Symptom Burden and Dietary Changes Among Older Adults with Cancer: A Cross-Sectional Study

Abstract

Background: Malnutrition has a direct impact on both the toxicities of cancer therapy and the overall survival of oncological patients. However, its prevalence amongst vulnerable groups such as older patients (age ≥ 65 years) is often underestimated. Screening tools recognizing patients at risk are well established, yet they do not take into account that cancer therapy may lead to changes in dietary habits or that therapy’s side effects may negatively influence nutritional status. Methods: To close this gap, we combined the validated Nutritional Risk Score 2002 (NRS-2002) and G8 screening tools with short questionnaires addressing diet changes and symptom load and screened 300 cancer inpatients between 12/2022 and 12/2023. Descriptive statistics (Fisher’s exact, Student’s t-test) as well as heat mapping were applied for data analysis. Results: Overall, two in three inpatients ≥65 years were at risk for malnutrition, and the majority of patients (87.67%) scored ≤14 points on the G8 and were considered frail. Surprisingly, the symptom complex of oral discomfort was most often mentioned by patients (xerostomia—178/300 patients, loss of appetite: 122/300 patients, dysgeusia: 93/300 patients). Diet changes were also common, with patients mainly avoiding certain foods (122/300 patients) or using dietary supplements (106/300 patients). Conclusions: Taken together, older cancer inpatients are frail and have a high risk of malnutrition. Screening should not only consider energy intake but also symptom burden and dietary changes to optimize supportive care.

1. Introduction

Over 60% of all individuals diagnosed with cancer are aged 65 and over, and the prevalence of malnutrition amongst cancer patients is estimated to be 20–70% [1,2,3]. Guidelines define malnutrition either as a qualitative insufficiency due to the inadequate intake of nutrients and food components and/or as a quantitative (energetic) problem due to an insufficient caloric intake [2]. The underlying genesis of malnutrition is variable, as cancer itself but also treatment-related side effects might promote malnutrition. The transition from malnutrition toward cachexia or sarcopenia is possible and can occur at any stage of the disease, sometimes even before diagnosis [4]. Older people with cancer are a particularly vulnerable group due to being prone to lower caloric intake and following a less varied diet. Especially reduced food intake is often present before the onset of the disease [5]. The variability of the triggers for malnutrition is often underestimated, and guidelines emphasize considering patients’ symptom burden, as local impairments (e.g., dysphagia) as well as general symptoms affect nutritional status [6]. Battling malnutrition is essential. Malnourished patients have a worse prognosis [7]. Thus, recognizing these patients and addressing malnutrition directly affects the success of tumor therapy and patient survival. It is therefore of utmost importance to provide help and counseling as early as possible to counteract a negative spiral [8]. A range of screening instruments and assessments is available to uncover nutritional deficiencies. The selection of the appropriate tool depends on the screening setting and the disease entity or patient population [9]. In an inpatient setting, the Nutritional Risk Score-2002 (NRS-2002) [10] is often used as a screening tool, as recommended by guidelines [11]. The NRS-2002 is a tool combining a simple prescreen (yes/no questions) using questions about weight loss, decreased calorie intake, low body mass index (BMI), and severity of disease. If one question is answered with “yes”, patients undergo more detailed screening. Depending on the severity of disease or decrease in nutritional intake, points are assigned. A score ≥ 3 points considers a patient at risk for malnutrition [10]. However, the tool gives no information on symptoms leading to a changed nutrition intake. Further, a previous study has demonstrated that dietary changes are common in cancer patients and may influence nutritional status [12]. To address this information gap concerning both symptom load and dietary changes, we adapted our screening for patients at risk: based on the NRS-2002 [10] in combination with the G8 questionnaire [13], we extended our screening tool-box by recording symptom burden and dietary changes, as previously proposed [14]. We aimed to identify the older patients (age ≥ 65 years) at risk and the extent of symptom burden as well as dietary changes. Knowledge of patients’ diet and symptom load should then enable medical staff to optimize both nutrition and supportive cancer care.

