A comparison of two pre-operative frailty measures in older surgical cancer patients
Article Outline
- Abstract
- 1. Introduction
- 2. Methods
- 3. Results
- 4. Discussion
- Funding
- Conflict of Interest
- Author Contributions
- References
- Copyright
Abstract
Background
Measuring frailty in older adults with cancer may identify patients with an increased risk of treatment complications. As it remains controversial how to identify frailty, the aim of this study was to compare a pre-operative multi-domain frailty measure based on a comprehensive geriatric assessment (CGA) to a modified version of the physical phenotype of frailty (PF) in a cohort of older adults with colorectal cancer, and to analyze the ability of the two classifications to predict post-operative complications and survival.
Methods
A prospective longitudinal study including 176 patients aged 70–94
years electively operated for colorectal cancer in three Norwegian hospitals. A pre-operative CGA, self-reported quality of life, and measurements of grip strength and gait speed were performed. CGA-frailty was defined as fulfilling one or more of the following criteria: dependency in activities of daily living, severe comorbidity, cognitive dysfunction, depression, malnutrition, or >seven daily medications. PF was defined with three or more of the following criteria: unintentional weight loss, exhaustion, low physical activity, impaired grip strength, or slow gait speed. Outcome measures were post-operative complications and survival.
Results
The agreement between the classifications was poor. CGA-frailty was identified in 75 (43%) patients, while PF was identified in 22 (13%) patients. Only CGA-frailty predicted post-operative complications [P=0.001]. Both measures predicted survival.
Conclusions
A multi-domain frailty measure based on a CGA was more useful than frailty identified from a modified version of the PF criteria in predicting post-operative complications. For overall survival, both frailty measures were predictive.
Keywords: Geriatric assessment, Frailty, Geriatric oncology, Geriatric surgery, Pre-operative evaluation, Post-operative complications, Colorectal cancer, Surgical risk
1. Introduction
Frailty is attracting attention from researchers and clinicians who deal with older patients. It is generally agreed that frailty describes “an elderly patient who is at heightened vulnerability to adverse health status change because of a multisystem reduction in reserve capacity”, but it remains controversial how to identify frailty in an individual patient.[1], [2] In older cancer patients, where treatment modalities such as surgery and chemotherapy may pose a high risk of complications and toxicity, identifying frailty is particularly important.3
Two vital parts of the decision making process in the older adult with cancer are assessments of treatment risks and remaining life expectancy, for which chronological age is an insufficient marker. Even though it is well known to surgeons that functional status is an important predictor of surgical outcomes, formal incorporation of preoperative functional status has not been established, and most surgeons do not measure baseline physical and cognitive function.4 Furthermore, when estimating remaining life expectancy, functional status, comorbidity, and frailty appear to be more important predictors than chronological age.[5], [6], [7]
Within geriatric oncology, a widely used concept of frailty is derived from criteria first described by Winograd and later modified by Balducci.[8], [9] Based on a comprehensive geriatric assessment (CGA), an older adult is considered frail when fulfilling any of the following criteria: dependency in activities of daily living (ADL), three or more comorbid illnesses, the presence of geriatric syndromes (for example dementia, malnutrition, depression, delirium, and falls), or age >85years. In this multi-component phenotype of frailty, deficits across different health domains, such as clinical, psychological, and functional, are considered to be predictors of treatment tolerance and life expectancy. This way of classifying frailty is somewhat similar to the accumulation of deficit definition that has been suggested by Rockwood and colleagues,10 where a frailty index is constructed based on a CGA.11 We have previously found that a multi-component frailty classification based on a CGA, but not increasing age or American Society of Anesthesiologists classification, predicted post-operative complications in a cohort of 178 older patients electively operated for colorectal cancer.12
Within the geriatric and biogerontological literature, a widely accepted definition of frailty is based on data from more than 5000 community dwelling individuals aged 65years and older who participated in the Cardiovascular Health Study.13 The physical phenotype of frailty (PF) based on data from this study was defined as fulfilling at least three of the following five criteria: unintentional weight loss, exhaustion, slow walking speed, low physical activity, and weakness. This approach to defining frailty does not include comorbidity and cognition, but highlights the association of frailty with physiological and metabolic changes leading to a loss of functional capacity. A physiologic loss of reserves is identified through clusters of physical impairments. PF predicts incident falls, hospitalizations, worsening mobility, and deaths in large cohorts.[13], [14] A recent publication found that PF predicted post-operative complications in 594 patients over 65years undergoing all types of elective surgery, but only when frailty was defined as fulfilling 4 to 5 of the proposed frailty criteria instead of the established cut-off of 3 to 5.15
To further investigate the clinical usefulness of two different approaches to identifying vulnerability in a homogeneous cohort of older adults with cancer undergoing surgery, we compared a multi-domain frailty classification based on CGA to a frailty classification based on a modified version of the PF in 176 patients aged 70 to 94years with stage I to IV colorectal cancer who were electively operated. A comparison of the ability of the two classifications to predict post-operative complications may indicate the most useful approach beyond chronological age and an established anesthesiology tool to identifying susceptible elderly cancer patients at risk for surgical morbidity.
