Nutrition, Metabolism & Cardiovascular Diseases (2014) 24, 554e562
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
High protein weight loss diets in obese subjects with type 2 diabetes mellitus E. Pedersen a,c, D.R. Jesudason a,c, P.M. Clifton a,b,c,* a
Commonwealth Scientific & Industrial Research Organisation (CSIRO), Australia University of South Australia, Australia c Clinical Research Excellence (CRE) in Nutritional Physiology, University of Adelaide, Discipline of Medicine, Australia b
Received 22 April 2013; received in revised form 7 November 2013; accepted 23 November 2013 Available online 2 December 2013
KEYWORDS Type 2 diabetes; Renal function; Albuminuria; Metabolic control; High protein diets; Weight loss
Abstract Background and aim: Diets where carbohydrate has been partially exchanged for protein have shown beneficial changes in persons with type 2 diabetes but no studies have enrolled people with albuminuria. We aim to determine if a high protein to carbohydrate ratio (HPD) in an energy reduced diet has a beneficial effect on metabolic control and cardiovascular risk factors without negatively affecting renal function. Method and results: Adult, overweight participants with type 2 diabetes, with albuminuria (30 e600 mg/24 h or an albumin-to-creatinine ratio of 3.0e60 mg/mmol), and estimated GFR of >40 ml/min/1.73 m2 were enrolled. Participants were randomized to an HPD or an SPD. Protein:fat:carbohydrate ratio was 30:30:40% of energy for the HPD and 20:30:50% for the SPD. Main outcomes were renal function, weight loss, blood pressure, serum lipids and glycaemic control. We recruited 76 volunteers and 45 (35 men and 10 women) finished. There were no overall changes in renal function at 12 months and no significant differences in weight loss between groups (9.7 2.9 kg and 6.6 1.4 kg HPD and SPD group respectively; p Z 0.32). Fasting blood glucose decreased significantly with no treatment effect. The decrease in HbA1c differed between treatments at 6 months (HPD 0.9 vs. SPD 0.3%; p Z 0.039) but not at 12 months. HDL increased significantly with no treatment effects. There were no changes in LDL or blood pressure overall but DBP was lower in the HPD group (p Z 0.024) at 12 months. Conclusion: Weight loss improved overall metabolic control in this group of well controlled participants with type 2 diabetes regardless of diet composition. ª 2013 Elsevier B.V. All rights reserved.
Introduction Short term diets replacing refined carbohydrate (CHO) with protein improve cardiovascular risk factors through increased weight loss and improved lipids. Epidemiological studies show an inverse association between dietary protein and hypertension [1]. In people with diabetes a
* Corresponding author. University of South Australia, Playford Building P5-16, University of South Australia, Adelaide 5000, Australia. Tel.: þ61 403197998. E-mail addresses: [email protected], peter.clifton@adelaide. edu.au (P.M. Clifton). 0939-4753/$ - see front matter ª 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.numecd.2013.11.003
high protein, lower carbohydrate diet has a beneficial effect on postprandial blood glucose (PBG) and HbA1c [2]. However, high saturated fat, high protein Atkins diets report increased LDL cholesterol [3]. Albuminuria is an early manifestation of diabetic nephropathy and is associated with risk factors including hypertension, impaired glycaemic control, and hyperlipidaemia [4]. It is also an early sign of microvascular disease in diabetes [5] and is associated with an increased risk of cardiovascular disease (CVD) [6]. Using albuminuria as inclusion criteria meant that this group was at much higher risk of cardiovascular events (equivalent to existing coronary disease) [5] and improving metabolic control
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Acronyms CHO BMI TC HDL LDL TG VLDL eGFR iGFR HPD SPD TE
carbohydrate body mass index total cholesterol high density lipoproteins low density lipoproteins triglycerides very low density lipoproteins estimated glomerular filtration rate isotope GFR high protein diet standard protein diet total energy
without harmful side effects is important. Improved metabolic control through improved diet and weight loss could prevent the progression of micro and macrovascular disease. High protein weight loss diets are commonly used, but concern has been raised about using this diet in diabetics due to the potential for a deleterious effect on renal function [7] as well as potential adverse effects on LDL cholesterol if saturated fat is not reduced. Our objective was to compare the effect on metabolic parameters of a high protein to carbohydrate ratio weight loss diet compared to a standard protein weight loss diet in volunteers with type 2 diabetes and an increased risk of CVD because of their albuminuria. We also wanted to investigate the relationship between changes in CGMS glucose values and markers of renal function, HbA1c, serum lipids and blood pressure to determine if using CGMS enhanced our understanding of the metabolic changes. Subjects and method Study design This was a parallel, randomized, 12 month dietary intervention study. Volunteers were blocked on sex, BMI and HbA1c before randomization. The randomization was performed by a trial manager not directly involved in the intervention and the analysis was conducted with the investigator blinded to the dietary allocation. Subjects Subjects were recruited via advertisement in local papers, radio and television. 76 overweight or obese (BMI 27 kg/m2), participants with type 2 diabetes aged 18e75 years, with albuminuria (30e600 mg/24 h or an albumin to creatinine ratio of 3.0e60.0 mg/mmol, with an estimated glomerular filtration rate (eGFR) of >40 ml/min/1.73 m2) were included. All potential volunteers underwent a physical examination by
555
AER albumin excretion rate alb/cr albumin to creatinine ratio FBG fasting blood glucose %T > 10 percentage blood glucose above 10 mmol/l Gmax peak blood glucose AUC area under the curve ANCOVAanalysis of covariance SEM standard error of the mean DXA dual-energy X-ray absorptiometry FM fat mass LBM lean body mass SBP systolic blood pressure DBP diastolic blood pressure CVD cardiovascular disease
an endocrinologist prior to randomization and medication was optimized to achieve good metabolic control (supplementary online data Appendix 2). Participants were excluded if they had impaired kidney function not due to diabetes. All experimental procedures were approved by the Human Ethics Committees of the Commonwealth Scientific and Industrial Research Organization. Participants provided written informed consent. ACTRN12608000045314. Diets A high protein diet (HPD) was compared to a standard protein diet (SPD). The two diets differed only in the proportion of protein and carbohydrate; total fat and saturated fat were similar in both diets. The planned protein:fat:carbohydrate ratio was 30:30:40% total energy (TE) for the HPD and 20:30:50 %TE for the SPD. The planned range of protein intake was 90e120 g/day in the HPD vs. 55e70 g/day in the SPD. Saturated fat intake was around 10%TE. All other nutrients were similar. Both diet regimes aimed at reducing body weight with energy content reduced to 6000 kJ. (For a full description see supplementary data Appendix 4). Outcome measures Height was measured at baseline. Weight was measured at all visits; with participants wearing light clothing and no shoes. Blood pressure was measured at baseline and at 4 month intervals. Renal function was measured directly by 99mTcediethylenetriamine-pentraacetic acid (99mTceDTPA) at baseline and at the end of the study. Fasting blood samples for fasting blood glucose (FBG), HbA1c, serum lipids (total cholesterol, HDL, LDL and triglyceride), and 24 h urine samples for albumin excretion rate (AER), albumin to creatinine ratio (alb/cr) and urine urea collected at baseline and at 4 month intervals. All samples were analysed at a commercial laboratory in Adelaide (IMVS).
