Physicians’ judgement and comprehensive geriatric assessment (CGA) select different patients as fit for chemotherapy

Critical Reviews in Oncology/Hematology 64 (2007) 1–9

Physicians’ judgement and comprehensive geriatric assessment (CGA) select different patients as fit for chemotherapy Ulrich Wedding a,b,∗ , Daphne K¨odding c , Ludger Pientka b , Hans T. Steinmetz c , Stefan Schmitz c a

Department of Internal Medicine II, Division of Haematology and Medical Oncology, Friedrich Schiller University, Erlanger Allee 101, 07747 Jena, Germany b Department of Geriatrics, Marienhospital Herne, Ruhr-University Bochum, Widumerstraat, Germany c Onkologische Praxis K¨ oln, Sachsenring 69, 50677 K¨oln, Germany Accepted 1 May 2007

Contents 1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Patients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Physicians’ rating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. CGA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Patients’ characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. CGA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reviewers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 2 2 2 3 3 3 4 8 8 8 9 9

Abstract Introduction: Elderly cancer patients are a very heterogeneous population. A comprehensive geriatric assessment (CGA) shall help to identify more precisely those patients who are fit, compared to those who are vulnerable or frail, when deciding on chemotherapeutical treatment. Methods: In a prospective trial, 200 cancer patients treated in an out-patient setting were judged by their physician for their fitness for chemotherapy as fit, vulnerable or frail. A CGA was performed thereafter. We determined the feasibility of a CGA in an out-patient setting and the frequency of changes within the different assessment tools and compared physicians’ judgement with the CGA results. Results: Physicians judged 64.3% of their patients as fit, 32.4% as vulnerable, and 3.2% as frail. A CGA was completed by 97.5% of patients and lasted 20 min per patients (range: 9–47 min). 26.5% of all patients had no deficits in the CGA. The CGA identified a mean of 1.7 problems per patient, 1.3 in patients judged as fit, 2.3 in those judged as vulnerable, and 4.2 in those judged as frail. A CGA is more sensitive in classifying patients as fit compared to vulnerable or frail than physicians’ judgement. Conclusion: A CGA is feasible and detects more elderly cancer patients as being unfit for chemotherapy than physicians’ judgement. Further trials including disease and treatment related end-points are needed. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Cancer; Geriatric oncology; Comorbidity; Geriatric assessment; Elderly; Decision-making; Classification and regression tree ∗ Corresponding author at: Department of Internal Medicine II, Division of Haematology and Medical Oncology, Friedrich Schiller University, Erlanger Allee 101, 07747 Jena, Germany. Tel.: +49 3641 9324216; fax: +49 3641 9324217. E-mail address: [email protected] (U. Wedding).

1040-8428/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.critrevonc.2007.05.001

2

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

1. Introduction

2. Methods

Cancer is the second commonest cause of death in Europe and Northern America [1]. About 60% of all people diagnosed with cancer are 65 years and older. Due to demographic changes, the number of people with cancer will increase substantially within the coming decades [2]. Elderly patients are a very heterogeneous population. Geriatric medicine has established a comprehensive geriatric assessment (CGA) to detect resources and deficits of the patients missed by a routine history and physical examination. CGA and treatment plans established on the basis of CGA improve functional abilities, decrease hospital readmission rates, decrease the number of patients who need institutional care, and some trials reported an improved survival [3–5]. The use of CGA in patients with cancer has demonstrated that CGA is feasible [6–8], and that it detects more problems than performance scores established in oncology such as the Eastern Cooperative Oncology PerformanceStatus (ECOG-PS) [9]. Within a very small group of patients with breast cancer, it has been demonstrated that results of CGA changed treatment plans [10]. Introduction of a nurse case management improved quality of care and the outcome of the treatment of women with breast cancer [11]. Other authors reported the prognostic relevance of CGA for QoL [12], for early termination of treatment [13], for postoperative morbidity [14], for survival in cancer patients in general [15], and in patients with ovarian cancer [16], or in patients with acute myeloid leukaemia [17] in particular. Although there are no prospective randomised trials supporting a survival or quality of life benefit, CGA is recommended for the use in elderly cancer patients [18,19]. For treatment decisions in elderly cancer patients, it is recommended to distinguish between three groups of patients, i.e. those who are fit, those who are vulnerable and those who are frail [20]. Balducci suggested a classification of the groups according to results of the CGA. Clear cut-off levels are however not yet defined by differences in outcome variables such as toxicity, quality of life or survival. As the use of CGA is time-consuming, information collected by CGA is only helpful if the same information would not be able to be obtained more easily by a physician or nurse based on experience or “gut feeling”. No trials comparing physicians’ judgement with the results of CGA have been reported to date. Against this background we conducted a prospective trial with the aim to determine the feasibility of CGA in an out-patient setting and the frequency of changes within the different assessment tools and to compare physicians’ judgements with the results of CGA.

