSE, calculated IADL and GDS were associated with POD based on

the unadjusted model as shown in Table 2. After adjusting for traditional risk factors of postoperative complications, such as type and duration of surgery, anesthesia type, or serum albumin, all of the CGA components were associated with POD as well as PS. After adjustment including PS, Barthel Index, MMSE, calculated IADL and GDS remained being statistically associated with POD (Barthel Index: adjusted HR 1.20, 95% CI 1.03–1.44, P = 0.0199; MMSE: adjusted HR 1.29, 95% CI 1.21–1.39, P < 0.0001; IADL: adjusted HR 1.46, 95% CI 1.22–1.77, P < 0.0001; GDS:

© 2015 The Authors. Geriatrics & Gerontology International published by Wiley Publishing Asia Pty Ltd on behalf of Japan Geriatrics Society

|

1039

Y Maekawa et al.

Table 2 Multiple logistic regression models for the factors associated with postoperative delirium Models

Factor Barthel Index ‡

Adjusted HR (95% CI) Adjusted P-value Comparison in HR with PS Adjusted HR (95% CI) Adjusted P-value

Vitality Index

1.20 (1.03–1.44)

1.19 (0.80–1.82)

0.0199

0.4003

0.67 (0.45–0.99)

0.53 (0.38–0.72)

0.0431

MMSE 1.29 (1.21–1.39)