Food intake during cancer therapy: a systematic review

Data Sources and Searches

This systematic review was conducted in accordance with PRISMA guidelines.¹⁴ The literature search occurred in July 2018. To identify studies for this review, we developed detailed strategies for each database searched with the help of a research informationist (LB). We based these on the search strategy developed for MEDLINE (PubMed), but revised appropriately for each of the following databases: Embase (Elsevier) and Cochrane Central Register of Controlled Trials (Cochrane Library). The search strategy used a combination of controlled vocabulary and free-text terms with truncation published from January 2008 through July 2018. Only studies in English were included. Animal studies were excluded. The following concepts and terms were used: (1) Neoplasms, Patients, Therapeutics; (2) Food, Diet, Diet Therapy; (3) Eating, Consume, Ingest, Intake, Feed; (4) Recurrence, Disease Progression, Second Primary Neoplasms, Mucositis, Nausea, Vomiting. The following concepts were excluded from the search strategy: Nutritional Support, Dietary Supplements, and Gastrointestinal Intubation. Concepts were combined using the AND operator. The complete PubMed strategy is available in Appendix 1, Supplemental Digital Content 1, http://links.lww.com/AJCO/A350. All search results were combined in a bibliographic management tool (EndNote) and duplicates were eliminated using the Bramer method for deduplication in Endnote.¹⁵ The final PubMed search strategy can be found in Figure 1 of the Appendix, Supplemental Digital Content 1, http://links.lww.com/AJCO/A350.

Study Selection

Nine studies were included in the final review. Risk of bias for the included studies was assessed independently by two review authors using the Cochran risk of bias tool.¹⁶ Figure 2 in the Appendix, Supplemental Digital Content 1, http://links.lww.com/AJCO/A350 shows the flow chart of the search strategy.

Studies were excluded if they focused solely on nutrient supplementation rather than food consumption. Studies with a focus on foods consumed prior to cancer diagnosis or following the conclusion of treatment were not included. In addition, studies that focused on foods used as topical therapy for mucositis were excluded.

Data Extraction and Quality Assessment

We included interventional and observational studies. All participants were adults. The types of cancers studied in the included papers were colorectal, prostate, esophageal, head and neck, breast, endometrial, ovarian, testicular, pelvic, or unspecified type. The interventions included comparisons between types of diets, consumption of a specific food, amount of food consumed, or energy and protein intakes. A summary of the studies can be found in Table 1 in the Appendix, Supplemental Digital Content 1, http://links.lww.com/AJCO/A350.

Analysis

Details on diet, study type, and outcomes were abstracted and reviewed.

Role of the Funding Source

Funds were provided by the Meg Berté Owen Fund and the NCI (P30CA008748). Funders had no role in the study design, conduct, or reporting.

Data Synthesis and Analysis

The articles included in this review are divided by study type and subdivided by focus on dietary patterns or a single food. Randomized controlled trials are outlined in Table 1, while additional studies are summarized in Supplemental Data.

Randomized Control Trials – Dietary Patterns

Artene et al. followed 165 breast cancer patients for 12 months to determine the effect of a high protein diet and isometric exercises on treatment-related weight loss during antiestrogenic treatment (Table 1). The patients were divided into either a high protein diet group (which consisted of foods naturally high in protein, omega-3 fatty acids, calcium, and pro- and prebiotics) or a high protein diet group with isometric exercises. Both groups experienced weight loss, while the diet and exercise group also experienced fat loss. Patients who underwent a mastectomy experienced weight and fat loss for both groups, while patients with breast-conserving surgery only reached significant weight and fat loss in the diet and exercise group. Chemotherapy type did not affect the measured outcomes in the diet and exercise group, while the diet only group saw some differences between adjuvant and neoadjuvant, with patients not experiencing fat loss who underwent the additional chemotherapies.¹⁷

Two studies focused on short-term fasting (STF) diets to determine the impact on treatment side effects. De Groot et al. conducted a randomized controlled trial (RCT) of 13 women with breast cancer receiving neoadjuvant chemotherapy to assess the impact of STF on treatment side effects and hematologic parameters. Patients were randomized to a STF group with a fast lasting 24-hours before and 24-hours after chemotherapy or to a non-STF group instructed to eat a healthy diet during the treatment period. No significant differences were found in self-reported toxicities between the two groups, although the fasting group had higher mean red blood cell and platelet counts, compared to the non-fasting group.¹⁸ A cross-over study by Bauersfeld et al. had a slightly larger sample size, in which 34 breast and ovarian cancer patients were recruited to determine the effect of STF on quality of life (QOL) and fatigue during chemotherapy. The patients were randomized into two groups: 60-hour fasting during the first 3 of 6 chemotherapy cycles (Group A) or 60-hour fasting during the last 3 of 6 chemotherapy cycles (Group B). Short-term fasting in both cohorts was associated with better tolerance to chemotherapy, with less compromised quality-of-life and fatigue. Patients in both studies reported that STF was feasible and well-accepted during treatment.