2.1. Patients Patients were recruited during their routine scheduled visit at the “Onkologische Schwerpunktpraxis” Cologne, an institution providing full out-patient care for oncology patients. Inclusion criteria were: age of 70 years or older and malignant disease. Written informed consent was obtained prior to inclusion into the trial. Patient-related data, i.e. age, sex, months since primary diagnosis, months since first visit to the institution, prior and current chemotherapy, comorbidity, medication used and type of cancer were retrospectively analysed from the patients’ charts. 2.2. Physicians’ rating The patients saw one of two physicians. Both were fulltime haematologists and oncologists and had more than 10 years in specialist haematology–oncology practise. After finishing the routine visit and prior to the CGA, each physician rated his patient according to one of the three following categories: • Fit: patients are medically fit, biologically younger than other persons of the same age, no or only mild comorbidities, no limitations in daily functioning, chemotherapy not restricted. • Vulnerable: patients are medically compromised, biologically old, one or more comorbidities, reversible limitations in daily functioning, adaptation of standard chemotherapy necessary. • Frail: patients are medically frail, biologically very old, severe comorbidities, irreversible limitations in daily functioning, treatment related stress will presumably lead to deterioration. 2.3. CGA CGA was conducted by a specially trained medical student, to whom results of physicians’ judgement were not known. It included the following variables: • Activities of daily living (ADL): the ADL score was assessed by means of the Barthel index [21]. This questionnaire comprises 10 items (‘eating’, ‘transferring from bed to chair’, ‘grooming’, ‘toilet use’, ‘bathing’, ‘walking on a corridor’, ‘ascending and descending stairs’, ‘dressing’, ‘bowel continence’, ‘urine continence’). Each single item was able to be assigned to the dichotomous outcome ‘with help’ or ‘independent’. Full credit (this means ‘independent’) for an activity/item was not given if the patient

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

•

•

•

•

•

•

needed help. A total score for all 10 items was calculated and patients were classified as those ‘without limitations’ having scored full marks (=100%) and those ‘with limitations’ having scores below <100%. Instrumental activities of daily living (IADL): the IADL score was assessed using the score published by Lawton and Brody in 1969 [22]. This questionnaire comprises eight items (‘ability to use telephone’, ‘shopping’, ‘food preparation’, ‘housekeeping’, ‘laundry’, ‘travelling via car or public transportation’, ‘medication use’, ‘ability to handle finances’). The IADL score is an extension of the ADL score describing everyday functional competence and the ability to adapt independently to the environment [23,24]. Analogous to the interpretation of the ADL, each single item of the IADL was able to be assigned to the dichotomous outcome ‘with help’ or ‘independent’. A total score for all eight items was calculated and patients were classified as those ‘without limitations’ having scored full marks (=8) and as those ‘with limitations’ having scored <8. If at least one single item was assessed ‘with help’ (score: =0), the dichotomous appraisal of this patient was ‘with limitations’. Mini nutritional assessment (MNA) is a widely used screening instrument to judge the nutritional status of a patient. We used a short version based on a two-step approach. In step one, six items are judged [25]. If the score is more than two points below the maximum score, a second step must be carried out. The score is judged as follows: ≥24 credits good nutritional status; 17–23.5 credits = risk for malnutrition, <17 poor nutritional situation, existing malnutrition. Mini-mental-status-examination (MMSE): the MMSE was first published by Folstein et al. [26]. The results according to credits can be grouped as follows: 24–30 credits = no cognitive impairment, 18–23 credits = mild cognitive impairment, and 0–17 credits moderate to severe cognitive impairment. The Tinetti test is a score for judging patient mobility. The Tinetti test judges stand, balance, rising, turning on a point and sitting down. The test provides credits for all categories. Maximum score for full mobility is 28 credits [27]. A score of less than 20 predicts an increased risk for falls. Number of deficits: ADL, IADL, MNA, MMSE and Tinetti results were used to calculate the number of deficits per patients. If ADL did not score 100%, IADL was less than eight credits, MNA < 24 credits, MMSE < 24 or Tinetti < 2 patients were considered to have a deficit in each test. Comorbidity was rated according to the Charlson score. The Charlson score is one of the most widely used and validated scores. Moreover, it yields high interrater and crosssource reliability. It consists of a score resulting from