556
E. Pedersen et al.
variables using the pre-test values as covariates. Data are presented as mean (SEM). Significance is assumed with a p-value less than 0.05 (see supplementary data Appendix 3).
As an additional measure of glucose control volunteers were asked to wear a Continuous Glucose Measuring System (CGMS), for at least 48 h at baseline, 4 and 12 months. Blood glucose was analysed as average 24 h blood glucose (24 h BG), peak blood glucose (Gmax), percentage time spent with a BG above 10 mmol/L (%T > 10) and area under the BG curve (AUC) computed as mmol/L per min over a 24 h period (mmol/L/min) (for full descriptions on method see supplementary data Appendix 1).
Results 76 volunteers commenced the study. 45 participants (35 men and 10 women) completed the one year dietary intervention (21 HPD and 24 SPD), and were included in the analysis. Metabolic control was similar between groups at baseline (Table 1). The groups differed only in the degree of AER, which was higher in the SPD compared to HPD.
Medication Four volunteers managed their diabetes with diet alone; all other volunteers were treated with oral BG lowering medication and/or insulin. All volunteers were treated with blood pressure lowering medication and all except two were treated with statins. Medication was adjusted to optimized glycaemic control, BP and serum lipids by an endocrinologist before the start of the study and monitored throughout the study. The volunteers own doctors also monitored them and made medication changes. All participants reported a moderate to low physical exercise level and they were asked to maintain this level throughout the study (see supplementary data Appendix 2 for changes in medication).
Dietary compliance Energy and macronutrient intake were similar at baseline. Compliance with the protein prescription was monitored by a daily food checklist and a food frequency questionnaire (at baseline, 4 and 12 months) and was also assessed by 24 h urine urea excretion. At baseline urea excretion did not differ significantly between groups (496 31 and 521 32 mmol/24 h in HPD and SPD respectively; p Z 0.41). At 12 months the UUE was not significantly different compared to baseline (p Z 0.13), however the adjusted urea excretion at 12 months was significantly different between groups (519 39, for the HPD and 456 25 for the SPD group; p Z 0.04) indicating compliance to the protein prescription. This was confirmed by self-reported diet data (supplement Appendix 5).
Statistical analysis Differences between randomized groups were tested at baseline using independent samples T-test. An analysis of covariance (ANCOVA) was used to compare the post-test
Table 1 Baseline characteristics. HPD
Gender (M/F) Age at study start Diabetes duration (years) Height (m) Weight (kg) BMI (kg/m2) FBG (mmol/L) HBA1C (%) Total Cholesterol (mmol/L) HDL cholesterol (mmol/L) LDL cholesterol (mmol/L) Triglycerides (mmol/L) SBP (mmHg) DBP (mmHg) iGFR eGFR AER (mg/min) Alb/Cr
15/6 59.4 2.2 12.4 2.5 1.7 0.02 108.1 5.0 36.7 8.3 0.4 7.5 0.2 3.9 0.2 1.1 0.0 1.8 0.2 2.4 0.3 127.2 3 74.7 1.6 113.7 9.2 99.0 5.9 41.3 10.6 6.1 1.4
SPD
20/4 62.4 1.7 7.9 1.0 1.7 0.02 104.7 3.8 35.4 8.3 0.4 7.1 0.2 3.6 0.1 1.1 0.1 1.7 0.1 2.0 0.2 125.5 2 70.1 1.93 89.6 6.0 88.6 6.0 81.8 13.2 11.4 2.5
p
0.35 0.1 0.91 0.59 0.44 0.94 0.19 0.12 0.23 0.31 0.31 0.64 0.13 0.03 0.22 0.02 0.08
ITT
ITT
HPD
SPD
21/10 58 12 1.7 104 36 7.9 7.5 3.8 1.0 1.8 2.3 127 74 109 100 50 4.0
23/10 61 8 1.7 104 35 8.1 7.2 3.6 1.1 1.6 2.0 128 72 92 90 79 7.0
Data are means (SEM). Significance assumed with p < 0.05. Participants were well matched at baseline, only iGFR and AER differed significantly in this group. Because of these significant differences all data was analysed adjusting for baseline values (ANCOVA). In the ITT group baseline values differed only for AER with a significantly lower AER in HPD (p Z 0.03) all other variables were similar. Abbreviations: HPD high protein diet and SPD standard protein diet. HDL high density lipoprotein, LDL low density lipoprotein, SBP systolic blood pressure, DBP diastolic blood pressure, iGFR isotope glomerular filtration rate, eGFR estimated GFR, AER albumin excretion rate, alb/cr albumin to creatinine ratio, ITT intension to treat.