2. Materials and Methods

Patients and study design. The study conducted is an anonymous cross-sectional study. Due to the exploratory character of the study, a timeframe for recruitment (one year) was predefined before starting the project. Data were collected prospectively between December 2022 and December 2023. A total of 300 patients treated as inpatients at the Department of Hematology and Medical Oncology at the University Medical Center Göttingen were included in the study and surveyed. Inclusion criteria were ≥65 years of age, an underlying oncological disease, an inpatient stay, and the ability to provide information as well as consent to participate. The aim was to offer participation to all inpatients who (1) fit the inclusion criteria and (2) were treated during the predefined recruitment time of one year. A yearlong screening was initiated to take into account seasonal fluctuation of patients. All patients were questioned by a previously trained member of the study group. The study’s protocol was approved by the Ethics Committee of the University of Göttingen (approval number 26/8/22). Patients were interviewed using an adapted screening tool. The components of this anamnesis questionnaire were the NRS-2002 [8], the Geriatric Screening G8 tool [13], a recently proposed questionnaire that considers dietary changes in cancer patients [14], and a question module on symptom burden. In addition, anthropometric data (age, sex, height, weight, BMI) and clinical data (initial diagnosis versus disease progression, tumor entity, and treatment modality) were collected. The NRS-2002 is suitable for screening patients with or at risk of malnutrition in an inpatient setting. It is recommended by the European Society for Clinical Nutrition and Metabolism for all hospitalized patients [4]. The Geriatric Screening G8 [13] is designed to identify patients who require more detailed geriatric assessment. A score ≤ 14 points is considered suspicious and should lead to reassessment or further investigation. Dietary changes and adherence to a cancer diet were assessed in a set of questions, as previously proposed by [12]. Patients were asked if they avoided or preferred certain foods, if they used dietary supplements, and if they followed a special or cancer diet. In addition to the established screening questionnaires, a symptom burden questionnaire was added. As part of this module, a list of symptoms that may be associated with malnutrition and oncological disease was discussed with the patients. Symptoms queried were as follows: oral mucositis, xerostomia, loss of appetite, dysgeusia, dysphagia, odynophagia, diarrhea, nausea, emesis, meteorism, abdominal discomfort, and constipation. Other symptoms could be listed in a free text box. The questionnaire (original German questionnaire, as well as a translated English version) is listed under Appendix A. The study has been reported in accordance with the STROBE guidelines [15], and the corresponding checklist is listed under Appendix B.

Statistical analysis. Due to smaller group sizes, some entities were summarized as follows: non-small lung cancer (NSCLC) and small-cell lung cancer (SCLC) as lung cancer; tumors of the cervix and vulva as other gynecological tumors; prostate, kidney, and urothelial tumors as urooncological tumors; and solid tumors of the jejunum, ileum, or colon as lower GI cancer. In addition, aggressive lymphoma and acute leukemia cases were grouped together, and chronic leukemia was grouped together with cases of indolent lymphoma, myelodysplastic (MDS), and myeloproliferative (MPN) neoplasms. Similarly, patients undergoing allogeneic or autologous stem cell therapy or CAR-T-cell therapy were considered participants receiving “cellular immunotherapy”. A treatment modality analysis was performed. A distinction was made between radiological therapy and systemic therapy. The subgroups were also analyzed with regard to different age cohorts. The following patient groups were analyzed: 65–74 years, 75–84 years, and ≥85 years. If data were missing for subgroup analysis, cases were excluded from analysis. The data were collected in Microsoft Excel. The statistical analysis was performed using Excel version 2021 and GraphPad Prism version 9.0. Descriptive statistics were applied to analyze demographic data. Fisher’s exact tests were used to test for differences. Generally, a p-value of p < 0.05 was considered statistically significant. A trend was defined as a p-value < 0.15. The following software was used to visualize data: GraphPad Prism (Version 9.3.0, GraphPad Software, Boston, MA, USA), Biorender [16], and wordcloud.com (accessed on on 28 September 2024) [17]. Heat-map analysis was conducted using the online freeware ClustVis [18].

3. Results

3.1. Risk of Malnutrition—An Ongoing Problem

Three hundred patients (136 female, 164 male) were included in this study. Hematological neoplasms were the most common diagnosis—187/300 patient cases (62.33%) versus 113/300 (37.67%) patients with solid tumors. Most patients with hematological neoplasms suffered from indolent (67/300, 22.33%) or aggressive lymphoma (45/300, 15%) and acute leukemia (39/300, 13.00%). Concerning solid tumors, lung cancer was the most common entity. A total of 25/300 (8.33%) patients received allogeneic, autologous stem cell therapy or CAR-T cell therapy. Further information on clinical baseline data is listed in

Table 1. Clinical characteristics.