3

a list of 19 medical conditions rated from 1 to 6 points. This score correlates to the risk of death within the next few years with reasonable accuracy [28]. • Classification by Balducci: Balducci suggested classifying patients without ADL, IADL limitations and without severe comorbidity as fit, those with IADL limitations but without ADL limitations and with one to two severe comorbidities as vulnerable, and those with ADL limitations and more than two comorbidities as frail [29]. 2.4. Statistics Data management and data analysis were performed using the statistical package SPSS® Version 11.5. To ensure a high quality of data in respect of completeness, correctness and consistency, plausibility checks were performed. Statistical measurements (frequency, relative frequency, mean and standard deviation) were calculated for the variables. Spearman’s correlation coefficient was used to test for correlation. To test statistical significance for non-parametric data, the Kruskal–Wallis test or for significance between groups the Fisher’s Exact Test were used. The outcome of a statistic test with p < 0.05 is called significant and with p < 0.10 a trend. A classification and regression tree was used to identify variables of the CGA contributing to physicians’ judgement. Age, ADL, IADL, MMSE, MNA, Tinetti, Charlson score and no. of medicine used, each as continuous variables, were included. Physicians’ judgement dichotomised as fit versus vulnerable or frail was the dependent variable.

3. Results 3.1. Patients’ characteristics Patients were enrolled between the 20 February 2003 and the 21 July 2003. Two hundred and twelve patients fulfilled the inclusion criteria. Thereof seven did not agree to participate and five were medically unfit for the test battery. Two hundred patients have therefore been studied and described. Five patients asked to stop the assessment after having agreed to participate and after the first tests were done. The mean age was 75.9 years (range 70–94 years). The types of primary carcinoma are presented in Fig. 1. Time since diagnosis: the time since cancer diagnosis ranged from a few days up to 31 years. The median time since diagnosis was 43 months. Seven patients (3.6%) were diagnosed within the same month and 47 (24.2%) within the previous 6 months. Previous and current chemotherapeutical treatment: one hundred and eleven (56%) patients had received previous chemotherapy and 86 patients had previously received

4

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

Fig. 1. Diagnosis of patients.

