Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Carbohydrates, glycemic index and diabetes mellitus

Effect of meal frequency on glucose and insulin levels in women with polycystic ovary syndrome: a randomised trial

A Corrigendum to this article was published on 04 May 2016

This article has been updated



The aim of the study was to compare the effect of two-meal patterns (three vs six meals per day) on glucose and insulin levels in women with polycystic ovary syndrome (PCOS).


In a randomised, crossover, 24-week study, 40 women with PCOS, aged 27±6 years, body mass index 27±6 kg/m2, followed a weight maintenance diet (% carbohydrates:protein:fat, 40:25:35), consumed either as a three- or a six-meal pattern, with each intervention lasting for 12 weeks. Anthropometric measurements, diet compliance and subjective hunger, satiety and desire to eat were assessed biweekly. All women underwent an oral glucose tolerance test (OGTT) with 75 g glucose for measurement of plasma glucose and insulin at the beginning and end of each intervention. HaemoglobinA1c (HbA1c), blood lipids and hepatic enzymes were measured at the beginning and end of each intervention.


Body weight remained stable throughout the study. Six meals decreased significantly fasting insulin (P=0.014) and post-OGTT insulin sensitivity (Matsuda index, P=0.039) vs three meals. After incorporation of individual changes over time, with adjustment for potential confounders, the only variable that remained significant was the Matsuda index, which was then used in multivariate analysis and general linear models. Six meals improved post-OGTT insulin sensitivity independently of age and body weight vs three meals (P=0.012). No significant differences were found between six and three meals for glucose, HbA1c, blood lipids, hepatic enzymes, subjective desire to eat and satiety.


Six meals had a more favourable effect on post-OGTT insulin sensitivity in women with PCOS compared with isocaloric three meals.


Polycystic ovary syndrome (PCOS) is the most common endocrine disorder in women, with prevalence up to 15%, and is the most common cause of infertility.1 Principal features of the syndrome include chronic anovulation, hyperandrogenism and polycystic ovaries.2 The majority of women with PCOS have hyperinsulinaemia and insulin resistance, which have a significant role in the pathogenesis of the syndrome3, 4 and in the long term may lead to impairment of glucose metabolism and type 2 diabetes.

Reduction in insulin resistance has been suggested as the principal goal of PCOS treatment. Several studies have assessed the effects of caloric restriction and/or alteration in the diet’s macronutrient composition on metabolic outcomes.5, 6, 7 Regardless of the dietary intervention or the study design used, the optimal dietary management is still unknown,8, 9 and lifestyle changes (diet plus physical activity), along with weight loss, are hitherto proposed as the first strategy for amelioration of insulin sensitivity.5, 10, 11, 12 However, despite the fact that the majority of women with PCOS are overweight or obese,13 many lean women with PCOS are considered to be also at increased risk for metabolic disorders.14 Thus, potential dietary interventions independent of weight loss may be of major importance,9 including meal frequency modification.

The daily distribution of energy and carbohydrates in the proposed diets is a topic that has not been adequately investigated and may also be important. A recent study reported that, in lean women with PCOS, higher energy intake at breakfast with lower intake at dinner significantly improved insulin resistance and androgen production.15 In contrast, another study in patients with type 2 diabetes showed most favourable post-oral glucose tolerance test (OGTT) glucose profile when the majority of carbohydrates (50% of energy) were consumed at lunch time vs breakfast and dinner or equally distributed throughout the day.16 Furthermore, the quality of carbohydrates may also be important. A recent study showed that an isocaloric low glycemic index diet improved insulin sensitivity in women with PCOS.17

The effects of meal frequency on glucose and insulin responses have been reported in several studies with healthy people following energy-deficit diets and people with type 2 diabetes, with conflicting results.18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 To our knowledge, there is no study examining the impact of meal frequency on glucose and insulin levels in women with PCOS, independently of weight loss. Therefore, the aim of the present study was to investigate any potential effect of a three- vs a six-meal pattern on glucose and insulin levels in women with PCOS, consuming a weight maintenance diet.

Materials and methods


Subjects were recruited from the outpatient unit of Attikon University Hospital and their initial assessment included a detailed history, full examination, ovaries ultrasound and biochemical and hormonal tests (within the last month before the beginning of the study), according to the routine clinical practice. Subjects who were eligible (according to the inclusion and exclusion criteria) and willing to participate were included in the study. The study’s exclusion criteria were as follows: (1) women on metformin, contraceptives, steroids or any medications known to affect glucose, insulin or reproductive hormones for at least 6 months; (2) known diabetes mellitus and/or pre-screened fasting glucose >126 mg/dl, abnormal thyroid function or prolactin levels; (3) compliance to a weight loss diet or the use of medications that affect body mass; and (4) change in body weight 4.5 kg or a change in physical activity within the preceding 6 months (that is, if they used to have organised physical activity and they stopped it or the opposite). PCOS was defined according to the Rotterdam criteria.1 The protocol and potential risks and benefits of the study were fully explained to each subject and written consent was obtained. The protocol was approved by the Bioethics Committee of Attikon University Hospital and was carried out in accordance with the Declaration of Helsinki (1997).

