Supportive Care in Cancer

, Volume 12, Issue 4, pp 246–252

Cancer: disease and nutrition are key determinants of patients’ quality of life

Authors

    • Centre of Nutrition and Metabolism, Institute of Molecular Medicine, Faculty of MedicineUniversity of Lisbon
  • Isabel Monteiro-Grillo
    • Centre of Nutrition and Metabolism, Institute of Molecular Medicine, Faculty of MedicineUniversity of Lisbon
    • Radiotherapy DepartmentSanta Maria University Hospital
  • Pedro Marques Vidal
    • Centre of Nutrition and Metabolism, Institute of Molecular Medicine, Faculty of MedicineUniversity of Lisbon
  • Maria E. Camilo
    • Centre of Nutrition and Metabolism, Institute of Molecular Medicine, Faculty of MedicineUniversity of Lisbon
Original Article

DOI: 10.1007/s00520-003-0568-z

Cite this article as:
Ravasco, P., Monteiro-Grillo, I., Vidal, P.M. et al. Support Care Cancer (2004) 12: 246. doi:10.1007/s00520-003-0568-z

Abstract

Goals of work

The aims of this study were (1) to evaluate quality of life (QoL), nutritional status and dietary intake taking into account the stage of disease and therapeutic interventions, (2) to determine potential interrelationships, and (3) to quantify the relative contributions of the cancer, nutrition and treatments on QoL.

Patients and methods

In this prospective cross-sectional study conducted in 271 head and neck, oesophagus, stomach and colorectal cancer patients, the following aspects were evaluated: QoL (EORTC-QLQ C30), nutritional status (percent weight loss over the previous 6 months), usual diet (comprehensive diet history), current diet (24-h recall) and a range of clinical variables.

Main results

Usual and current intakes differed according to the site of the tumour (P=0.02). Patients with stage III/IV disease showed a significant reduction from their usual energy/protein intake (P=0.001), while their current intakes were lower than in patients with stage I/II disease (P=0.0002). Weight loss was greater in patients with stage III/IV disease than in those with stage I/II disease (P=0.001). Estimates of effect size revealed that QoL function scores were determined in 30% by cancer location, in 20% by nutritional intake, in 30% by weight loss, in 10% by chemotherapy, in 6% by surgery, in 3% by disease duration and in 1% by stage of disease. Likewise in the case of symptom scales, 41% were attributed to cancer location, 22% to stage, 7% to nutritional intake, 7% to disease duration, 4% to surgery, 1% to weight loss and 0.01% to chemotherapy. Finally for single items, 30% were determined by stage, 20% by cancer location, 9% by intake, 4% by surgery, 3% by weight loss, 3% by disease duration and 1% by chemotherapy.

Conclusions

Although cancer stage was the major determinant of patients’ QoL globally, there were some diagnoses for which the impact of nutritional deterioration combined with deficiencies in nutritional intake may be more important than the stage of the disease process.

Keywords

CancerDietNutritional statusStageQuality of life

Introduction

Malnutrition in cancer is likely to be multifactorial [1, 2], although tumour location and presenting symptoms, e.g. anorexia, taste changes, dysphagia, nausea, vomiting and diarrhoea, may further compromise nutrition and functional ability [2, 3, 4]. The interaction between nutritional status and intake, and the above-mentioned symptoms and/or disease and treatment-related factors, is a complex combination which may dictate a patient’s quality of life (QoL).

QoL is a subjective multidimensional construct reflecting functional status, psychosocial wellbeing, health perception and disease- and treatment-related symptoms [5]. Despite the suggested association between worse overall wellbeing, morbidity and nutritional deterioration [6], the interaction between nutrition and QoL remains underestimated [7]. Although nutritional care has been suggested to be beneficial to cancer patients [8], to date there is scant evidence to support an interaction between nutrition and QoL.

Within this framework, this prospective cross-sectional study conducted in head and neck (HN), oesophageal (OES), stomach (STO) and colorectal (CR) cancer patients was designed to explore the potential interaction(s) between various disease-related and diet-related factors likely to be implicated in such patients’ QoL. Our specific aims were: (1) to evaluate patients’ nutritional status, nutrient intake and QoL, taking into account disease stage and previous therapeutic interventions, (2) to determine the potential interrelationships, and (3) to quantify the relative impact of the cancer and its treatment and nutrition-related factors on QoL.

