There is conflicting evidence from observational studies examining potato consumption and predictors of obesity, such as increases in weight, body mass index (BMI) and waist circumference (WC). Mozaffarian et al. reported a small weight gain of 0.71 lb (95% CI: 0.53–0.89) over a four year period for every 1 serving/day increase in boiled, baked, or mashed potato, and a larger weight gain (4.11 lb; 95% CI: 3.46–4.76) for every 1 serving/day increase in French fries [54
]. French fries have also been associated with weight gain in women but not in men [55
]. Consumption of French fries, but not other potato types, has been associated with increased BMI [56
]. Some have observed an association between total potato intake and increased WC in women [57
], whereas others have found no association between total potato intake and WC [58
]. Details of these studies are reported in Table 2
. A systematic review of these observational studies concluded that there was no evidence for an association between potato consumption and obesity, however, there may be some evidence for an association between French fries and obesity. Overall there were few studies and those that were included were of relatively poor quality [18
Whilst these studies, and those discussed in later sections on T2DM and CVD, attempted to control for other dietary and lifestyle factors, utilising multivariate linear regression models, it is possible that other unidentified factors are confounding their results. For example, few studies included socioeconomic status (SES) in their analysis, although other factors associated with SES, such as activity levels, fruit and vegetable intake, smoking and alcohol intake were included in most analyses. Dietary intake was measured by food frequency questionnaires (FFQs), often containing a very restricted number of food items. The use of these questionnaires whilst practical for studies involving large numbers of participants, limits the amount of information that can be obtained. When considering potato consumption they appear to have been limited to very few options, sometimes only reporting on total potato consumption [63
], or grouping baked, boiled and mashed potato as a single item [66
]. Where more detail is provided, it is still insufficient with regard to cooking methods. For example, boiled potatoes will differ in nutrient content, particularly fibre, depending on whether or not they are cooked and consumed with their skins on. Additionally, French fries are often cited as producing a different effect from other preparation methods, however what is classed as a French fry varies from country to country. In the US any type of potato that has been sliced into batons and fried is classed as a French fry, regardless of size of the baton, whereas in the UK and parts of Europe the size of the baton determines the name; a very thin baton would be classed as a skinny chip, a wide-cut baton as a chip and only a medium sized baton referred to as a French fry. An additional point of confusion is that in the USA and other countries, ‘chips’ would be a thin, fried potato snack sold in bags, known as ‘crisps’ in the UK. When deep-frying, the surface area-to-volume ratio affects the amount of fat absorbed, with thinner-sliced batons absorbing more oil than their thicker counterparts. Furthermore, there is no distinction between oven-baked and deep-fried French fries, two preparation methods which could vary widely in fat content.
When food intake is assessed by dietary recall methods, answers may not be very accurate, particularly if a person’s diet has changed in the intervening years. Reverse causality can also be a concern when interpreting results as it is common for people to make changes to their diet after learning they are at increased risk of a particular disease, for example reducing saturated fat intake after receiving a high cholesterol report. If these dietary changes are not made in time to affect their health, this could lead to people with apparently healthy diets being erroneously reported to be more likely to develop a particular disease.
