Abstract

Context: Evidence of an association between dietary patterns and individual foods and the risk of overall mortality among cancer survivors has not been reviewed systematically. Objective: The aim of this meta-analysis of cohort studies was to investigate the association between food intake and dietary patterns and overall mortality among cancer survivors. Data Sources: The PubMed and Embase databases were searched. Study Selection: A total of 117 studies enrolling 209 597 cancer survivors were included. Data Extraction: The following data were extracted: study location, types of outcome, population characteristics, dietary assessment method, risk estimates, and adjustment factors. Results: Higher intakes of vegetables and fish were inversely associated with overall mortality, and higher alcohol consumption was positively associated with overall mortality (RR, 1.08; 95%CI, 1.02–1.16). Adherence to the highest category of diet quality was inversely associated with overall mortality (RR, 0.78; 95%CI, 0.72–0.85; postdiagnosis RR, 0.79; 95%CI, 0.71–0.89), as was adherence to the highest category of a prudent/healthy dietary pattern (RR, 0.81; 95%CI, 0.67–0.98; postdiagnosis RR, 0.77; 95%CI, 0.60–0.99). The Western dietary pattern was associated with increased risk of overall mortality (RR, 1.46; 95%CI, 1.27–1.68; postdiagnosis RR, 1.51; 95%CI, 1.24–1.85). Conclusion: Adherence to a high-quality diet and a prudent/healthy dietary pattern is inversely associated with overall mortality among cancer survivors, whereas a Western dietary pattern is positively associated with overall mortality in this population.

INTRODUCTION

The number of cancer survivors in the United States and Europe is growing rapidly.1,2 The needs and metabolic functions of this group, who have experienced a malignant process and its corresponding treatment, might differ from those of healthy populations. To date, however, the evidence to formulate specific dietary recommendations for this group has been considered insufficient,3 despite the widely held conviction that better insight into both the role of diet after cancer and the relation of diet to survival might offer further opportunities to improve long-term treatment options for this growing population. Whereas recent meta-analyses of observational studies clearly demonstrate reduced risks of different types of cancer among healthy populations that follow specific dietary regimens,4–7 a comparable synthesis of empirical data for cancer survivors, such as a systematic review and meta-analysis, is available only for breast cancer survivors, and even then, the literature is sparse.8 Furthermore, prospective cohort studies that have investigated the association between composition of diet and overall mortality among cancer survivors show contradicting results.9 A further complicating factor is that some studies evaluated the relationship between mortality and nutrients instead of between mortality and food patterns,9,10 the latter being more easily understood and reflecting more accurately the lifestyles of the targeted population. Therefore, the current dietary advice for cancer survivors is to receive personalized nutritional care from a trained professional and to follow current recommendations for cancer prevention, if not otherwise advised.3

Most of the single cohort studies on cancer survivors have focused on associations between prediagnosis dietary intake, especially alcohol intake, and overall mortality and/or cancer recurrence, while only a few studies have assessed associations with postdiagnosis dietary intake11–13 Clinical trials enrolling subjects with existing cancer have the advantage of possessing data on the postdiagnosis diet; however, such studies often focus on nutrients rather than dietary patterns or food groups and are weakly suited for studying long-term successful dietary interventions.

A better understanding of the effect of food intake and food-based dietary patterns on overall mortality and recurrence among cancer survivors can inform the development of tailored prevention strategies in the important area of modifiable lifestyle factors such as diet.

Therefore, the aim of this study was to conduct a meta-analysis of all cohort studies that investigated the effects of adherence to diet-quality indices, dietary patterns, and food and beverage consumption on overall mortality and cancer recurrence among cancer survivors.

METHODS

The review protocol has been registered in the PROSPERO International Prospective Register of Systematic Reviews (no. CRD42015023684).

Data sources and searches

The literature search was performed using the electronic databases PubMed (until May 17, 2016) and Embase (until May 17, 2016), with no restrictions on calendar date, using the following search terms: (“diet” OR “dietary” OR “food” OR “foods” OR “alcohol”) AND (“cancer”) AND (“survivors” OR “survivor” OR “recurrence”) AND (“prospective” OR “cohort” OR “longitudinal” OR “follow up”). The search strategy had no language restrictions. Moreover, the reference lists of reviews and retrieved articles were checked as well. The literature search was conducted by 2 authors (C.S. and L.S.); disagreements were resolved by discussion. MOOSE guidelines14 were followed when planning, conducting, and reporting the meta-analysis.

Study selection

Studies were included in the systematic review and meta-analyses if they met the following criteria (Table 1): (1) cohort study design (prospective and retrospective); (2) exposure to individual food components such as fruits, vegetables, dairy, meat, fish, cereals, and bread; (3) exposure to beverages such as alcohol, coffee, and tea; (4) use of diet-quality indices (Mediterranean diet, DASH [Dietary Approaches to Stop Hypertension], Healthy Eating Index, WCRF/AIRC [World Cancer Research Fund/American Institute for Cancer Research] dietary guidelines adherence score) that were determined a priori; (5) examination of data-driven dietary patterns (healthy/prudent, unhealthy/Western); (6) overall mortality and/or cancer recurrence included in reported outcomes; and (7) study participants limited to adult populations (≥18 years old) of survivors of various cancer types (the definition of “cancer survivor” does not include survivors of cervical lesions or adenomas in the colon). When a study seemed to have been published in duplicate, the article containing the most comprehensive information (longest follow-up period) was selected. Finally, studies reporting cancer-specific mortality (but no overall mortality) were excluded.

Table 1

PICOS criteria for inclusion and exclusion of studies

Criteria Description 
Participants Survivors of various types of cancer, adult populations (aged >18 y) only. The definition of “cancer survivor” here does not include survivors of cervical lesions or adenomas in the colon 
Interventions/exposure 
  • Intake of individual foods: fruits, vegetables, dairy, meat, fish, cereals, and bread

  • Intake of beverages: alcohol, coffee, tea

  • A priori–based diet-quality indices (Mediterranean diet, DASH, Healthy Eating Index, WCRF/AIRC dietary guidelines adherence score)

  • Adherence to data-driven dietary patterns (healthy/prudent, unhealthy/Western)

 
Comparison Highest vs lowest categories of exposure 
Outcome Overall mortality and/or cancer recurrence 
Study design Cohort studies (prospective and retrospective) 
Criteria Description 
Participants Survivors of various types of cancer, adult populations (aged >18 y) only. The definition of “cancer survivor” here does not include survivors of cervical lesions or adenomas in the colon 
Interventions/exposure 
  • Intake of individual foods: fruits, vegetables, dairy, meat, fish, cereals, and bread

  • Intake of beverages: alcohol, coffee, tea

  • A priori–based diet-quality indices (Mediterranean diet, DASH, Healthy Eating Index, WCRF/AIRC dietary guidelines adherence score)

  • Adherence to data-driven dietary patterns (healthy/prudent, unhealthy/Western)

 
Comparison Highest vs lowest categories of exposure 
Outcome Overall mortality and/or cancer recurrence 
Study design Cohort studies (prospective and retrospective) 

Abbreviations: DASH, Dietary Approaches to Stop Hypertension; WCRF/AIRC, World Cancer Research Fund/American Institute for Cancer Research.