Patients and methods

Study design and patient sample

This prospective cross-sectional study, approved by the University Hospital Ethics Committee and conducted in accordance with the Helsinki Declaration of 1975 as revised in 1983, was designed to investigate the relationships between cancer and its treatment, and nutrition-related factors and patients’ QoL. Between July 2000 and September 2002, all consecutive ambulatory patients with HN, OES, STO and CR cancer referred to the Radiotherapy Department were considered eligible; patients with other chronic diseases were excluded. All participants gave their informed consent to enter the study. For every patient prior to radiotherapy planning the medical staff recorded clinical variables, duration of disease, cancer location, presence of distant metastases, and tumour burden according to TNM stage [9] determined by local and whole-body imaging methods. The duration of the disease confirmed by histology was defined as the length of time (in months) between symptomatic manifestations and study entry. In order to evaluate differences between cancer stages, patients were clinically and physiologically grouped into two classes: stage I/II (in situ or local disease) and stage III/IV (locally advanced disease with or without lymph node invasion and/or distant metastases) [10]. Data were recorded on individual sheets constructed for statistical analysis.

Study measures

Nutritional assessment

Weight was determined using a Jofre floor scale. Nutritional status was assessed by calculating the percentage weight loss in comparison with the patient’s reported usual weight, and classified as severe when the loss was >10% over the previous 6 months [11].

Nutritional requirements and dietary assessment

Basal energy requirements were estimated using the World Health Organisation formulae for patients aged ≤60 years [12] or by the formulae of Owen et al. for those aged >60 years [13, 14], given their better performance in predicting resting metabolic rate [15]. To estimate patients’ daily energy requirements (EER), basal requirements were multiplied by an activity factor of 1.5 [16]. Daily protein requirements were estimated by comparison with age- and sex-standardised reference values, which ranged between 0.8 and 1.0 g/kg per day [16].

Usual (prior to diagnosis) nutritional intake was derived from a diet history [17, 18] and current intake was assessed by a 24-h recall food questionnaire [19]. The software DIETPLAN version 5 for Windows (Forestfield Software 2003, Horsham, UK) was used to analyse the nutrient contents of foodstuffs and meals.

QoL instrument

The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire version 3.0 (EORTC-QLQ C30) was used in all patients to assess QoL. This instrument is a 30-item cancer-specific questionnaire including six function scales (physical, emotional, cognitive, social, role, and global health/QoL), three symptom scales (fatigue, pain, nausea/vomiting), and six single items assessing symptoms and the financial impact of the disease [20]. Higher scores on the function scales indicate better functioning whilst higher scores on the symptom scales and single items denote increased symptomatology or worse financial impairment. Original scores were linearly transformed to obtain quantified scores within the range 0 to 100. In addition, for better validation in the clinical context, overall scores derived from function scales, symptom scales and single items, were calculated on the basis of the very high statistical significance of interscale correlations according to EORTC’s guidelines [20].

Statistical analysis

Statistical analysis was conducted using SPSS 10.0 (SPSS, Chicago, USA) and EPI–Info 2000 (CDC, Atlanta, USA). Qualitative data, and cancer location and stage, were expressed as number and percentage, while age, disease duration, weight loss, nutritional intake and QoL were expressed as median or mean and standard deviation. Between-group comparisons were performed by one-way analysis of variance (ANOVA) for continuous variables, with Bonferroni or Dunn adjustment because of multiple comparisons. Paired comparisons were performed using Student’s t-test, and categorical variables were compared using the Chi-squared test. Correlations were assessed by nonparametric (Spearman’s) test. A multivariate general linear model was used to identify variables that were significantly related to patients’ QoL. For all statistics, significance was accepted at the 5% probability level.

Results

Patient sample

This study included 271 free-living patients (173 male, 98 female), mean age 54±12 years (range 32–87 years), referred for radiotherapy (primary, adjuvant to surgery, combined with chemotherapy or with palliative intent). Table 1 shows the locations and stages of the cancer: there were 65 stage I/II and 206 stage III/IV. Duration of disease was longer in patients with stage III/IV disease (6±13 months) than in those with stage I/II disease (3.6±5 months; P=0.002).
Table 1