2. Methods
2.1. Setting and Patients
The eligibility criteria for this study included: age 70years or older, being scheduled for surgery of a suspected or confirmed colorectal cancer in three Norwegian hospitals from November 2006 through June 2008, and ability to provide a written informed consent. The Regional Committee for Medical and Health Research Ethics in East Norway approved the study. The principal investigator, a medical doctor with training in geriatrics, performed the pre-operative assessments 0 to 14days prior to surgery at the hospitals.
2.2. CGA Measurements and CGA-classification
Functional dependence in ADL and instrumental ADL (IADL) were assessed using the Barthel Index (BI) and the Nottingham Extended Activities of Daily Living Scale (NEADL).[16], [17] Comorbidity was registered from hospital records, supplied with information from the patient interview, and scored by using the revised Cumulative Illness Rating Scale (CIRS) manual from 2008.[18], [19] The CIRS assesses fourteen organ systems, and comorbidity in each organ system is scored on a five-point scale ranging from grade 0 (no problem) to grade 4 (extremely severe/immediate treatment required/end organ failure/severely impaired function). The Mini Nutritional Assessment (MNA), Mini Mental State Examination, and the Geriatric Depression Scale were used to assess nutritional, cognitive, and emotional status, respectively.[20], [21], [22] The number of systemic drugs in daily use was recorded from the hospital records. A previous publication from the same cohort provides more detailed descriptions of the tools used in the CGA assessment and the reasoning behind the selected cut-off values.12
Based on the pre-operative CGA assessment, patients were categorized into the three groups “fit”, “intermediate” and “frail” using a modification of criteria proposed by Balducci and Extermann.[9], [23] CGA-frailty was defined as fulfilling one or more of the following criteria: dependency in ADL, severe comorbidity, cognitive dysfunction, depression, malnutrition, or >seven daily medications. The CGA-frailty defining criteria are presented in Table 1.
Table 1. Comparison of Balducci-criteria for frailty and frailty-criteria used in the current study.
| Characteristic | Balducci-criteria9 | CGA-frailty in current study |
|---|---|---|
| Age | ≥85y | Not used |
| ADL | Dependency for 1 or more | Dependency for 1 or more |
| Comorbidity | 3 or more | Severe comorbidity according to CIRS-scores |
| Geriatric syndromes | 1 or more of the following: | 1 or more of the following: |
| • Dementia | • MMSE score<24 (range 0–30) | |
| • Depression | • GDS score>13 (range 0–30⁎) | |
| • Malnutrition | • MNA score<17 (range 0–30) | |
| • Delirium | • Incontinence (ADL) | |
| • Osteoporosis | ||
| • Incontinence | ||
| • Falls | ||
| • Neglect and abuse | ||
| • Failure to thrive | ||
| Polypharmacy | Not used | >7 daily medications |
⁎Higher score indicates more serious disease. |
2.3. Frailty Indicators Used to Operationalize the Physical Frailty Phenotype13
In order to classify the patients according to PF, we used information from the CGA assessment, self-reported quality of life, and physical measurements. Information about weight loss was obtained from MNA; and a loss of greater than 3kg during the last 3months was considered a positive weight loss criterion. Maximal grip strength was measured on the dominant side with a handheld Jamar dynamometer, and the same cut-offs as described by Fried et al. according to gender and body mass index were used.13 Subjective exhaustion was based on the European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire C30, a widely used oncology tool measuring self-reported health during the last week.24 Questions 12 (“Have you felt weak?”) and 18 (“Were you tired?”) were used to measure exhaustion, while question 4 (“Do you need to stay in bed or a chair during the day?”) was used to measure low physical activity. Slowness was defined as a time of ≥19s on the timed up and go test, based on the cutoff that approximately identified the slowest quintile among the non-institutionalized participants in the Canadian Study of Health and Aging.25 In the original physical frailty phenotype definition, slowness was defined as the slowest 20% of the population walking 15ft. Frailty was subsequently defined as fulfilling three or more of these five criteria, while patients fulfilling one or two were considered pre-frail. Those with no characteristics were considered robust. As in the original paper by Fried and colleagues, patients considered evaluable for frailty had three or more non-missing frailty components.13 Table 2 compares the modified PF criteria to the original criteria presented by Fried and colleagues.
