A new score for quantifying adherence to a cancer-preventive Mediterranean diet

Abstract

Herein a Mediterranean Cancer Preventive Diet Score (MCAP Score) is proposed to quantify adherence to both traditional Mediterranean fat intakes and the current dietary recommendations for cancer prevention. The scoring uses research-backed cut-off values, unlike other scores that are based on a population-specific median value. The MCAP score awards positive points for seven preventive food categories, including Mediterranean fats (monounsaturated fats, ω-3 fatty acids) associated with reduced adiposity, and negative points for four food categories associated with increased cancer risk, including ultra-processed foods. In a randomized trial of 120 persons at increased risk of colon cancer, the baseline MCAP Score averaged 7 of 22 possible points. Counseling for a Healthy Diet or a Mediterranean Diet improved the score to either 11 or 13 points, respectively, and the highest score observed in any individual was 20 points. The MCAP Score was correlated with serum carotenoids and serum ω-3 fatty acids, and improvements in the score were associated with weight loss over six months of study. The MCAP Score is therefore proposed as a new method to assess adherence to a Mediterranean type of diet for cancer prevention using absolute criteria that will facilitate comparisons of dietary intakes across studies.

Introduction

Diet and obesity contribute to the risks of many types of cancer [1]. Unfortunately, the average U.S. diet falls well short of the dietary recommendations for cancer prevention, and over 70% of Americans are overweight or obese [2–4]. Diet is estimated to contribute up to one-third of all cancers, and at least 13 cancer types have been identified to be obesity-associated [5, 6].

Many dietary factors have potential for cancer prevention, and this stimulates interest in identifying dietary patterns for prevention that provide exposure to multiple preventive compounds. The simultaneous absence of foods associated with increased risk together, with the consumption of preventive foods, results in a combination that can provide a multi-faceted, preventive milieu. The importance of avoiding foods associated with increased risks of cancer cannot be over-looked since simply adding preventive foods to the diet would likely result in inappropriate weight gain from the caloric excess. The action of preventive compounds thus may depend on the overall “package”.

Both the American Cancer Society (ACS) and the American Institute for Cancer Research/World Cancer Research Fund (AICR/WCRF) have developed dietary guidelines for cancer prevention [7, 8]. These guidelines for cancer prevention have many similarities with the U.S. Department of Agriculture (USDA) Dietary Guidelines for Americans but include specific emphasis on certain foods found to be associated with increased or decreased risks of cancer. The cancer prevention guidelines from both the ACS and AICR also recommend maintaining a healthy weight [7, 8].

A diet that meets the cancer prevention guidelines can be attained using a variety of eating patterns and foods. One type of food pattern of interest for cancer prevention that has emerged from initial research on prevention of cardiovascular diseases and diabetes is the Mediterranean type of diet [9]. In addition to favorable nutrient intakes, the relatively high monounsaturated fat content of the Mediterranean diet has been associated with weight control and lower visceral adiposity. Epidemiological findings show that individuals who adhere most closely to a Mediterranean type of diet exhibit relatively less adult weight gain and have a smaller waist circumference [10–14]. A Mediterranean diet therefore could be utilized to meet the AICR and ACS guidelines to avoid excess weight gain/adiposity.

We previously developed an exchange list method to achieve a Greek-Mediterranean diet using foods readily available in the U.S. We tested this method in a clinical trial of persons at increased colon cancer risk who were randomized to either the Mediterranean diet or a standard Healthy Eating diet based on the USDA recommendations. In that study, the Mediterranean diet intervention resulted in a modest but significant weight loss in the overweight or obese subgroup of study participants. The weight loss occurred despite the explicit design of food group goals to maintain baseline energy intakes [15]. This modest weight loss is potentially important. A slow rate of weight loss is indicated for cancer prevention, and this is unlike the standard obesity treatment approach that targets a 10% loss of body weight over 6 months [16–21].

The purpose of this present investigation was to develop a dietary cancer prevention score based on the traditional Greek-Mediterranean eating pattern and the AICR and ACS guidelines for cancer prevention. The Mediterranean Cancer Preventive Diet Score (MCAP) Score that was derived includes positive points for consuming quantities of foods associated with prevention, based on published recommendations and/or research findings. The score also incorporates negative points for consumption of Western foods associated with increased cancer risk, namely red meats and ultra-processed foods (UPF). Although UPF is newly defined food category, there are now data showing that higher consumption of UPF is associated with increased cancer risk [22]. UPF account for 25-60% of energy intakes in industrialized countries, and diets high in UPF have been shown to lead to weight gain [23, 24]. Here the method and rationale for calculating the MCAP Score is given, and an example of changes in the score after intervention with either a Mediterranean diet or a Healthy Eating diet is shown.