chemotherapy at the same institution. Fifty-three patients (27%) were currently on treatment. Physicians’ rating: the patients consulted one of two physicians, physician 1 saw 124 patients (62%), physician 2 76 patients (38%). Physicians’ ratings were available for 185 patients (92.5%). One hundred and nineteen (64.3%) patients were rated as fit, 60 (32.4%) patients as vulnerable and 6 (3.2%) patients as frail. 3.2. CGA Duration of assessment: the median duration of assessment was 20 min (S.D. 6 min, range 9–47 min). ADL: an ADL score was available for 199 patients. Fifty percent scored the highest possible score (100%). The lowest score was 35 in three patients (1.5%) and the mean score 92.2% (S.D. 13.1). Male patients scored slightly higher than female patients, 93.5 (S.D. 12.1) versus 90.8 (S.D. 13.9), but the difference was not significant. IADL: data on IADL was available for all 200 patients. Median score was 6.8 credits (S.D. 1.7). 53.5% patients had a full score with eight credits. Four patients had a score of only one credit. There were no differences between the two sexes. MNA: data on MNA were available for 199 patients. Median score was 25 credits (S.D. 5.2). Ninety-two patients (46%) had a full score of 30. The lowest score was 10 credits. Small differences between the two sexes existed: male patients, mean MNA 25.7 (S.D. 5.3) versus female patients, mean 24.5 (S.D. 5.1). According to the categories, 57% were of good nutritional status; 34% were at risk of malnutrition, and 9% were of poor nutritional status. MMSE: data on cognitive function measured by the MMSE were available for 196 patients. The score ranged from 14 to 30 points. Only 28 patients (14%) had a full score. The mean score was 27.4 (S.D. 2.5). No differences between the two sexes were able to be observed. According to the

criteria mentioned above, 91% of all patients had no cognitive impairment, 8% were slightly cognitively impaired, and none of the patients were suffering from moderate or severe cognitive impairment. Tinetti: full data for Tinetti were available for 195 patients. The minimum score was 1, the maximum 28. The mean score was 22.9 (S.D. 5.7). Female patients had a slightly lower scale than male patients: 22 (S.D. 6.1) versus 23.8 (S.D. 5.0). According to the cut-off level of <20, 46 patients (23%) had an increased risk for falls. Charlson score: patients’ charts for rating of comorbidity were available for 192 patients. One hundred and sixteen (60.4%) had no additional morbidity according to the Charlson score in addition to their cancer diagnosis. Forty-five patients (23.4%) had one additional comorbidity, 20 patients (10.4%) two and 11 patients (5.7%) three or more. Number of medicines used: information on the number of medicines used was available for 183 patients. A median of three drugs per patient was reported, range 0–9. Thirty-nine patients (21.3%) did not take any drugs at all. Overall deficits in CGA: eight patients (4%) had deficits in all five evaluated categories of CGA. Fifty-three patients (27%) had no deficits in any category of the CGA. The mean number of deficits per patient was 1.7 (S.D. 1.5). The number of deficits was higher in female than in male patients, 1.9 versus 1.5. IADL were the items of CGA affected most often by deficits. No patient judged as frail, had a low number of deficits (<3), but 20% of patients judged as fit had three or more deficits. Classification by Balducci: according to the classification suggested by Balducci, 50 (25%) patients were fit, 51 (25.5%) vulnerable, and 99 (49.5%) frail. Age group dependent results of CGA are presented in Table 1 as well as mean scores, upper and lower limits in Fig. 2. Type of cancer and CGA: no association was found between the specific type of cancer and results of CGA (data not shown). Comorbidity and CGA: a small correlation was found between comorbidity and results of ADL (r = −0.23), but none between IADL assessment (r = −0.14), MNA (r = 0.01), MMSE (r = −0.04), Tinetti (r = −0.15), and the number of deficits (r = 0.16). Number of medicines used and CGA: a small correlation was found between the number of medicines used and ADL (r = −0.21), none between IADL assessment (r = −0.16), MNA (r = −0.07), MMSE (r = −0.04), Tinetti (r = −0.16), and the number of deficits (r = 0.19) (Table 2). Prior or current chemotherapy and CGA: no association existed between the results of different items of CGA and the fact that patients had received previous chemotherapy or were currently receiving chemotherapy, when CGA was performed. Duration of disease and CGA: no association was seen between the time since primary diagnosis and results of different items of CGA.

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

5

Table 1 Mean scores in ADL, IADL, MNA, MMSE, Tinetti, number of deficits, Charlson score, number of medicines used, and duration of CGA for different age groups Age group

ADL (S.D.)

IADL (S.D.)

MNA (S.D.)

MMSE (S.D.)

Tinetti (S.D.)

No. of deficits (S.D.)

Charlson score (S.D.)

No. of medication (S.D.)

Duration of CGA (S.D.)