Table 2. Frailty-defining criteria adapted from the criteria based on the cardiovascular health study (physical frailty phenotype criteria).
| Characteristic | Physical Frailty Phenotype criteria | Criteria used in current study |
|---|---|---|
| Weight loss | Lost>10 lbs unintentionally in last y | Weight loss greater than 3kg during the last 3months (question B from MNA) |
| Weakness | Grip strength: lowest 20% (by gender, BMI) | Same as in physical frailty phenotype criteria |
| Exhaustion | Self-report of either: | Score 3 (“quite a bit”) or 4 (“very much”) on either question 12 (“Have you felt weak?”) or 18 (“Were you tired?”) from EORTC QLQ-C30 |
| • Feeling everything I did was an effort in the last wk | ||
| • Could not get going in the last wk | ||
| Slowness | Walking 4.57m: | TUG≥19s |
| • Time≥7s for height<160cm | ||
| • Time≥6s for height>159cm | ||
| Low activity | MLTA questionnaire (short version) | Score 3 or 4 on question 4 (“Do you need to stay in bed or a chair during the day?”) from EORTC QLQ-C30 |
| • Evaluating all 18 items | ||
| • <270kcals per wk on activity scale | ||
| Overall frailty status | Robust: 0 criteria | Same as in physical frailty phenotype criteria |
| Pre-frail: 1 or 2 criteria | ||
| Frail: 3 or more criteria |
2.4. Post-operative Measurements
Complications were broadly defined as any event occurring within 30days of surgery requiring treatment measures that are not routinely applied post-operatively for colorectal cancer. In order to present the morbidity data in a standardized and reproducible manner we classified complications as minor (grade I), potentially life-threatening without (grade II) or with (grade III) lasting disability, or fatal (grade IV) based on the morbidity grading system developed by Clavien et al.26 This classification of post-operative complications has been used in other studies of gastrointestinal surgery in the elderly,[27], [28] and allows for separate outcome variables according to the severity of morbidity. Examples of how to classify complications are available in the supplementary material. A comprehensive collection of morbidity data was ensured by extracting data from all available medical records and charts. In uncertain cases the medical team, nursing homes, caregivers, or patients were consulted. Post-operative mortality was defined as death within 30days after surgery.
Information on patients' deaths was retrieved from the National Registry of Norway. All deaths identified as of August 10, 2009 were included in the current analysis.
2.5. Statistical Analyses
For the frailty categorizations, the degrees of agreement between the two measures were tested with the Kappa Measure of Agreement.
Two dichotomized outcome variables were created for 30-day post-operative morbidity: “severe” complications (grade II or higher according to the classification by Clavien et al.26) versus “no/mild” complication, and “any” complication versus “no” complication. Chi-square tests for trend were used to compare the associations between these outcomes and the potentially predictive factors CGA-classification (fit, intermediate, or frail) and modified PF (robust, pre-frail, or frail). Rectal cancer was an independent predictor of severe post-operative complications in this dataset, and tumor location was therefore included in the multivariable models.12
Survival curves were estimated by the Kaplan–Meier method and compared by the log rank test. Overall survival was compared between age groups (70–79years compared to 80–94years), according to cancer stage (grouped into stages 0–II, III, and IV), according to CGA-classification (fit, intermediate, or frail), and according to the modified PF (robust, pre-frail, or frail). In order to calculate whether frailty predicted survival after correcting for cancer stage and age, these variables were entered into a Cox proportional hazards model with CGA-classification and the modified PF, respectively. As the proportionality assumption was not fulfilled for the CGA-classification, this variable was dichotomized into frail versus non-frail.