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Weight loss Weight decreased significantly in both groups over time (p < 0.001) with no significant difference between groups (Supplementary figure S1). Total weight lost was 9.7 2.9 kg in the HPD group and 6.6 1.4 kg in the SPD group (p Z 0.32 between groups). The average percentage body weight lost was 8.7% (range þ5.5% to 34.7%) in the HPD and 6.3% (range þ4.0% to 17.3%) in the SPD. 24 participants lost more than 5% body weight (10 HPD and 14 SPD) and 15 volunteers lost more than 10% body weight (8 HPD and 7 SPD). Weight change between 6 and 12 months was insignificant (þ0.5 kg in HPD and 1.4 kg in SPD; p Z 0.29). A total of 7 volunteers gained an average of 3.1 kg overall (4.6 kg HPD vs. 2.5 kg SPD). Body composition was measured using dual-energy X-ray absorptiometry (DXA). The percentage fat mass (FM) decreased significantly with time (p < 0.01), with no difference between groups. The percentage lean body mass (LBM) increased significantly in both groups with time (p < 0.01) with no significant treatment effect (Fig. 1). Renal function iGFR decreased in the HPD by 9.3 4.8 and increased by 2.4 3.9 ml/min/1.73 m2 in the SPD (p Z 0.34 for time and p Z 0.38 between treatments). eGFR decreased in the
557
HPD from 98 to 97 ml/min/1.73 m2 while it decreased in the SPD from 91 to 90 ml/min/1.73 m2 (p Z 0.6 for time and p Z 0.9 for treatment). Albumin excretion rate decreased in the HPD by 8.9 8.6 and increased by 1.6 13.7 mg/min in the SPD (p Z 0.82 for time and p Z 0.13 between treatments). A decrease in hyperfiltration in the group with eGFR > 120 ml/min (n Z 12) of 15 ml/min (p Z 0.001) was seen. Participants with an eGFR < 120 ml/min (n Z 33) improved their eGFR by 4 ml/min (p Z 0.03). This improvement was directly related to weight loss (r Z 0.43, p Z 0.01) in this group but not in the participants with hyperfiltration. Dietary treatment group and protein intake were not related to the change in eGFR. Baseline eGFR remained a predictor of the change even after adjusting for weight loss (p Z 0.001). The full report on renal function has been published elsewhere [8]. Glycaemic control Fasting blood glucose decreased significantly and remained lower than baseline throughout the study period (1.0 0.3 and 1.5 0.5 mmol/L in HPD and SPD respectively at 12 months (p < 0.001 for time, NS for diet)). HbA1c decreased significantly with time (p < 0.01). The change in HbA1c was significantly different between groups (HPD 0.9 vs. SPD 0.3%; p Z 0.039) at 6 months but not at 12 months (HPD 0.4 vs. SPD 0.3%) (Fig. 2).
a
b
a
Figure 1 Body composition. Data are means (SEM). a) is weight change in kg, b) is % fat mass change and c) is % lean body mass change, HPD high protein diet (dark grey), SPD standard protein diet (light grey). Weight decreased significantly from baseline to 6 months and from baseline to 12 months (*p < 0.001). The change between 6 and 12 months was not significant. The change in weight was unrelated to diet (p Z 0.17 at 6 m and p Z 0.39 at 12 m). The %FM decreased significantly with time (p < 0.01), with no difference between groups (p Z 0.67). The LBM increased significantly in both groups with time (p < 0.01) with no significant treatment effect (p Z 0.86). Paired samples T-test was used to determine significance of change over time and treatment effect was analysed with values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). p Values are 6 and 12 months for body weight and 12 month values for %FM and %FFM.
558
E. Pedersen et al.
Figure 2 Absolute changes in glycaemic control over time. %T > 10 decreased significantly between baseline and 4 months (p Z 0.003) with no treatment effect (p Z 0.14). At 12 months the change was no longer significantly different from baseline (p Z 0.46 for time and p Z 0.29 between groups). Gmax decreased significantly between baseline and 4 months but not between baseline and 12 months (p Z 0.02 and p Z 0.56 respectively). AUC decreased significantly with time (p Z 0.002) with no treatment effect (p Z 0.08) at 4 months but at 12 months the difference from baseline was no longer significant (p Z 0.42). Mean 24 h BG decreased with time (p Z 0.001) with a significant treatment effect at 4 months (#p Z 0.02) but at 12 months the effect was no longer significant (p Z 0.27 for time and p Z 0.12 for treatment). Data are means (SEM). p Values are change over time using paired samples T-test and 4 and 12 months values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). Abbreviations: HPD high protein diet (dark grey), SPD standard protein diet (light grey), %T > 10 percentage of a 24 h period with a blood glucose above 10 mmol/L, Gmax the maximum blood glucose level over a 24 h period, AUC 24 h area under the blood glucose curve, 24 h BG mean blood glucose over 24 h.
CGMS A total of 39 volunteers (26 male and 13 female) were recruited to the CGMS sub-study. Baseline characteristics for all participants who completed both baseline and 12 month CGMS measurements were similar and not significantly different from the main group of volunteers. Results are reported for the 29 volunteers who completed all measurements. The 24 h average blood glucose decreased significantly with time (p Z 0.001) by 1.6 0.5 and 0.7 0.4 mmol/L in HPD and SPD respectively at 4 months. Between 4 and 12 months the 24 h BG increased and was not different from baseline in either group. Adjusting for baseline values revealed a treatment effect at 4 months (p Z 0.023) but not at 12 months (p Z 0.12). %T > 10 decreased significantly between baseline and 4 months (p Z 0.003). The HPD group decreased from 22% to 6% (relative change 73%) and the SPD group decreased from 19% to 10% (relative change 47%) with no significant treatment effect. At the end of the study %T > 10 was not
significantly different from baseline (HPD 16% and SPD 20%). At 4 months Gmax had decreased by 2.2 1.1 HPD and 1.1 1.0 mmol/L SPD (p Z NS adjusted for baseline). Gmax increased between 4 and 12 months and did not differ significantly from baseline. AUC decreased in HPD by 2155 and in SPD by 1004 (mmol/L/min) p Z 0.001 for time. AUC increased in both groups between 4 and 12 months and was no longer significantly different from baseline. Adjusting for baseline values there was no significant effect of treatment at 12 months. The change in AER was significantly correlated with the change in Gmax at 12 months (r Z 0.389; p Z 0.037) but after adjusting for baseline AER and weight loss (r2 Z 0.57; p Z 0.028); change in Gmax was not significant (p Z 0.14). All CGMS variables were strongly correlated with HbA1c at both 4 and 12 months (Gmax r Z 0.574, p Z 0.001; % T > 10 r Z 0.695, p < 0.00; AUC r Z 0.692, p < 0.00; mean 24 h BG r Z 0.651, p < 0.00 at 12 months). SBP was significantly correlated with 24 h BG (r Z 0.38; p Z 0.04) at 4 months but not at 12 months.
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Serum lipids At baseline all volunteers were well treated for dyslipidaemia. Total cholesterol (TC), low density lipoproteins (LDL), high density lipoproteins (HDL) and triglycerides (TG) were all within the recommended range (TC < 4.0, TG < 2.0, LDL < 2.0 and HDL >0.9; all measured as mmol/L) with no significant difference between groups at baseline. Total cholesterol did not change significantly between baseline and 12 months or between groups but showed significant changes in response to caloric restriction at 1 month (Supplementary data figure S2). TG decreased significantly with time (p < 0.001) with no treatment effect. HDL increased significantly in both groups with time (p Z 0.001) with no significant treatment effect. There were no significant changes in LDL. Adjusting for baseline values, there was a borderline significant difference in LDL cholesterol between the randomized groups at 12 months (2.02 in the HPD and 1.68 in the SPD group p Z 0.049). Statin use decreased in the HPD and increased in the SPD. When the analysis was confined to those who did not change medication no effect of diet was seen in LDL cholesterol. Non-HDL cholesterol differed significantly between groups at baseline (p Z 0.041). There was a significant change in non-HDL cholesterol with time (p < 0.005), but no treatment effect when adjusting for baseline (Fig. 3). Blood pressure The change in SBP and DBP between baseline and 12 months was not significant. However, when adjusting DBP for baseline values there was a significant difference between treatments, with lower DBP in the HPD at twelve months (Fig. 4, p Z 0.024).