Data extraction

The following data were extracted from each study: (1) author’s last name; (2) country of origin and name of the study; (3) types of outcome (overall mortality, cancer recurrence); (4) sample size of the population and length of the study (mean follow-up duration in years); (5) age of participants at entry; (6) sex of participants; (7) tumor characteristics; (8) assessment of recurrence; (9) exposure assessment, including timeframe; (10) components of diet-quality indices/dietary patterns (including score range); (11) adjustment factors; and (12) multivariate-adjusted risk estimates (risk ratio [RR] or hazard ratio for the highest vs the lowest category) along with their corresponding 95%CIs. If only separate risk estimates were available for specific foods (ie, processed meat and unprocessed meat) or beverages within a food/beverages category (ie, meat), the risk estimates were pooled (using a fixed-effects model) and were treated as 1 study. Moreover, if different categories of alcohol consumption (ie, current and ex-drinkers) were combined (the lowest intake category was considered the reference category), these estimates were pooled (using a fixed-effects model). When a study provided several risk estimates, the multivariate-adjusted model was chosen. Risk estimates were pooled if at least 3 studies reported data for the same exposures (eg, diet-quality indices, dietary patterns, and intakes of fruit, vegetables, dairy, meat, fish, cereals and bread, and beverages [alcohol, tea, coffee]) for overall mortality and cancer recurrence.

Statistical analysis

The meta-analysis was performed by combining the multivariable-adjusted RRs or hazard ratios of the highest dietary categories compared with the lowest dietary categories using a random-effects model (using the Der Simonian-Laird method15), which incorporated both within- and between-study variability. This model was chosen because exposures were not rare, and the heterogeneity modeling was deemed important. For clarity, studies of each exposure were grouped by type of cancer (analyzed jointly and separately) when calculating overall mortality and cancer recurrence. As an initial approach, the prediagnosis and the postdiagnosis assessments of exposure were combined and then stratified.

Main analyses assessed food categories before and after diagnosis for all types of cancer. Separate subgroup analyses were performed for studies that assessed postdiagnosis diet and for all studies grouped by type of cancer.

Heterogeneity was estimated by the Cochrane Q test, together with the I2 test. An I2 of >50% was considered an indicator of substantial heterogeneity across studies, as previously reported.16,17 The heterogi command in Stata software (Stata Corp; College Station, TX, USA) was used to calculate the 95%CI for the heterogeneity estimates. Funnel plots were used to assess small-study effects if at least 10 studies were available, as recommended by the Cochrane Collaboration.18 Moreover, Egger’s test was performed to test for small-study effects.19 The presence of publication bias was assessed by the symmetry of contour-enhanced funnel plots by plotting risk estimates against their corresponding standard error; shaded areas indicate whether studies with significant or nonsignificant effects seem to be missing.20

All analyses were conducted using Review Manager version 5.3 (Nordic Cochrane Collection; Copenhagen, Denmark) and Stata 13.0.

RESULTS

Literature search and study characteristics

The detailed steps of the systematic search and selection process are shown as an adapted PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram in Figure 1. Table S1 in the Supporting Information online provides a list of full-text articles that were excluded, with reasons.

Figure 1

Flow diagram of the literature search process. Number of records by dietary pattern/indices, foods/food groups, and beverages may not add up because some studies were listed for multiple exposures (dietary patterns, foods, beverages).

Figure 1

Flow diagram of the literature search process. Number of records by dietary pattern/indices, foods/food groups, and beverages may not add up because some studies were listed for multiple exposures (dietary patterns, foods, beverages).

Overall, 117 cohort studies extracted from 1658 studies (excluding duplicates) were identified and included in the qualitative synthesis (112 included in the quantitative synthesis) (Table S2 in the Supporting Information online).

General study characteristics are summarized in Table S2 in the Supporting Information online. Overall, there were 60 134 mortality cases and 9297 recurrence cases. This information was not available for 14 studies, which provided survival or disease-/recurrence-free survival rates (11 provided no information on explicit cases of mortality, and 6 provided no information on cases of cancer recurrence). The total number of participants in the included studies was 209 597. The breakdown of studies according to cancer type is as follows: 41 studies on breast cancer survivors, 18 studies on colorectal cancer survivors, 16 studies on head and neck cancer survivors, and 13 studies on gastroesophageal cancer survivors. The sample sizes varied between 57 in a study of head and neck cancer patients from a single institution in Switzerland to 22 890 in a large breast cancer cohort from the United States. The follow-up periods varied from 1.2 to 16 years.

Nine of the included studies used diet-quality scores defined a priori as the exposure measure. Seven of these studies used the Healthy Eating Index (2010 or 2005) or an adapted version of it.21 A Mediterranean diet score was used in 3 studies and the DASH diet as a score in 2 studies. A WCRF/AIRC dietary guidelines adherence score was used in 1 study, assigning a point to every dietary recommendation met. Furthermore, another study used 2 healthy-diet scores (the Diet Quality Index Revised [DQIR]22 and the Recommended Food Score23) that had been developed and used in previous studies. Similar to the other scores mentioned, these indices assign points to a diet high in whole grains, fruit, and vegetable consumption and lower in fat consumption. The DQIR additionally awards points for diet diversity. For the quantitative synthesis, the following diet-quality indices were combined: the alternate Healthy Eating Index (2010),21 the Healthy Eating Index (2005),24 the WCRF/AIRC (dietary) guidelines adherence score,25 the alternate Mediterranean diet,26,27 the DASH diet,28 the DQIR,22 and the Recommended Food Score.23 All indices for dietary quality focus on high intake of fruit and vegetables, the inclusion of whole grains, legumes and nuts, and low intake of meat (mainly red and processed meat). Eight of these 9 studies were included in the meta-analysis. The ninth study was not included because another study on the same cancer cohort provided a longer follow-up period.