All statistical analyses were performed using SPSS 15.0 software (Chicago, IL).
3. Results
Of the 187 patients included pre-operatively, 176 were evaluable for frailty according to both CGA-classification and the modified PF. Age ranged from 70 to 94years, with a mean and median age of 80years, and there were 101 (57%) women. A total of 171 (97%) patients lived in their own home, and 84 (48%) lived alone. Thirty-six patients (21%) had public help, while 46 (26%) received help from relatives or friends on a regular basis. The cancer was localized to colon in 125 (71%) patients, while the remainder had rectal cancer. A total of 21 (12%) patients had metastatic disease.
Based on a CGA, the number of patients categorized as fit, intermediate, and frail was 21 (12%), 80 (45%), and 75 (43%), respectively. According to the modified PF, 70 (40%) were robust, 84 (48%) were pre-frail, and 22 (13%) were frail. As seen in Table 3, the two measurements selected different patients as frail, and the Kappa Measure of Agreement value was 0.05, representing poor agreement. Of the 75 patients classified as frail according to the CGA, only 17 were frail according to the PF. Of the 101 patients classified as non-frail according to the CGA, 96 were classified as non-frail according to the PF. When comparing the relationships between the two frailty classifications and the occurrence of post-operative complications, we have previously shown that frailty defined from CGA significantly predicted any post-operative complication (p=0.001) as well as severe complications (p=0.002).12 The most common severe complications in the CGA-frail group were pulmonary complications in 18 patients (24%), cardiac complications in 17 patients (23%), and delirium in 10 patients (13%). Anastomotic leakage was observed in 7 patients who were frail according to CGA, compared to 2 patients who were not frail according to CGA (relative risk 5.1, 95% confidence interval 1.03 to 25.28). Increasing frailty identified from the modified PF was neither a significant predictor of any complication (p=0.18) nor of severe complications (p=0.23). Of the five physical frailty indicators, only low activity was associated with the outcome severe complications (p=0.10). When this indicator was included in a multivariable model with the elements of CGA that were bivariately associated with severe post-operative complications (IADL, co-morbidity, and nutritional status), low activity was the first variable to be removed from the risk prediction model at a p-level of 0.21.
Table 3. Comparison of frailty-defining criteria and relations to severe post-operative complications.
| CGA-classification | Modified physical phenotype of frailty | Total | No. severe morbidity (%) | ||
|---|---|---|---|---|---|
| Robust | Pre-frail | Frail | |||
| Fit | 12 | 9 | 0 | 21 | 7 (33) |
| Intermediate | 42 | 33 | 5 | 80 | 29 (39) |
| Frail | 16 | 42 | 17 | 75 | 46 (61) |
| Total/no. severe morbidity (%) | 70/28 (40) | 84/43 (51) | 22/11 (50) | 176 | 82 (47) |
There were no differences in morbidity according to cancer stage in this cohort. Three patients (1%) died within 30days of surgery.
Overall, the median follow-up time was 20months (interquartile range, 15 to 25). Both frailty classifications were associated with overall survival, and the results are displayed in Table 4.
Table 4. Bivariate analyses of age, cancer stage, CGA-classification, modified PF and overall survival.
| Variables | No. of patients | P (Log-rank test) | |
|---|---|---|---|
| Died | Total | ||
| Age, years | 0.260 | ||
| 70–79 | 26 | 87 | |
| 80–94 | 20 | 89 | |
| Cancer stage⁎ | <0.001 | ||
| TNM 0–II | 14 | 108 | |
| TNM III | 15 | 43 | |
| TNM IV | 17 | 21 | |
| CGA-classification | 0.001 | ||
| Fit | 3 | 21 | |
| Intermediate | 13 | 80 | |
| Frail | 30 | 75 | |
| Modified PF | 0.056 | ||
| Robust | 12 | 70 | |
| Pre-frail | 25 | 84 | |
| Frail | 9 | 22 | |
⁎4 patients were unclassified according to TNM. |
When correcting for cancer stage and age in the Cox regression model, frailty classified by CGA remained an independent predictor of mortality. When correcting for cancer stage in the Cox regression model of the modified PF, both pre-frailty and frailty were significant predictors of survival. The results are displayed in Table 5.
