A mini geriatric assessment helps treatment decision in elderly patients with digestive cancer. A pilot study

1. Introduction 

Most digestive tract cancer occur in elderly patients [1]. Until recently, elderly patients have been underrepresented in clinical trials [2]. Recent studies show that selected elderly patients can benefit from standard chemotherapy [3], [4], [5], [6], and also from high-risk surgery such as hepatic or pancreatic resection [7], [8], [9], [10]. Nevertheless, a large proportion of elderly patients do not receive the standard cancer treatments [11], [12], [13]. The reasons for therapeutic abstention are frequently unclear [13]. We can infer that numerous patients are inappropriately under treated [14].

Geriatricians have validated standardized tools to identify high-risk elderly patients. Comprehensive geriatric assessment (CGA) is a compilation of valid tools assessing cognitive function, psychological status, functional status, nutritional status, co-morbidities, and medication review. The use of CGA has shown its effectiveness to improve care of elderly patients [15], [16], [17]. CGA implicates a specialized and time-consuming geriatric consultation, which is not usually available within a short delay in every hospital. Recently shorter versions based on CGA had been proposed so as to allow oncologist to select patients for standard treatments [18], [19], [20]. The evaluation of a mini geriatric assessment (MGA) for digestive cancer before treatment decision had never been realised.

We conducted a pilot study to evaluate the feasibility of MGA in gastroenterology consultation or during hospitalisation and its reliability with CGA.

2. Patients and methods 

A gastroenterologist performed MGA in consultation or during hospitalisation before the treatment decision for cancer cure was made, unless colic surgery had previously been planned. Patients self-evaluated their quality of life with a 10cm scale. MGA includes 8 simplified modules evaluating cognitive status, psychological status, functional status, nutritional status, co-morbidities, medication review, social support and measuring hemoglobin level and creatinin clearance (Table 1). The modules performed in GMA derived from Balducci and Extermann [21]. MGA was considered during the cancer multidisciplinary team meeting before taking decision for treatment.

Table 1. Mini geriatric assessment, 8 simplified modules.

Threshold used: (1) loss of more than 5% of weight and/or BMI<23; (2) 7 medications or more; (3) Hb<10g/dl, clearance<50ml/min.

After the therapeutic decision, CGA was realised during ambulatory hospitalisation in the geriatric department. The patient was evaluated with his caregiver as much as possible. CGA included the following tools: CIRS-G [22], ADL [23], IADL [24], MMSE [25], MNA [26], GDS [27], standing on one leg, number of falls in the 6 previous months, medication review and social support. The cuts-off for each module are presented in Table 2. The geriatrician then proposed adaptation of oncological treatment, non-oncological treatment and/or social support. If a change in the oncological treatment was suggested the cancer multidisciplinary team had to reconsider the decision taken at the first meeting.

Table 2. Threshold of the comprehensive geriatric assessment.

	Screening	Test	Threshold
	Mental status	MMSE	≤24
	Depression	GDS simplified	≥1
	Dependence	ADL	<6
		IADL	<100% of total applicable
	Nutrition	MNA	≤23.5
	Co-morbidities	CIRS-G	System≥3
	Polymedication	Collected	1 anticoagulant or 2 cardiovascular or 2 psychotropic medication or ≥10 medications
	Environment	Monopodal support Story of fall Social service advice	Social intervention
	Biology	Hemoglobin	(<10g/dl)
		Creatinin clearance	(<50ml/min)

The normality or abnormality of each module of MGA was compared to the score of the corresponding items of CGA to access the reliability of MGA to CGA.

A second consultation with the geriatrician was proposed at 6 months to evaluate the recommended adaptation of treatment. If necessary the geriatrician recommended a specific geriatric follow-up. The patient follow-up was performed until death.

3. Results 

3.2. Mini geriatric assessment 

MGA was performed for each patient before cancer multidisciplinary team meetings. An abnormality was detected for at least one module of MGA in every patient. A median of 3 modules of the MGA was abnormal for the whole population. Nutritional status was most frequently altered followed by biology and co-morbidities (Table 3). The environment was the less frequently involved module but concerned more than 30% of the patients.

Table 3. Frequency of abnormality for the modules of MGA and CGA.