559
Adding weight loss as covariate did not change the significance (p Z 0.03). Nine volunteers decreased their blood pressure medication. When the data was analysed excluding these participants both SBP (p Z 0.036) and DBP (p Z 0.054) were lower in HPD compared with SPD at 12 months.
Discussion This study showed a significant improvement in weight which reached a plateau after 6 months with no differences between treatments (as planned). The 13% difference in urinary urea at 12 months confirms that the planned dietary protein separation was maintained. It has been suggested that an HPD will have a negative effect on renal function in T2DM participants with renal impairment; but the evidence is scant and there is not sufficient evidence to recommend a protein intake lower than usual intake in T2DM with mild renal impairment nor is there evidence of harm with a higher intake [9,10]; however in this study renal function showed a trend towards a beneficial effect of weight loss in both those patients with hyperfiltration and those with slightly impaired GFR regardless of diet allocation. Some researchers have shown that weight loss is greater in subjects using a high protein diet [11,12] although others have reported similar weight loss in HPD and SPD when using controlled energy restricted diets [13e15]. By design we achieved matched weight loss in both groups so that we could tease out the effect of the different macronutrient prescription on metabolic parameters. Even a modest weight loss has been shown to improve metabolic control [16]. Müller et al. in their structured diabetes treatment and teaching program were
Figure 3 Changes in serum lipids (mmol/L). Data are means (SEM). p Values at 6 months (6) and 12 months (12) are changes between baseline and 6 months and changes between baseline and 12 months. Time effect was analysed using paired samples T-test (time *p < 0.001; yp < 0.005). Significant differences between treatments were analysed using analysis of covariance (ANCOVA) with baseline values as covariates. Only LDL was significantly different between treatments at 12 month (treatment #p Z 0.049). Abbreviations: HPD high protein diet and SPD standard protein diet. TC total cholesterol, HDL high density lipoprotein, LDL low density lipoprotein, Non-HDL-c non-HDL cholesterol, TG triglyceride. All values are in mmol/L.
560
E. Pedersen et al.
Figure 4 Changes in blood pressure over time. Data are means (SEM). p Values are 6 and 12 months values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). DBP was significantly lower with HPD at 12 months #p Z 0.024. Abbreviations: HPD high protein diet (dark grey), SPD standard protein diet (light grey), SBP systolic blood pressure, DBP diastolic blood pressure.
able to show continued benefit on HbA1c over 24 months with a modest weight loss of less than 2 kg [17]. While FBG decreased significantly from baseline and did not increase significantly between 6 and 12 months, the weight loss of more than 6 kg did not result in a sustained improvement in HbA1c. There are several explanations. Firstly, the initial optimizing of glycaemic control the volunteers went through in the run in period would have limited the scope for further reductions in HbA1c. Müller et al. showed a significant decrease in HbA1c only in the group with elevated HbA1c (>7.5%) [17]. Major improvements in HbA1c would demand a very high level of compliance to the planned energy prescription in which this study was not strictly maintained beyond 4 months. Secondly, some subjects reduced the number or dosage of diabetes medications as they lost weight. This will have attenuated the improvements seen in glycaemic control. Although we did not find any evidence of hypoglycaemia as control improved, those subjects who did not reduce medications may have been more cautious with dietary restriction so as to prevent hypoglycaemia. It has been suggested that postprandial blood glucose is an independent predictor of decline in GFR [18]. In this study we found a significant correlation between changes in albumin excretion and changes in Gmax but the linear regression did not show a significant independent contribution of Gmax. This may be explained by the very strict control of HbA1c and BP (7% and 125/74 mmHg respectively) in this group. It has been reported in a group of T2DM with HbA1c < 7.5% and a BP < 140/90 mmHg that postprandial blood glucose did not predict a change in GFR whereas the group with HbA1c > 7.5e8% had a marked decline in GFR with postprandial levels above 10 mmol/L [18]. Exchanging carbohydrate with protein or fat has been shown to be beneficial in lowering triglycerides, and increasing HDL [19,20] in clinical trials. However the
Nurses’ Health study found no association with low carbohydrate diet scores and total, HDL or LDL cholesterol. They did however report a beneficial effect on triglyceride levels [21]; but we failed to observe this in this study probably because the carbohydrate changes were too small at 12 months and the triglyceride levels were very normal. Higher HDL has been proposed to decrease the risk of incident chronic kidney disease in type 2 diabetes [22]. In the present study HDL increased throughout and was significantly higher than baseline in both groups at the end of the study. It was surprising that HDL increased from the start of the study and no further decrease was seen in response to the rather significant weight loss which is what is usually observed [23]. It has been proposed that non-HDL cholesterol is a better predictor of CVD risk as it may include a wider array of atherogenic particles such as intermediate density lipoproteins (IDL), very low density lipoproteins (VLDL), lipoprotein (a) and LDL and it can be calculated in a nonfasting sample [24,25]. In a population based study Lu et al. [25] found a strong association between non-HDL cholesterol and CVD in both men and women. The hazard ratio in the highest quartile of non-HDL cholesterol was 2.23 for men and 1.80 for women, higher than either LDL cholesterol alone or total cholesterol-to-HDL ratio. In our study, non-HDL cholesterol decreased significantly with time in both groups (p < 0.05) with weight loss alone with no effect of diet. The small increase in LDL cholesterol in the HPD was due to greater reductions in statin prescribing in this group. BP was not significantly different from baseline at the end of the study; however DBP was significantly lower in the HPD compared to the SPD at the end of the study after adjustment for baseline values. It has been shown that partially exchanging carbohydrate with protein may reduce blood pressure [26]. In a meta-analysis it was
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