As changes in weight, BMI and WC take place over a relatively long period of time, it would be difficult to design a well-controlled, longer-term intervention examining their effect on these markers, particularly as potatoes are not consumed alone, but within the context of a mixed diet. Instead of measuring weight gain directly, several acute studies have compared the effects of consuming different starchy carbohydrates, including potatoes, on satiety and subsequent energy intake [67
]. A variety of methodologies have been implemented, such as matching for carbohydrate or energy content or allowing ad libitum intake; some served the carbohydrate on its own, whereas others served it within the context of a mixed meal. These studies are summarised in Table 3
When isoenergetic portions of starchy carbohydrates are consumed, potatoes have been reported to be more satiating than pasta, rice and bread [67
]. Furthermore, when 38 different test foods were compared, boiled potatoes were reported to have the highest satiety index of all test foods, even when compared to protein and fat-rich foods [67
]. When different preparation/serving methods were compared, Leeman et al. reported both boiled and mashed potatoes to be more satiating than French fries when meals were energy matched, but not when matched for carbohydrate content [69
]. This is likely due to boiled potatoes being less energy dense and, therefore, having a larger portion size than French fries when energy matched, as feelings of fullness and satiety are affected by stomach distension and capacity [73
]. Indeed, a 1000 kJ portion of boiled potatoes weighs 368 g in comparison to a matched portion of French fries which weighs only 93 g [74
]. Geliebter et al. compared isoenergetic (~1000 kJ) amounts of instant mashed potato, peeled baked potato, pasta, and brown rice [70
]. Each test meal was accompanied by a variable amount of water designed to bring the total meal water content to 400 g, which could potentially ameliorate some of the effects of meal volume. However, despite this, they found that both potato meals reduced appetite compared to pasta and rice. This may be because water served alongside a meal has been demonstrated to have no effect on satiety in contrast to water incorporated into a meal [75
In contrast to these studies, Diaz-Toledo et al. reported higher satiety ratings for French fries compared to an energy-matched pasta control, with no differences between baked potato or mashed potato and the pasta control for any satiety measure [72
]. There were several differences between their study and the other isoenergetic studies. Their participants were given a personalised breakfast 3 h prior to the test meal and so were not fasted when they consumed the test meal. Furthermore, their test meal was a mixed meal, containing meatballs in tomato sauce, salad and dressing; the starchy test food was not consumed in isolation. The mashed potato was a pre-prepared dish, which included a comparable amount of fat to the French fries, resulting in a smaller difference in portion weight and energy density between the two. Finally the overall energy content of the test meal was 1883 kJ, almost double that of other isoenergetic studies. These differences in energy density, macronutrient content and total energy may contribute to the contradictory results from this study.
When adult participants were given ad libitum amounts of boiled pasta, rice or potatoes, along with a fixed amount of meat, and instructed to eat until they were no longer hungry, equivalent amounts of the carbohydrate element were eaten (353–372 g), however, because of the lower energy density of the potatoes, less total energy was consumed in that meal [68
]. After 4 h, satiety and hunger levels had returned to baseline for those consuming the potato meal, whereas they had not for the pasta and rice meals. Plasma insulin was lower after the potato meal, with no difference between meals for glucose, despite potatoes having a reported higher glycaemic index than pasta and rice [76
]. This discrepancy is likely due to the smaller amount of carbohydrate in the potato meal.
Similarly, Akilen et al. served ad libitum amounts of potato (either boiled and mashed, oven fries or French fries), boiled pasta or rice with a fixed portion of meatballs to a group of normal-weight children [71
]. Lower weights of oven fries and French fries were consumed compared to pasta, however when energy intake was compared, energy from the boiled mashed potato meal was lower than all other meals. In this study, there was no difference between meals for appetite scores until they were adjusted for energy intake, which resulted in a lower mean appetite rating/kcal for the boiled mashed potato meal in the 2 h following the meal.
Evidence for a relationship between satiety scores and energy intake at a subsequent ad libitum meal is mixed. When isoenergetic portions are compared, some have reported no difference in subsequent energy intake either between different potato preparation methods or in comparison to other starchy carbohydrates, despite higher satiety scores for boiled potatoes [70
]. Holt et al., however, reported an inverse association between satiety score and energy intake at a second meal [67
] across their 38 test foods, with a tendency for a lower total energy intake across the whole day from the most satiating foods. When participants were permitted ad libitum amounts of potato, pasta or rice at a first meal, the lower energy intake and lower 4 h satiety score from the potato meal did not translate into a greater energy intake at a subsequent ad libitum sandwich meal [68
Of the studies that measured postprandial glucose and insulin responses [67
], neither reported any direct correlation between glycaemic or insulinaemic response and satiety score, although an indirect relationship between insulin response and satiety was suggested by Holt et al., who reported inverse associations between both insulin score and satiety score with subsequent ad libitum energy intake.
In summary, isoenergetic portions of potatoes, in particular boiled potatoes, appear to be more satiating than other starchy carbohydrates when eaten in isolation. When ad libitum consumption is permitted, less energy is consumed in mixed meals containing potato, with no compensatory increase in energy intake at a subsequent meal, despite lower satiety ratings. It should be noted that the evidence is limited as there have been few studies of this type, particularly those examining the effects of ad libitum consumption of potatoes in the context of a mixed meal. Despite this, the results from studies so far do not support a link between potato consumption and risk of overweight and obesity.