Total [N]		300
Sex	Female	136 (45.33%)
	Male	164 (54.67%)
Age	Age cohort 65–74 years	190 (63.33%)
	Age cohort 75–84 years	97 (32.33%)
	Age cohort ≥ 85 years	13 (4.33%)
Entity	Malignant hematology	187 (62.33%)
	Lymphoma (aggressive)	45 (15.00%)
	Lymphoma (indolent)	67 (22.33%)
	Acute leukemia	39 (13.00%)
	Chronic leukemia	6 (2.00%)
	Myeloproliferative neoplasia	6 (2.00%)
	Myelodysplastic syndrome	7 (2.33%)
	Solid Tumor	113 (37.67%)
	Lung cancer (NSCLC)	31 (10.33%)
	Lung cancer (SCLC)	12 (4.00%)
	Breast cancer	2 (0.67%)
	Sarcoma	9 (3.00%)
	Lower gastrointestinal tract tumor	6 (2.00%)
	Other gynecological cancers	7 (2.33%)
	Urooncological cancer	10 (3.33%)
	Other	54 (18.00%)
Treatment modality [N]	213
	Systemic treatment	140 (65.73%)
	Radiation therapy	73 (34.27%)
Cellular immunotherapy	Allogenic, autologous stem cell therapy + CAR-T-cell therapy	25/300 (8.33%)

CAR-T-cell therapy—chimeric antigen receptor-T-cell therapy; NSCLC—non-small cell lung cancer; SCLC—small-cell lung cancer.

.

At the time of inclusion, 43/300 (14.3%) patients were recently diagnosed with cancer. The average BMI was 24.99 kg/m². There were no significant differences in the mean BMI between the age cohorts. A total of 239/300 (79.67%) of the hospitalized patients tested positive in the pre-screening using the NRS-2002, and a risk of malnutrition was confirmed in 209/300 (69.67%) in the main screening. Subgroup analysis was possible for the following groups: hematological neoplasm, solid tumors, aggressive lymphoma/acute leukemia, indolent lymphoma/chronic leukemia/MDS/MPN, lung cancer, and cellular immunotherapy. The malnutrition risk rates were as follows: 73.80% (138/187, hematological neoplasms), 62.83% (71/113, solid tumors), 73.81% (62/84, aggressive lymphoma/acute leukemia), 72.62% (61/84, indolent lymphoma/chronic leukemia/MDS/MPN), 61.36% (27/44, lung cancer), and 64.00% (16/25, cellular immunotherapy). Comparing different cancer entities, we observed a trend toward a higher risk for malnutrition in patients with hematological neoplasms (

Table 2. Analysis of the NRS-2002 screening results by cancer entity.

Entity	Total	Pre-Screen NRS-2002 Positive	NRS-2002 Positive	p-Value
	[N]	[N]	[N]
Hematologocal neoplasm	187	153	138	0.1044
Solid tumors	113	86	71
Lymphoma (aggressive)/acute leukemia	84	67	62	0.7638
Lymphoma (indolent)/chronic leukemia/MDS/MPN	84	68	61
Lymphoma (aggressive)/acute leukemia	84	67	62	0.1917
Others	216	172	147
Lung (NSCLC/SCLC)	44	34	27	0.1519
Others	256	205	182
NSCLC	31	23	20	1.0000
Others	269	216	189
SCLC	13	11	7	0.0357
Others	287	228	202
Cellular immunotherapy	25	22	16	0.0432
Others	272	214	190
ED	43	36	29	0.1816
Others	255	201	178

NRS-2002—Nutritional Risk Score-2002, MDS—myelodysplastic syndrome, MPN—myeloproliferative neoplasia, NSCLC—non-small cell lung cancer, SCLC—small-cell lung cancer.

).

Furthermore, patients suffering from SCLC or undergoing cellular immunotherapy have a significantly lower risk of malnutrition.

To consider both radiation and systemic therapy, treatment modality was recorded separately after piloting. The distribution of patient cases was as follows: 140 patients received systemic therapy and 73 received radiation therapy. A total of 113/140 (80.71%) patients undergoing systemic therapy were positive in the pre-screening of the NRS-2002; the risk of malnutrition was observed in 93/113 (66.43%) patients. A total of 56/73 (76.71%) of radiotherapy patients were positive in pre-screening, and 48/56 (65.75%) were also at risk of malnutrition after completing the NRS-2002. There were no significant differences in the rate of patients at risk between both groups.