Methods

Healthy Eating Study Data

De-identified data from the Healthy Eating Study for Colon Cancer Prevention were utilized to derive an example of the performance of the MCAP Score. The Healthy Eating Study enrolled 120 adults at increased risk of colon cancer defined as either a personal history of adenoma or cancer or a strong family history of colorectal cancer in one first degree relative or two second degree relatives. Eligible adults were ages 25 and older and had a BMI of 18.5-35 kg/m². All subjects gave written informed consent to participate. The study was approved by the University of Michigan Institutional Review Board (HUM00007622). The study was listed on the ClinicalTrials.gov website maintained by the National Institutes of Health (registration number NCT00475722).

Details of the study methodology have been published [15, 25]. A total of 93 subjects completed the 6-month study, and one additional subject completed the intervention and dietary assessment for the 6-month visit without completing the clinic visit for sample collection. Serum carotenoids and serum fatty acids were measured as previously described [26]. Briefly, carotenoids and tocopherols were extracted from serum with hexane after adding Tocol as an internal standard. Micronutrients were then analyzed using HPLC with both visible (450 nm) and electrochemical detection, and quantitation was done using standard curves for each compound. Total serum fatty acids were extracted with Folch reagent after adding fatty acid 17:0 as the internal standard. Fatty acids were then measured as fatty acid methyl esters by gas chromatography with mass spectral detection using standard curves for each fatty acid quantified.

The percent of study dietary goals met was based on ten criteria in the Mediterranean arm and five criteria in the Healthy Eating arm, as previously described [15]. The ten Mediterranean dietary goals entailed increases in mono-unsaturated fatty acids (MUFA), ω-3 fatty acids, whole grains, and seven specific categories of fruits and vegetables. In the Mediterranean intervention, low-fat dairy (or dairy substitute) was allowed, and in the Healthy Eating arm dairy intakes were not addressed. The Healthy Eating diet had five dietary goals: saturated fatty acids (SFA), fruit, vegetables, dark green or orange vegetable and whole grains. These goals were based on the USDA Healthy People 2010 recommendations. The intervention was designed to be isocaloric with baseline.

The dietary intervention was conducted using an exchange list approach and administered by counseling with a registered dietitian that occurred mainly by telephone. After six months, participants met 82% of the diet goals in the Mediterranean arm and 88% of the diet goals in the Healthy Eating arm [15]. Diet in the study was assessed at baseline and 6 months using two 24-hour recalls and two days of food records (four days per time point). The food records and recalls were analyzed using the Nutrition Data System for Research (NDS-R) software (version 2010, Nutrition Coordinating Center, University of Minnesota) to generate average daily food servings and average daily nutrient intakes for each participant at each time point.

Basis of the Mediterranean Cancer Preventive Diet (MCAP) Score

The MCAP score was developed relying on recommendations for cancer prevention from the ACS and the AICR as given on their websites [7, 8]. The MCAP scoring system also took into account the 2015-2020 USDA Dietary Guidelines for Americans [27]. Finally, we relied on aspects of traditional Greek-Mediterranean eating patterns. This was modeled after the Mediterranean score developed by Sofi et al. that uses set cut-offs that were informed by a meta-analysis of Mediterranean diet studies [28]. Two other Mediterranean scores also utilized defined cut-off values for scoring but did not differ substantially from that of Sofi et al. [29, 30]. In contrast, the widely used score of Trichopoulou et al. scores for amounts above/below the gender-specific median for the population being studied [31].

Points were awarded within the MCAP scoring system for higher consumption levels of foods identified as being preventive. This included six categories: 1) Total fruits and vegetables, 2) Non-starchy vegetables, 3) Whole grains, 4) Legumes, nuts, and seeds, 5) MUFA:SFA ratio, and 6) Long-chain ω-3 fatty acids. Also included was scoring for foods to limit: There were four categories for which negative points were awarded for high consumption levels: 1) Foods made with refined flour, 2) Foods with added sugars, 3) Processed meats and 4) Red meats. The MCAP Score uses values of 0,1 or 2 for each of those 11 categories, and this is shown in Table 1. In general, the optimal intake of above was assigned a score 2, half that was assigned a score of 1, and less than half of the optimal intake received a score of zero. With 11 categories, the maximal MCAP Score is 22.

Table 1.

Scoring of Nutrients or Food Group Categories for the Mediterranean Cancer-Preventive Diet (MCAP) Score