shown that SBP decreased by 1.76 mmHg (95% CI: 2.33, 1.20) and DBP decreased by 1.15 mmHg (95% CI: 1.59, 0.71) in the HPD compared to the SPD diets [26]. The median energy intake from protein in these trials was 27% TE which is comparable to the 26.6% of total energy ingested from protein in our trial. In the weight stable state, as well as in weight reduction, protein rich diets have a well described effect on both SBP and DBP [16,27e29]. It is not clear why in this study protein and weight loss had no effect on SBP although the fact that subjects had excellent blood pressure control to begin with led to dose reductions in antihypertensive agents and when analysing the data excluding participants who reduced their antihypertensive medication both SBP and DBP were lower in HPD compared to SPD at 12 months. Body composition may be a better predictor of risk for cardiovascular complications compared to weight loss per se in T2DM. In this study we saw a significant decrease in total weight, %FM and significant increases in %LBM over time with no significant difference between diets. This is in contrast to results reported in a recent meta-analysis [30] of 24 short term RCTs comparing an energy restricted SPD to an energy restricted HPD which showed greater reduction in weight and FM together with a lower reduction in FFM resulting in decreased triglycerides and sustained resting energy metabolism. It could be speculated that the length of the study would have an effect and that longer duration would show declining effect with declining compliance. However the meta analysis showed no difference between studies of 12weeks however only one study in the meta analysis was long term (52 weeks) and was excluded from the analysis due to lack of data. Our study had several strengths. Subjects had their cardiovascular risk factors optimized before study start therefore any change from baseline was likely due to different dietary composition between groups rather than a dramatic response to energy restriction in poorly controlled subjects. We achieved similar weight loss in both groups and had biochemical evidence that we achieved differences in protein intake. The study was designed to assess the effect of the different diets on the full array of cardiovascular risk factors and we were able to look at both short and longer term effects of our interventions on glucose by measuring HbA1c as well as using CGMS and measuring body composition. Our study would have been strengthened by more subjects. Because of concerns about using high protein diets in renal dysfunction, we were only permitted to recruit subjects with early renal dysfunction and did not obtain as many subjects as we would have liked. Although the benefits were modest on an individual level, they are clinically significant at a population level of obese subjects with diabetes and renal dysfunction. Conclusion This study shows that weight loss is the driving force in the improvement in metabolic control in this group of
561
participants with type 2 diabetes. Overall the HPD had a beneficial effect on metabolic control comparable to the benefits achieved with the SPD in this weight loss study with no adverse effects demonstrated. Whilst the metabolic improvements were modest further research could focus on less well controlled subjects where weight loss and dietary interventions may show a greater benefit. Disclosure PMC is the co-author of The CSIRO Total Wellbeing Diet Book. No other conflict of interest reported. Acknowledgements Thank you to all the volunteers and clinic staff for their hard work. Special thanks to Julia Weaver, Lindy Lawson, Kathryn Bastiaans, Mark Mano, Candita Sullivan, Rosemary McArthur, Xenia Cleanthous, Anne McGuffin, Vanessa Russell, Pennie Taylor and Jennifer Keogh. Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.numecd.2013.11.003. References [1] Hu FB. Protein, body weight, and cardiovascular health. Am J Clin Nutr 2005;82:242Se7S. [2] Gannon MC, Nuttall FQ, Saeed A, Jordan K, Hoover H. An increase in dietary protein improves the blood glucose response in persons with type 2 diabetes. Am J Clin Nutr 2003;78:734e41. [3] Hu T, Mills KT, Yao L, Demanelis K, Eloustaz M, Yancy Jr WS, et al. Effects of low-carbohydrate diets versus low-fat diets on metabolic risk factors: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol 2012;176(Suppl. 7):S44e54. [4] Cerasola G, Cottone S, Mule G. The progressive pathway of microalbuminuria: from early marker of renal damage to strong cardiovascular risk predictor. J Hypertens 2010;28:2357e69. [5] Schmiedel O, Schroeter ML, Harvey JN. Microalbuminuria in type 2 diabetes indicates impaired microvascular vasomotion and perfusion. Am J Physiol Heart Circ Physiol 2007;293:H3424e31. [6] Freedman BI, Langefeld CD, Lohman KK, Bowden DW, Carr JJ, Rich SS, et al. Relationship between albuminuria and cardiovascular disease in type 2 diabetes. J Am Soc Nephrol 2005;16: 2156e61. [7] American Diabetes A, Bantle JP, Wylie-Rosett J, Albright AL, Apovian CM, Clark NG, et al. Nutrition recommendations and interventions for diabetes: a position statement of the American Diabetes Association. Diabetes Care 2008;31(Suppl. 1):S61e78. [8] Jesudason DR, Pedersen E, Clifton PM. Weight-loss diets in people with type 2 diabetes and renal disease: a randomized controlled trial of the effect of different dietary protein amounts. Am J Clin Nutr 2013;98(2):494e501. [9] Martin WF, Armstrong LE, Rodriguez NR. Dietary protein intake and renal function. Nutr Metab (Lond) 2005;2:25. [10] Mann JI, De Leeuw I, Hermansen K, Karamanos B, Karlstrom B, Katsilambros N, et al. Evidence-based nutritional approaches to the treatment and prevention of diabetes mellitus. Nutr Metab Cardiovasc Dis 2004;14:373e94. [11] Skov AR, Toubro S, Ronn B, Holm L, Astrup A. Randomized trial on protein vs carbohydrate in ad libitum fat reduced diet for the treatment of obesity. Int J Obes Relat Metab Disord 1999;23: 528e36. [12] Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA. A randomized trial comparing a very low carbohydrate diet and a calorie-restricted
562
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
E. Pedersen et al. low fat diet on body weight and cardiovascular risk factors in healthy women. J Clin Endocrinol Metab 2003;88:1617e23. Clifton PM, Bastiaans K, Keogh JB. High protein diets decrease total and abdominal fat and improve CVD risk profile in overweight and obese men and women with elevated triacylglycerol. Nutr Metab Cardiovasc Dis 2009;19:548e54. Johnston CS, Tjonn SL, Swan PD. High-protein, low-fat diets are effective for weight loss and favorably alter biomarkers in healthy adults. J Nutr 2004;134:586e91. Krebs JD, Elley CR, Parry-Strong A, Lunt H, Drury PL, Bell DA, et al. The Diabetes Excess Weight Loss (DEWL) Trial: a randomised controlled trial of high-protein versus high-carbohydrate diets over 2 years in type 2 diabetes. Diabetologia 2012;55:905e14. Wing RR, Lang W, Wadden TA, Safford M, Knowler WC, Bertoni AG, et al. Benefits of modest weight loss in improving cardiovascular risk factors in overweight and obese individuals with type 2 diabetes. Diabetes Care 2011;34:1481e6. Müller N, Stengel D, Kloos C, Ristow M, Wolf G, Muller UA. Improvement of HbA(1c) and stable weight loss 2 years after an outpatient treatment and teaching program for patients with type 2 diabetes without insulin therapy based on urine glucose selfmonitoring. Int J Gen Med 2012;5:241e7. Nosadini R, Tonolo G. Relationship between blood glucose control, pathogenesis and progression of diabetic nephropathy. J Am Soc Nephrol 2004;15(Suppl. 1):S1e5. Nordmann AJ, Nordmann A, Briel M, Keller U, Yancy Jr WS, Brehm BJ, et al. Effects of low-carbohydrate vs low-fat diets on weight loss and cardiovascular risk factors: a meta-analysis of randomized controlled trials. Arch Intern Med 2006;166:285e93. Lichtenstein AH. Thematic review series: patient-oriented research. Dietary fat, carbohydrate, and protein: effects on plasma lipoprotein patterns. J Lipid Res 2006;47:1661e7. Halton TL, Willett WC, Liu S, Manson JE, Albert CM, Rexrode K, et al. Low-carbohydrate-diet score and the risk of coronary heart disease in women. N Engl J Med 2006;355:1991e2002.