Nine studies assessed the effects of dietary patterns on overall mortality or cancer recurrence. All of these studies used principal component analysis to derive the patterns, which, in most cases, were termed whole-foods, prudent, or healthy if they described a pattern of intake high in fruit and vegetables and whole grains but low in red and processed meat, refined grains, and high-fat foods. Otherwise, patterns describing a diet perceived as unhealthy were mostly labeled as Western, high-fat or high-sugar, snacks, or unhealthy. For the purpose of this study, patterns were grouped into prudent/healthy and Western/unhealthy dietary patterns. The prudent/healthy food pattern includes high intakes of fruit and vegetables, whole grains, poultry, and low-fat dairy products, whereas the Western/unhealthy dietary pattern is based on high intakes of red and processed meat, refined grains, sweets and desserts, and high-fat dairy products (Table S2 in the Supporting Information online).

Enumerative data for food groups and beverages are shown in Tables 2 and 3.

Table 2

Risk of overall mortality comparing the highest vs the lowest category of pre-/post diagnosis dietary exposure (random effects analyses data only)

Exposure No. of studies Risk ratio (95%CI) I2 (95%CI) 
Diet-quality indices 0.78 (0.72–0.85) 0% (0–68%) 
 Breast cancer 0.74 (0.60–0.90) 6% (0–90%) 
 Postdiagnosis 0.79 (0.71–0.89) 0% (0–79%) 
Prudent/healthy dietary pattern 0.81 (0.67–0.98) 44% (0–75%) 
 Breast cancer 0.76 (0.60–0.95) 4% (0–90%) 
 Colorectal cancer 1.06 (0.83–1.36) 0% (0–90%) 
 Postdiagnosis 0.77 (0.60–0.99) 56% (0–82%) 
Western dietary pattern 1.46 (1.27–1.68) 0% (0–68%) 
 Breast cancer 1.44 (1.17–1.77) 0% (0–90%) 
 Colorectal cancer 1.55 (1.13–2.13) 35% (0–79%) 
 Postdiagnosis 1.51 (1.24–1.85) 17% (0–62%) 
Fruit consumption 19 0.94 (0.87–1.01) 5% (0–52%) 
 Breast cancer 0.93 (0.75–1.17) 33% (0–75%) 
 Non-Hodgkin lymphoma 1.00 (0.88–1.14) 0% (0–90%) 
 Postdiagnosis 0.96 (0.64–1.45) 49% (0–88%) 
Vegetable consumption 21 0.86 (0.79–0.94) 43% (5–66%) 
 Breast cancer 0.97 (0.84–1.13) 37% (0–73%) 
 Postdiagnosis 0.88 (0.59–1.30) 78% (39–92%) 
Dairy consumption 13 1.08 (0.94–1.23) 66% (38–81%) 
 Breast cancer 1.06 (0.74–1.53) 78% (27–93%) 
 Colorectal cancer 0.93 (0.70–1.26) 76% (22–93%) 
 Postdiagnosis 1.02 (0.75–1.37) 80% (56–91%) 
Meat consumption 17 0.91 (0.83–1.01) 53% (18–73%) 
 Breast cancer 0.97 (0.73–1.28) 48% (0–83%) 
 Colorectal cancer 1.10 (0.84–1.43) 54% (0–85%) 
 Postdiagnosis 0.93 (0.75–1.17) 35% (0–77%) 
Fish consumption 0.85 (0.78–0.93) 0% (0–71%) 
Cereals and bread consumption 1.03 (0.87–1.22) 0% (0–75%) 
Egg consumption 1.10 (0.84–1.44) 0% (0–85%) 
Alcohol consumption 63 1.08 (1.02–1.16) 70% (61–77%) 
 Breast cancer 21 0.94 (0.85–1.04) 59% (34–75%) 
 Colorectal cancer 0.95 (0.86–1.06) 5% (0–80%) 
 Hepatocellular carcinoma 1.21 (1.07–1.36) 0% (0–90%) 
 Non-Hodgkin lymphoma 1.33 (1.10–1.63) 46% (0–82%) 
 Laryngeal and pharyngeal cancer 1.48 (1.08–2.02) 49% (0–78%) 
 Head and neck cancer 1.39 (1.10–1.76) 53% (0–83%) 
 Gastric and esophageal cancer 12 1.14 (0.98–1.34) 69% (44–83%) 
 Postdiagnosis 15 0.94 (0.81–1.11) 63% (36–79%) 
Tea consumption 0.78 (0.52–1.19) 33% (0–76%) 
Exposure No. of studies Risk ratio (95%CI) I2 (95%CI) 
Diet-quality indices 0.78 (0.72–0.85) 0% (0–68%) 
 Breast cancer 0.74 (0.60–0.90) 6% (0–90%) 
 Postdiagnosis 0.79 (0.71–0.89) 0% (0–79%) 
Prudent/healthy dietary pattern 0.81 (0.67–0.98) 44% (0–75%) 
 Breast cancer 0.76 (0.60–0.95) 4% (0–90%) 
 Colorectal cancer 1.06 (0.83–1.36) 0% (0–90%) 
 Postdiagnosis 0.77 (0.60–0.99) 56% (0–82%) 
Western dietary pattern 1.46 (1.27–1.68) 0% (0–68%) 
 Breast cancer 1.44 (1.17–1.77) 0% (0–90%) 
 Colorectal cancer 1.55 (1.13–2.13) 35% (0–79%) 
 Postdiagnosis 1.51 (1.24–1.85) 17% (0–62%) 
Fruit consumption 19 0.94 (0.87–1.01) 5% (0–52%) 
 Breast cancer 0.93 (0.75–1.17) 33% (0–75%) 
 Non-Hodgkin lymphoma 1.00 (0.88–1.14) 0% (0–90%) 
 Postdiagnosis 0.96 (0.64–1.45) 49% (0–88%) 
Vegetable consumption 21 0.86 (0.79–0.94) 43% (5–66%) 
 Breast cancer 0.97 (0.84–1.13) 37% (0–73%) 
 Postdiagnosis 0.88 (0.59–1.30) 78% (39–92%) 
Dairy consumption 13 1.08 (0.94–1.23) 66% (38–81%) 
 Breast cancer 1.06 (0.74–1.53) 78% (27–93%) 
 Colorectal cancer 0.93 (0.70–1.26) 76% (22–93%) 
 Postdiagnosis 1.02 (0.75–1.37) 80% (56–91%) 
Meat consumption 17 0.91 (0.83–1.01) 53% (18–73%) 
 Breast cancer 0.97 (0.73–1.28) 48% (0–83%) 
 Colorectal cancer 1.10 (0.84–1.43) 54% (0–85%) 
 Postdiagnosis 0.93 (0.75–1.17) 35% (0–77%) 
Fish consumption 0.85 (0.78–0.93) 0% (0–71%) 
Cereals and bread consumption 1.03 (0.87–1.22) 0% (0–75%) 
Egg consumption 1.10 (0.84–1.44) 0% (0–85%) 
Alcohol consumption 63 1.08 (1.02–1.16) 70% (61–77%) 
 Breast cancer 21 0.94 (0.85–1.04) 59% (34–75%) 
 Colorectal cancer 0.95 (0.86–1.06) 5% (0–80%) 
 Hepatocellular carcinoma 1.21 (1.07–1.36) 0% (0–90%) 
 Non-Hodgkin lymphoma 1.33 (1.10–1.63) 46% (0–82%) 
 Laryngeal and pharyngeal cancer 1.48 (1.08–2.02) 49% (0–78%) 
 Head and neck cancer 1.39 (1.10–1.76) 53% (0–83%) 
 Gastric and esophageal cancer 12 1.14 (0.98–1.34) 69% (44–83%) 
 Postdiagnosis 15 0.94 (0.81–1.11) 63% (36–79%) 
Tea consumption 0.78 (0.52–1.19) 33% (0–76%) 
Table 3