Category	Definition	Dietary Criteria For Scoring			Summary of the Rationale for Maximal Score
		0	1	2
Total Fruits and Vegetables	½ cup servings except for 1 cup leafy greens, ¼ cup dried fruit, ¼ cup tomato paste; 4 oz. juice, 1 medium fruit piece	≤ 3.5 servings/day	> 3.5 to < 7.0 servings/day	≥ 7 servings/day	Average USDA and ACS recommendations
Non-starchy vegetables	Sum of dark green, deep yellow, tomato and other vegetables, using serving size above	≤ 2.5 servings/day	> 2.5 to < 5.0 servings/day	≥ 5 servings/day	AICR recommendation; adds weight to vegetable intake in the score; excludes fried potatoes
Whole Grains	1 oz./serving (1 slice bread, ⅓ cup brown rice)	≤ 1.5 servings/day	> 1.5 to < 3.0 servings/day	≥ 3 servings/day	ACS “choose whole grains”; AICR “at every meal”; USDA 3 oz. whole grains/day
Legumes, Nuts, and Seeds	Legumes ½ cup servings; Nuts and seeds ½ ounce (14 g or 3 TB) servings; Nut butters 1 TB (20g) servings	≤ 0.5 servings/day	> 0.5 to < 1.0 servings/day	≥ 1 servings/day	USDA ≥ 5 oz/week, MED daily handful AICR “at every meal
MUFA:SFA ratio	Gram ratio	≤ 1.0	> 1.0 to < 2.0	≥ 2	USDA limits SFA, epidemiological data on MUFA to decrease visceral adipose stores
Long-chain ω-3 Fatty Acids or Equivalenta	Grams/day of long-chain ω-3 fatty acids plus 5% of linolenic acid intake	≤ 0.18 g/day	> 0.18 to < 0.36 g/day	≥ 0.36 g/day	USDA & MED 8 oz./week fish
Added Sugar	Percent of calories from all added mono-, di- and poly-saccharides	≥ 10 % of energy	< 10 to > 5 % of energy	≤ 5 % of energy	USDA ≤10% of energy, ACS & AICR “avoid”
Red Meat	Ounces/week, includes beef, ???	≥ 2.6 oz./day	> 2.6 to < 1.3 oz./day	≤ 1.3 oz/day	AICR ≤18 oz/wk, ACS avoid, USDA <26 oz/wk meat & eggs
Processed Meat	1 oz. serving size	≥ 0.5 servings/day	> 0 to < 0.5 servings/day	none	ACS & AICR recommendations to “avoid”
Processed Snack Foods	Servings/day of snack bars (40 g), deserts (30-125 g/serving), savory snacks such as chips (1 oz.)	≥ 1 servings/day	> 0 to < 1 servings/day	none	AICR “avoid”, USDA limits on sodium and sugar

^aFish high in ω-3 fatty acids contains 0.5-2 g/3 oz. serving: for two servings/week this comes to an average of 2.5 g long chain ω-3 fatty acids/week or 0.36 g/day. A serving of flax contains about 4-7g linolenic acid which is 0.2-0.35 g long chain ω-3 fatty acid equivalents assuming 5% metabolic conversion to eicosapentaenoic acid.

Rationale for Scoring Preventive Food Categories

Fruits and Vegetables.

For fruits and vegetables, total fruit and vegetable intakes and non-starchy vegetables were scored separately using two categories. Vegetables are the major source of carotenoids, and increased carotenoid intakes have been associated with reduced risks of many cancers across studies [32]. The maximal score for 3.5 cups or more of fruits and vegetables per day (seven ½ cup servings) was based on averaging the USDA recommendation for 4.5 cups/day and the ACS recommendation of 2.5 cups/day of total fruits and vegetables. A separate category for non-starchy vegetables only was based on the AICR recommendation for non-starchy vegetables [8]. Non-starchy vegetables are a low calorie option for obtaining phytonutrients. The MCAP score therefore weights intakes of non-starchy vegetables by including them in both categories.

Servings consumed were enumerated within the NDS-R program using a serving size of ½ cup for fresh, frozen, canned or cooked fruit or vegetables, 1 cup for raw leafy vegetables, 1 medium piece fruit (or ½ a grapefruit), ¼ cup tomato paste, or ¼ cup dried fruit. The categories summed were dark green, deep yellow, tomato and other vegetables to create a non-starchy vegetable serving total. The total of fruit and vegetables summed non-starchy vegetables, citrus fruits, other fruits, juices, white potatoes, and other starchy vegetables such as cassava, corn, green peas, and jicama. Fried potatoes were omitted since this is not regarded as a healthful food for cancer prevention by either the AICR or ACS.

Whole Grains.

High fiber diets are associated with reduced BMI in addition to reduced cancer risk in epidemiological studies [33–36]. However, in intervention studies, fiber supplementation has not shown protection from colorectal cancer risk, indicating that the preventive effects many other components of whole grain foods [33, 37]. The AICR recommendation is to include whole grains at every meal and to limit refined grain foods to 3 servings/day. The ACS guideline is to “choose whole grains”. Lastly, the USDA My Plate goal is to make “half your grains whole” (U.S. adults consume about 6 grain servings/day) [7, 8, 38]. The maximal score of two was therefore set at 3 or more servings/day of whole grains, a score of one for 1.5 to <3 servings/day, and zero for less than 1.5 servings/day of whole grains.