Patients and disease stage

Location

Total no. of patients

Stage

I

II

III

IV

Head and neck

  Base of tongue

11

3

4

4

  Salivary gland

6

1

5

  Tonsil

4

4

  Nasopharynx

11

2

9

  Oropharynx

22

5

17

  Larynx

33

4

3

10

16

Gastrointestinal tract

  Oesophagus

14

3

6

5

  Stomach

26

2

4

11

9

  Colorectum

144

15

19

76

34

Nutritional intake

Usual and current energy and protein intakes were respectively compared with EER and the protein median reference values, taking into account disease location (Fig. 1). In patients with stage III/IV disease, the current protein intake was significantly lower than the reference value (P=0.001), whilst energy intake remained within the EER. Conversely, in those with stage I/II disease, current energy and protein intakes were still significantly higher than the reference values (P=0.005). Moreover, current energy and protein intakes were lower in patients with stage III/IV disease (P=0.0002 and P=0.001, respectively). Table 2 summarizes the median intake reductions for each diagnosis and disease stage and shows that decreases in energy and protein intakes followed a similar pattern and tended to be proportional (P=0.05). The largest decreases, both in energy and protein intake, were seen in HN and OES cancer patients with stage III/IV disease (P=0.02). Stratified analyses further emphasized the differences: patients with stage III/IV disease showed a significant decrease from their usual energy intake (P=0.001) and protein intake (P=0.0002) in contrast to those with stage I/II disease.
Fig. 1

a Energy intake; b protein intake ( white median estimated requirement, black median usual intake, hatched median current intake; HN head-neck, OES oesophagus, STO stomach, CR colorectal)

Table 2

Energy and protein intake: median reduction from usual intake

Diagnosis

Energy (kcal/day)

Protein (g/day)

Stage I/II

Stage III/IV

Stage I/II

Stage III/IV

(n=65)

(n=206)

(n=65)

(n=206)

Head and neck (n=87)

−50

−910

−0.8

−89

Oesophagus (n=14)

−64

−1095

−1

−94

Stomach (n=26)

−25

−491

−0,2

−64

Colorectum (n=144)

−20

−652

−0,2

−68

Nutritional status

The median percentage weight loss for each patient and diagnosis are shown in Fig. 2. OES and STO cancer patients showed a higher percentage weight loss than HN and CR cancer patients (P=0.04). Overall, weight loss was significantly greater in patients with stage III/IV disease than in those with stage I/II disease (P=0.001). In the latter group, only 7 of 65 patients (10%) had lost more than 10% of their usual weight, whereas 175 of 206 patients (85%) with stage III/IV disease had a weight loss greater than 10%.
Fig. 2

Percentage weight loss over the previous 6 months (median) in stage I/II disease (□) and stage III/IV disease (■) according to cancer location (HN head-neck, OES oesophagus, STO stomach, CR colorectal)

Quality of life

The median QoL dimension scores are summarized in Table 3. There was a distinct pattern between diagnosis (P<0.03) and the QoL function scales which were poorer in HN, OES and STO cancer patients (P≤0.008). Overall, symptom scales were worse in stage III/IV than in stage I/II (P<0.003). However, fatigue was significantly higher in stage I/II HN and OES cancer patients than in stage I/II STO or CR cancer patients (P=0.02), whereas nausea/vomiting was worse in stage I/II CR cancer patients (P=0.03), and pain was not significantly different between diagnoses. In all diagnoses, dyspnoea, insomnia and anorexia were worse in stage III/IV than in stage I/II cancer patients (P=0.002). Diarrhoea was more prevalent in CR cancer patients (P=0.001) and more severe in stage III/IV cancer (P=0.03). Financial limitations associated with social/economic conditions were prevalent in HN cancer patients (P=0.002).
Table 3

QoL dimension scores according to cancer location and stage. The results are expressed as median values (HN head and neck, OES oesophagus, STO stomach, CR colorectal). Higher scores on the function scales indicate better functioning; higher scores on the symptom scales and single items denote increased symptomatology

Parameter

HN (n=87)

OES (n=14)

STO (n=26)

CR (n=144)