Trial registration: numberNCT02248272.

Study design

The study had a randomised crossover design, and subjects were randomly enrolled to the interventions using a single allocation ratio. Volunteers followed a weight maintenance diet (40% carbohydrates, 25% protein and 35% fat), consumed either as a three- or a six-meal pattern. The first meal pattern (3 or 6) was followed for 12 weeks and then participants switched to the other meal pattern for another 12 weeks. A meal was defined as an eating occasion of the day containing >150 kcal, occurring at morning (‘breakfast’), mid-day (‘lunch’) or evening (‘dinner’).31 Snacks were defined as eating episodes, containing <150 kcal, consumed at other than recognised ‘meal’ times.31 The carbohydrate distribution for the three-meal pattern was 20% at breakfast, 50% at lunch and 30% at dinner, whereas for the six-meal pattern was 20% at breakfast, 10% at morning snack, 30% at lunch, 10% at afternoon snack, 20% at dinner and 10% at before bedtime snack. Table 1 gives an example of a 1900 kcal diet with the two-meal pattern distributions.32 All volunteers were asked to be consistent with mealtimes throughout the intervention. Most women were sedentary at baseline and were asked to maintain their usual physical activity levels. Participants were also advised not to drink more than one unit of alcohol (1 small glass of wine, ½ pint of ordinary strength beer or 1 single measure of spirits) per week throughout the study.

Table 1 Example of a 1900 kcal daily regime with either three or six meals. Bayesian Information Criterion

Dietary and physical activity assessment

At baseline, dietary habits were assessed through a semi-quantitative food frequency questionnaire, and the quality of the background diet was assessed through the Mediterranean diet score that assesses adherence to the Mediterranean dietary pattern.33 The range of the diet score is 0–55, with higher values indicating greater adherence to the Mediterranean diet.

Subjects recorded their caloric intake on a daily basis. Subjects were asked to record the type and amount of any food and beverage consumed throughout the intervention periods and detailed instructions were given on how to record the quantity of food eaten, using standard household and other measures. To avoid bias, the diaries were checked for misreporting or other possible problems by the dieticians, and if needed food models and photographs were employed to clarify discrepancies in portion size. Food records were analysed using Diet Analysis Plus software (version 6.1, ESHA Research, Salem, OR, USA). The database was extensively modified to include new foods and recipes. Food records were monitored biweekly and dietary adjustments were made if needed. At biweekly visits, subjects completed three 10-point visual analogue scales to record their subjective feelings of hunger, satiety and desire to eat. During these evaluations participants were not asked to be in a fasting state and were advised to follow their meal pattern according to the intervention arm.

Physical activity of the participants was assessed through a validated brief self-reported questionnaire (the Harokopio Physical Activity Questionnaire (HPAQ)).34 This questionnaire collects the previous week’s self-reported physical activity and examines the time spent in light, moderate, high-intensity activities and sleep. On the basis of the metabolic equivalents of all activities, the mean daily energy expenditure and the physical activity level were estimated.

Anthropometric measurements

Height, body weight, and waist and hip circumference were measured. Body mass index (BMI) was calculated as weight (kg) divided by height in metres squared (m2).

Biochemical analyses and an oral glucose tolerance test

Subjects visited the clinic between 0900 and 0930 hours at weeks 0, 12 and 24 after a 12-hour fast and avoidance of alcohol and exercise. The night before each one of these visits to the clinic, participants were asked to ensure that their evening meal containing 40–60 grams of carbohydrate was consumed by 2100 hours. For those following the six-meal intervention, their evening meal had to be consumed by 1900 hours, with the ‘bed time snack’ consumed by 2100 hours as the last meal of the evening and then to remain fasted. Fasting venous blood samples were collected, and, immediately after, all subjects underwent an OGTT with 75 g glucose. Blood samples were taken every 30 for a total of 120 min and were immediately frozen at −80 °C. All analyses were performed at the end of both interventions.