3.2. Type 2 Diabetes Mellitus (T2DM)
An association between total potato consumption and risk of developing T2DM has been reported [66
], with the highest risk associated with consumption of French fries. Muraki et al., in an analysis of data from three prospective cohort studies [66
], reported that, for every three servings/week of boiled, mashed or baked potatoes there was an increased risk of T2DM (HR, 1.04; 95% CI 1.01–1.08), with a greater risk associated with French fries (HR, 1.19; 95% CI 1.13–1.25).
It has been suggested that the high glycaemic index (GI) of potatoes may be a contributory factor, as high GI diets have been associated with an increased risk of T2DM [79
]. GI is a measure of how much a carbohydrate-containing food raises blood glucose in relation to a control (glucose):
Foods with a GI > 70 are classed as high GI, whereas those with a GI < 55 are classed as low GI. Various factors affect the GI of the potato, such as variety, cooking method, and length of cooking. Shorter boiling times, in particular, may lead to incomplete gelatinization of the starch, with residual RS2 contributing to a lower GI. This is well demonstrated with the Carisma potato, which has been labelled low GI [81
]. This is likely due to its higher onset of gelatinization temperature than other cultivars [82
], as this would result in less extensive gelatinization, and therefore more resistant starch, when cooked for the same length of time as other varieties. Using GI as a predictor of a food’s effect on blood glucose is also problematic. It is calculated based on consumption of a fixed amount of CHO, usually 50 g; it does not take portion size into account. Thus, a food may have a high GI but have little effect on blood glucose because the carbohydrate density is low, resulting in a small amount of carbohydrate being consumed in a standard portion. This has led to the suggestion that glycaemic load (GL) may give a better representation of the actual effect a food has on the glycaemic response as it considers portion size along with GI:
A GL less than 10 is considered low, with high GL categorized as a GL > 20. Typical GI and GL values, along with the amount of available CHO in a standard portion, for some common methods of preparing and serving potatoes are shown in Table 4
. To put these values in context with other starchy CHO, boiled/steamed rice typically has a GI of 68–87 and a GL of 25–33 (150 g portion), depending on rice type and cooking time, whereas pasta has a GI of 51–61 and a GL of 24–29 (180 g portion) [83
It appears from the reported GI values that GI alone cannot explain the association observed between French fries and T2DM risk, as French fries typically have a lower GI than other potato preparations. They do have a higher GL, which may partially explain the reported associations, however, other factors such as fat content and other unidentified, unhealthy lifestyle choices cannot be discounted. It should also be considered that potatoes are not usually eaten in isolation; other foods in the meal will affect the overall GI/GL of the meal. For example, serving a baked potato with cheese reduced the GI from 93 to 39 [84
] and serving chicken breast, salad, and oil with mashed potato resulted in a reduction in GI from 108 to 54 compared to mashed potato served alone [85
Furthermore, not all studies agree, with some reporting either no association [64
] or an inverse association between potato consumption and development of T2DM [63
]. Farhadnejad et al. reported a lower incidence of T2DM in those who consumed higher amounts of potatoes (55.5 g/day) compared to the lowest (7.3 g/day) consumption range [87
]. A significant inverse association was observed for both total potato consumption and boiled potatoes with a trend observed for fried potatoes. These studies are summarized in Table 5
A recent systematic review and meta-analysis [19
] reported a slightly increased risk of T2DM (RR: 1.09, 95% CI 1.01, 1.18) for every 150 g/day increase in boiled, baked and mashed potatoes, with a stronger association reported for French fries (RR: 1.66, 95% CI 1.43, 1.94). The authors reported, however, that the quality of evidence was low for total potato consumption and moderate for French fries. The contradictory and limited evidence from the epidemiology does not support any recommendation to reduce total potato intake, with the possible exception of French fries, at this time.