Risk of cancer recurrence comparing the highest vs lowest category of pre-/postdiagnosis dietary exposure (random effects analyses data only)

Exposure No. of studies Risk ratio (95%CI) I2 (95%CI) 
Prudent/healthy dietary pattern 0.87 (0.68–1.11) 24% (0–88%) 
 Postdiagnosis 0.94 (0.71–1.24) 19% (0–92%) 
Western dietary pattern 1.21 (0.69–2.13) 81% (51–93%) 
 Postdiagnosis 1.34 (0.61–2.92) 85% (54–95%) 
Vegetable consumption 0.99 (0.74–1.33) 69% (0–91%) 
Alcohol consumption 17 1.17 (1.05–1.31) 38% (0–65%) 
 Breast cancer 1.21 (1.06–1.39) 23% (0–66%) 
 Hepatocellular carcinoma 1.34 (0.73–2.46) 73% (25–90%) 
 Postdiagnosis 1.31 (1.04–1.66) 54% (0–85%) 
Tea consumption 0.76 (0.58–1.01) 0% (0–90%) 
Exposure No. of studies Risk ratio (95%CI) I2 (95%CI) 
Prudent/healthy dietary pattern 0.87 (0.68–1.11) 24% (0–88%) 
 Postdiagnosis 0.94 (0.71–1.24) 19% (0–92%) 
Western dietary pattern 1.21 (0.69–2.13) 81% (51–93%) 
 Postdiagnosis 1.34 (0.61–2.92) 85% (54–95%) 
Vegetable consumption 0.99 (0.74–1.33) 69% (0–91%) 
Alcohol consumption 17 1.17 (1.05–1.31) 38% (0–65%) 
 Breast cancer 1.21 (1.06–1.39) 23% (0–66%) 
 Hepatocellular carcinoma 1.34 (0.73–2.46) 73% (25–90%) 
 Postdiagnosis 1.31 (1.04–1.66) 54% (0–85%) 
Tea consumption 0.76 (0.58–1.01) 0% (0–90%) 

Main outcomes

The different clinical outcomes documented in the extracted studies were distributed as follows: overall mortality was evaluated in 117 studies (112 included in the meta-analysis) and cancer recurrence in 28 studies (27 included in the meta-analysis) (see Figures S1–S28 in the Supporting Information online).

Overall mortality

When food groups were combined, an inverse association for risk of mortality could be observed for the highest intakes of vegetables (RR, 0.86, 95%CI, 0.79–0.94; I2 = 43%) and fish (RR, 0.85, 95%CI, 0.78–0.93; I2 = 0%). Alcohol consumption was associated with increased risk of overall mortality (RR, 1.08; 95%CI, 1.02–1.16; I2 = 70%). With respect to the remaining food groups, no significant associations could be observed for intakes of dairy, meat, cereals and bread, eggs, tea, red meat, or processed meat. When only studies providing data on postdiagnosis food intakes were retained in the analysis, the results showed no statistically significant associations (Table 2).

Adherence to a high-quality diet was inversely associated with overall mortality (RR, 0.78; 95%CI, 0.72–0.85, I2 = 0%) (Figure 227,29–34 and Table 2). These effects remained consistent following analysis of only the postdiagnosis diet-quality indices (RR, 0.79; 95%CI, 0.71–0.89; I2 = 0%). The highest category of prudent healthy dietary pattern was inversely associated with overall mortality, both in the main analysis (RR, 0.81; 95%CI, 0.67–0.98; I2 = 44%) and when taking only postdiagnosis diet into account (RR, 0.77; 95%CI, 0.60–0.99; I2 = 56%). Conversely, the highest category of adherence to a Western dietary pattern was associated with increased risk of overall mortality (RR, 1.46; 95%CI, 1.27–1.68; I2 = 0%), which could also be found when synthesizing data of postdiagnosis Western dietary pattern only (RR, 1.51; 95%CI, 1.24–1.85; I2 = 17%) (Figure 327,35–41 and Table 2).

Figure 2

Forest plot showing pooled risk ratios (RRs) with 95%CIs for overall risk of mortality when comparing the highest vs the lowest category of adherence to diet-quality indices. Abbreviations: I2, inconsistency; SE, standard error; tau, estimate between study variance.

Figure 2

Forest plot showing pooled risk ratios (RRs) with 95%CIs for overall risk of mortality when comparing the highest vs the lowest category of adherence to diet-quality indices. Abbreviations: I2, inconsistency; SE, standard error; tau, estimate between study variance.

Figure 3

Forest plot showing pooled risk ratios (RRs) with 95%CIs for overall risk of mortality when comparing the highest vs the lowest category of adherence to a Western dietary pattern. Abbreviations: I2, inconsistency; SE, standard error; tau, estimate between study variance.

Figure 3

Forest plot showing pooled risk ratios (RRs) with 95%CIs for overall risk of mortality when comparing the highest vs the lowest category of adherence to a Western dietary pattern. Abbreviations: I2, inconsistency; SE, standard error; tau, estimate between study variance.

Cancer recurrence

When data for pre- and postdiagnosis levels of alcohol consumption were combined, an increased risk of cancer recurrence was found (RR, 1.17; 95%CI, 1.05–1.31; I2 = 38%). When only data for postdiagnosis alcohol intake were retained in the analysis, the results remained unchanged (RR, 1.31; 95%CI, 1.04–1.66; I2 = 54%). No significant associations between cancer recurrence and vegetable or tea intake could be observed. The intake of other food groups such as fruit, dairy, and coffee could not be meta-analyzed because of the limited number of eligible studies (minimum of 3) (Table 3).