Plasma glucose measurements were based on the enzymatic method with glucose oxidase (BIOSIS, Athens, Greece). Plasma insulin was measured by a commercially available human ELISA kit (Invitrogen, Frederick, MA, USA; intra-assay CV: 4.8–6.0%; sensitivity: 0.17 μIU/ml). HaemoglobinA1c (HbA1c) was measured with RXDaytona analyser (Randox Laboratories, Kearneysville, WV, USA). Hepatic enzymes (SGOT, SGPT, γ-GT, ALP), total cholesterol, triglycerides and high-density lipoprotein cholesterol were measured by the AEROSET/ARCHITECT c8000 System (Abbott, Chicago, IL, USA). Low-density lipoprotein cholesterol was calculated using the Friedewald formula.35 Fasting insulin resistance was assessed by the calculation of the homeostasis model assessment of insulin resistence.36 Post-OGTT insulin sensitivity was assessed by the Matsuda index.37

Statistical Analysis

Normality of the variables was tested using P–P plots. Normally distributed continuous variables (age, BMI and Mediterranean diet score) are presented as mean values±s.d. and categorical variables (family history of diabetes mellitus) as frequencies. Associations between categorical variables were tested by the calculation of Pearson’s χ2-test or the McNemar test when the samples were paired. Comparisons of mean values of normally distributed continuous variables by clinical outcome were performed using the paired t-test and for the skewed one using the non-parametric test Mann–Whiney. All reported P-values were based on two-sided statistical tests. The effects of treatments and time were assessed using 2 × 2 repeated measures analysis of variance. In the absence of normality, variables were ranked and then the Friedman non-parametric statistical test was used. Treatment to period interactions were used as covariates in the repeated measures analysis of variance to test for the carry-over effect. Paired sample t-tests and related samples Wilcoxon ranks test were performed, to determine differences in absolute changes in parameters within a period. A linear mixed model for three different time points was used: baseline, end of 3-month three-meal intervention and end of 3-month six-meal intervention. The study had 80% power (α=0.05) to detect differences between dietary groups of 0.28 mmol/l in fasting plasma glucose. Significance was set at P<0.05. SPSS 18.0 software (SPSS Inc., Chicago, IL, USA) was used for all the statistical calculations.


Forty-five women with PCOS entered the study, but five dropped out because of pregnancy during the course of the study (n=3), personal issues (n=1) and non-compliance to weight stability (n=1). Forty women, aged 27±6 years, with mean BMI of 27±6 kg/m2, completed the study (Table 2). Participants’ BMI ranged from 20 to 40 kg/m2, with half of the participants classified as normal weight (BMI <24.9 kg/m2) and half classified as overweight/obese (25>BMI>29.9, n=7; 30>BMI>40, n=13, respectively). Overweight/obese women were older (P=0.004), had higher mean daily energy expenditure (P<0.001) and higher physical activity level (P=0.05) compared with normal weight women (Table 2). Almost half of all women (49%) reported a positive family history of diabetes mellitus, with the majority of overweight/obese women (65%) reporting a positive family history of diabetes (Table 2). According to the OGTT test performed at the beginning of the study, none of the participants had diabetes mellitus; two women had impaired fasting glucose and two women had impaired glucose tolerance. The rest had normal glucose levels. Mean Mediterranean diet score estimating long-term adherence to a Mediterranean-type diet was 32±5, similar to the general population, implying a moderate adoption (Table 2). No significant differences were found for fasting glucose and insulin between groups who did the three- or six-meal arms first or second (P=0.31 and P=0.08, respectively).

Table 2 Participants’ baseline characteristics in the whole sample (n=40) and according to body mass index categories

Dietary intervention regardless meal frequency was well tolerated, with no adverse effects reported. There were no significant differences noted when we analysed the 3- or 7-day dietary records in a subgroup of volunteers, and, as the 3-day food records are equally reliable to the 7-day ones, we chose to present the 3-day food records (two consecutive weekdays and one weekend day). The energy and macronutrient contents derived from the subjects’ 3-day diet weighed food records are shown in Table 3 and did not differ from those of the prescribed diets (P=0.6). No difference was reported between the three- and the six-meal pattern intervention regarding total daily energy intake (P=0.14) and macronutrients intake, that is, carbohydrates intake (P=0.27), protein (P=0.13) and fat (P=0.93; Table 3). Subjective satiety and desire to eat did not differ significantly between the three- and the six-meal pattern interventions (P=0.35 and P=0.06, respectively; Table 3). Those following the six-meal pattern reported lower subjective hunger compared with the three-meal pattern (P=0.03).