[22] Zoppini G, Targher G, Chonchol M, Perrone F, Lippi G, Muggeo M. Higher HDL cholesterol levels are associated with a lower incidence of chronic kidney disease in patients with type 2 diabetes. Nutr Metab Cardiovasc Dis 2009;19:580e6. [23] Noakes M, Clifton PM. Changes in plasma lipids and other cardiovascular risk factors during 3 energy-restricted diets differing in total fat and fatty acid composition. Am J Clin Nutr 2000;71: 706e12. [24] Saenger A. Cardiovascular risk assessment beyond LDL cholesterol: non-HDL cholesterol, LDL particle number, and apolipoprotein B. Communiquet Mayo Clinik; 2011. pp. 1e9. [25] Lu W, Resnick HE, Jablonski KA, Jones KL, Jain AK, Howard WJ, et al. Non-HDL cholesterol as a predictor of cardiovascular disease in type 2 diabetes: the strong heart study. Diabetes Care 2003;26: 16e23. [26] Rebholz CM, Friedman EE, Powers LJ, Arroyave WD, He J, Kelly TN. Dietary protein intake and blood pressure: a meta-analysis of randomized controlled trials. Am J Epidemiol 2012;176(Suppl. 7): S27e43. [27] Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med 1997;336:1117e24. [28] Appel LJ, Sacks FM, Carey VJ, Obarzanek E, Swain JF, Miller 3rd ER, et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. J Am Med Assoc 2005;294:2455e64. [29] Santesso N, Akl EA, Bianchi M, Mente A, Mustafa R, HeelsAnsdell D, et al. Effects of higher- versus lower-protein diets on health outcomes: a systematic review and meta-analysis. Eur J Clin Nutr 2012;66:780e8. [30] Wycherley TP, Moran LJ, Clifton PM, Noakes M, Brinkworth GD. Effects of energy-restricted high-protein, low-fat compared with standard-protein, low-fat diets: a meta-analysis of randomized controlled trials. Am J Clin Nutr 2012;96:1281e98.
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
High protein weight loss diets in obese subjects with type 2 diabetes mellitus E. Pedersen a,c, D.R. Jesudason a,c, P.M. Clifton a,b,c,* a
Commonwealth Scientific & Industrial Research Organisation (CSIRO), Australia University of South Australia, Australia c Clinical Research Excellence (CRE) in Nutritional Physiology, University of Adelaide, Discipline of Medicine, Australia b
Received 22 April 2013; received in revised form 7 November 2013; accepted 23 November 2013 Available online 2 December 2013
KEYWORDS Type 2 diabetes; Renal function; Albuminuria; Metabolic control; High protein diets; Weight loss
Abstract Background and aim: Diets where carbohydrate has been partially exchanged for protein have shown beneficial changes in persons with type 2 diabetes but no studies have enrolled people with albuminuria. We aim to determine if a high protein to carbohydrate ratio (HPD) in an energy reduced diet has a beneficial effect on metabolic control and cardiovascular risk factors without negatively affecting renal function. Method and results: Adult, overweight participants with type 2 diabetes, with albuminuria (30 e600 mg/24 h or an albumin-to-creatinine ratio of 3.0e60 mg/mmol), and estimated GFR of >40 ml/min/1.73 m2 were enrolled. Participants were randomized to an HPD or an SPD. Protein:fat:carbohydrate ratio was 30:30:40% of energy for the HPD and 20:30:50% for the SPD. Main outcomes were renal function, weight loss, blood pressure, serum lipids and glycaemic control. We recruited 76 volunteers and 45 (35 men and 10 women) finished. There were no overall changes in renal function at 12 months and no significant differences in weight loss between groups (9.7 2.9 kg and 6.6 1.4 kg HPD and SPD group respectively; p Z 0.32). Fasting blood glucose decreased significantly with no treatment effect. The decrease in HbA1c differed between treatments at 6 months (HPD 0.9 vs. SPD 0.3%; p Z 0.039) but not at 12 months. HDL increased significantly with no treatment effects. There were no changes in LDL or blood pressure overall but DBP was lower in the HPD group (p Z 0.024) at 12 months. Conclusion: Weight loss improved overall metabolic control in this group of well controlled participants with type 2 diabetes regardless of diet composition. ª 2013 Elsevier B.V. All rights reserved.