Table 5. Frailty-classifications as predictors of mortality by Cox regression analyses. Final models are corrected for cancer stage and age.
| Variables | Hazard ratio | 95% CI | P |
|---|---|---|---|
| Model with CGA-frailty | |||
| Cancer stage | |||
| TNM 0, I, and II | 1.00 | ||
| TNM III | 3.14 | 1.51–6.53 | .002 |
| TNM IV | 10.62 | 5.04–22.37 | <.001 |
| Age | |||
| 70–79 | 1.00 | ||
| 80–94 | 0.74 | 0.41–1.34 | .32 |
| CGA-classification | |||
| Non-frail | 1.00 | ||
| Frail | 3.39 | 1.82–6.29 | <.001 |
| Model with modified PF | |||
| Cancer stage | |||
| TNM 0, I, and II | 1.00 | ||
| TNM III | 2.77 | 1.30–5.90 | .008 |
| TNM IV | 11.87 | 5.52–23.74 | <.001 |
| Age | |||
| 70–79 | 1.00 | ||
| 80–94 | 0.78 | 0.43–1.41 | .40 |
| Modified PF | |||
| Robust | 1.00 | ||
| Pre-frail | 2.33 | 1.16–4.67 | .018 |
| Frail | 2.67 | 1.11–6.83 | .029 |
4. Discussion
This study of older adults with colorectal cancer demonstrated poor agreement between two different ways of identifying frailty — one based on geriatric assessment criteria and the other based on a modified version of the PF criteria. There was a marked inconsistency between the cases each method considered frail: CGA-frailty was found in 43% of the patient cohort, while PF was identified in 13%. Interestingly, only the frailty measure based on a CGA predicted post-operative complications. Frailty identified from the PF was neither a significant predictor of any post-operative complication nor of the severity of complications, whereas the CGA-based frailty measure predicted post-operative complications as well as severe complications. Both frailty classifications were predictive of overall survival.
Frailty is defined by a limited reserve capacity and a heightened vulnerability to adverse events. Thus, aggressive treatment modalities may expose a frail patient to complications and toxicity. Assessing frailty in older cancer patients may identify patients who would benefit from targeted interventions aimed at improving the patient's tolerance to therapy, or alert the physician that the risk versus benefit ratio of therapy may be too high. The PF does not necessarily identify remediable conditions that may be optimized pre-operatively, such as malnutrition and depression, and in our cohort this classification did not identify the patients with an increased operative risk. We have previously shown that comorbidity, IADL-dependency, and depression are CGA-elements that independently predict post-operative complications.29 As PF does not include comorbidity or psychological factors, its usefulness in the pre-operative setting may be limited. Contrary to our results, a recent publication found that the criteria comprising PF predicted post-operative complications in a heterogeneous cohort of surgical patients aged 65years or older. However, the standardized cut-off for defining PF was not used in that study: intermediate frailty was defined as fulfilling 2 to 3 of the five criteria, while frailty was defined as fulfilling 4 to 5 criteria.15 Thus, the results cannot be directly compared.
It seems that the concept of frailty cannot be universally applied for all kinds of outcomes in the geriatric population. In geriatric oncology, it would perhaps be more appropriate and less ambiguous to speak about risk factors rather than frailty. However, as the word “frailty” is now well established in the geriatric oncology setting, it is necessary to point out that the concept of frailty is still under debate. Thus, the term always needs to be clearly defined. In older patients with cancer, our efforts are focused on identifying accurate parameters that may select older patients who are candidates to a specific treatment option. It is likely that these variables may differ according to the cancer primary site and stage and the treatment modalities, such as surgery, chemotherapy, and combined therapy. Further studies of geriatric assessment parameters in homogeneous patient cohorts would serve to unveil these questions.
The prevalence of CGA-frailty in our cohort was much higher than the prevalence of frailty identified by the modified PF. The population-based United States Health and Retirements Study also found higher estimates of frailty when using the accumulation of deficit definition compared to PF.30 Severe comorbidity was the most frequently appearing CGA-frailty defining criterion in our cohort,12 while comorbidity is not among the criteria in the definition of PF.