Table 3 Mean dietary intake based on weekly 3-day weighed food records completed throughout each period and mean subjective appetite scores

The differences between the two interventions are presented in Table 4. BMI and physical activity level remained stable throughout the interventions (P=0.52 and P=0.71, respectively). Waist circumference was significantly decreased during both interventions, but there was no significant difference between the interventions (P=0.163, Table 4). Compared with the three-meal intervention, fasting plasma insulin was significantly lower at the end of the six-meal intervention (P=0.03) and fasting insulin resistance (homeostasis model assessment of insulin resistence) decreased at the end of the six-meal period but was not statistically significant (P=0.06). Post-OGTT insulin sensitivity (Matsuda index) was significantly increased at the end of the six-meal intervention (P=0.02). The above mentioned differences between the two-meal patterns were then adjusted for a family history of diabetes mellitus and only the post-OGTT insulin sensitivity (Matsuda index) remained significant (P=0.01). Therefore, this variable, the Matsuda index, was used in a multivariable analysis and the general linear model in order to incorporate individual changes over time, with adjustment for potential confounders (Table 5). The six-meal intervention resulted in a higher mean Matsuda index vs the three-meal intervention (P=0.01). Existence of a family history of diabetes mellitus was independently associated with a lower mean Matsuda Index, and the same association was reported for higher BMI (Table 5). When BMI was taken into account in the multivariate model (Table 5), it was positively associated with the Matsuda index, suggesting an independent association. Thus, in order to further investigate this finding, an interaction term (obesity status yes–no with Matsuda index) was explored but was not significant (P for interaction=0.19), and therefore no further analysis was performed to investigate the differences between normal weight and overweight or obese women. No significant differences were found before and after the three- and six-meal patterns for fasting glucose, incremental area under the curve for glucose, and incremental area under the curve for insulin, HbA1c, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides and hepatic enzymes (P>0.05 for all pairwise comparisons after Bonferroni correction).

Table 4 Anthropometric, physical activity and laboratory variables before and after the intervention with three- and six-meal pattern, n=40
Table 5 Linear mixed model of the Matsuda index across the 6-month intervention period


This randomised crossover 24-week study examined, for the first time to our knowledge, the effect of meal frequency on glucose and insulin levels in weight stable women with PCOS. A comparison of the effect of a three- vs a six-meal pattern of a weight maintenance diet revealed a beneficial effect of the six-meal pattern on post-OGTT insulin sensitivity.

Our results on the beneficial effect of six meals on insulin sensitivity are in agreement with data from seven other randomised clinical trials in samples of either healthy individuals, obese or with diagnosed type 2 diabetes. Of those, short-term trials lasting from 8 to 12 h up to 14 days showed that six eating occasions produced lower peak insulin responses and improved insulin sensitivity without significant differences in glucose responses.20, 21, 22, 23, 26, 27 Similarly, a longer-term trial lasting 3 months that tested the effect of isocaloric, energy-deficit diets consumed as six vs five meals in people with type 2 diabetes (N=66), found that six meals produced lower HbA1c, without significant differences on fasting and post-OGTT glucose and insulin concentrations.28 However, three trials with a duration from 24 weeks up to 1 year found no association between six meals vs three meals and glucose or insulin responses in healthy overweight or obese individuals following energy-deficit diets.38, 39, 40 Contrary to the aforementioned results, two trials, one short-term lasting 12 h and one long-term lasting 6 months, did not find a beneficial effect of six meals on insulin sensitivity in healthy, non-obese individuals and in people with type 2 diabetes on antidiabetic treatment.25, 41 However, the results of the aforementioned studies cannot be directly compared with ours because of the different study populations (healthy, obese, people with type 2 diabetes on medication) and the use of energy-deficit diets and weight loss, which are strong confounding factors affecting insulin sensitivity in the long term.

Women with PCOS show unique features and needs in regard to increased insulin resistance and hyperinsulinaemia, which in addition to increased androgen synthesis and disrupted folliculogenesis lie at the pathophysiological basis of PCOS. Previous studies suggest that women with PCOS show higher insulin resistance, but also increased insulin secretion, compared with age-, BMI- and insulin resistance-matched controls, which appears to be at least in part independent of the effect of obesity, with both lean and obese women with PCOS showing decreased insulin sensitivity.3, 42 However, obesity exacerbates insulin resistance in PCOS. Previous studies provide evidence that there is a post-binding defect in receptor signalling, possibly due to increased receptor and insulin receptor substrate-1 serine phosphorylation that selectively affects the metabolic pathways in both insulin target tissues and in the ovary.42, 43 Constitutive activation of serine kinases in the MAPK-ERK pathway may contribute to insulin resistance in muscle.42, 43 Insulin functions as a co-gonadotropin through its cognate receptor to modulate ovarian steroidogenesis; genetic disruption of insulin signalling in the brain has indicated that this pathway is important for ovulation and body weight regulation.42, 43 In addition, androgens contribute to insulin resistance in PCOS.42, 43 Therefore, any dietary modulation that has a beneficial effect on insulin sensitivity in the absence of weight loss is of great importance.