3.2. Nutritional Deficiency—An Age-Independent Challenge

A total of 263/300 (87.67%) patients scored a value ≤ 14 points in the G8 and were therefore considered conspicuous. The mean G8 score was 10.55 points (+/− 2.97). Items of the G8 were categorized into the following groups: “nutrition”, “mobility”, and “others”. Overall, we observed no significant difference in the points scored between patients of the age cohort < 75 years compared to patients ≥ 75 years for the categories of nutrition and others. Younger (<75 years) patients, however, scored better on the mobility items (Figure 1).

3.3. Dietary Change and Cancer Diet

As part of the screening, we also interviewed three hundred patients using a nutrition module for diet changes. A total of 122/300 (40.67%) of the hospitalized patients stated that they were avoiding certain foods. Abstaining from alcohol was mentioned most frequently by 24/122 (19.67%) patients. This was closely followed by the avoidance of meat products and foods labeled as “solid/hard”, each with 23/122 (18.85%) mentions. After being diagnosed with cancer, 49/300 (16.33%) patients preferred certain foods. Most frequently listed were liquid and pureed food 22/49 (44.90%) or soft food 10/49 (20.41%). Patients were also asked about the intake of dietary supplements. Overall, 106/300 (35.33%) used dietary supplements. A cancer diet was denied by the majority (278/299, 92.98%) of patients. No significant difference between age cohorts or gender was observed concerning diet changes. Patients of the ≥ 85 years age cohort used nutritional supplements more often than younger patients. However, the sample size (13 patients ≥ 85 years) was very small. Subgroup analysis did not reveal differences between cancer entities except for lung cancer. Compared to other patients, lung cancer patients are significantly more prone to avoiding certain foods (p = 0.018). All subgroups are depicted in

Table 3. Dietary changes sub-group analysis of age cohorts, sex, and cancer entity.

Age Cohort	Avoid Specific Food Yes	Avoid Specific Food No	p-Value
≥65–74 years	83	107	0.2540
75–84 years	35	62
75–84 years	35	62	1.0000
≥85 years	4	9
≥65–74 years	83	107	0.4032
≥85 years	4	9
Age cohort	Prefer specific food Yes	Prefer specific food No	p-value
≥65–74 years	33	157	1.0000
75–84 years	16	81
75–84 years	16	81	0.2084
≥85 years	0	13
≥65–74 years	33	157	0.1331
≥85 years	0	13
Age Cohort	Use of Additional Supplements Yes	Use of Additional Supplements No	p-Value
≥65–74 years	65	125	0.7916
75–84 years	31	66
75–84 years	31	66	0.0040
≥85 years	10	3
≥65–74 years	65	125	0.0050
≥85 years	10	3
Age Cohorts	Cancer Diet Yes	Cancer Diet No	p-Value
≥65–74 years	14	175	0.3179
75–84 years	4	93
75–84 years	4	93	0.0346
≥85 years	3	10
≥65–74 years	14	175	0.0836
≥85 years	3	10
Sex	Avoid Specific Food Yes	Avoid Specific Food No	p-Value
m	72	93	0.2878
f	50	85
	Prefer Specific Food Yes	Prefer Specific Food No	p-Value
m	24	141	0.4328
f	25	110
	Use of Additional Supplements Yes	Use of Additional Supplements No	p-Value
m	59	106	0.1904
f	47	88
	Cancer Diet Yes	Cancer Diet No	p-Value
m	10	154	0.5046
f	11	124
Cancer Entity	Avoid Specific Food Yes	Avoid Specific Food No	p-Value
Lymphoma (aggressive)/acute leukemia	34	50	1.0000
Others	87	128
Lymphoma (indolent)/chronic leukemia/MDS/MPN	32	52	0.6943
Others	89	126
Lung cancer (NSCLC/SCLC)	10	33	0.0179
Others	111	145
	Prefer Specific Food Yes	Prefer Specific Food No	p-Value
Lymphoma (aggressive)/acute leukemia	10	74	0.2257
Others	39	176
Lymphoma (indolent)/chronic leukemia/MDS/MPN	12	72	0.6051
Others	37	178
Lung cancer (NSCLC/SCLC)	5	38	0.5043
Others	44	212
	Use of Additional Supplements Yes	Use of Additional Supplements No	p-Value
Lymphoma (aggressive)/acute leukemia	25	59	0.2270
Others	81	134
Lymphoma (indolent)/chronic leukemia/MDS/MPN	31	53	0.7884
Others	75	140
Lung cancer (NSCLC/SCLC)	15	28	1.0000
Others	91	165
	Cancer Diet Yes	Cancer Diet No	p-Value
Lymphoma (aggressive)/acute leukemia	6	78	0.8022
Others	14	200
Lymphoma (indolent)/chronic leukemia/MDS/MPN	4	80	0.6069
Others	16	198
Lung cancer (NSCLC/SCLC)	1	42	0.3277
Others	19	236