The highest category of adherence to a prudent/healthy dietary pattern was not associated with cancer recurrence when compared with the lowest category of adherence (RR, 0.87; 95%CI, 0.68–1.11; I2 = 24%). These results were confirmed when the postdiagnosis pattern was analyzed (RR, 0.94; 95%CI, 0.71–1.24; I2 = 19%). Moreover, the Western dietary patterns were not associated with any significant effects on cancer recurrence data (RR, 1.21; 95%CI, 0.69–2.13; I2 = 81%) (postdiagnosis: RR, 1.34; 95%CI, 0.61–2.92; I2 = 85%) (Table 3).

Subgroup analyses for types of cancer

Breast cancer.

No evidence of an association with overall mortality was observed for individual foods, including fruits, vegetables, dairy, meat, and alcohol. Adherence to high diet-quality indices and a prudent/healthy dietary pattern was associated with a reduced risk of overall mortality ([diet-quality indices: RR, 0.74; 95%CI, 0.60–0.90; I2 = 6%] [prudent/healthy dietary pattern: RR, 0.76; 95%CI, 0.60–0.95; I2 = 4%]) among breast cancer survivors. In contrast, high adherence to a Western dietary pattern was associated with an increased overall mortality risk among breast cancer survivors (RR, 1.44; 95%CI, 1.17–1.77; I2 = 0%). Moreover, alcohol intake was associated with increased risk of breast cancer recurrence (RR, 1.21; 95%CI, 1.06–1.39; I2 = 23%).

Other cancers.

A Western dietary pattern (RR, 1.55; 95%CI, 1.13–2.13; I2 = 35%) was associated with an increased risk of overall mortality among colorectal cancer survivors. Higher intakes of alcohol were associated with increased mortality rates among survivors of hepatocellular carcinoma (RR, 1.21; 95%CI, 1.07–1.36; I2 = 0%), non-Hodgkin lymphoma (RR, 1.33; 95%CI, 1.10–1.63; I2 = 46%), laryngeal and pharyngeal cancer (RR, 1.48; 95%CI, 1.08–2.02; I2 = 49%), and head and neck cancer (RR, 1.39; 95%CI, 1.10–1.76; I2 = 53%).

Subgroup analyses for adjustment factors and effect measures

Overall, 71% (for diet-quality indices/dietary patterns), 29% (for foods), and 32% (for beverages) of the included studies were adjusted for cancer stage and treatment. A subgroup analysis of only these studies confirmed the primary results that showed associations of diet-quality indices and dietary patterns with overall mortality and cancer recurrence as well as the association between alcohol intake and cancer recurrence. On the other hand, the significant associations between vegetable and alcohol intake and overall mortality shown in the main analysis could not be confirmed (data not shown). Additionally, a subgroup analysis that included only studies that reported hazard ratios as effect estimates (96 of 112 studies) was performed. All results of the main analysis could be confirmed except the significant inverse association between adherence to a prudent/healthy dietary pattern and overall mortality (data not shown).

Small-study effects and publication bias

The Egger’s linear regression test was performed for overall mortality as well as for the following exposures: pre- and postdiagnosis intakes of fruits, vegetables, meat, and alcohol. There was no evidence of small-study effects of alcohol intake (P = 0.716) on risk of cancer recurrence or of fruit intake (P = 0.112), vegetable intake (P = 0.266), dairy intake (P = 0.961), or meat intake (P = 0.520) on overall mortality. However, a potential risk of bias for alcohol consumption (P = 0.002) and overall mortality was observed.

The funnel plots for associations between risk of overall mortality and intakes of fruit, vegetable, dairy, and meat indicate little asymmetry, and the plot for alcohol intake indicates moderate asymmetry, suggesting that small-study effects cannot be completely excluded as a factor of influence in the present meta-analysis (see Figures S29–S34 in the Supporting Information online). Assessment of the contour-enhanced funnel plots suggests that the asymmetry observed is more likely caused by factors other than publication bias (Figures S35–S40 in the Supporting Information online).

DISCUSSION

In summary, higher diet-quality indices and prudent/healthy dietary patterns were associated with a decreased risk of mortality in cancer survivors, whereas Western dietary patterns resulted in the opposite outcome. The food groups of vegetables and fish were associated with lower mortality, while alcohol consumption was associated with higher mortality. In breast cancer survivors, higher adherence to diet-quality indices and prudent/healthy dietary patterns were associated with a decreased risk of overall mortality, but higher fruit and vegetable intakes were not. Similarly, Western dietary patterns were associated with increased risk of mortality in breast cancer and colorectal cancer survivors, but meat consumption was not. The number of studies assessing cancer recurrence was low, but 1 study suggests that a Western dietary pattern is associated with an increased risk of cancer recurrence among colorectal cancer survivors.38 Finally, consumption of alcoholic beverages was associated with an increased risk of cancer recurrence. This association was also seen among breast cancer survivors and in postdiagnosis alcohol intake.

With increasing numbers of cancer patients, a growing willingness of patients to participate in check-up screenings, and technical improvements in diagnostic and therapeutic tools, it is reasonable to speculate that the number of cancer survivors will increase.1,2,42 In this context, lifestyle measures will gain importance because they represent an option for patients to make self-controlled decisions. In addition, a switch to a healthier lifestyle can be expected to have benefits even if the lifestyle change takes place after the onset of disease. It is thus critical to provide evidence-based recommendations for cancer patients to prevent unsuitable or even deleterious lifestyle practices in this population.

Classification of the various dietary indices and patterns into their respective components results in a number of food groups. The constituents of these food groups might, at least in part, explain the biochemical mechanisms by which the dietary matrices affect tumor pathogenesis. For example, dietary patterns associated with beneficial effects are regularly high in antioxidants (eg, a prudent or healthy pattern high in olive oil, fruits and vegetables, and fish contains phenolic compounds, flavonoids, and n-3 fatty acids) or dietary fiber. On the other hand, a Western diet is high in food groups with potentially detrimental ingredients, eg, red and processed meats, which contain nitrite and nitrate or proinflammatory fatty acids.43 High consumption of fruit, vegetables, fish, and whole-grain products has usually been associated with a reduced risk of chronic diseases,44–47 while abundant consumption of red meat was associated with poor outcomes of colorectal cancer.48 Inconsistent results have been reported for dairy product consumption and risk of prostate cancer, presumably due to differences in the type of product and source of calcium.49 Consumption of alcoholic beverages has been reported to be associated with increased incidence rates of cancers of the mouth, larynx, pharynx, esophagus, liver, and colon/rectum as well as with pre- and postmenopausal breast cancer.50

Another distinctive feature of this systematic review is the separate evaluation of postdiagnosis subgroups whenever a sufficient amount of data was available. With respect to overall mortality, subgroup analyses confirmed the main results (pre- and postdiagnosis results combined) obtained for diet-quality indices, prudent/healthy dietary patterns, and Western dietary patterns. This shows that the overall results (pre-and postcancer exposures combined) can be used to support lifestyle changes after a diagnosis of cancer.