		MGA (%)	CGA (%)
	Mental status	9/21 (43)	8/21 (38)
	Depression	9/21 (43)	9/21 (43)
	Dependence	10/21 (48)	7/21 (33)
	Nutrition	14/21 (67)	15/21 (71)
	Co-morbidities	13/21 (62)	11/21 (52)
	Polymedication	8/21 (38)	10/21 (48)
	Environment	7/21 (33)	7/21 (33)
	Biology	13/20 (65)	10/20 (50)

MGA: mini geriatric assessment; CGA: comprehensive geriatric assessment.

The therapeutic strategies were then discussed during the cancer multidisciplinary team meeting. The recommended treatments were standard treatments in 9/21 (41%) of cases. In 10/21 (47%) patients, the recommended treatment was adapted from the standard treatment: chemotherapy dose reduction in 5 patients and rejection of surgery in 5 patients which received chemotherapy and/or radiotherapy. Moreover, solely best supportive care was proposed in 2 (12%) patients.

3.3. Comprehensive geriatric assessment 

CGA was performed for each patient with a median delay of 11 days [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27] after the MGA. The abnormalities detected by CGA according to the threshold are presented in Table 3. The concordance between MGA and CGA was good for the assessment of cognitive function, psychological status and functional status. The concordance between the two evaluations was less accurate for nutritional status and co-morbidity (Fig. 1). A severe undernutrition (MNA<17) was present in 10% of the patients. CGA led to adaptation of the non-oncological treatment in 15/21 (72%) of the patients and changes in social care management for 8/21 (38%) of patients. The cancer treatment strategy was never modified by CGA.

View Large Image Download to PowerPoint

Fig. 1. Concordance between the MGA and the CGA.

The non-oncological treatment adaptation concerned co-morbidity medications in 10 cases, psychotropic medication treatment in 5 cases, nutrition in 4 cases, anaemia in 3 cases, anticoagulation in 1 case and dementia in 1 case. The social care propositions were home caretaker in 4 cases, visiting nurse in 2 cases and institutionalization in 2 cases. A specific geriatric follow-up was recommended in 8 (38%) patients.

3.5. Factors associated with full completion of therapeutic strategy 

Patient characteristics were analysed according to the full completion of the cancer treatment determined by the multidisciplinary team meeting (Table 4). The small number of the patient included did not allow a multivaried analysis. Nevertheless, the patients that did not complete the recommended treatment appeared to be more vulnerable, having significantly more abnormal modules in the MGA and a worst ADL score. Interestingly, among these 6 patients, 5 have had a proposition of standard treatment. This suggests that the endurance of these patients was overestimated.

Table 4. Patient's characteristics according the full completion of the cancer treatment.

	Patient's characteristic	Full treatment completed n=15	Treatment not completed n=6
	Medium age (extreme)	79 (75–87)	83 (78–85)
	Sex: men (%)	9 (60)	2 (33)
	Women (%)	6 (40)	4 (67)
	Metastases (%)	6 (40)	5 (83)
	WHO: 0–1 (%)	10 (67)	3 (50)
	≥2 (%)	5 (33)	3 (50)
	Charlson: 0–1 (%)	6 (40)	3 (50)
	≥2 (%)	9 (60)	3 (50)
	Quality of life (cm)	7.5	6.35
	Median number of abnormal module of the MGA (extreme)	3 (1–7)	5 (3–6)a
	Cancer treatment recommended
	Standard (%)	4 (27)	5 (83)
	Adapted (%)	11 (73)	1 (17)
	Median MMSE (extreme)	27 (18–30)	26.5 (19–30)
	GDS≥1 (%)	7 (46)	3 (50)
	Median MNA (extreme)	20 (11–27)	20 (13–23)
	ADL<6 (%)	2 (15)	4 (67)b
	IADL<100% total applicable (%)	3 (23)	4 (67)
	Standing on one leg (%)	5 (38)	1 (17)
	Fall (%)	4 (30)	2 (33)

a p=0.01 (Student's test). b p=0.04 (Fisher test).

4. Discussion 

The adaptation of standard cancer treatment in elderly patients is a major challenge. A randomized study has showed that CGA improves prognosis of elderly patients with advanced cancer [28]. CGA has also allowed to identify preliminary depression and dependence as predictive factors of toxicity in elderly patients receiving chemotherapy for treatment of advanced ovarian cancer [29]. However if most studies in geriatric oncology have focused on identifying prognostic factors detected by CGA, they did not include an intervention in their design [30]. The lack of intervention studies limits conclusions about the value of CGA in oncology.