The scoring utilized only foods made with whole grains (eg. whole wheat, wholemeal corn, brown rice, oats), to the exclusion of foods made with some whole grain since it was difficult to discern how much whole grain was used in the latter case. The food categories summed were breads, crackers, pasta, all types of cereals, snack bars, snack chips, popcorn, and baked deserts such as cookies and cakes. A serving was generally defined as one ounce or 28-45 grams, depending on the food (eg. 1 slice bread, ½ cup rice, or ½ English muffin). Serving sizes for whole grain deserts varied based on usual serving sizes in the U.S.: doughnut 55 g, cookie 30 g, pie and heavy cake at 125 g/serving.

Legumes, nuts and seeds.

This category is based on both the AICR recommendation to include high fiber foods such as “legumes, whole grains, fruit, vegetables, nuts and seeds”, and the consideration that these foods have a favorable impact on the gut microbiome as one mechanism behind their cancer-preventive effects [8]. The ACS also recommends including legumes as part of the desired variety in vegetable intakes (“a variety of vegetables – dark green, red and orange, fiber-rich legumes (beans and peas) [7]. Mediterranean diets also included legumes and “a handful of nuts and seeds daily” [28, 31]. The NDS-R program includes peanut butter in the “nut and seed butters” category, and both peanuts and cashews are included with “nuts”. This made it difficult to enumerate legumes and nuts separately, and since NDS-R is widely used in nutrition research, the categories were simply combined.

Servings sizes for scoring purposes were the same as within the NDS-R program: ½ cup of cooked beans (any type), ½ ounce of nuts and seeds, and 1 TB for nut or seed butters. The maximal score of two was fixed for consuming for at least 1 serving/day of legumes, nuts/seeds and/or nut/seed butters, a score of one for at least 0.5 serving/day but less than 1 serving/day, and a score of zero for less than 0.5 serving per day.

MUFA:SFA ratio.

Traditional Greek Mediterranean diet was high in olive oil. Estimates of the MUFA:SFA ratio from a review of the literature average from 1.6 to 2.7 for the Mediterranean diet [39]. In the Mediterranean intervention utilized within the Healthy Eating Study, the MUFA:SFA goal was 2.5, and the actual ratio achieved after 6 months was 1.7 [15]. Utilizing a MUFA:SFA ratio, instead of MUFA alone, is consistent with the USDA recommendation to limit SFA [27]. It is also consistent with the observations that high MUFA diets are helpful for weight management [10–14]. The MUFA:SFA ratio was calculated from average daily MUFA and SFA intakes from the NDS-R program. A score of two was assigned to MUFA:SFA intakes of 2 or greater, and score of one for ratios of 1 to <2, and a score of zero for ratios <1.

Long chain ω-3 fatty acids.

In Greece in the 1960’s, ω-3 fatty acid intake were much higher relative to a Western diet [40]. The notable decline in ω-3 fatty acid intakes observed in Westernized diets is worrisome since ω-3 fatty acid intakes have potential for prevention of many cancers [41, 42]. The ω-3 fats also appear to contribute to lower visceral fat accumulation, weight control, and inhibition of endogenous de-novo lipogenesis [43–45].

Since seafood varies widely in ω-3 content, we used a calculated value of g/day of ω-3 fatty acids for the MCAP Scoring criteria. A Mediterranean diet containing 2-3 portions of fish per week using a 3 oz. serving size is similar to the USDA dietary recommendation to consume at least 8 oz. fish/week [27]. Another source of long chain ω-3 fatty acids is the metabolic conversion of linolenic acid. The efficiency of conversion varies depending on many factors (sex, age, intake of other fats), but here 5% conversion was used as an approximate value for calculating the contribution of linolenic acid to long chain ω-3 fatty acids [46, 47].

To approach Greek intakes of ω-3 fatty acids in the score, the daily long chain ω-3 fatty acid intake was estimated using 1 g long chain ω-3 fatty acids per 3-ounce serving which is an average value among different kinds of popular fish. A recommendation of 2.5 servings of fish/week averages to 0.36 g/day of long chain ω-3 fatty acids. Intakes of eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), docosahexaenoic acid (DHA), and 5% of linolenic acid (18:3, ω-3) intake were summed. The maximal score of two was then set for intakes of 0.36 g/day or greater, a score of one for intakes of 0.18-0.359 g/day, and zero points for intakes that are less than half the recommendation (<0.18 g/day).

Scoring of Foods to Limit

Red meat.

The association of high red meat and processed meat intakes with increased risks of colon and other cancers is well established [7, 8]. AICR recommends consuming less than 18 ounces/week, which is equivalent to 2.6 ounces/day on average [8]. Average daily intakes above 2.6 oz./day were therefore scored with zero points, ≥ 1.3 to < 2.6 oz./day with one point, and below 1.3 oz./day as two points.

Percent of calories from added sugar.