Although the mechanisms by which increased meal frequency may exert a beneficial effect on glucose levels are unclear, it has been proposed that this may be due to the nutrient load spreading, which produces lower postprandial insulin concentrations, reduces hunger and suppresses the inhibitory effects of free fatty acids on glucose uptake, leading to better glucose clearance from the circulation with a significant economy in insulin secretion.24, 44

In regard to hunger sensation, our results showed that the six-meal pattern, assessed in postprandial conditions, reduced subjective hunger and tended to reduce subjective desire to eat, compared with the three meal. These findings are in agreement with some studies45, 46 but not others.27, 47 As recently reviewed by Mattes,48 the impact of meal frequency on appetitive ratings, primarily based on preload studies and short-term feeding trials, remains controversial, with some studies associating greater meal frequency with lower fullness ratings and others lacking association. In general, increased meal frequency (>3meals per day) has been associated with prolonged gastric emptying, increased concentrations of gut hormones involved in satiety in response to food intake and improved insulin and glucose levels,49, 50 especially when examined in the acute postprandial period. However, a major confounder to the aforementioned findings may be the energy intake conditions (hypocaloric vs eucaloric). Briefly, it has been suggested that postprandial responses (for example, lower peaks in glucose, insulin and ghrelin)51 leading to changes in appetite are observed only in the eucaloric conditions, where the quantity of food provided at each eating episode is adequate to exert these effects. The eucaloric conditions of the present study may be at least partly responsible for the results regarding hunger and desire to eat.

Although our study has allowed insights into the effects of meal frequency on glucose metabolism in lean and overweight/obese women with PCOS, independently of weight loss, some limitations should be born in mind when considering the results. First of all, we have investigated fasting concentrations and indices of glucose metabolism derived from an OGTT and we do not have measurements before or after the meals during the two intervention periods. Although it would be interesting to have some data on the pre-and postprandial values (during each intervention period), the results could not be comparable between subjects (because of differences in the amount of carbohydrates consumed in one meal and the quantity of each meal, as each subject followed a weight maintenance diet). Second, the OGTT was performed in the morning after an overnight fast. Although the OGTT is a well-established method to obtain information on glucose and insulin levels using the same glucose load in every subject, the data obtained from an OGTT may differ from the overall daily metabolic profile, as generally suggested by the second-meal phenomenon. However, according to this phenomenon, the prior meal has an effect in decreasing the rise in blood glucose after a subsequent meal.52 In this point of view, a dietary pattern with a favourable effect on the first meal of the day could possibly be day-long beneficial. Further studies with day-long measurements are warranted to investigate whether the favourable effect of six- compared with the three-meal pattern on postprandial insulin sensitivity, investigated with an OGTT performed in the morning, is maintained during the rest of the day. Furthermore, dietary habits at baseline were assessed only through a food frequency questionnaire and not with 24 h recalls or food records, and therefore we could not assess the macronutrient composition of the participants’ diets before entering the study. Moreover, according to our data analysis, we believe that an inclusion of a wash-out period would not have influenced our results.

In conclusion, our study showed that the six- vs the three-meal pattern improved post-OGTT insulin sensitivity in women with PCOS under conditions of weight stability, which may be a beneficial non-pharmacologic adjunct to PCOS treatment. Moreover, during the six-meal pattern participants reported less subjective hunger and a trend for less desire to eat. Future studies are needed to investigate these beneficial effects of the six-meal pattern along with weight loss in overweight and obese women with PCOS.

Change history

  • 04 May 2016

    This article has been corrected since Advance Online Publication and a corrigendum is also printed in this issue


  1. 1

    Rotterdam ESHRE/ASRM-Sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 2004; 19: 41–47.

    Article  Google Scholar 

  2. 2

    Azziz R, Carmina E, Dewailly D, Diamanti-Kandarakis E, Escobar-Morreale HF, Futterweit W et al. The Androgen Excess and PCOS Society criteria for the polycystic ovary syndrome: the complete task force report. Fertil Steril 2009; 91: 456–488.

    Article  Google Scholar 

  3. 3

    Manco M, Castagneto-Gissey L, Arrighi E, Carnicelli A, Brufani C, Luciano R et al. Insulin dynamics in young women with polycystic ovary syndrome and normal glucose tolerance across categories of body mass index. PloS One 2014; 9: e92995.