I/II

III/IV

I/II

III/IV

I/II

III/IV

I/II

III/IV

Function scales

  Global QoL

73

50

69

52

70

56

75

68

  Physical

80

50

65

42

55

40

74

69

  Activity

77

55

68

53

62

42

78

62

  Emotional

64

51

63

51

45

36

65

65

  Social

86

56

74

48

58

55

69

69

  Cognitive

72

53

65

54

55

41

58

38

Symptom scales

  Fatigue

52

67

51

64

19

68

26

46

  Pain

13

60

22

58

29

52

25

49

  Nausea and vomiting

18

43

25

45

24

78

48

58

Symptoms and single items

  Dyspnoea

18

25

38

56

2

2

5

5

  Insomnia

23

53

25

45

25

35

19

39

  Anorexia

19

73

41

55

19

79

28

68

  Constipation

2

2

2

2

1

1

4

15

  Diarrhoea

2

2

2

2

0

0

44

79

  Financial impact

38

38

4

4

1

1

8

8

The analysis of nutrition-related factors and their relationships with QoL showed that energy and protein intake were correlated with function scales and with some symptoms: global QoL (r=0.53, P=0.001), physical (r=0.26, P=0.02) and emotional (r=0.29, P=0.01); anorexia (r=−0.52, P=0.001), fatigue (r=−0.60, P=0.001), pain (r=−0.55, P=0.003), nausea/vomiting (r=−0.51, P=0.003) and diarrhoea (r=−0.60, p=0.001). Malnutrition in these patients was associated with poorer function scales and with some symptoms: global QoL (P=0.05), physical (P=0.01), role (P=0.02), cognitive (P=0.02), emotional (P=0.01) and social (P=0.01); anorexia (P=0.001), increased fatigue (P=0.03), dyspnoea, insomnia and diarrhoea (P=0.04).

Given the strong interaction between QoL (dependent variable) and cancer stage and nutrition (independent variables), a nonparametric correlation analysis stratified by diagnosis was conducted. This analysis showed a distinct QoL pattern between diagnoses, and identified which variables were significantly associated with individual QoL global scores (Fig. 3, in which the vertical axes denote the global scores of function, symptom scales and single items derived from interpatient median values). Functional capacity for all diagnoses was significantly influenced by current nutritional intake deficit and recent weight loss, but was not affected by cancer stage (Fig. 3a, b). Both OES and STO cancer patients showed poorer global function scores than HN and CR cancer patients (P=0.02). Global symptom scores were strongly associated with cancer stage and were not significantly different between diagnoses nor were they influenced by nutritional parameters (Fig. 3c). In a similar manner to the global symptom scores, poorer global single item scores were only associated with stage III/IV cancer (Fig. 3d). Although there were no significant differences between HN, CR and OES cancer patients, they showed worse single item scores than STO cancer patients (P=0.03).
Fig. 3

a Global function scores according to diagnosis stratified by quartile of energy and protein intakes (□ 0–24%, P=0.003; ▲ 25–50%, P=0.01; ● 51–75%, P=0.04; ■ 76–100%, P=0.05). b Global function scores according to diagnosis stratified by category of significant percent weight loss (□ >10%, P=0.001; ■ 5–10%, P=0.06). c Global symptom scores according to diagnosis stratified by cancer stage (□ stage IV, P=0.001; ▲ stage III, P=0.002; ● stage II, P=0.04; ■ stage I, P=0.04). d Global single item scores according to diagnosis stratified by cancer stage (□ stage IV, ▲ stage III, IV+III P=0.001; ● stage II, ■ stage I, II+I P=0.05. The vertical axes denote the scores derived from patients’ median values (HN head-neck, OES oesophagus, STO stomach, CR colorectal)

Table 4 shows the results of a general linear model that included global QoL scores, nutritional parameters and cancer- and treatment-related variables in order to calculate effect size estimates and the respective statistics. Cancer location, chemotherapy and surgery were significantly associated with all QoL scores whilst stage was only associated with symptom scores and single item scores. Nutritional intake and weight loss were significantly associated only with function scores, although there was a trend for an association with symptom scores and single items (P=0.06).
Table 4

Interrelationships and estimates of effect size (relative weights) of nutritional parameters and cancer- and treatment-related variables on QoL (results from general linear model analysis). Columns denote dependent variables, and rows independent variables. Each of the scales and single items were linearly transformed and grouped to obtain global scores before inclusion in the analytical model

Variable

Global function scores

Global symptom scoresa

Global single item scoresa

F-test

Estimates of effect size (%)b

P value

F-test

Estimates of effect size (%)b

P value

F-test

Estimates of effect size (%)b

P value

Stage

1.6

1

0.18

56.5

22

0.001

103.7

30

0.0001

Location

111.2

30

0.0001

77.2

41

0.0001

49.2

20

0.001

Energy intake

27.2

10

0.01

1.0

3

0.35

3.9

4

0.07

Protein intake

27.2

10

0.01

1.0

4

0.25

4.2

5

0.07

Weight loss

133.7

30

0.0001

0.05

1

0.82

1.2

3

0.10

Duration of disease

1.5

3

0.14

10.0

7

0.06

1.2

3

0.30

Chemotherapy

35.3

10

0.001

2.1

4

0.22

1.3

1

0.25

Surgery

6.1

6

0.01

1.4

1

0.86

3.0

4

0.09

aDue to the potential association between symptoms and diagnoses, associations were adjusted for cancer location