70–74 75–79 80–84 85+

95.5 (10.0) 90.1 (14.5) 89.5 (15.6) 85.5 (13.2)

7.3 (1.4) 6.5 (1.8) 6.4 (1.7) 5.1 (2.6)

25.4 (5.1) 24.7 (5.4) 25.9 (5.0) 21.6 (4.9)

27.7 (2.3) 26.9 (2.8) 27.7 (2.3) 26.8 (2.6)

25.1 (4.2) 21.2 (6.4) 21.9 (5.9) 19.2 (5.2)

1.2 (1.2) 2.1 (1.6) 1.9 (1.5) 3.1 (1.6)

0.5 (0.8) 0.8 (1.0) 0.6 (0.9) 0.9 (1.1)

2.4 (2.1) 2.9 (2.2) 3.0 (2.2) 3.2 (2.6)

18 min (5) 22 min (7) 19 min (6) 23 min (5)

Fig. 2. Box and whisker graphs of activities of daily living (ADL according to Barthel index), instrumental activities of daily living (IADL), mini-nutritional assessment (MNA), mini-mental-status-examination (MMS), Tinetti test, and Charlson score.

6

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

Table 2 Median score of age, ADL, IADL, MNA, MMSE, Tinetti, number of deficits, and Charlson score according to the different groups of patients by physicians’ judgement Physicians’ judgement

Age (years)

ADL (%)

IADL

MNA

MMSE

Tinetti

No. of deficits

Charlson score

Fit Vulnerable Frail

75 77 82

94 90 65

7 6 3

26 24 18

28 27 23

25 20 16

1.3 2.3 4.2

0.5 0.7 2.0

Table 3 Number of patients according to physicians’ judgement and results of different domains of CGA Domains of CGA

Category

Frail

Vulnerable

Fit

All n=

ADL

<100 =100

6 0

38 22

49 69

93 91

IADL

<8 =8

6 0

36 24

46 73

88 97

MNA

Poor At risk Good

2 4 0

6 26 28

8 32 78

16 62 106

MMSE

Moderate to severe Mild No

1 3 2

1 3 55

0 9 107

2 15 164

Tinetti

Increased risk <20 No increased risk

3 3

26 34

14 100

43 137

No. of deficits

0 1 2 3 4 5

0 0 0 2 1 3

9 10 11 14 14 1

39 36 16 14 7 3

48 46 27 30 22 7

Table 4 Sensitivity and specificity of number of deficits in CGA, CGA classification according to Balducci and physicians’ judgement compared to ADL, IADL, MNA, MMSE, Tinetti, age, number of medicines used, and Charlson score

ADL (100 vs. <100) IADL (8 vs. <8) MNA (≥24 vs. <24) MMSE (≥24 vs. <24) Tinetti (≥20 vs. <20) Age (≤75 years vs. >75 years) No. of medication (0–2 vs. >2) Charlson (0–2 vs. >2) No. of deficits (none vs. 1 or more) CGA (fit vs. vulnerable or frail) Physicians’ judgement (fit vs. vulnerable or frail) a

No. of deficits (none vs. 1 or more)

CGA according to Balducci (fit vs. vulnerable or frail)

Physicians’ judgement (fit vs. vulnerable or frail)

Sensitivity

Specificity

Sensitivity

Specificity

Sensitivity

Specificity

0.68 0.63 0.58 0.12 0.32 0.51

1.00a 1.00a 1.00a 1.00a 1.00a 0.66

0.66 0.62 0.49 0.10 0.31 0.54

1.00a 1.00a 0.77 0.77 1.00 0.75

0.66 0.63 0.56 0.56 0.38 0.56

0.53 0.61 0.66 0.66 0.88 0.58

0.51

0.54

0.56

0.70

0.51

0.50

0.44 –

0.70 –

0.53 –

1.00a –

0.43 0.86

0.63 0.33

–

–

–

–

0.88

0.31

0.41

0.81

0.43

0.80

–

–

Part of the definition of the true value.