Introduction Short term diets replacing refined carbohydrate (CHO) with protein improve cardiovascular risk factors through increased weight loss and improved lipids. Epidemiological studies show an inverse association between dietary protein and hypertension [1]. In people with diabetes a
* Corresponding author. University of South Australia, Playford Building P5-16, University of South Australia, Adelaide 5000, Australia. Tel.: þ61 403197998. E-mail addresses: [email protected], peter.clifton@adelaide. edu.au (P.M. Clifton). 0939-4753/$ - see front matter ª 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.numecd.2013.11.003
high protein, lower carbohydrate diet has a beneficial effect on postprandial blood glucose (PBG) and HbA1c [2]. However, high saturated fat, high protein Atkins diets report increased LDL cholesterol [3]. Albuminuria is an early manifestation of diabetic nephropathy and is associated with risk factors including hypertension, impaired glycaemic control, and hyperlipidaemia [4]. It is also an early sign of microvascular disease in diabetes [5] and is associated with an increased risk of cardiovascular disease (CVD) [6]. Using albuminuria as inclusion criteria meant that this group was at much higher risk of cardiovascular events (equivalent to existing coronary disease) [5] and improving metabolic control
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Acronyms CHO BMI TC HDL LDL TG VLDL eGFR iGFR HPD SPD TE
carbohydrate body mass index total cholesterol high density lipoproteins low density lipoproteins triglycerides very low density lipoproteins estimated glomerular filtration rate isotope GFR high protein diet standard protein diet total energy
without harmful side effects is important. Improved metabolic control through improved diet and weight loss could prevent the progression of micro and macrovascular disease. High protein weight loss diets are commonly used, but concern has been raised about using this diet in diabetics due to the potential for a deleterious effect on renal function [7] as well as potential adverse effects on LDL cholesterol if saturated fat is not reduced. Our objective was to compare the effect on metabolic parameters of a high protein to carbohydrate ratio weight loss diet compared to a standard protein weight loss diet in volunteers with type 2 diabetes and an increased risk of CVD because of their albuminuria. We also wanted to investigate the relationship between changes in CGMS glucose values and markers of renal function, HbA1c, serum lipids and blood pressure to determine if using CGMS enhanced our understanding of the metabolic changes. Subjects and method Study design This was a parallel, randomized, 12 month dietary intervention study. Volunteers were blocked on sex, BMI and HbA1c before randomization. The randomization was performed by a trial manager not directly involved in the intervention and the analysis was conducted with the investigator blinded to the dietary allocation. Subjects Subjects were recruited via advertisement in local papers, radio and television. 76 overweight or obese (BMI 27 kg/m2), participants with type 2 diabetes aged 18e75 years, with albuminuria (30e600 mg/24 h or an albumin to creatinine ratio of 3.0e60.0 mg/mmol, with an estimated glomerular filtration rate (eGFR) of >40 ml/min/1.73 m2) were included. All potential volunteers underwent a physical examination by
555
AER albumin excretion rate alb/cr albumin to creatinine ratio FBG fasting blood glucose %T > 10 percentage blood glucose above 10 mmol/l Gmax peak blood glucose AUC area under the curve ANCOVAanalysis of covariance SEM standard error of the mean DXA dual-energy X-ray absorptiometry FM fat mass LBM lean body mass SBP systolic blood pressure DBP diastolic blood pressure CVD cardiovascular disease
an endocrinologist prior to randomization and medication was optimized to achieve good metabolic control (supplementary online data Appendix 2). Participants were excluded if they had impaired kidney function not due to diabetes. All experimental procedures were approved by the Human Ethics Committees of the Commonwealth Scientific and Industrial Research Organization. Participants provided written informed consent. ACTRN12608000045314. Diets A high protein diet (HPD) was compared to a standard protein diet (SPD). The two diets differed only in the proportion of protein and carbohydrate; total fat and saturated fat were similar in both diets. The planned protein:fat:carbohydrate ratio was 30:30:40% total energy (TE) for the HPD and 20:30:50 %TE for the SPD. The planned range of protein intake was 90e120 g/day in the HPD vs. 55e70 g/day in the SPD. Saturated fat intake was around 10%TE. All other nutrients were similar. Both diet regimes aimed at reducing body weight with energy content reduced to 6000 kJ. (For a full description see supplementary data Appendix 4). Outcome measures Height was measured at baseline. Weight was measured at all visits; with participants wearing light clothing and no shoes. Blood pressure was measured at baseline and at 4 month intervals. Renal function was measured directly by 99mTcediethylenetriamine-pentraacetic acid (99mTceDTPA) at baseline and at the end of the study. Fasting blood samples for fasting blood glucose (FBG), HbA1c, serum lipids (total cholesterol, HDL, LDL and triglyceride), and 24 h urine samples for albumin excretion rate (AER), albumin to creatinine ratio (alb/cr) and urine urea collected at baseline and at 4 month intervals. All samples were analysed at a commercial laboratory in Adelaide (IMVS).
556
E. Pedersen et al.
variables using the pre-test values as covariates. Data are presented as mean (SEM). Significance is assumed with a p-value less than 0.05 (see supplementary data Appendix 3).
As an additional measure of glucose control volunteers were asked to wear a Continuous Glucose Measuring System (CGMS), for at least 48 h at baseline, 4 and 12 months. Blood glucose was analysed as average 24 h blood glucose (24 h BG), peak blood glucose (Gmax), percentage time spent with a BG above 10 mmol/L (%T > 10) and area under the BG curve (AUC) computed as mmol/L per min over a 24 h period (mmol/L/min) (for full descriptions on method see supplementary data Appendix 1).
Results 76 volunteers commenced the study. 45 participants (35 men and 10 women) completed the one year dietary intervention (21 HPD and 24 SPD), and were included in the analysis. Metabolic control was similar between groups at baseline (Table 1). The groups differed only in the degree of AER, which was higher in the SPD compared to HPD.
Medication Four volunteers managed their diabetes with diet alone; all other volunteers were treated with oral BG lowering medication and/or insulin. All volunteers were treated with blood pressure lowering medication and all except two were treated with statins. Medication was adjusted to optimized glycaemic control, BP and serum lipids by an endocrinologist before the start of the study and monitored throughout the study. The volunteers own doctors also monitored them and made medication changes. All participants reported a moderate to low physical exercise level and they were asked to maintain this level throughout the study (see supplementary data Appendix 2 for changes in medication).
Dietary compliance Energy and macronutrient intake were similar at baseline. Compliance with the protein prescription was monitored by a daily food checklist and a food frequency questionnaire (at baseline, 4 and 12 months) and was also assessed by 24 h urine urea excretion. At baseline urea excretion did not differ significantly between groups (496 31 and 521 32 mmol/24 h in HPD and SPD respectively; p Z 0.41). At 12 months the UUE was not significantly different compared to baseline (p Z 0.13), however the adjusted urea excretion at 12 months was significantly different between groups (519 39, for the HPD and 456 25 for the SPD group; p Z 0.04) indicating compliance to the protein prescription. This was confirmed by self-reported diet data (supplement Appendix 5).