bThe sum of percentages may not equal 100% due to the corrected error size

In order to evaluate which diagnosis was most strongly associated with poorer QoL, individual dimensions were grouped and valued according to their relative weights. STO cancer patients had the worst QoL although not significantly different from OES cancer patients. HN and CR cancer patients had a better QoL (P=0.02), with CR patients having a better QoL than HN patients (NS). Overall, the stage of disease was identified as the major determinant of patient QoL (P=0.002), closely followed by deterioration in nutritional status (P=0.005) and dietary intake (P=0.007).

Discussion

To be meaningful, QoL assessment must include the impact of the disease together with therapeutic interventions, expectations and personal satisfaction. Therefore, the EORTC-QLQ was chosen as the most effective tool [20]. The present study clearly showed that cancer patients’ QoL is multifactorial and that it is distinctively influenced by the disease, therapeutic interventions and various nutritional parameters.

Cancer-related nutritional deterioration has been traditionally attributed to anorexia and metabolic derangement [3, 21, 22]. Despite the fact that nutritional deterioration is associated with functional impairment [6], the interaction between nutrition and QoL is as-yet unexplored [7]. Artificial nutrition in cancer patients has been suggested to maintain nutritional status and QoL [23, 24]. Indeed, fatigue, anorexia and emotional stress, common in cancer patients, may further aggravate, but also be worsened by, poor nutritional intake and QoL [25, 26].

This study focused on the evaluation of the potential and relative weight of nutritional baseline data amongst other potential QoL determinants. In cancer, longstanding energy and substrate deficits have not previously been investigated nor adjusted by the patients’ disease stage. Our results demonstrated marked nutritional intake deficits in the advanced stages of disease. There was not only a significant decrease from the usual energy and protein intake in stage III/IV disease, but current intakes were also markedly lower than in stage I/II disease. Stage III/IV HN and OES cancer were shown to be associated with the most severe energy and protein depletion. An advanced stage of disease was, indeed, the common denominator of patients’ nutritional deterioration. Weight loss and reduced energy and protein intake were associated (P=0.06), although there was no consistent pattern. Our results corroborate and expand previous observations that progressive nutritional intake deficit may be associated with cancer location [3, 4, 27, 28], and may eventually be proportional to the extent of the disease [29].

Concerning QoL dimensions, cancer stage mainly influenced the severity of symptom scales and single items, while energy and protein intake deficits and weight loss were detrimental to QoL function scales. Overall, patients’ functional capacity was affected by cancer location and nutritional factors (in about 40% each), and by the stage of disease in only 1%. The relative contributions ascribed to chemotherapy, surgery and duration of disease were 10%, 6% and 3%, respectively, as previously suggested in different patient groups [30, 31, 32]. The symptom scales had an inverse pattern compared with the function scales: 41% were attributed to cancer location, 22% to stage, 7% to nutritional intake, 7% to the duration of the disease, 4% to surgery, 1% to weight loss and 0.01% to chemotherapy. Likewise, cancer stage and location were the major determinants of QoL single items, which were worse in stage III/IV. Altogether, although the stage of disease was the major determinant of patient QoL, in some diagnoses the impact of nutritional deterioration combined with deficient intake may be more clinically important.

This study of 271 patients with HN, OES, STO and CR cancer provides objective evidence that cancer, dietary deficits, nutritional depletion and therapeutic interventions are determinants of patient’s QoL, but with distinct relative weights. Whereas chemotherapy and surgery were viewed by the patients as of minor relevance, our data are consistent with the hypothesized relationship between wasting and progressive disease [29]. Although nutritional deficits and/or deterioration were intrinsic to the site and stage of disease, reduced energy and protein intakes and weight loss were independent determinants of QoL. Our results concur with those of the landmark study by Keys et al. which showed that semistarvation impairs functional and psychological abilities [33]. It is of clinical relevance that individualized nutritional counselling and education appears to effectively maintain and improve nutritional intake and status, and to significantly improve patients’ overall QoL [34, 35].

Acknowledgements

We are indebted to the medical, nursing and technical staff of the Radiotherapy Department of the Hospital of Santa Maria. We express our gratitude to Ms. Pat Howard for the insights and thorough revision of the text. This study was supported by a grant from Fundação para a Ciência e Tecnologia (RUN 437).

Copyright information

© Springer-Verlag 2004