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

Does physicians’ judgement and CGA based criteria identify the same patients as fit?: the number of patients classified as fit, vulnerable and frail compared to results of CGA is shown in Table 3. We made the following three different assumptions, to compare the sensitivity and specificity of the different assessment tools within the CGA. (1) CGA is standard and the number of deficits (none versus 1 or more) is the cut-off level, (2) CGA is standard and the suggested classification by Balducci is appropriate,

7

(3) Physicians’ judgement currently is the standard of care as CGA has not yet been validated. The results of analysis of sensitivity and specificity for the different assumptions are presented in Table 4. CART: a Tinetti score of ≤25 or of >25 best reflected physicians’ grouping of the patients as fit versus vulnerable or frail. Results are presented in Fig. 3. Those patients with a Tinetti score of 25 or below were able to be further classified by age and IADL score. As results of CGA are part of the definition of the number of deficits or the classifica-

Fig. 3. Classification and regression analysis of physicians’ judgement of the patients as fit vs. vulnerable or frail.

8

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

tion according to Balducci, CART analysis was performed for physicians’ judgement only.

4. Discussion Five domains of a CGA, assessment of the number of medicines used and comorbidities were chosen (1) to test the feasibility of a CGA in an out-patient setting, (2) to describe the prevalence of changes and (3) to compare physicians’ judgement with the results of CGA. The selection of ADL, IADL, MMSE, Tinetti, and MNA reflects the recommendations of national and international experts [18,19]. The assessment of depression and of the social situation was omitted to reduce the burden of tests for the patients. CGA was completed by 97% of all patients and as the mean duration of CGA was 20 min, we consider the CGA as feasible within an out-patient setting. The time needed for CGA in our population is in line with the 27 min reported by Hurria et al. [8]. In addition, the results mainly reflect those reported by other authors, apart from the ADL and Tinetti results. The higher rate of deficits in ADL (50%) compared to IADL (47%) was not reported by other groups [7,9,30]. Extermann et al. reported in their trials that 21% of patients had deficits in the ADL and 56% in the IADL score. In the presented study, the high prevalence of ADL deficits is mainly attributed to incontinence. Overcash et al. identified ‘bathing’, ‘continence’ and ‘transfer’ as the most important items for identifying patients with ADL deficits [31]. R¨ohrig et al. reported climbing stairs as the single most affected items in 21% of all patients. Urine incontinence however was the next significant item to identify those patients with deficits in ADL. Their patients were younger, but were hospitalised [32]. To date, Tinetti score data are not available for elderly cancer patients. Consequently, the results cannot be compared with those of other reports. Comorbidity and functional status show only a weak correlation, as also reported by Extermann et al. [7]. Results of the assessment of comorbidity are difficult to compare. Extermann et al. included patients >65 years and used the cumulative illness rating scale for measurement of comorbidity [7], and Repetto et al. measured comorbidity using a modified Satariano score [6]. This is the first report comparing physicians’ judgement of the patients’ fitness for chemotherapy with results of a CGA. In a trial on the referral of patients with cancer to hospice care, the extent of physicians’ experience was associated with a better estimation of the patients’ life expectancy [33]. However physicians’ judgement was far from precise. The value of physicians’ judgement in our cohort may therefore even be overestimated, as most of the patients were known to the physicians prior to inclusion in the study.

In addition, CGA provides a repeatable and objective method to assess patients’ fitness, unlike physicians’ judgement. It should be the aim to base decision-making on such repeatable and objective methods, to make decisions comprehensible and comparable. Cut-off levels for the diagnosis of fit, vulnerable or frail patients regarding tolerance to chemotherapy have been suggested, but not yet validated. We used two possible cut-off levels, firstly patients with or without deficits in CGA and secondly a definition suggested by Balducci. As both use results of CGA, a true comparison between the test in respect of sensitivity and specificity is only possible for domains, not included into the definition. All in all, physicians’ judgement is less sensitive but more specific than the number of deficits or the classification by Balducci for diagnosing fit versus unfit elderly patients. Even 10 of 29 (34%) patients with four or five deficits in CGA were classified as fit by physicians’ judgement. First data demonstrate that patients with intermediate to high grade non-Hodgkin’s lymphoma, who are classified as frail according to the suggested definition by Balducci, experience a very high rate of toxicity [34]. Certainly further trials are necessary to validate these data. As demonstrated in the CART analysis, physicians’ judgement seems to be predominantly based on the issue of patients’ mobility. Whether this is appropriate remains to be answered. The questions as to whether or not physicians’ rating or data from CGA are more relevant for prognosis and decision-making in elderly cancer patients cannot be decided from the presented data. The application of CGA was not restricted to patients starting with chemotherapy. The diagnoses were quite heterogeneous. Further trials are therefore needed to decide whether CGA identifies those patients who are at increased risk for chemotherapy-related morbidity and toxicity more precisely than physicians’ judgement. The presented data demonstrate that CGA is feasible, not identical to and more sensitive than physicians’ judgement.