The Cox regression analysis demonstrated that cancer stage and frailty (according to both classifications) were independent predictors of mortality after a median follow-up time of 20months. Interestingly, increasing age actually resulted in a (non-significant) reduced hazard ratio for mortality. As increasing age was neither a predictor of post-operative morbidity nor of overall survival in our cohort, it appears that selected octa- and nonagenarians tolerate colorectal cancer surgery well. Even the oldest old, when fit, should be offered treatment when colorectal cancer impacts on remaining life expectancy or in the likelihood of local tumor complications.
In a previous publication from the same cohort, we studied the relationship between biomarkers of frailty and the two frailty classifications.31 We found that pre-operative levels of C-reactive protein and interleukin-6 were significantly higher in frail compared to non-frail patients within both frailty measures. The significant results with regards to both prediction of mortality and levels of frailty biomarkers suggest a potential relevance of PF in this study, even if PF did not predict post-operative complications.
A limitation of our study is that PF was not measured in the exact same way as in the original publication.13 Furthermore, our hospital-based cohort consists of elderly patients who underwent surgery for cancer, while Fried and colleagues included subjects from the general population and excluded patients with active cancer from the CHS study. Even though the resection rate is high even in advanced age in colorectal cancer,32 our patients have already been selected through primary care and the pre-operative surgical evaluation, and are not representative of the general population. In addition, our cohort is considerably older than the CHS-cohort where 67% of the patients were aged 65–74years. In contrast, 75% of our cohort was 75years and older. This may explain the higher prevalence of PF in our study (13% versus 7%). Another limitation is that data regarding post-operative complications were recorded retrospectively.
An important strength of our dataset is that the cohort is truly elderly, with a median age of 80years. Furthermore, all these patients have the same index disease, and they were prospectively assessed before a major surgical procedure — an elective bowel resection. This allowed us to directly compare two frailty measures in a homogeneous cohort. Data regarding the clinical usefulness of identifying vulnerability in geriatric oncology are lacking, and we think our data are a useful contribution.
Successful surgical care of elderly cancer patients depends on several factors, such as pre-operative patient selection, avoiding emergency surgery, and careful peri-operative handling of patients. Growing evidence indicates that post-operative complications, in addition to decreasing well being of patients and increasing the risk of post-operative mortality, also may have important effects on long-term survival and recovery to pre-operative levels of independence.[33], [34], [35] According to our findings, a pre-operative CGA may identify elderly patients with a high risk of post-operative complications. Furthermore, elements from a CGA, such as severe comorbidity, IADL-dependency, and depression were found to be better predictors than chronological age, anesthesiological risk scores, and objective functional measurements such as gait speed and grip strength.29 A prospective randomized trial of susceptible elderly surgical patients would yield information about whether geriatric assessment and targeted intervention can improve surgical oncology care by decreasing the rate of post-operative complications.
In conclusion, it seems that the optimal tool to measure frailty will vary according to the specific outcome in question. PF has been validated in several population-based studies as identifying those at high risk for disability, falls, hospitalizations, hip fracture, and mortality.[36], [37], [38], [39] For prediction of treatment tolerance, in this case post-operative complications, our results indicate that a multi-component classification of frailty based on geriatric assessment criteria is more useful. Furthermore, CGA may unmask remediable problems, and consequently allow for pre-treatment patient optimization. Future studies are needed to test the ability of CGA-guided interventions to decrease surgical morbidity in frail older adults with cancer.
Funding
The study is supported by a research grant from the Norwegian Cancer Society (to SRK). The Norwegian Cancer Society had no role in the study design, data collection, analysis, and interpretation of data; in writing the manuscript; or in the decision to submit the manuscript for publication.
Conflict of Interest
SRK, BR, ES, MSJ, and AN report no conflict of interest. AH has received research support from Abraxis Bioscience, and has worked as a consultant for AMGEN. TBW holds honoraria for lectures from Pfizer, Lundbeck, and Roche.
Author Contributions
Concept and design: SRK, TBW, ES, AN, MSJ, Data collection: SRK, Analysis and interpretation of data: SRK, TBW, ES, BR, AH, MSJ, Manuscript drafting: SRK, Discussion and final approval: SRK, TBW, AH, BR, MSJ, ES, AN.