Because previous studies had shown that many patients undergoing treatment for breast cancer have treatment-related weight gain, Villarini et al. conducted a dietary intervention RCT in 94 breast cancer patients with the aim of preventing weight gain during treatment. The intervention group received dietary recommendations with cooking classes and common meals at least twice per week; the control group received general recommendations based on the Mediterranean diet and macrobiotic recipes. The intervention group had significant reductions in body weight, waist and hip circumferences, and skinfold measurements, compared to the control group.¹⁹ Contrary to the investigators’ expectations, however, the control group also lost weight, suggesting that general dietary recommendations may be sufficient to prevent weight gain during breast cancer treatment.

The Women’s Intervention Nutrition Study (WINS) was a RCT of 2,437 breast cancer patients undergoing cancer therapy to determine the effect of a low-fat diet during treatment on cancer recurrence. The intervention started with a four-month intensive phase with biweekly in-person counseling sessions followed by a maintenance phase including individual counseling sessions every three months and optional monthly group sessions. Dietary fat intake was significantly reduced in the intervention group, compared to the control, (percent energy from fat at 60 months 23.2%±8.4% vs 31.2%±8.9%, respectively, P<0.0001) and was associated with a mean 6.1 pound weight difference between the groups (P=0.005).²⁰ In long-term analyses, the low-fat diet had no impact on relapse-free survival, although a post-hoc exploratory analysis suggested that women with estrogen receptor negative tumors may have benefitted.^{21, 22}

Soto-Lugo et al. conducted a RCT among 26 cervical or endometrial cancer patients to determine if a diet low in fermentable oligosaccharides, disaccharides, monosaccharides, and polyols (FODMAP) can decrease gastrointestinal toxicity during pelvic external beam radiotherapy. The main symptoms reported were nausea, vomiting, and diarrhea. Patients were assigned to a FODMAP or standard Mexican diet group. The researchers did not find significant differences in toxicities of any grade between the groups and determined that the FODMAP diet did not result in improved outcomes compared to a standard diet.²³

Wedlake et al. conducted a RCT of 166 pelvic cancer patients receiving radiotherapy to ascertain if fiber intake influences radiation-induced gastrointestinal toxicities; the study was motivated by anecdotal reports of patients being advised to restrict fiber intake prior to pelvic radiation in order to reduce gastrointestinal toxicities. Patients were randomized into a low-fiber group, a habitual fiber group, or a high-fiber group. There was a significant reduction in severity of bowel symptoms experienced in the high fiber group compared to the habitual fiber group, but there was no gradient of effect across the diets.²⁴ There was a trend towards improvement in the low-fiber group compared to the habitual fiber group, but this was not statistically significant; the authors concluded that the recommendation to restrict fiber intake in this population be should therefore be abandoned.

Randomized Control Trials – Specific Food Consumption

Jafarpour-Sadegh et al. investigated fresh yellow onion consumption in an RCT of 56 breast cancer patients undergoing chemotherapy to assess its effect on treatment efficacy and related toxicities, as determined by serum measurements of carcinoembryonic antigen (CEA), cancer antigen 125 (CA125), AST, ALT, and ALP. Patients were assigned to a high onion intervention group of 100-160g/day or a low onion control group of 30-40g/day. Mean changes in serum AST, ALT, and ALP did not statistically differ between the groups, though there was a trend towards fewer hepatotoxic effects in the high onion group. Both groups had mean decreases in CEA and CA125 serum levels from baseline, though the drop was only statistically significant in the high onion group and clinical significance was not demonstrated.²⁵

Observational Studies – Dietary Patterns

De Vries et al. conducted an observational study of chemotherapy-related symptoms and diet characteristics (Supplemental Table 1). Their study included 117 breast cancer patients receiving chemotherapy and 88 women without cancer. Food frequency questionnaires (FFQ) and 24-hour recalls were used; the baseline diet was similar for both groups. During treatment, breast cancer patients had less energy, protein, fat, and alcohol intake than the comparison group. Patients undergoing chemotherapy reported more side effects than the comparison group, including dry mouth, lack of energy, nausea, and difficulty chewing, all of which were associated with lower energy intake.²⁶

