The major finding of this study is that consumption of 4 servings/d of low-fat dairy milk and yogurt products under free-living conditions for 6 months reduced fasting plasma insulin (9%) and improved insulin resistance (11%) in overweight and obese adults. Although the long-term clinical implications of these results are unclear, fasting insulin and HOMA-IR have been shown to be sensitive predictive markers of diabetes, ischemic stroke, and coronary heart disease risk in the general population as well as in at-risk subjects [28–32]. Previous data from the Verona Diabetes Complications Study suggest that every 1-unit increase in HOMA-IR is associated with an odds ratio of CVD incidence of 1.56 in Type II diabetics . This improvement in insulin resistance may be protective in this population of overweight and obese individuals.
Our results are supported by dairy intervention studies under both controlled feeding and free-living conditions. Compared to a low dairy diet (<0.5 servings/d), a 12-week study by Stancliffe et al. (2011) reported reductions in fasting insulin and insulin resistance in a similar population of overweight and obese subjects consuming dairy as milk and yogurt (3.5 serving/d) . Similar results were recently reported by Nikooyeh et al. (2011)  in type 2 diabetics consuming 250 mL/day of plain, vitamin D-fortified, or a vitamin D + calcium-fortified yogurt drink. Moreover, results from a recent meta-analysis and previous observational and prospective cohort studies support long-term dairy consumption in lowering the incidence of type II diabetes and metabolic syndrome [11, 35–37].
Alternatively, a recently published 12-month cross-over study examining the cardiometabolic health responses to increased dairy consumption by Crichton et al.  did not observe an insulin sensitizing response to HD vs. LD intake. This study was similar to the present study in terms of design (crossover), length (6 month), dairy servings (HD, 4 servings/d vs. LD, ≤ 1 servings/d), and study population (overweight and obese). However, while our study limited dairy products to low fat milk and yogurt, the study design by Crichton et al. was more inclusive by incorporating low fat flavored milk and yogurt, vanilla custard, and a limited amount (no more than 7 servings/week) of cheese (sliced, cottage, ricotta, cream), ice cream, butter, and margarine. We deliberately utilized milk and yogurt as treatments because these dairy products have reasonably long expiration dates, are easily portable, and can be readily incorporated into the diet. Depending on ingredient composition, formulation, and manufacturing process, dairy products have a broad range of textural and physiochemical properties. These differences greatly influence the bioavailability of dairy-derived nutrients and bioactive compounds and may ultimately affect the magnitude and direction of health responses following their consumption. Although our results suggest that consumption of dairy foods in the form of milk and yogurt may have beneficial cardiometabolic effects, it is difficult to speculate if similar responses would be observed in response to the consumption of other dairy foods. Ivey et al. (2011)  recently reported that, compared to milk and cheese, yogurt consumption decreased carotid artery intima-media thickness in elderly women. Thus, studies designed to specifically examine health responses to different whole dairy foods and isolated dairy food bioactive components are clearly warranted.
Numerous milk-derived bioactive compounds may contribute to the insulin-sensitizing effects of dairy including calcium, vitamin D, whole whey protein, and fractionated peptides [39, 40]. Vitamin D is well recognized in sensitizing insulin responses through multiple purported mechanisms which include regulation of insulin receptor expression and stimulation of insulin release by pancreatic β-cells [41, 42]. Although both milk and yogurt products used in the present study were vitamin D fortified as stipulated under Canadian federal law, we observed no difference in vitamin D intake between the LD and HD phases, suggesting that the insulin sensitizing response to increased dairy intake was independent of vitamin D consumption. However, as this study was designed as free-living and assessed nutrient intake using food records, it is subject to design limitations including poor dietary recall [43, 44]. Limitations of dietary recall during the 3-day food record may also be responsible for the tendency (p = 0.1) observed in calcium intake between the HD (1222 mg) and LD (1024 mg) groups.
Given that the present study was designed to examine the combined health response to low-fat milk and yogurt whole foods, we cannot partition the effects of the specific dairy foods or identify the particular bioactive component(s) that may be responsible for the protective effects against increased insulin resistance. The vast majority of studies examining the influence of dairy consumption on biomarkers of metabolic syndrome have been conducted with whole dairy foods with very few studies designed to partition the specific health effects of isolated dairy components. One notable exception is a recent study by Palacios et al. (2011)  who reported no change in body composition or serum lipids in obese subjects instructed to consume low calorie diets providing ~1200 mg/d calcium from dairy or calcium supplements.
Concerns have been raised that high dairy consumption may be associated with an increased risk of dyslipidemia and obesity . However, the results of this study lend further support to a growing clinical trial database suggesting that low fat dairy foods can be incorporated into diets without adversely raising blood lipids [38, 45–47]. Similarly, there is no clear consensus of the long-term effects of dairy consumption on body weight and body composition, with reports of both increased , decreased [49–51], and neutral [52, 53] body weight changes following dairy consumption. A recent systematic literature review by Kratz et al. examining 16 studies  found little support for the hypothesis that increased dairy fat or high fat dairy food consumption contributes to increased adiposity. Similarly, a meta-analysis from Chen et al.  using 29 RCT recently examined the effects of dairy consumption on body weight and body fat mass. Their results suggest that dairy consumption may not be associated with weight gain, but may have a modest effect in facilitating weight loss in energy restricted and short-term (<1 yr) intervention studies. Results from the present 6 month study demonstrate that, under free-living conditions and without energy restriction, long-term low-fat dairy consumption does not alter body weight, body composition or energy expenditure. These results further suggest that the insulin sensitizing effects of dairy consumption are not related to body fat loss or factors that modulate energy metabolism.
There are several limitations to this study. First, it proved difficult to sustain volunteer interest over such a long trial duration and only 23 subjects of the original 39 fully completed each of the study phases (41% dropout rate); however, we were able to retain our desired target of 20 subjects. Interestingly, the previously discussed study by Crichton et al.  with a similar design to ours also had a high drop-out rate (49%), which suggests that future studies involving high dairy consumption may benefit from a parallel arm design. Until a similar study is repeated with a larger study population, our results should be interpreted with caution. Second, although volunteers were given log books to record their daily dairy intake, these records were not reviewed by the study staff until the end of the study, at which point it was determined that they were incomplete. Therefore, the lack of a compliance evaluation and actual dairy intakes are a major limitation of the current study. Third, subjects were provided with dairy products during the HD phase but were not provided with dairy-free foods of comparable macronutrient composition during the LD phase. This could have resulted in differences in food intake and/or macronutrient balance between the two phases beyond the sensitivity of the dietary analysis conducted.