AICR dietary recommendations and the USDA Dietary Guidelines both recommend avoiding foods with added sugar [8, 38]. The NDS program enumerates added sugars to foods, which includes monosaccharides, disaccharides and polysaccharides. USDA recommends limiting added sugars to less than 10% of energy, while the AICR recommends limiting sugar-sweetened drinks and “foods that are high in fat, starches, or sugars” since these foods can contribute to weight gain [8]. The score of zero was therefore used for added sugar intakes of 10% or more, a score of one for intake of 5-9.9%, and a score of 2 for intake below 5% of energy from added sugars.

Processed meat.

Avoiding processed meat intake is consistent with both the AICR and ACS recommendations [7, 8]. The maximal score of two was therefore set for no consumption of processed meats. More than none but less than ½ serving/day was scored as one point, and more than ½ serving/day, on average, was scored as zero.

Statistical Analyses

All statistical analyses were conducted in SPSS version 24 (PASW Statistics, Release 24.0.0, Chicago: IBM Corporation). The MCAP Score was generated by summing the individual scores for each of the 11 dietary categories. Since each category was scored 0-2, the maximal score possible was 22. To exemplify utility of the MCAP score, descriptive statistics were generated for the MCAP score using the Healthy Eating Study data. Differences in proportions between the two intervention groups were evaluated using 2-sided, Pearson Chi-Square tests. Relationships of the MCAP Score with serum measures of diet were made using Spearman correlations.

Results

Comparisons of the MCAP with Other Dietary Scores

Table 2 shows the differences between dietary scoring using the MCAP Score, the AICR dietary recommendations, the ACS dietary recommendations, the USDA recommendations and three other published Mediterranean diet scores. Unlike the ACS and AICR scores, the MCAP Score includes MUFA, ω-3 fatty acids and nut/seed intakes. In addition, the MCAP Score enumerates non-starchy vegetables intakes in a separate category. For the total fruit and vegetable score, the maximal score within MCAP for total fruits and vegetables was an average of the ACS and USDA criteria. A total of 3.5 cups/day corresponds roughly to 7 servings/day, which is the USDA recommendation for a 1600 kcal/day diet. Nuts and seeds were enumerated with fruit in both Mediterranean scores shown in Table 2 but not within the MCAP. When grams of nuts/seeds and grams of fruits were summed, subjects consumed a mean of 2.35 g/day and 4.34 g/day at baseline and 6 months, respectively.

Table 2.

Comparison of the MCAP Score with other Mediterranean scores, dietary recommendations for cancer prevention and dietary recommendations for Americans from the U.S. Department of Agriculture (USDA).

Food or nutrient	Mediterranean score, Sofi et al. [28], scores of 0, 1 or 2 for each of 9 categories	Mediterranean Diet Score, Trichopoulou et al.[31], score 1 vs. 0 for 9 categories with noted modification for aMED [51]	American Cancer Society (ACS) recommendation [7]	American Institute for Cancer Research (AICR) recommendation [8]	USDA for 2000 kcal/day [27]	MCAP Score, scores of 0,1 or 2 for each category
Fruit and vegetables	>300 g/day (1 cup) fruit, >250 g (2.5 cups) veg./day	Above the median for fruit plus nuts (separate categories in aMED)	≥ 2.5 cups/day	-	4.5 cups/day	≥ 3.5 cups/day
Vegetables	-	Above the median for all vegetables (excludes potatoes in aMED)	-	≥ 5 servings/day non-starchy fruit & vegetables.	-	≥ 5 servings/day non-starchy vegetables (½ cup/serv.)
Whole grains	> 195 g (~6 oz.)/day Cereals	Above the median for cereals (whole grain cereals in aMED)	Choose whole grains (as opposed to refined grains)	Whole grains with every meal	3 oz./day	≥ 3 oz./day
Legumes	>140 g/day (0.8 cup)/day	Above the median for total legumes	-	Pulses (legumes) with every meal	1½ cups/week	≥ ½ cup/day
Nuts and seeds	Count with fruit, handful daily	Enumerated with fruit (counted separately in aMED)	-	-	5 oz./week nuts, seeds, soy	Included with legumes
MUFA:SFA	Regular use of olive oil as the main fat	Above the median MUFA:SFA ratio	-	-	SFA <10% energy, use oils not solid fats	≥ 2 (typically achieved with 7 tsp./day)
Fish and seafood or ω-3 fats	>250 g/week fish	Above the median for all seafood	-	-	8 oz./week fish (230 g)	ω-3 fatty acid score, roughly ≥ 6 oz./week of salmon
Added sugars		-	Avoid	Avoid sweetened beverages and high energy foods	Limit, <10% of energy	≤ 10% of energy
Meat	< 80 g/day (< 20 oz./week) meat and meat products	Below the median for all meats except seafood (only red and processed meat for aMED)	Avoid	≤ 18 oz/week =2.6 oz./d	26 oz./week meat, poultry & eggs	< 2.6 oz./day of red meat (<18 oz./week)
Processed meat	Included in meat	Counted with red meat	Avoid	Avoid/Limit	Sodium and SFA limits	none for maximal score
Processed foods: savory snacks, chips, cereal bars, sweets	-	-	-	Avoid/Limit	Sodium and sugar limits	none for maximal score
Alcohol	12-24 g/day	10-50 g/day men, 5-25 g/day women (1 point for 5-15 g/day)	Avoid or <1-2/day	Avoid or <7 g/day	Avoid or <1-2/day	-
Dairy	<180 g/day	Below the median (eliminated in aMED)	-	-	3 cups/day	-