MDS—myelodysplastic syndrome, MPN—myeloproliferative neoplasia, NSCLC—non-small cell lung cancer, SCLC—small-cell lung cancer.

.

3.4. Symptom Burden—An Underestimated Player

Side effects of cancer therapy may have an influence on nutritional status. To estimate symptom load, we added a simple symptom burden module to our survey. The frequency of single symptoms related to nutrition was as follows: xerostomia—178/300 (59.33%) patients, loss of appetite—122/300 (40.67%) patients, dysgeusia—93/300 (31.00%) patients, constipation—80/300 (26.67%) patients, meteorism—70/300 (23.33%) patients, nausea—64/300 (21.33%) patients, dysphagia—63/300 (21.00%), diarrhea—49/300 (16.33%) patients, oral mucositis—48/300 (16.00%) patients, abdominal discomfort—38/300 (12.67%) patients, odynophagia—30/300 (10.00%) patients, and emesis—17/300 (5.67%) patients (Figure 2).

Symptom burden has an influence on nutritional status and risk of malnutrition: an analysis of symptoms and malnutrition risk revealed that xerostomia in particular is significantly (p = 0.008) associated with being at risk of malnutrition (positive NRS-2002). Patients with dysgeusia showed at least a trend toward a risk of malnutrition (p = 0.100). Other symptoms showed no association with the risk of malnutrition. A detailed overview of subgroup analysis is depicted in

Table 4. Symptom burden in relation to nutritional status.

Nutritional Status	Symptom Positive [N]	Symptom Negative [N]	p-Value
	Mucositis
NRS-2002 positive	39	170	1.000
NRS-2002 negative	5	25
	Dysgeusia
NRS-2002 positive	76	133	0.0997
NRS-2002 negative	7	23
	Dysphagia
NRS-2002 positive	50	159	0.8183
NRS-2002 negative	6	24
	Diarrhea
NRS-2002 positive	34	175	1.0000
NRS-2002 negative	5	25
	Meteorism
NRS-2002 positive	51	157	0.6554
NRS-2002 negative	6	24
	Nausea
NRS-2002 positive	47	162	0.6441
NRS-2002 negative	8	22
	Xerostomia
NRS-2002 positive	138	71	0.0082
NRS-2002 negative	12	18
	Loss of appetite
NRS-2002 positive	104	105	0.3362
NRS-2002 negative	12	18
	Odynophagia
NRS-2002 positive	23	186	1.0000
NRS-2002 negative	3	27
	Constipation
NRS-2002 positive	64	144	0.2863
NRS-2002 negative	6	24
	Abdominal discomfort
NRS-2002 positive	31	178	0.3929
NRS-2002 negative	2	28
	Emesis
NRS-2002 positive	14	195	1.0000
NRS-2002 negative	2	28

Nutritional Risk Score-2002—NRS-2002.

.

Certain cancer entities are associated with a distinct symptom profile. Comparing patients with hematological neoplasms and solid tumors showed that dysphagia (p < 0.001), odynophagia (p = 0.010), and constipation (p = 0.032) occur significantly more frequently in patients with solid tumors. In contrast, patients with hematological diseases reported diarrhea significantly more frequently (p = 0.016). They also suffered significantly more frequently from nausea (p = 0.042) but not from emesis. There was a trend for patients with hematological neoplasms to report dysgeusia more frequently. Furthermore, patients reported an increased loss of appetite (p = 0.069,

Table 5. Symptom burden comparing patients with either hematological neoplasms or solid tumors.