    Article  Google Scholar 

  4. 4

    Trakakis E, Basios G, Peppa M, Simeonidis G, Labos G, Creatsa M et al. The prevalence of glucose metabolism abnormalities in Greek women with polycystic ovary syndrome. Gynecol Endocrinol 2012; 28: 867–870.

    CAS  Article  Google Scholar 

  5. 5

    Moran LJ, Pasquali R, Teede HJ, Hoeger KM, Norman RJ . Treatment of obesity in polycystic ovary syndrome: a position statement of the Androgen Excess and Polycystic Ovary Syndrome Society. Fertil Steril 2009; 92: 1966–1982.

    Article  Google Scholar 

  6. 6

    Moran LJ, Noakes M, Clifton PM, Tomlinson L, Galletly C, Norman RJ . Dietary composition in restoring reproductive and metabolic physiology in overweight women with polycystic ovary syndrome. J Clin Endocrinol Metab 2003; 88: 812–819.

    CAS  Article  Google Scholar 

  7. 7

    Goss AM, Chandler-Laney PC, Ovalle F, Goree LL, Azziz R, Desmond RA et al. Effects of a eucaloric reduced-carbohydrate diet on body composition and fat distribution in women with PCOS. Metabolism 2014; 63: 1257–1264.

    CAS  Article  Google Scholar 

  8. 8

    Marsh K, Brand-Miller J . The optimal diet for women with polycystic ovary syndrome? B J Nutr 2005; 94: 154–165.

    CAS  Article  Google Scholar 

  9. 9

    Moran LJ, Ko H, Misso M, Marsh K, Noakes M, Talbot M et al. Dietary composition in the treatment of polycystic ovary syndrome: a systematic review to inform evidence-based guidelines. J Acad Nutr Diet 2013; 113: 520–545.

    Article  Google Scholar 

  10. 10

    Salley KE, Wickham EP, Cheang KI, Essah PA, Karjane NW, Nestler JE . Glucose intolerance in polycystic ovary syndrome—a position statement of the Androgen Excess Society. J Clin Endocrinol Metab 2007; 92: 4546–4556.

    CAS  Article  Google Scholar 

  11. 11

    Moran LJ, Hutchison SK, Norman RJ, Teede HJ . Lifestyle changes in women with polycystic ovary syndrome. Cochrane Database Syst Rev 2011; (2): CD007506.

    Google Scholar 

  12. 12

    Bates GW, Legro RS . Longterm management of Polycystic Ovarian Syndrome (PCOS). Mol Cell Endocrinol 2013; 373: 91–97.

    CAS  Article  Google Scholar 

  13. 13

    Stankiewicz M, Norman R . Diagnosis and management of polycystic ovary syndrome: a practical guide. Drugs 2006; 66: 903–912.

    Article  Google Scholar 

  14. 14

    Holte J, Bergh T, Berne C, Lithell H . Serum lipoprotein lipid profile in women with the polycystic ovary syndrome: relation to anthropometric, endocrine and metabolic variables. Clin Endocrinol 1994; 41: 463–471.

    CAS  Article  Google Scholar 

  15. 15

    Jakubowicz D, Barnea M, Wainstein J, Froy O . Effects of caloric intake timing on insulin resistance and hyperandrogenism in lean women with polycystic ovary syndrome. Clin Sci 2013; 125: 423–432.

    CAS  Article  Google Scholar 

  16. 16

    Pearce KL, Noakes M, Keogh J, Clifton PM . Effect of carbohydrate distribution on postprandial glucose peaks with the use of continuous glucose monitoring in type 2 diabetes. Am J Clin Nutr 2008; 87: 638–644.

    CAS  Article  Google Scholar 

  17. 17

    Barr S, Reeves S, Sharp K, Jeanes YM . An isocaloric low glycemic index diet improves insulin sensitivity in women with polycystic ovary syndrome. J Acad Nutr Diet 2013; 113: 1523–1531.

    Article  Google Scholar 

  18. 18

    Arnold L, Mann JI, Ball MJ . Metabolic effects of alterations in meal frequency in type 2 diabetes. Diabetes Care 1997; 20: 1651–1654.

    CAS  Article  Google Scholar 

  19. 19

    Palmer MA, Capra S, Baines SK . Association between eating frequency, weight, and health. Nutr Rev 2009; 67: 379–390.

    Article  Google Scholar 

  20. 20

    Bertelsen J, Christiansen C, Thomsen C, Poulsen PL, Vestergaard S, Steinov A et al. Effect of meal frequency on blood glucose, insulin, and free fatty acids in NIDDM subjects. Diabetes care 1993; 16: 4–7.