3.3 Cardiovascular Disease (CVD) and CVD Risk Factors
Several cohort studies have examined associations between potato consumption and CVD and its risk factors; a summary of these studies is presented in Table 6
. Larsson et al., investigated associations between potato consumption and risk of myocardial infarction, heart failure, stroke or mortality from CVD in Swedish men and women [94
]. They found no significant association between total potato intake and risk of major CVD event or mortality from CVD. Nor did they report any associations between boiled, fried, or French-fried potato consumption and any CVD outcome. In a large cohort study investigating the relationship between fruit and vegetable consumption and risk of ischemic stroke, total potato consumption was not associated with ischemic stroke risk, although individual preparation methods were not explored [95
Studies examining the relationship between hypertension (HT) and potato intake have reported mixed results. In a Chinese cohort, total potato consumption, stir-fried and non-stir-fried potato consumption were all associated with increased risk of developing HT [96
]. However, when non-potato-consumers were excluded from the analysis, higher intakes of total potato and stir-fried potato were associated with lower risk of HT. Borgi et al. also reported an association between total potato consumption and HT for those consuming ≥ 1 serving/day and an increased risk of HT for those consuming ≥ 4 servings/week of French fries. Consuming ≥ 4 servings/week of boiled, baked and mashed potatoes was associated with increased HT risk in women but not men [97
]. In contrast to these two studies, Hu et al. reported no association between total potato consumption and change in blood pressure (BP) or HT risk in either the PREDIMED or SUN cohort over four years [98
In their systematic review and dose-response meta-analysis, Schwingshackl et al. examined associations between potato consumption and risk of chronic disease [19
]. They reported no association between total potato consumption and risk of coronary heart disease or stroke, even for the highest total potato intake (150 g/day). However, high consumption of French fries (150 g/day), but not other preparation methods (boiled, mashed or baked) was associated with increased risk of HT. Again, the authors noted that quality of evidence was low for boiled, mashed, and baked potatoes and moderate for French fries. They also stated that the studies’ results were confounded by only reporting total potato consumption in the majority of cases.
There have been few interventional-type studies examining the effect of potato consumption on CVD risk factors. One explanation may be that any intervention examining these measures, would have to be carried out over a longer period of time, unlike, for example, those investigating effects on postprandial glucose metabolism. Furthermore, in order to maintain energy balance, an intervention would have to remove some other component of the diet in order to incorporate potatoes. This in itself would confound the results, because any effect may be the result of what has been removed, rather than what has been added. If potatoes were added to the diet, without removing anything then overall energy intake could increase, potentially resulting in weight gain, although this was not observed in one of the studies discussed here [99
Arterial stiffness is an independent risk factor for the development of CVD [100
]. Tsang et al. explored the effects of an anthocyanin-rich potato, Purple Majesty (PM), on pulse wave velocity (PWV), a clinical measure of arterial stiffness [101
]. They found that consumption of 200 g/day of PM potatoes, for 14 days, significantly reduced carotid-femoral PWV, in healthy individuals, whereas consumption of an equivalent amount of white potato had no effect. They reported no change in blood pressure, fasted glucose, insulin, triacylglycerol or HDL-, LDL-, and total cholesterol for either potato variety. They hypothesized that anthocyanins in the PM potatoes contributed to the observed results as anthocyanin intake has been associated with reduced arterial stiffness [102
Vinson et al. also investigated the effects of the PM potato in two separate trials [99
]. In an acute study, they investigated the effects of the PM potato on plasma antioxidant activity and urinary polyphenols compared to a control biscuit containing an equivalent amount of potato starch. Plasma antioxidant capacity, measured by ferric reducing antioxidant power, was non-significantly higher following the PM potato meal and urinary polyphenols were increased by 92% (p
= 0.09 for trend) following PM consumption, compared to controls. Urinary polyphenols are a marker for polyphenol intake, with higher concentrations associated with reduced risk of HT [103
]. In a second study, they investigated the effect of PM potatoes on BP in 18 individuals, 14 of whom were hypertensive, of which 13 were taking antihypertensive medication. In this crossover trial they compared the effects of four weeks consumption of PM potatoes at lunch and dinner with no potatoes for the same time period. They reported a significant (4 mmHg) reduction in diastolic blood pressure (DBP) following PM consumption, with no significant effects on systolic blood pressure (SBP), body weight, glucose, HDL- or total cholesterol and triacylglycerol. These interventional studies are reported in Table 7
In summary, the epidemiology generally reports no associations between potato consumption and the risk of CVD, with the possible exception of HT, where some, but not all, have reported increased risk from both total potato consumption and French fries. Potatoes are a rich source of potassium, which has been associated with reduced risk of CVD [104
], however, French fries are often consumed with salt which could attenuate any beneficial effect of potassium as high salt intake is associated with HT and could increase the risk of CVD [105
]. In contrast, interventional studies have demonstrated some beneficial effects from an anthocyanin-rich pigmented potato variety on PWV and DBP. Whilst these results are interesting, it should be noted that these results are confined to a single pigmented potato cultivar and no effect on PWV was observed following consumption of white potatoes. Clearly, further research is required, utilizing different, commonly consumed potato cultivars, before conclusions can be drawn.