This review has several strengths (inclusion of a large number of studies and patients, focus on cancer survivors, analyses of postdiagnosis data and cancer recurrence), as well some limitations. The present analyses are based exclusively on cohort studies. Cohort study designs have disadvantages such as unproven across-group gradients in intake because of bias in dietary estimates and confounding caused by other risk factors.51,52 Sometimes, the number of studies was rather small (especially for studies of cancer recurrence, specific cancer sites, and postdiagnosis dietary assessment). Independent of the power of individual studies, there was – to some extent – high heterogeneity between the different studies (see I2 values in Tables 1 and 2). This can be explained by several reasons, such as differences in sex and age of participants; time point of diagnosis; types of cancer; and diet-quality indices.

The fact that different diet-quality indices were summarized into one category might pose a problem, particularly since some of these indices are not assessed by standardized methods (eg, different scores for adherence to a Mediterranean diet). However, all of these indices rely on a combination of potentially protective food groups. It is therefore unlikely that the benefit indicated by the resulting effect estimate is attenuated by a single detrimental index. Special attention might be required for the Mediterranean diet, which was shown to exert favorable effects on tumor incidence and mortality in a number of studies.7 Usually, however, the highest scores for adherence to this type of diet can only be obtained with moderate alcohol consumption. Given the inconsistent results of studies investigating the effects of moderate alcohol intake on cancer,53,54 recommending alcohol consumption even in small quantities might not be sensible.

An important distinction of the studies is the point during the course of disease at which diet was measured. Most of the included studies analyzed the prediagnosis diet. However, the potential of the postdiagnosis diet to improve survival is of particular interest to cancer survivors.

The development of tumors on different locations is a highly heterogeneous process, and even the pathogenesis of cancer at the same site can vary considerably. Heterogeneity in study design and participant characteristics made it especially difficult to identify clinical factors with an important impact on survival and cancer recurrence when various studies were synthesized. The stage of cancer diagnosis has been recognized as an important factor of disease progression. Patients with tumors diagnosed at an earlier stage show higher survival rates.55,56 The treatment method and type of hospital can also greatly influence the outcome.57 Similarly, age and comorbidity could also be relevant predictors of survival in some types of cancer.58 Nevertheless, most studies included in the present analyses adjusted at least for age, tumor stage, and treatment method.

Despite the advantages offered by a large number of studies and cancer occurrences, different populations are exposed to different potential risks and/or protective factors. Likewise, dietary habits and dietary patterns differ between populations. In the present analysis, most of the data summarized are from studies of US populations (55 studies), followed by studies of European populations (33 studies) and Asian populations (21 studies), with the remaining 9 studies conducted in populations in Canada, Australia, or other countries). With a higher proportion of US populations and an underrepresentation of other populations, such as those from South America and Africa, the results of this review were weighted more heavily toward the US population. Furthermore, sociopolitical factors might have an effect on outcome, too. For instance, the health insurance system can be a predicting factor for the stage at which cancer is diagnosed and the treatment options available.59

CONCLUSION

This systematic review and meta-analysis of cohort studies suggests that adherence to a healthy dietary pattern (assessed predominantly prior to cancer diagnosis) is inversely associated with overall mortality, whereas an unhealthy Western dietary pattern is positively associated with risk of overall mortality among cancer survivors. More research is warranted to assess the role of postdiagnosis diet in cancer survival and site-specific cancer recurrence.

Acknowledgments

Author contributions. L.S. designed the research; C.S. and L.S. conducted the research; L.S. and C.S. performed statistical analysis; C.S., L.S., and G.H. wrote the paper; L.S. and C.S. had primary responsibility for the final content; and H.B. and K.A. provided other contributions (revised the paper critically for important intellectual content and made contributions to the final version). All authors read and approved the final manuscript.

Funding/support. No external funds supported this work.

Declaration of interest. The authors have no relevant interests to declare.

SUPPORTING INFORMATION

The following Supporting Information is available through the online version of this article at the publisher’s website.

Table S1Full-text articles excluded, with reasons

Table S2Characteristics of the cohort studies included in the present meta-analysis by type of exposure (dietary patterns, food groups, beverages)

Figure S1Forest plot showing pooled risk ratios (RRs) with 95%CIs for overall mortality risk comparing the highest vs lowest adherence to prudent/healthy dietary pattern category.

Figures S2–S12Forest plots showing pooled risk ratios (RRs) with 95%CIs for overall mortality risk comparing the highest vs lowest intake category of different food and beverage groups.

Figures S13–S14Forest plots showing pooled risk ratios (RRs) with 95%CIs for cancer recurrence comparing the highest vs lowest adherence to dietary patterns.

Figures S15–S17Forest plots showing pooled risk ratios (RRs) with 95%CIs for cancer recurrence comparing the highest vs lowest intake category of different food and beverage groups.

Figures S18–S20Forest plots showing pooled risk ratios (RRs) with 95%CIs for overall mortality comparing the highest vs lowest postdiagnosis adherence to diet-quality indices and dietary patterns.

Figures S21–S25Forest plots showing pooled risk ratios (RRs) with 95%CIs for overall mortality comparing the highest vs lowest postdiagnosis intake category of different food and beverage groups.

Figures S26–S27Forest plots showing pooled risk ratios (RRs) with 95%CIs for cancer recurrence comparing the highest vs lowest postdiagnosis adherence to dietary patterns.

Figure S28Forest plot showing pooled risk ratios (RRs) with 95%CIs for cancer recurrence comparing the highest vs lowest postdiagnosis alcohol intake category.

Figures S29–S33Funnel plots showing study precision against the relative risk with 95%CIs for intake of different food and beverage groups and overall mortality.

Figure S34Funnel plot showing study precision against the relative risk with 95%CIs for alcohol intake and cancer recurrence.

Figures S35–S39Contour-funnel plots showing study precision against the relative risk with 95%CIs for intake of different food and beverage groups and overall mortality.

Figure S40Contour-funnel plot showing study precision against the relative risk with 95%CIs for alcohol intake and cancer recurrence.