Cancer Entity	Symptom Positive [N]	Symptom Negative [N]	p-Value
	Mucositis
Hematological neoplasm	32	155	0.5212
Solid tumor	16	97
	Dysgeusia
Hematological neoplasm	65	122	0.0730
Solid tumor	28	85
	Dysphagia
Hematological neoplasm	26	161	0.0002
Solid tumor	37	76
	Diarrhea
Hematological neoplasm	38	149	0.0161
Solid tumor	11	102
	Meteorism
Hematological neoplasm	43	143	0.8887
Solid tumor	27	86
	Nausea
Hematological neoplasm	47	140	0.0423
Solid tumor	17	96
	Xerostomia
Hematological neoplasm	115	71	0.3317
Solid tumor	63	50
	Loss of appetite
Hematological neoplasm	84	103	0.0686
Solid tumor	38	75
	Odynophagia
Hematological neoplasm	12	175	0.0098
Solid tumor	18	95
	Constipation
Hematological neoplasm	42	145	0.0319
Solid tumor	38	74
	Abdominal discomfort
Hematological neoplasm	26	161	0.4759
Solid tumor	12	101
	Emesis
Hematological neoplasm	10	177	0.7995
Solid tumor	7	106

).

Depicting symptom patterns by heat-map (Figure 3), we were able to observe certain symptom patterns that were more typical for patients with hematological neoplasms (e.g., loss of appetite, dysgeusia, nausea) or patients suffering from solid cancers (e.g., dysphagia, odynophagia, constipation).

A significant correlation was found between radiotherapy and dysphagia (p < 0.001) and odynophagia (p = 0.017). We observed a trend toward less frequent dysgeusia in patients who were recently diagnosed compared to cancer patients under prolonged therapy (p = 0.074). The symptom burden in relation to treatment modality is depicted in

Table 6. Symptom burden in relation to treatment modality.

Treatment Modality	Symptom Positive [N]	Symptom Negative [N]	p-Value
	Mucositis
Systemic treatment	23	117	0.8481
Radiation therapy	13	60
	Dysgeusia
Systemic treatment	49	91	0.2836
Radiation therapy	20	53
	Dysphagia
Systemic treatment	19	121	0.0002
Radiation therapy	27	46
	Diarrhea
Systemic treatment	20	120	0.3904
Radiation therapy	7	66
	Meteorism
Systemic treatment	29	110	0.7271
Radiation therapy	17	56
	Nausea
Systemic treatment	25	115	0.7018
Radiation therapy	11	62
	Xerostomia
Systemic treatment	83	56	0.6611
Radiation therapy	41	32
	Loss of appetite
Systemic treatment	55	85	0.2296
Radiation therapy	22	51
	Odynophagia
Systemic treatment	12	128	0.0168
Radiation therapy	15	58
	Constipation
Systemic treatment	37	102	1.0000
Radiation therapy	20	53
	Abdominal discomfort
Systemic treatment	16	124	0.8176
Radiation therapy	7	66
	Emesis
Systemic treatment	6	134	0.7387
Radiation therapy	4	69

.

In regard to symptom burden, there was a significant difference between the age cohorts (<75 years; ≥75 years). The cohort of patients over 75 years was significantly more likely to have xerostomia (p = 0.015), dysphagia (p = 0.110), and constipation (p = 0.08). In contrast, dysgeusia (p = 0.120) and diarrhea (p = 0.003) were described more frequently in the cohort of younger patients (age < 75 years, also refer to

Table 7. Symptom burden in relation to age cohorts.

Age Cohort	Symptom Positive [N]	Symptom Negative [N]	p-Value
	Oral Mucositis
<75 a	29	161	0.744
≥75 a	19	91
	Dysgeusia
<75 a	65	125	0.1219
≥75 a	28	82
	Dysphagia
<75 a	34	156	0.1052
≥75 a	29	81
	Diarrhea
<75 a	40	150	0.0034
≥75 a	9	101
	Meteorism
<75 a	44	145	1.0000
≥75 a	26	84
	Nausea
<75 a	43	147	0.5590
≥75 a	21	89
	Xerostomia
<75 a	103	87	0.0146
≥75 a	75	34
	Loss of appetite
<75 a	75	115	0.6262
≥75 a	47	63
	Odynophagia
<75 a	17	173	0.4308
≥75 a	13	97
	Constipation
<75 a	44	145	0.0799
≥75 a	36	74
	Abdominal discomfort
<75 a	27	163	0.3684
≥75 a	11	99
	Emesis
<75 a	11	179	1.0000
≥75 a	6	104

).