References
- . Frailty: different tools for different purposes?. Age Ageing. 2008;37:129–131
- . The I.A.N.A Task Force on frailty assessment of older people in clinical practice. J Nutr Health Aging. 2008;12:29–37
- The frailty syndrome: a critical issue in geriatric oncology. Crit Rev Oncol Hematol. 2003;46:127–137
- . Functional status assessment in the preoperative evaluation of older adults. JAMA. 2009;302:1582–1583
- . Cancer screening in elderly patients: a framework for individualized decision making. JAMA. 2001;285:2750–2756
- . In: The biology of aging: observations and principles. Oxford New York: Oxford University Press; 2006;p. 63–67
- . Performance status is the single most important prognostic factor in lymphoma patients aged greater than 75 overriding other prognostic factors such as histology. Leuk Lymphoma. 2008;49:149–151
- . Screening for frailty: criteria and predictors of outcomes. J Am Geriatr Soc. 1991;39:778–784
- . Management of cancer in the older person: a practical approach. Oncologist. 2000;5:224–237
- . Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci. 2007;62:722–727
- . Operationalizing a frailty index from a standardized comprehensive geriatric assessment. J Am Geriatr Soc. 2004;52:1929–1933
- Comprehensive geriatric assessment can predict complications in elderly patients after elective surgery for colorectal cancer: a prospective observational cohort study. Crit Rev Oncol Hematol. 2010;76:208–217
- Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146–M156
- Frailty and risk of falls, fracture, and mortality in older women: the study of osteoporotic fractures. J Gerontol A Biol Sci Med Sci. 2007;62:744–751
- Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg. 2010;210:901–908
- . Functional evaluation: the Barthel Index. Md State Med J. 1965;14:61–65
- . The extended activities of daily living scale: a further validation. Disabil Rehabil. 1992;14:41–43
- Rating chronic medical illness burden in geropsychiatric practice and research: application of the Cumulative Illness Rating Scale. Psychiatry Res. 1992;41:237–248
- A manual of guidelines to score the modified cumulative illness rating scale and its validation in acute hospitalized elderly patients. J Am Geriatr Soc. 2008;56:1926–1931
- . “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–198
- Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res. 1982;17:37–49
- . The Mini Nutritional Assessment (MNA) review of the literature — what does it tell us?. J Nutr Health Aging. 2006;10:466–485
- . Management of the frail person with advanced cancer. Crit Rev Oncol Hematol. 2000;33:143–148
- The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst. 1993;85:365–376
- . A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci. 2007;62:738–743
- . Proposed classification of complications of surgery with examples of utility in cholecystectomy. Surgery. 1992;111:518–526
- . Evaluation of Charlson-Age Comorbidity Index as predictor of morbidity and mortality in patients with colorectal carcinoma. J Gastrointest Surg. 2004 Dec;8:1061–1067
- . Liver resection in octogenarians. HPB (Oxford). 2006;8:206–210
- Which elements of a comprehensive geriatric assessment (CGA) predict post-operative complications and early mortality after colorectal cancer surgery?. J Geriatr Oncol. 2010;1:57–65
- . Comparing models of frailty: the Health and Retirement Study. J Am Geriatr Soc. 2009;57:830–839
- Frailty measures, inflammatory biomarkers and post-operative complications in older surgical patients. Age Ageing. 2010;39:758–761
- . Prognostic impact of increasing age and co-morbidity in cancer patients: a population-based approach. Crit Rev Oncol Hematol. 2005;55:231–240
- Functional independence after major abdominal surgery in the elderly. J Am Coll Surg. 2004;199:762–772
- . Determinants of long-term survival after major surgery and the adverse effect of postoperative complications. Ann Surg. 2005;242:326–341
- . Which octogenarians do poorly after major open abdominal surgery in our Asian population?. World J Surg. 2006;30:547–552
- Phenotype of frailty: characterization in the women's health and aging studies. J Gerontol A Biol Sci Med Sci. 2006;61:262–266
- Lower plasma vitamin E levels are associated with the frailty syndrome: the InCHIANTI study. J Gerontol A Biol Sci Med Sci. 2006;61:278–283
- Frailty in older men: prevalence, progression, and relationship with mortality. J Am Geriatr Soc. 2007;55:1216–1223
- . Prevalence of frailty in middle-aged and older community-dwelling Europeans living in 10 countries. J Gerontol A Biol Sci Med Sci. 2009;64:675–681
PII: S1879-4068(11)00062-2
doi:10.1016/j.jgo.2011.09.002
© 2011 Elsevier Inc. All rights reserved.