References

1
DeSantis
CE
Lin
CC
Mariotto
AB
et al
.
Cancer treatment and survivorship statistics, 2014
.
CA Cancer J Clin
 .
2014
;
64
:
252
271
.
2
Rowland
JH
Kent
EE
Forsythe
LP
et al
.
Cancer survivorship research in Europe and the United States: where have we been, where are we going, and what can we learn from each other?
Cancer
 .
2013
;
119
(
suppl 11
):
2094
2108
.
3
World Cancer Research Fund/American Institute for Cancer Research
.
Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective
 .
Washington, DC
:
American Institute for Cancer Research
;
2007
.
4
Schwingshackl
L
Hoffmann
G
.
Diet quality as assessed by the Healthy Eating Index, the Alternate Healthy Eating Index, the Dietary Approaches to Stop Hypertension Score, and health outcomes: a systematic review and meta-analysis of cohort studies
.
J Acad Nutr Diet
 .
2015
;
115
:
780
800
.
5
Schwingshackl
L
Hoffmann
G
.
Adherence to Mediterranean diet and risk of cancer: an updated systematic review and meta-analysis of observational studies
.
Cancer Med
 .
2015
;
4
:
1933
1947
.
6
Schwingshackl
L
Hoffmann
G
.
Adherence to Mediterranean diet and risk of cancer: a systematic review and meta-analysis of observational studies
.
Int J Cancer
 .
2014
;
135
:
1884
1897
.
7
Schwingshackl
L
Hoffmann
G
.
Does a Mediterranean-type diet reduce cancer risk?
Curr Nutr Rep
 .
2016
;
5
:
9
17
.
8
Continuous Update Project Team, World Cancer Research Fund International
.
Systematic Review on Diet, Nutrition, Physical Activity and Survival and Second Cancers in Breast Cancer Survivors
 .
London, UK
:
World Cancer Research Fund International
;
2014
.
9
Rock
CL
Doyle
C
Demark-Wahnefried
W
et al
.
Nutrition and physical activity guidelines for cancer survivors
.
CA Cancer J Clin
 