Reviewers Hans Wildiers, MD, PhD, Department of Medical Oncology, University Hospital Gasthuisberg, Herestraat 49, Leuven B-3000, Belgium. Brian Stein, MBBS, FRACP, Ashford Cancer Center, 48 Marlestone Avenue, Ashford, SA 5035, Australia.

Acknowledgements U.W. is currently research fellow at the “Forschungskolleg Geriatrie” of the Robert Bosch Foundation, Stuttgart, Germany. The authors thank Ulrich Thiem for his careful review of the manuscript and his most helpful discussion.

U. Wedding et al. / Critical Reviews in Oncology/Hematology 64 (2007) 1–9

References [1] Jemal A, Siegel R, Ward E, et al. Cancer statistics 2006. CA Cancer J Clin 2006;56(2):106–30. [2] Edwards BK, Howe HL, Ries LA, et al. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 2002;94(10):2766–92. [3] Stuck AE, Siu AL, Wieland GD, et al. Comprehensive geriatric assessment: a meta-analysis of controlled trials. Lancet 1993;342(8878):1032–6. [4] Rubenstein LZ, Josephson K, Wieland GD, et al. Geriatric assessment on a subacute hospital ward. Clin Geriatr Med 1987;3(1):131–43. [5] Ellis G, Langhorne P. Comprehensive geriatric assessment for older hospital patients. Br Med Bull 2004;71:45–59. [6] Repetto L, Venturino A, Vercelli M, et al. Performance status and comorbidity in elderly cancer patients compared with young patients with neoplasia and elderly patients without neoplastic conditions [see comments]. Cancer 1998;82(4):760–5. [7] Extermann M, Overcash J, Lyman GH, et al. Comorbidity and functional status are independent in older cancer patients. J Clin Oncol 1998;16(4):1582–7. [8] Hurria A, Gupta S, Zauderer M, et al. Developing a cancer-specific geriatric assessment: a feasibility study. Cancer 2005;104(9):1998–2005. [9] Repetto L, Fratino L, Audisio RA, et al. Comprehensive geriatric assessment adds information to Eastern Cooperative Oncology Group performance status in elderly cancer patients: an Italian Group for Geriatric Oncology Study. J Clin Oncol 2002;20(2):494–502. [10] Extermann M, Meyer M, McGinnis M, et al. A comprehensive geriatric intervention detects multiple problems in older breast cancer patients. Crit Rev Oncol Hematol 2004;49(1):69–75. [11] Goodwin JS, Satish S, Anderson, et al. Effect of nurse case management on the treatment of older women with breast cancer. J Am Geriatr Soc 2003;51(9):1252–9. [12] Wedding U, Rohrig B, Klippstein A, Brix C, Pientka L, Hoffken K, Co-morbidity and functional deficits and independently contribute to quality of life before chemotherapy in elderly cancer patients, Support Care Cancer, 2007, Feb 21. [13] Frasci G, Lorusso V, Panza N, et al. Gemcitabine plus vinorelbine versus vinorelbine alone in elderly patients with advanced non-small-cell lung cancer. J Clin Oncol 2000;18(13):2529–36. [14] Audisio RA, Ramesh HS, Gennari R, et al. Can preoperative assessment of cancer in the elderly (PACE) predict 30-days postoperative outcomes? Annual Meeting Proceedings Part I 24 18S(8537). J Clin Oncol 2006. [15] Zagonel V, Fratino L, Piselli P, et al. The comprehensive geriatric assessment (CGA) predicts mortality among elderly cancer patients (ECP). Proc Am Soc Clin Oncol 2002;21:1458, abstract. [16] Freyer G, Geay JF, Touzet S, et al. Comprehensive geriatric assessment predicts tolerance to chemotherapy and survival in elderly patients with advanced ovarian carcinoma: a GINECO study. Ann Oncol 2005;16(11):1795–800. [17] Wedding U, Rohrig B, Klippstein A, Frick HJ, Sayer HG, Hoffken K. Impairment in functional status and survival in patients with acute myeloid leukaemia. J Cancer Res Clin Oncol 2006;132(10): 665–71. [18] Friedrich C, Kolb G, Wedding U, et al. Comprehensive geriatric assessment in the elderly cancer patient. Onkologie 2003;26(4):355–60. [19] Extermann M, Aapro M, Bernabei R, et al. Use of comprehensive geriatric assessment in older cancer patients: recommendations from the