4. Discussion

Older patients with cancer are a diverse group of people. Guidelines recommend considering biological and not chronological aging as well as health status when stratifying cancer patients for treatment [19]. The G8 screening tool is able to identify frail older patients. In addition to items on mobility, the neuropsychological situation, and medication intake, the screening primarily covers nutritional status (three question items, [13]). However, using the G8, we identified more than four out of five patients (87%) as frail, and patients mainly showed a decrease in points in nutrition items. This merits combining the G8 with established tools used for malnutrition screening. Recognizing the risk of malnutrition in older patients is important. Malnourished older patients are more prone to developing disease-associated complications, which in turn leads to longer hospitalization and poorer outcomes [20]. Accordingly, the European Society for Clinical Nutrition and Metabolism (ESPEN) recommends (1) routine and regular malnutrition screening for all patients diagnosed with cancer and argues for (2) repeated screening and more specific assessments for patients with a high-risk profile [4]. In 2016, the Global Leadership Initiative on Malnutrition (GLIM) established the “GLIM criteria”, a well-defined definition of malnutrition. These criteria comprise low BMI, non-volitional weight loss, reduced muscle mass, inflammation, disease burden, and decreased food intake [21]. The NRS-2002 covers a major part of these criteria and is also recommended by the German Society for Nutritional Medicine for screening inpatients [22]. Therefore, we combined the NRS-2002 with the G8 for our study.

While the German Society for Nutritional Medicine estimates that around half of all older people are at risk of malnutrition [23], our data demonstrates that more than two out of three older cancer patients are at risk. This higher risk of malnutrition fits the previous data of Shaw et al. [3], who described malnutrition in 71% of cancer patients. The risk of malnutrition is also higher in our cohort of older individuals compared to data from a previous study. While Döring et al. reported an overall risk of malnutrition of 37% in oncological patients and specifically described a risk of 44% in patients with hematological neoplasm [12], we observed a risk of 70% (all patients) and 74% (hematological neoplasm) respectively. As both studies used the NRS-2002 to assess the risk of malnutrition, this demonstrates that older people are an even more vulnerable patient group benefiting from nutritional screening.

Reflecting the risk of malnutrition in patients with solid tumors and hematological neoplasms, we observed a trend toward a higher risk of malnutrition in the latter. This is not surprising since different cancer entities are associated with an entity-specific risk, as each oncological treatment modality is characterized by a specific pattern of side effects (e.g., locoregional—radiation therapy/surgery, systemic—oral mucositis, [24,25]). Interestingly, we did not observe a significant difference in malnutrition risk rates when comparing patients with aggressive hematological neoplasms (aggressive lymphoma/acute leukemia) to patients with more chronic/indolent hematological disease (indolent lymphoma/chronic leukemia/MDS/MPN). The lower malnutrition risk rate in patients undergoing cellular immunotherapy compared to the overall population and patients with hematological neoplasms is probably due to a selection bias—only patients in good health with low comorbidities are considered for cellular immunotherapy [26].

Neither the aforementioned GLIM criteria [21] nor the NRS-2002 [10] take into account, that cancer patients may change their diet during oncological treatment. Cancer diets are a controversy [27,28,29], and oncologists are aware that cancer diets often diverge from official dietary guidelines such as the American Cancer Society or the American Institute for Cancer Research/World Cancer Research Fund [28]. Cancer diets are not only restricted to regimes offered by alternative medicine. During recent years, data showed that dietary restrictions prescribed by conventional medicine such as a neutropenic diet do not have a positive impact on infection rates in patients undergoing stem cell transplantation [30]. Fortunately, recent data from a German study showed that only a small minority of cancer patients follow a specific cancer diet [12], which fits the observation we made in our study. However, avoiding or preferring certain foods as well as using nutritional supplements may also influence nutritional status. To close this information gap, our group previously proposed a four-item questionnaire assessing cancer diet/dietary changes and the intake of micronutrients [14]. A pilot trial revealed that up to a third of cancer patients showed dietary changes after being diagnosed with cancer [12]. In this study, we observe that older patients are more prone to avoiding specific food (41%) compared to the younger cohort of Döring et al. (27%, [12]). Overall, abstaining from hard food or preferring soft/pureed nutrition seems to be a recurring topic. The latter could be interpreted as an adaption to the side effects of cancer therapy (e.g., no hard foods—oral mucositis, dysphagia, or odynophagia). These symptoms are often interconnected in patients undergoing radio(chemo)therapy. If a change of dietary habits is caused by therapy and thus influences nutritional status, we should (1) be aware of the side effects in order to (2) ameliorate supportive care and alleviate symptom burden. Further, a lower symptom burden also affects quality of life independent of weight loss [31], and supportive care provides patients with relief regardless of disease stage [32]. Following this rationale, we added a symptom burden questionnaire to our survey. In doing so, we aimed to identify the individual needs of patients, enabling oncological staff to ensure a patient-centered multimodal supportive therapy.