    CAS  Article  Google Scholar 

  21. 21

    Farshchi HR, Taylor MA, Macdonald IA . Regular meal frequency creates more appropriate insulin sensitivity and lipid profiles compared with irregular meal frequency in healthy lean women. Eur J Clin Nutr 2004; 58: 1071–1077.

    CAS  Article  Google Scholar 

  22. 22

    Farshchi HR, Taylor MA, Macdonald IA . Beneficial metabolic effects of regular meal frequency on dietary thermogenesis, insulin sensitivity, and fasting lipid profiles in healthy obese women. Am J Clin Nutr 2005; 81: 16–24.

    CAS  Article  Google Scholar 

  23. 23

    Heden TD, Liu Y, Sims LJ, Whaley-Connell AT, Chockalingam A, Dellsperger KC et al. Meal frequency differentially alters postprandial triacylglycerol and insulin concentrations in obese women. Obesity 2013; 21: 123–129.

    CAS  Article  Google Scholar 

  24. 24

    Jenkins DJ, Wolever TM, Ocana AM, Vuksan V, Cunnane SC, Jenkins M et al. Metabolic effects of reducing rate of glucose ingestion by single bolus versus continuous sipping. Diabetes 1990; 39: 775–781.

    CAS  Article  Google Scholar 

  25. 25

    Kahleova H, Belinova L, Malinska H, Oliyarnyk O, Trnovska J, Skop V et al. Eating two larger meals a day (breakfast and lunch) is more effective than six smaller meals in a reduced-energy regimen for patients with type 2 diabetes: a randomised crossover study. Diabetologia 2014; 57: 1552–1560.

    Article  Google Scholar 

  26. 26

    Kanaley JA, Heden TD, Liu Y, Fairchild TJ . Alteration of postprandial glucose and insulin concentrations with meal frequency and composition. Br J Nutr 2014; 112: 1484–1493.

    CAS  Article  Google Scholar 

  27. 27

    Leidy HJ, Armstrong CL, Tang M, Mattes RD, Campbell WW . The influence of higher protein intake and greater eating frequency on appetite control in overweight and obese men. Obesity 2010; 18: 1725–1732.

    CAS  Article  Google Scholar 

  28. 28

    Salehi M, Kazemi A, Hasan Zadeh J . The effects of 6 isocaloric meals pattern on blood lipid profile, glucose, hemoglobin a1c, insulin and malondialdehyde in type 2 diabetic patients: a randomized clinical trial. Irani J Med Sci 2014; 39: 433–439.

    Google Scholar 

  29. 29

    House BT, Shearrer GE, Miller SJ, Pasch KE, Goran MI, Davis JN . Increased eating frequency linked to decreased obesity and improved metabolic outcomes. Int J Obes 2015; 39: 136–141.

    CAS  Article  Google Scholar 

  30. 30

    Carlson O, Martin B, Stote KS, Golden E, Maudsley S, Najjar SS et al. Impact of reduced meal frequency without caloric restriction on glucose regulation in healthy, normal-weight middle-aged men and women. Metabolism 2007; 56: 1729–1734.

    CAS  Article  Google Scholar 

  31. 31

    Nicklas TA, Demory-Luce D, Yang SJ, Baranowski T, Zakeri I, Berenson G . Children's food consumption patterns have changed over two decades (1973-1994): The Bogalusa heart study. J Am Diet Assoc 2004; 104: 1127–1140.

    Article  Google Scholar 

  32. 32

    Wheeler ML, Daly A, Evert A, Franz MJ, Geil P, Holzmeister LA et al Choose Your Foods: Exchange Lists for Diabetes, Sixth Edition, 2008: Description and Guidelines for Use J Am Diet Assoc 2008; 108: 883–888.

  33. 33

    Panagiotakos DB, Pitsavos C, Stefanadis C . Dietary patterns: a Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk. Nutr Metab Cardiovas Dis 2006; 16: 559–568.

    Article  Google Scholar 

  34. 34

    Kollia M, Gioxari A, Maraki M, Kavouras SA Development, validity and reliability of the Harokopio Physical Activity Questionnaire in Greek adults. Proceedings of the 8th Panhellenic Congress on Nutrition and Dietetics. Beta Medical Publishing: Greece, 2006.

  35. 35

    Friedewald WT, Levy RI, Fredrickson DS . Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.

    CAS  PubMed  Google Scholar 

  36. 36

    Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC . Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–419.

    CAS  Article  Google Scholar 

  37. 37

    Matsuda M, DeFronzo RA . Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999; 22: 1462–1470.