[20] [21] [22]

[23] [24]

[25]

[26]

[27] [28]

[29] [30]

[31]

[32]

[33]

[34]

9

task force on CGA of the International Society of Geriatric Oncology (SIOG). Crit Rev Oncol Hematol 2005;55(3):241–52. Balducci L, Stanta G. Cancer in the frail patient. A coming epidemic. Hematol Oncol Clin North Am 2000;14(1):235–50, xi. Mahoney FI, Barthel DW. Functional evaluation: the Barthel index. Md State Med J 1965;14:61–5. Lawton MP, Brody EM. Assessment of older people: selfmaintaining and instrumental activities of daily living. Gerontologist 1969;9(3):179–86. Garman KS, Cohen HJ. Functional status and the elderly cancer patient. Crit Rev Oncol Hematol 2002;43(3):191–208. Spector WD, Katz S, Murphy JB, et al. The hierarchical relationship between activities of daily living and instrumental activities of daily living. J Chronic Dis 1987;40(6):481–9. Rubenstein LZ, Harker JO, Salva A, et al. Screening for undernutrition in geriatric practice: developing the short-form mini-nutritional assessment (MNA-SF). J Gerontol A: Biol Sci Med Sci 2001;56(6):M366–72. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12(3):189–98. Tinetti ME. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986;34(2):119–26. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40(5):373–83. Balducci L, Extermann M. Management of cancer in the older person: a practical approach. Oncologist 2000;5(3):224–37. Serraino D, Fratino L, Zagonel V. Prevalence of functional disability among elderly patients with cancer. Crit Rev Oncol Hematol 2001;39(3):269–73. Overcash JA, Beckstead J, Extermann M, et al. The abbreviated comprehensive geriatric assessment (aCGA): a retrospective analysis. Crit Rev Oncol Hematol 2005;54(2):129–36. Rohrig B, Hoffken K, Pientka L, Wedding U. How many and which items of Activities of Daily Living (ADL) and Instrumental Activities of Daily Living (IADL) are necessary for screening. Crit Rev Oncol Hematol 2007;62(2):164–71. Lamont EB, Christakis NA. Physician factors in the timing of cancer patient referral to hospice palliative care. Cancer 2002;94(10): 2733–7. Monfardini S, Aversa SM, Zoli V, et al. Vinorelbine and prednisone in frail elderly patients with intermediate-high grade non-Hodgkin’s lymphomas. Ann Oncol 2005;16(8):1352–8.

Biography Ulrich Wedding is a physician and specialist in internal medicine, haematology, oncology, and palliative care. He serves as a consultant at the Department of Haematology and Medical Oncology at the University Hospital of the Friedrich Schiller University Jena in Germany. Currently, he is a research fellow at the Robert Bosch Foundation. His main interest in clinical research is geriatric oncology. He is an active member of national and international Working Parties in the field of geriatric oncology.