The MCAP Score is consistent with the AICR recommendation to avoid processed meats and processed snack foods. These types of foods are classified as ultra-processed foods that have been shown to be associated with increased cancer risks [22]. Refined sugar is also an ultra-processed food. The decision to use scoring of added sugars as percentage of calories, versus for avoidance, was based on the observation that none of the 120 subjects in the Healthy Eating Study achieved complete avoidance of added sugars. Mean intake of added sugars was 13.2% of calories (SD 5.8, range 1.4 - 38.8) at baseline. This improved after intervention decreasing to 11.9% and 9.6% of calories in the Healthy Eating and Mediterranean arms, respectively.

Alcohol and dairy were not included in the MCAP Score. It was difficult to reconcile the potential benefit of a limited daily intake of wine in Mediterranean diets versus the ACS and AICR recommendations to avoid or limit alcohol. Differences in recommendations for dairy foods also was difficult to reconcile. In the historical Greek diet, dairy foods generally were not used in reduced fat form and would have been a significant source of saturated fat, which is limited in the USDA recommendations. Dairy products do, however, provide calcium, and fermented dairy products could have beneficial effects on cancer prevention via modulation of the microbiome [48, 49]. These foods were therefore omitted from the MCAP Score.

For comparison with the MCAP Score, Table 2 also shows scoring criteria developed by Sofi et al. and Trichopoulou et al. [28, 31, 50]. The widely used Mediterranean Diet Score developed by Trichopoulou et al. provides a score of either 0 or 1 for intakes above or below the median intake of nine food groups, with a higher score indicating a beneficial intake (Table 2). In the Alternative Mediterranean Score (aMED) developed by Fung et al., the score of 1 point below median diary intakes was omitted, only red and processed meats were included, only whole grain intakes were included, and a score of 1 was assigned for a moderate alcohol intake [51].

Effects of Dietary Intervention on MCAP Score

The MCAP Score improved after intervention in both the Healthy Eating and Mediterranean study arms in the Healthy Eating Study for Colon Cancer Prevention. The Healthy Eating Study enrolled 120 individuals at increased risk for colon cancer [15]. As shown in Figure 1, the distribution of MCAP scores kept improving from 3 to 6 months in the Healthy arm but deteriorated somewhat in the Mediterranean arm, perhaps due to the difficulty in tracking a greater number of goals. The mean score was, however, significantly higher in the Mediterranean arm versus the Healthy Eating arm after 6 months of intervention (Table 3).

Figure 1.

Distribution of MCAP Scores in the Healthy Eating Study at baseline and after intervention utilizing counseling for either: A. a standard Healthy Eating diet or B. a Mediterranean diet.

Table 3.

The MCAP Score and individual food components of the score before and after dietary intervention. Data shown is mean, SD and percent of subjects with a maximal score. For each diet sub-score, the values range 0-2 with higher score indicating a more optimal diet. For the total MCAP score, the maximum value reached by any individual after 6 months of intervention was 20 of 22 total points possible.

Dietary Score	Baseline, n=120	Healthy Eating at 6 Months, n=47	Mediterranean at 6 Months, n=47
MCAP Score	7.1 (2.8, 0%)	10.9 (3.3, 0%)	12.7 (3.4, 0%) a
Total Fruits and Vegetables	0.84 (0.61, 11.7%)	1.51 (0.65, 45.9%)	1.55 (0.62, 49.2%)
Non-starchy vegetables	0.43 (0.50, 42.5%)	1.00 (0.66, 16.4%)	1.04 (0.72, 22.0%)
Whole Grains	0.67 (0.75, 16.7%)	1.51 (0.66, 59.6%)	1.34 (0.73, 48.9%)
Legumes, Nuts, and Seeds	1.01 (0.90, 41%)	1.11 (0.87, 32.8%)	1.55 (0.77, 57.6%) b
MUFA:SFA ratio	0.71 (0.49, 1.7%)	0.89 (0.48, 4.9%)	1.60 (0.50, 47.5%) b
Long-chain ω-3 Fatty Acid Equivalent c	0.54 (0.77, 16.7%)	0.79 (0.83, 19.7%)	1.02 (0.87, 30.5%)
Added Sugar	0.34 (0.53, 2.5%)	0.57 (0.68, 8.2%)	0.74 (0.64, 8.5%)
Red Meat	1.28 (0.83, 51.7%)	1.64 (0.64, 55.7%)	1.64 (0.64, 57.6%)
Processed Meat	0.63 (0.78, 18.3%)	0.96 (0.86, 26.2%)	1.02 (0.82, 27.1%)
Processed Snack Foods	0.62 (0.61, 6.7%)	0.92 (0.75, 18%)	1.19 (0.80, 33.9%)

^aMean MCAP score between the two diet groups at 6 months differs with p ≤ 0.01, as determined from a two-sided t-test.