Di Fiore et al. collected nutrition parameters on 101 patients with esophageal cancer to study whether nutritional status during treatment was associated with greater treatment-related toxicity, recurrence, or mortality. Nutrition parameters were retrospectively collected at baseline, at 5-8 weeks follow up, and at 11 weeks follow up. Median overall survival among participants that were undernourished during treatment was 25 months, compared to 42 months among well-nourished participants. Undernourished participants showed a significantly lower treatment response rate than their well-nourished counterparts. Among participants who responded to chemotherapy, underweight or undernourished participants had worse median survival than well-nourished participants (33 vs 59 months, p<0.001 and 29 vs 61 months, p=0.001, respectively).²⁷

Zahn et al. conducted a prospective, single-arm, observational study of 40 head and neck cancer patients receiving radiation therapy to assess the impact of nutrition on oral mucositis. A registered dietician established nutrition goals for patients and oral mucositis symptom severity was measured weekly. Meeting the protein goal of 1.5g/kg/day in the same or previous week was associated with lower oral mucositis severity; no significant associations with other measured nutrition parameters were identified.²⁸

IJpma et al. analyzed side effects of changes in taste and smell, food preferences, dietary intake, and body composition among 21 testicular cancer patients undergoing cisplatin-based chemotherapy compared to 48 healthy controls with similar demographics. Testicular cancer patients reported lower taste function, appetite, and hunger, than controls. Food preference varied over the study duration, but cancer patients described a drop in preference for high protein savory foods and an increase in preference for high protein sweet foods. Among controls, no changes were found in dietary intake or food preference.²⁹

Ganzer et al. conducted a cross-sectional study on the relationship between oral symptom burden (xerostomia, thick secretions, and mucosal sensitivity), energy and protein intake, and weight change in 43 participants with head and neck cancer. Energy and protein intake were collected using 24-hour dietary recalls excluding ten participants using feeding tubes, whose intake was determined by formula type and volume. Xerostomia and mucosal sensitivity were associated with decreased energy and protein intake at the mid-recovery stage (4.0-9.9 months post-completed chemoradiation), whereas thick phlegm was not. Overall, the authors found that the symptom burden was high at the beginning of the recovery process and lessened over time, but was never completely gone among any of the participants, regardless of dietary intake or weight change.³⁰

The impact of energy and protein intake on fatigue and 6-month mortality was in investigated in a 285-patient prospective observational study by Stobaus et al.³⁷ Mortality was significantly higher among advanced cancer patients undergoing chemotherapy who were found to have a low protein intake compared to high protein intake. Protein intake of less than one gram per kilogram body weight was the factor most strongly associated with fatigue and mortality, even after adjustment for treatment-related weight loss.

Ingersoll et al. investigated whether Concord grape juice would decrease the severity and incidence of chemotherapy-induced nausea and vomiting among 76 participants with cancer by lessening the severity of chemotherapy-induced cellular damage. The authors measured all participant food intake to account for flavonoid-enhancing or -depleting foods. Participants consumed grape juice or a grape-flavored placebo for each week following four chemotherapy treatments, in addition to the standard supportive care. While the study suffered from a high attrition rate of 50%, which prevented any statistically significant results, the intervention group had a trend towards reduction in nausea frequency, nausea distress, and vomiting distress compared to the placebo group.³¹

Observational Studies - Specific Food Consumption

Richman et al. conducted a large-scale prospective study on the intake of fruits and vegetables following a diagnosis of non-metastatic prostate cancer among 1,560 men to ascertain if specific food groups have varying effects on cancer prognosis. The authors used FFQs to determine dietary intake, and prostate progression was defined as recurrence, secondary treatment, metastasis, or prostate cancer-related death. Cruciferous vegetable intake was found to have an inverse relationship with prostate cancer progression: HR 0.41 (95% CI 0.22, 0.76) when comparing the highest quartile of cruciferous vegetable intake to the lowest quartile.³²

Three separate prospective studies investigated specific food intake among patients with stage III colon cancer undergoing treatment through a National Cancer Institute clinical trial. All three studies used FFQs to determine dietary intake, and the outcomes measured were disease-free survival (DFS), recurrence-free survival (RFS), and overall survival (OS). In a prospective observational study of 826 patients with stage III colon cancer, Fadelu et al determined that increased consumption of tree nuts was associated with both improved disease-free and overall survival after a median follow up 6.5 years.³³ Fuchs et al. conducted a prospective observational study of sugar-sweetened beverages and recurrence or mortality among 1,011 patients with stage III colon cancer. After a median follow-up of 7.3 years, consuming 2 or more servings of sugar sweetened beverages per day was associated with a greater risk of cancer recurrence and mortality (HR 1.67, CI 1.04-2.68), with the strongest association among patients who were both overweight and sedentary.³⁴ In a study with 953 participating colon cancer patients, Guercio et al. found that increased total coffee intake was associated with a reduction in cancer recurrence and mortality and improvement in OS. These results did not hold for decaffeinated coffee or herbal tea.³⁵