The study included consecutive patients treated for digestive cancer in our institution, so selection bias was avoided.

MGA detected most of the abnormalities found by the CGA and could allow gastroenterologist to a rapid adaptation of therapeutic strategy. Every patient had at least one abnormal module, so the presence of only one abnormality could not be used to select patient for CGA. Nutritional status was the most frequently altered module. It must be pointed out that nutritional status was recently identified as the main prognostic factor for death as detected by CGA in elderly patients treated with chemotherapy [31]. The impact of CGA on cancer treatment decision making has rarely been studied. In a recent study involving mostly breast cancer patients, CGA was followed by changes in the planned cancer treatment in 38% of the patients [32]. In Girre et al.’s work, only body mass index and the absence of depressive symptoms were associated with a modification of the treatment plan. In our study there are numerous reasons for adapting anticancer treatment and the small sample size does not allow a statistical analysis. A severe undernourishment was described in 10% of our patients, which is comparable to the 15% observed in the study of Girre et al. In another pilot study performed on 15 breast cancer patients, extensive CGA influenced cancer treatment in 36% of the patients [33]. As in our study, social intervention was predominant to cancer treatment modification. These data suggest that a geriatric evaluation is necessary to adapt cancer treatment in a large number of cases.

In our study, the cancer treatment recommended by the multidisciplinary team meeting was adapted from standard treatment in almost half of the cases. This adaptation was done after the MGA but before the CGA. This suggests that the MGA could help to adapt cancer treatment. However, our study could not assert the contribution of the MGA because this was only an observational pilot study. It must be pointed out that there is no randomized study in geriatric oncology demonstrating the value of MGA.

The choice of the most suitable tools for MGA is still a matter for debate. Several scales similar to ours are currently used, from the work of Balducci and Extermann [21] and the Vulnerable Elders’ Survey-13 [19]. Two studies aiming to validate a MGA in cancer patients are ongoing in France (OLD study from GERCOR and ONCODAGE study from FNCLCC). A large French trial on metastatic colorectal in elderly patients will soon be completed and their accrual and preliminary results have been already presented [34]. The analysis of geriatric items collected in this study is planned. The validation of the most accurate tool is still necessary in large prospective trials.

Nevertheless, MGA used in this study shows a good concordance with CGA. Multidimensional evaluation is necessary to detect vulnerability, it has been shown that solely scoring ADL and IADL is not enough [35]. Obviously, MGA could not be a substitute to CGA for specific geriatric decision making. The geriatrician modified non-oncological treatment for most of the patients. Some pathologies, such as anaemia and denutrition should be considered for treatment earlier by the gastroenterologist or oncologist to allow cancer treatment. Social care was modified in one-third of the patients after CGA, and can be an important factor for successful cancer treatment. On the other hand, the cancer treatment based on the GMA was never modified after CGA. This suggests that GMA performed before decision making for treatment helps to define the most accurate strategy. This two-step approach is recommended by the Society of Geriatric Oncology [36].

In our study, we had a high rate (72%) of achievement of the recommended cancer treatment, half of which had been adapted. Moreover, in 5 out of the 6 cases of treatment failure the treatment offered was the standard treatment. These six patients had a worst ADL score and more abnormal modules in the MGA than the other patients. This suggests that an adapted cancer treatment should be offered frailest patients.

In our study, most of the patients died of cancer; this points out that digestive cancer even in elderly patients is generally the most severe disease and rules the prognosis. This should be taken into account when carefully discussing a treatment plan for every patient in multidisciplinary team meeting including geriatric evaluation.

In summary, our study reveals that MGA is routinely feasible even in gastroenterologic consultation. MGA helps for cancer treatment decision and allows a majority of completed treatment. The characteristics of the patients that should have a specialized geriatric follow-up remain to be determined.

Conflict of interest 

None of the authors had conflict of interest concerning this work.

Reviewers 

Dr. Martine Extermann, H Lee Moffitt Cancer Center USF, 12902 Magnolia Drive, Tampa, FL 33612, United States.