^bThe sub-score distributions at 6 months differ between the two diet groups, as determined from the Pearson Chi Square test, with p<0.014 in each case. For the Legumes, nuts, and seeds category, the difference between study arms was mainly due to nuts and seeds.

^cThe long-chain ω-3 fatty acid equivalent in the diet was the sum of eicosapentaenoic acid, docosahexaenoic acid, docosapentaenoic acid and 5% of linolenic acid intakes.

Evaluating the individual sub-scores summed to create the MCAP Score, the main differences between the two arms were in the dietary MUFA:SFA ratios and in the Legumes, nuts and seeds sub-scores (Table 3), which were both significantly increased in the Mediterranean arm. Other differences that were not statistically significant included a somewhat worse “whole grain” sub-score and somewhat better “ω-3 fatty acid” and “added sugar” sub-scores for the Mediterranean arm at 6 months. The total of dietary ω-3 fatty acids was largely composed of linolenic acid intakes. Fish intake was not very well correlated with intakes of ω-3 fatty acids in this study population, likely due to low fish intakes at baseline and a modest goal for fish intake in the Mediterranean arm of two 3-ounce servings/week (not shown).

Correlations of the MCAP Score with Serum Biomarkers

We evaluated the relationship of the MCAP Score with two serum measures of dietary intakes using the Healthy Eating Study data [15, 52]. In that study, serum samples were analyzed for analysis of carotenoids and fatty acids. Overall, body weight changed in the study, even though the dietary goals were designed to maintain baseline body weight. The MCAP Score was negatively correlated with BMI, indicating that a higher score was associated with lower BMI. This was significant at baseline but not after intervention when MCAP Scores improved in most study participants regardless of BMI (Table 4). However, change in the MCAP Score, as a difference of the value at 6 months minus the value at baseline, was significantly correlated with change in BMI (ρ = −0.238, p = 0.021) indicating that an increase in the MCAP score was related to increased weight loss after intervention.

Table 4.

Correlation of the MCAP Score with select measures in the Healthy Eating Study at baseline and after 6 months of dietary intervention.

Measure	Spearman Correlation (ρ) a
	Baseline, n=120	6 Months, n=93
BMI (kg/m2)	−0.209*	−0.094
Weight Change (pounds)	-	−0.203 b
% Diet goals met	-	0.372**
Serum ω-3 fatty acids (% of total fatty acids) c	0.340**	0.210*
Serum carotenoids (ng/mL)	0.198*	0.276**

^aStatistically significant p-values for the Spearman correlations are marked as

^*p<0.05,

^**p<0.01

^bThis correlation approached significance with p=0.051. Weight change was, however, significantly correlated with change in the MCAP Score with ρ = − 0.238, p = 0.021.

^cSerum ω-3 fatty acids were calculated as % of total fatty acids and the ω-3 fatty acids summed were linolenic acid, eicosapentaenoic acid and docosahexaenoic acid.

The MCAP Score at 6 months also was correlated with the percent of the study-set diet goals met (ρ = 0.372, p < 0.01). Both diet interventions emphasized fruit and vegetable intakes. Irrespective of diet arm assignment, MCAP Score was positively correlated with serum total carotenoids and with serum total ω-3 fatty acids at the pre-intervention and post-intervention time points (Table 4).

Discussion

There is a large body of data to support a Greek-Mediterranean style of diet for the prevention of cancer at multiple organ sites [53–55]. However, not all prospective studies have observed an association between adherence to a Mediterranean diet and decreased risks of cancer [56–58]. One of the reasons behind these inconsistencies may stem from the use of Mediterranean diet scores that are based on the median intakes for the population being studied. This approach can be problematic, especially in regions of the world where the average diet deviates from a preventive ideal by a large extent. In such populations, the median intake is likely not at all close to that of the traditional Greek diet and, as such, a high Mediterranean score could be misleading [59]. For example, good sources of MUFA are both meat and olives.

Trichopoulou et al. initially proposed a Mediterranean diet scoring system that has been widely utilized. The score is based on dietary intakes above or below the median for nine food categories [31]. Sofi et al. designed a Mediterranean dietary score that is also based on nine food categories, but with this approach, defined cutoff values derived from a comprehensive meta-analysis synthesizing the effect of Mediterranean diet interventions on multiple health outcomes (mortality, cardiovascular disease and neoplasia) [28]. The score of Sofi et al. is similar to other scores that have been developed to characterize a Mediterranean eating pattern [29, 30, 60, 61]. We therefore designed the MCAP score relying on the dietary cutoffs proposed by Sofi et al. in addition to incorporating the ACS and AICR guidelines for cancer prevention.