Surprisingly, patients’ chief complaints were xerostomia, loss of appetite, and dysgeusia. While this could be partly explained by treatment modality (radio(chemo)chemotherapy), treatment modality alone does not explain the high percentage of symptom burden. However, xerostomia is known to be more common in older people. Additionally, therapies such as radiation therapy but also chemo- or immunotherapy further decrease the salivation rate. Dysgeusia in turn often accompanies xerostomia, and sufficient salivation also aids digestion [33]. Digestion troubles of the upper gastrointestinal tract are not always a clear-cut symptom, and patients tend to describe the symptom as a “loss of appetite” or abdominal discomfort. Recognizing and addressing xerostomia is important, as we were able to observe a significant association between xerostomia and patients at risk of malnutrition. In addition to these symptoms, which are common in both patients with solid tumors or hematological neoplasm, we also observed specific symptom patterns in these groups. Some symptom associations are easily explained by the treatment modality: dysphagia or odynophagia are common side effects of radiation-induced esophagitis [25], and most of our patients with solid cancer (73/103) underwent radiotherapy.

Anticipating and (in the best case) preventing the side effects of oncological treatment is one of the core pillars of supportive care. An adapted, dietary, and symptom-based screening for malnutrition offers a chance to detect patients at risk before they develop cancer cachexia. As the screening proposed by our group also touches on the topics of diet and symptom burden, it could be also used as a standardized tool to enter a conversation about nutrition itself, educating patients and raising awareness. Understanding the diagnosis of “cancer” and the various phases of the disease as well as the necessity of a balanced diet can be of central importance [34,35]. The knowledge of symptoms and nutrition enables patients to become active themselves, thus empowering self-efficacy.

Limitations

We conducted a cross-sectional study including 300 older inpatients with cancer (age ≥ 65 years). Our cohort was a mix of patients with solid cancer and hematological neoplasm. While subgroup analysis was possible for patients with hematological disease, the sample size was too small to compare different entities of solid cancers (except for lung cancer). Overall, subgroup analysis should be interpreted carefully due to the small sample size. While patients ≥ 85 years were included, we only surveyed 13 individuals, which is not representative. Here, a larger sample size is required to make assumptions on nutritional risk and symptom burden in this particular age group. No distinction between patients with curative or palliative concepts was made. We also have to consider that our study’s design only identified patients at risk but did not confirm manifesting malnutrition by nutritional assessment. The screening results were forwarded to the treating physicians. A shortcoming of this anonymous study is that we could not follow up on whether treating physicians initiated a deeper nutritional assessment of patients at risk or not. As the data analysis was mainly descriptive, we did not consider adjusting for multiple testing. Overall, our screening tool is reductive. Screening results should be used as a starting point for a deeper, more detailed assessment and the counseling of patients at risk. No information on comorbidities such as gastrointestinal diseases or other factors influencing nutrition (surgery, irradiated fields) was assessed.

5. Conclusions

Malnutrition is common in cancer patients, and older people are especially at risk. The reasons for developing malnutrition are diverse, and both systemic and local cancer therapy have an impact on patients’ nutritional status. Both the G8 and the NRS-2002 are good screening tools to identify patients at risk. However, they do not offer any information on possible underlying causes. By considering both diet changes and symptom burden, the treating oncological staff gains more insight into factors that may contribute to malnutrition and subsequently is able to address these specifically, e.g., by nutritional intervention or optimizing supportive care (Figure 4).