2012
;
62
:
243
274
.
10
Jones
LW
Demark-Wahnefried
W
.
Diet, exercise, and complementary therapies after primary treatment for cancer
.
Lancet Oncol
 .
2006
;
7
:
1017
1026
.
11
Fahey
PP
Mallitt
KA
Astell-Burt
T
et al
.
Impact of pre-diagnosis behavior on risk of death from esophageal cancer: a systematic review and meta-analysis
.
Cancer Causes Control
 .
2015
;
26
:
1365
1373
.
12
Ferronha
I
Bastos
A
Lunet
N
.
Prediagnosis lifestyle exposures and survival of patients with gastric cancer: systematic review and meta-analysis
.
Eur J Cancer Prev
 .
2012
;
21
:
449
452
.
13
Gou
YJ
Xie
DX
Yang
KH
et al
.
Alcohol consumption and breast cancer survival: a meta- analysis of cohort studies
.
Asian Pac J Cancer Prev
 .
2013
;
14
:
4785
4790
.
14
Stroup
DF
Berlin
JA
Morton
SC
et al
.
Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group
.
JAMA
 .
2000
;
283
:
2008
2012
.
15
DerSimonian
R
Laird
N
.
Meta-analysis in clinical trials
.
Control Clin Trials
 .
1986
;
7
:
177
188
.
16
Higgins
JP
Thompson
SG
.
Quantifying heterogeneity in a meta-analysis
.
Stat Med
 .
2002
;
21
:
1539
1558
.
17
Higgins
JPT
Thompson
SG
Deeks
JJ
et al
.
Measuring inconsistency in meta-analyses
.
BMJ
 .
2003
;
327
:
557
560
.
18
Mega
JL
Stitziel
NO
Smith
JG
et al
.
Genetic risk, coronary heart disease events, and the clinical benefit of statin therapy: an analysis of primary and secondary prevention trials
.
Lancet
 .
2015
;
385
:
2264
2271
.
19
Egger
M
Smith
GD
Schneider
M
et al
.
Bias in meta-analysis detected by a simple, graphical test
.
BMJ
 .
1997
;
315
:
629
634
.
20
Peters
JL
Sutton
AJ
Jones
DR
et al
.
Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry
.
J Clin Epidemiol
 .
2008
;
61
:
991
996
.
21
Chiuve
SE
Fung
TT
Rimm
EB
et al
.
Alternative dietary indices both strongly predict risk of chronic disease
.
J Nutr
 .
2012
;
142
:
1009
1018
.
22
Newby
PK
Hu
FB
Rimm
EB
et al
.
Reproducibility and validity of the Diet Quality Index Revised as assessed by use of a food-frequency questionnaire
.
Am J Clin Nutr
 .
2003
;
78
:
941
949
.
23
McCullough
ML
Feskanich
D
Stampfer
MJ
et al
.
Diet quality and major chronic disease risk in men and women: moving toward improved dietary guidance
.
Am J Clin Nutr
 .
2002
;
76
:
1261
1271
.
24
Guenther
PM
Reedy
J
Krebs-Smith
SM
Reeve
BB
.
Evaluation of the Healthy Eating Index-2005
.
J Am Diet Assoc
 .
2008
;
108
:
1854
1864
.
25
Inoue-Choi
M
Robien
K
Lazovich
D
.
Adherence to the WCRF/AICR guidelines for cancer prevention is associated with lower mortality among older female cancer survivors
.
Cancer Epidemiol Biomarkers Prev
 .
2013
;
22
:
792
802
.
26
Kim
EH
Willett
WC
Fung
T
Rosner
B
Homes
MD
.
Diet quality indices and postmenopausal breast cancer survival
.
Nutr Cancer
 .
2011
;
63
:
381
388
.
27
Fung
TT
Kashambwa
R
Sato
K
et al
.
Post diagnosis diet quality and colorectal cancer survival in women
.
PLoS One
 .
2014
;
9
:
e115377
.
doi:10.1371/journal.pone.0115377
.
28
Fung
TT
Chiuve
SE
McCullough
ML
Rexrode
KM
Logroscino
G
Hu
FB
.
Adherence to a DASH-style diet and risk of coronary heart disease and stroke in women
.
Arch Intern Med
 .
2008
;
168
:
713
720
.
29
George
SM
Irwin
ML
Smith
AW
et al
.
Postdiagnosis diet quality, the combination of diet quality and recreational physical activity, and prognosis after early-stage breast cancer
.
Cancer Causes Control
 .
2011
;
22
:
589
598
.
30
George
SM
Ballard-Barbash
R
Shikany
JM
et al
.
Better postdiagnosis diet quality is associated with reduced risk of death among postmenopausal women with invasive breast cancer in the Women's Health Initiative
.
Cancer Epidemiol Biomarkers Prev
 .
2014
;
23
:
575
583
.
31
Izano
MA
Fung
TT
Chiuve
SS
et al
.
Are diet quality scores after breast cancer diagnosis associated with improved breast cancer survival?
Nutr Cancer
 .
2013
;
65
:
820
826
.
32
Pelser
C
Arem
H
Pfeiffer
RM
et al
.
Prediagnostic lifestyle factors and survival after colon and rectal cancer diagnosis in the National Institutes of Health (NIH)-AARP Diet and Health Study
.
Cancer
 .
2014
;
120
:
1540
1547
.
33
Kenfield
SA
DuPre
N
Richman
EL
et al
.
Mediterranean diet and prostate cancer risk and mortality in the Health Professionals Follow-up Study
.
Eur Urol
 .
2014
;
65
:
887
894
.
34
Thomson
CA
Crane
TE
Wertheim
BC
et al
.
Diet quality and survival after ovarian cancer: results from the Women's Health Initiative
.
J Natl Cancer Inst
 .
2014
;
106
.
doi:10.1093/jnci/dju314
.
35
Kroenke
CH
Fung
TT
Hu
FB
et al
Dietary patterns and survival after breast cancer diagnosis
.
J Clin Oncol
 .
2005
;
23
:
9295
9303
.
36
Kwan
ML
Weltzien
E
Kushi
LH
et al
Dietary patterns and breast cancer recurrence and survival among women with early-stage breast cancer
.
J Clin Oncol
 .
2009
;
27
:
919
926
.
37
Vrieling
A
Buck
K
Seibold
P
et al
.
Dietary patterns and survival in German postmenopausal breast cancer survivors
.
Br J Cancer
 .
2013
;
108
:
188
192
.
38
Meyerhardt
JA
Niedzwiecki
D
Hollis
D
et al
.
Association of dietary patterns with cancer recurrence and survival in patients with stage III colon cancer
.
JAMA
 .
2007
;
298
:
754
764
.
39
Zhu
Y
Wu
H
Wang
PP
et al
.
Dietary patterns and colorectal cancer recurrence and survival: a cohort study
.
BMJ Open
 .
2013
;
3
:
e002270
.
doi:10.1136/bmjopen-2012-002270
.
40
Arthur
AE
Peterson
KE
Rozek
LS
et al
.
Pretreatment dietary patterns, weight status, and head and neck squamous cell carcinoma prognosis
.
Am J Clin Nutr
 .
2013
;
97
:
360
368
.
41
Yang
M
Kenfield
SA
Van Blarigan
EL
et al
.
Dietary patterns after prostate cancer diagnosis in relation to disease-specific and total mortality
.
Cancer Prev Res (Philadelphia)
 .
2015
;
8
:
545
551
.
42
Jemal
A
Bray
F
Center
MM
et al
.
Global cancer statistics
.
CA Cancer J Clin
 .
2011
;
61
:
69
90
.
43
Okreglicka
K
.
Health effects of changes in the structure of dietary macronutrients intake in Western societies
.
Rocz Panstw Zakl Hig
 .
2015
;
66
:
97
105
.
44
Aune
D
Chan
DS
Lau
R
et al
.
Dietary fibre, whole grains, and risk of colorectal cancer: systematic review and dose-response meta-analysis of prospective studies
.
BMJ
 .
2011
;
343
:
d6617
.
doi:10.1136/bmj.d6617
.
45
Turati
F
Rossi
M
Pelucchi
C
et al
.
Fruit and vegetables and cancer risk: a review of southern European studies
.
Br J Nutr
 .
2015
;
113
(
suppl 2
):
S102
110
.
46
Wang
X
Ouyang
Y
Liu
J
et al
.
Fruit and vegetable consumption and mortality from all causes, cardiovascular disease, and cancer: systematic review and dose-response meta-analysis of prospective cohort studies
.
BMJ
 .
2014
;
349
:
g4490
.
doi:10.1136/bmj.g4490
.
47
Zheng
JS
Hu
XJ
Zhao
YM
et al
.
Intake of fish and marine n-3 polyunsaturated fatty acids and risk of breast cancer: meta-analysis of data from 21 independent prospective cohort studies
.
BMJ
 .
2013
;
346
:
f3706
.
doi:10.1136/bmj.f3706
.
48
Alexander
DD
Weed
DL
Miller
PE
et al
.
Red meat and colorectal cancer: a quantitative update on the state of the epidemiologic science
.
J Am Coll Nutr
 .
2015
;
34
:
521
543
.
49
Aune
D
Navarro Rosenblatt
DA
Chan
DS
et al
.
Dairy products, calcium, and prostate cancer risk: a systematic review and meta-analysis of cohort studies
.
Am J Clin Nutr
 .
2015
;
101
:
87
117
.
50
Bagnardi
V
Rota
M
Botteri
E
et al
.
Alcohol consumption and site-specific cancer risk: a comprehensive dose–response meta-analysis
.
Br J Cancer
 .
2015
;
112
:
580
593
.
51
National Research Council (US) Committee on Diet and Health
.
Methodological considerations in evaluating the evidence
. In:
Diet and Health: Implications for Reducing Chronic Disease Risk
 .
Washington, DC
:
National Academy Press
;
1989
:
23
40
.
52
Schwingshackl
L
Knüppel
S
Schwedhelm
C
et al
.
NutriGrade: a scoring system to assess and judge the meta-evidence of randomized controlled trials and cohort studies in nutrition research
.
Adv Nutr
 .
2016
;
7
.
53
Bagnardi
V
Rota
M
Botteri
E
et al
.
Light alcohol drinking and cancer: a meta-analysis
.
Ann Oncol
 .
2013
;
24
:
301
308
.
54
Jin
M
Cai
S
Guo
J
et al
.
Alcohol drinking and all cancer mortality: a meta-analysis
.
Ann Oncol
 .
2013
;
24
:
807
816
.
55
Jemal
A
Siegel
R
Ward
E
et al
.
Cancer statistics, 2008
.
CA Cancer J Clin
 .
2008
;
58
:
71
96
.
56
Richards
MA
Westcombe
AM
Love
SB
et al
.
Influence of delay on survival in patients with breast cancer: a systematic review
.
Lancet
 .
1999
;
353
:
1119
1126
.
57
Lee-Feldstein
A
Anton-Culver
H
Feldstein
PJ
.
Treatment differences and other prognostic factors related to breast cancer survival. Delivery systems and medical outcomes
.
JAMA
 
1994
;
271
:
1163
1168
.
58
van Eeghen
EE
Bakker
SD
van Bochove
A
et al
.
Impact of age and comorbidity on survival in colorectal cancer
.
J Gastrointest Oncol
 .
2015
;
6
:
605
612
.
59
Lukavsky
R
Sariego
J
.
Insurance status effects on stage of diagnosis and surgical options used in the treatment of breast cancer
.
Southern Med J
 .
2015
;
108
:
258
261
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact journals.permissions@oup.com