A Mediterranean style of diet generally does meet with the ACS and AICR dietary recommendations, and it has the potential to support successful weight management in the absence of explicit weight loss goals. There is plentiful data to support the roles of high fruit/vegetable intakes, high MUFA, and high ω-3 polyunsaturated fatty acids (PUFA) in avoidance of adult weight gain and reduced adult adiposity. Diets high in plant-based foods have been shown to promote maintenance of a lower body weight versus diets high in sugar-sweetened beverages, refined grains, potatoes, and sweets in many studies [62–64]. Phytochemicals promote mitochondrial fat oxidation to prevent lipid accumulation, as one mechanism behind their anti-obesogenic effects [62]. Similarly, randomized trials show that diets high in MUFA or ω-3 PUFA also increase fatty acid oxidation, inhibit de-novo lipogenesis, and result in reduced visceral fat vs. diets high in either saturated fat (SFA) or ω-6 PUFA [10–12, 14, 44, 45, 65, 66].

Both the ACS and AICR stress the importance of healthy BMI and decreased adiposity in reducing risk of many cancers [7, 8]. Experimental data show that energy balance has strong effects on tumor development and may even overshadow the effect of diet composition. The use of energy restriction in cancer prevention has been supported by many animal studies [67]. Clinical trials are currently testing the relative benefit (or harm) of intermittent energy restriction versus continuous calorie restriction on cancer endpoints [68]. Unfortunately, a major problem in conventional treatment of overweight and obesity is weight regain, even after bariatric surgery, perhaps due to compensatory physiological changes [69–71]. This then points to the potential utility of focusing on dietary quality that leads to long-term weight management instead of acute weight loss per se to achieve long-term weight management and avoid weight cycling.

A major difference between the MCAP Score shown here and the ACS and AICR dietary recommendations is in the dietary fat intakes. Increased intakes of ω-3 fatty acids and MUFA are not targeted by the ACS and AICR dietary recommendations at the present time, but the USDA Dietary Recommendations for Americans do encourage a weekly fish intake and limits on intakes of SFA (Table 2). Fish intake is a key component of Mediterranean diets that could contribute to the observed cancer preventive effects of the diet. In addition to possible effects on fat metabolism, mentioned above, there is abundant experimental evidence that dietary fish intake and ω-3 fatty acids are associated with cancer prevention via multiple mechanisms, including decreased production of pro-inflammatory eicosanoids [72].

As an example of the potential utility of the MCAP Score, we calculated the MCAP scores in a randomized dietary intervention trial that we conducted in persons at increased risk of colon cancer, the Healthy Eating Study. The MCAP score increased after counseling for either a Healthy Eating or a Mediterranean diet, but the increase was greater with the Mediterranean diet (Table 3). Further, the MCAP score correlated with blood carotenoids and ω-3 fatty acids, and an overall increase in the MCAP Score at 6 months was significantly correlated with a decrease in BMI (Table 4). This is consistent with other studies demonstrating that shifts in fat intakes have favorable effects on body weight and waist circumference in the absence of weight loss goals [10–14].

A focus on attaining a diet that increases the MCAP score instead of attaining an energy deficit could provide individuals with a method that they can use life-long. A focus on diet quality has been suggested as a new approach to weight management [69]. This is unlike the conventional weight loss approach of targeting weight loss first followed by targeting behaviors for weight loss maintenance second. A focus on food goal achievement also provides a more immediate and tangible goal for individuals vs. focus on weight loss that is secondary to diet change over a period of time. Lastly, in an important high risk sub-group of cancer survivors, a weight loss rate that is typically prescribed for obese individuals (10% in 6 months) is contraindicated, and increased biomarkers of cancer cell proliferation and adipose inflammation have been detected when weight loss is rapid [16–21].

Important limitations of the present study that need to be recognized are the modest sample size and the lack of a cancer endpoint. Despite this, we were able to show that following a Mediterranean diet led to an increased MCAP Score. The MCAP score was correlated with desirable increases in serum carotenoids and serum ω-3 fatty acids, and with decreases in weight after dietary intervention (Table 4).

In summary, the Mediterranean diet pattern is useful to consider for interventions aimed at cancer prevention. The MCAP score we developed expands upon the recommendations of the ACS and AICR to include a focus on monounsaturated fatty acids and on ω-3 polyunsaturated fatty acids that have preventive properties as well as effects on weight control. Importantly, the MCAP score incorporates defined dietary category cut-off values based on the literature to ensure that research-backed, preventive intakes are reached in scoring. The MCAP Score is therefore potentially useful for 1) facilitating the measurement of this dietary pattern in epidemiological studies, and 2) guiding how dietary recommendations are formulated for individuals at increased risk of cancer to optimize dietary intakes and produce modest weight loss. Future research to validate the MCAP Score should be conducted in larger cohorts that ideally have cancer incidence endpoints available.