    CAS  Article  Google Scholar 

  38. 38

    Finkelstein B, Fryer BA . Meal frequency and weight reduction of young women. Am J Clin Nutr 1971; 24: 465–468.

    CAS  Article  Google Scholar 

  39. 39

    Poston WS, Haddock CK, Pinkston MM, Pace P, Karakoc ND, Reeves RS et al. Weight loss with meal replacement and meal replacement plus snacks: a randomized trial. Int J Obes 2005; 29: 1107–1114.

    CAS  Article  Google Scholar 

  40. 40

    Berteus Forslund H, Klingstrom S, Hagberg H, Londahl M, Torgerson JS, Lindroos AK . Should snacks be recommended in obesity treatment? A 1-year randomized clinical trial. Eur J Clin Nutr 2008; 62: 1308–1317.

    CAS  Article  Google Scholar 

  41. 41

    Holmstrup ME, Owens CM, Fairchild TJ, Kanaley JA . Effect of meal frequency on glucose and insulin excursions over the course of a day. Eur e-J Clin Nutr Metab 2010; 5: e277–e280.

    Article  Google Scholar 

  42. 42

    Diamanti-Kandarakis E, Dunaif A . Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocrine Rev 2012; 33: 981–1030.

    CAS  Article  Google Scholar 

  43. 43

    Randeva HS, Tan BK, Weickert MO, Lois K, Nestler JE, Sattar N et al. Cardiometabolic aspects of the polycystic ovary syndrome. Endocrine Rev 2012; 33: 812–841.

    CAS  Article  Google Scholar 

  44. 44

    Jenkins DJ . Carbohydrate tolerance and food frequency. Br J Nutr 1997; 77 (Suppl 1), S71–S81.

    CAS  Article  Google Scholar 

  45. 45

    Bachman JL, Raynor HA . Effects of manipulating eating frequency during a behavioral weight loss intervention: a pilot randomized controlled trial. Obesity 2012; 20: 985–992.

    CAS  Article  Google Scholar 

  46. 46

    McCrory MA, Campbell WW . Effects of eating frequency, snacking, and breakfast skipping on energy regulation: symposium overview. J Nutr 2011; 141: 144–147.

    CAS  Article  Google Scholar 

  47. 47

    Ohkawara K, Cornier MA, Kohrt WM, Melanson EL . Effects of increased meal frequency on fat oxidation and perceived hunger. Obesity 2013; 21: 336–343.

    Article  Google Scholar 

  48. 48

    Mattes R . Energy intake and obesity: ingestive frequency outweighs portion size. Physiol Behav 2014; 134: 110–118.

    CAS  Article  Google Scholar 

  49. 49

    Solomon TP, Chambers ES, Jeukendrup AE, Toogood AA, Blannin AK . The effect of feeding frequency on insulin and ghrelin responses in human subjects. Br J Nutr 2008; 100: 810–819.

    CAS  Article  Google Scholar 

  50. 50

    Capasso R, Izzo AA . Gastrointestinal regulation of food intake: general aspects and focus on anandamide and oleoylethanolamide. J Neuroendocrinol 2008; 20 (Suppl 1), 39–46.

    CAS  Article  Google Scholar 

  51. 51

    Leidy HJ, Campbell WW . The effect of eating frequency on appetite control and food intake: brief synopsis of controlled feeding studies. J Nutr 2011; 141: 154–157.

    Article  Google Scholar 

  52. 52

    Jovanovic A, Gerrard J, Taylor R . The second-meal phenomenon in type 2 diabetes. Diabetes Care 2009; 32: 1199–1201.

    Article  Google Scholar 

Download references


We are grateful to Vasso Fragaki, head nurse of the Diabetes Center of Attikon University Hospital, for conducting the participants’ OGTT procedure and to Maria-Assimina Gerama, clinical dietitian, for her valuable help during the conduction of the intervention. EP and MK conceptualised and designed the study; EP and IK collected the data and served as the dietitians of the team; IK conducted the biochemical analysis of blood samples; ET, DV and PM were responsible for patient screening, medical diagnosis and completed the medical examination; AZ and GD served as scientific counsellors to the project; EG conducted the statistical analysis; all authors contributed to the writing and editing of this manuscript according to their area of expertise and all authors approved it.

Author information



Corresponding author

Correspondence to E Papakonstantinou.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Papakonstantinou, E., Kechribari, I., Mitrou, P. et al. Effect of meal frequency on glucose and insulin levels in women with polycystic ovary syndrome: a randomised trial. Eur J Clin Nutr 70, 588–594 (2016).

Download citation

Further reading


Quick links