Statistical analysis Differences between randomized groups were tested at baseline using independent samples T-test. An analysis of covariance (ANCOVA) was used to compare the post-test
Table 1 Baseline characteristics. HPD
Gender (M/F) Age at study start Diabetes duration (years) Height (m) Weight (kg) BMI (kg/m2) FBG (mmol/L) HBA1C (%) Total Cholesterol (mmol/L) HDL cholesterol (mmol/L) LDL cholesterol (mmol/L) Triglycerides (mmol/L) SBP (mmHg) DBP (mmHg) iGFR eGFR AER (mg/min) Alb/Cr
15/6 59.4 2.2 12.4 2.5 1.7 0.02 108.1 5.0 36.7 8.3 0.4 7.5 0.2 3.9 0.2 1.1 0.0 1.8 0.2 2.4 0.3 127.2 3 74.7 1.6 113.7 9.2 99.0 5.9 41.3 10.6 6.1 1.4
SPD
20/4 62.4 1.7 7.9 1.0 1.7 0.02 104.7 3.8 35.4 8.3 0.4 7.1 0.2 3.6 0.1 1.1 0.1 1.7 0.1 2.0 0.2 125.5 2 70.1 1.93 89.6 6.0 88.6 6.0 81.8 13.2 11.4 2.5
p
0.35 0.1 0.91 0.59 0.44 0.94 0.19 0.12 0.23 0.31 0.31 0.64 0.13 0.03 0.22 0.02 0.08
ITT
ITT
HPD
SPD
21/10 58 12 1.7 104 36 7.9 7.5 3.8 1.0 1.8 2.3 127 74 109 100 50 4.0
23/10 61 8 1.7 104 35 8.1 7.2 3.6 1.1 1.6 2.0 128 72 92 90 79 7.0
Data are means (SEM). Significance assumed with p < 0.05. Participants were well matched at baseline, only iGFR and AER differed significantly in this group. Because of these significant differences all data was analysed adjusting for baseline values (ANCOVA). In the ITT group baseline values differed only for AER with a significantly lower AER in HPD (p Z 0.03) all other variables were similar. Abbreviations: HPD high protein diet and SPD standard protein diet. HDL high density lipoprotein, LDL low density lipoprotein, SBP systolic blood pressure, DBP diastolic blood pressure, iGFR isotope glomerular filtration rate, eGFR estimated GFR, AER albumin excretion rate, alb/cr albumin to creatinine ratio, ITT intension to treat.
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Weight loss Weight decreased significantly in both groups over time (p < 0.001) with no significant difference between groups (Supplementary figure S1). Total weight lost was 9.7 2.9 kg in the HPD group and 6.6 1.4 kg in the SPD group (p Z 0.32 between groups). The average percentage body weight lost was 8.7% (range þ5.5% to 34.7%) in the HPD and 6.3% (range þ4.0% to 17.3%) in the SPD. 24 participants lost more than 5% body weight (10 HPD and 14 SPD) and 15 volunteers lost more than 10% body weight (8 HPD and 7 SPD). Weight change between 6 and 12 months was insignificant (þ0.5 kg in HPD and 1.4 kg in SPD; p Z 0.29). A total of 7 volunteers gained an average of 3.1 kg overall (4.6 kg HPD vs. 2.5 kg SPD). Body composition was measured using dual-energy X-ray absorptiometry (DXA). The percentage fat mass (FM) decreased significantly with time (p < 0.01), with no difference between groups. The percentage lean body mass (LBM) increased significantly in both groups with time (p < 0.01) with no significant treatment effect (Fig. 1). Renal function iGFR decreased in the HPD by 9.3 4.8 and increased by 2.4 3.9 ml/min/1.73 m2 in the SPD (p Z 0.34 for time and p Z 0.38 between treatments). eGFR decreased in the
557
HPD from 98 to 97 ml/min/1.73 m2 while it decreased in the SPD from 91 to 90 ml/min/1.73 m2 (p Z 0.6 for time and p Z 0.9 for treatment). Albumin excretion rate decreased in the HPD by 8.9 8.6 and increased by 1.6 13.7 mg/min in the SPD (p Z 0.82 for time and p Z 0.13 between treatments). A decrease in hyperfiltration in the group with eGFR > 120 ml/min (n Z 12) of 15 ml/min (p Z 0.001) was seen. Participants with an eGFR < 120 ml/min (n Z 33) improved their eGFR by 4 ml/min (p Z 0.03). This improvement was directly related to weight loss (r Z 0.43, p Z 0.01) in this group but not in the participants with hyperfiltration. Dietary treatment group and protein intake were not related to the change in eGFR. Baseline eGFR remained a predictor of the change even after adjusting for weight loss (p Z 0.001). The full report on renal function has been published elsewhere [8]. Glycaemic control Fasting blood glucose decreased significantly and remained lower than baseline throughout the study period (1.0 0.3 and 1.5 0.5 mmol/L in HPD and SPD respectively at 12 months (p < 0.001 for time, NS for diet)). HbA1c decreased significantly with time (p < 0.01). The change in HbA1c was significantly different between groups (HPD 0.9 vs. SPD 0.3%; p Z 0.039) at 6 months but not at 12 months (HPD 0.4 vs. SPD 0.3%) (Fig. 2).
a
b
a
Figure 1 Body composition. Data are means (SEM). a) is weight change in kg, b) is % fat mass change and c) is % lean body mass change, HPD high protein diet (dark grey), SPD standard protein diet (light grey). Weight decreased significantly from baseline to 6 months and from baseline to 12 months (*p < 0.001). The change between 6 and 12 months was not significant. The change in weight was unrelated to diet (p Z 0.17 at 6 m and p Z 0.39 at 12 m). The %FM decreased significantly with time (p < 0.01), with no difference between groups (p Z 0.67). The LBM increased significantly in both groups with time (p < 0.01) with no significant treatment effect (p Z 0.86). Paired samples T-test was used to determine significance of change over time and treatment effect was analysed with values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). p Values are 6 and 12 months for body weight and 12 month values for %FM and %FFM.
558
E. Pedersen et al.
Figure 2 Absolute changes in glycaemic control over time. %T > 10 decreased significantly between baseline and 4 months (p Z 0.003) with no treatment effect (p Z 0.14). At 12 months the change was no longer significantly different from baseline (p Z 0.46 for time and p Z 0.29 between groups). Gmax decreased significantly between baseline and 4 months but not between baseline and 12 months (p Z 0.02 and p Z 0.56 respectively). AUC decreased significantly with time (p Z 0.002) with no treatment effect (p Z 0.08) at 4 months but at 12 months the difference from baseline was no longer significant (p Z 0.42). Mean 24 h BG decreased with time (p Z 0.001) with a significant treatment effect at 4 months (#p Z 0.02) but at 12 months the effect was no longer significant (p Z 0.27 for time and p Z 0.12 for treatment). Data are means (SEM). p Values are change over time using paired samples T-test and 4 and 12 months values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). Abbreviations: HPD high protein diet (dark grey), SPD standard protein diet (light grey), %T > 10 percentage of a 24 h period with a blood glucose above 10 mmol/L, Gmax the maximum blood glucose level over a 24 h period, AUC 24 h area under the blood glucose curve, 24 h BG mean blood glucose over 24 h.
CGMS A total of 39 volunteers (26 male and 13 female) were recruited to the CGMS sub-study. Baseline characteristics for all participants who completed both baseline and 12 month CGMS measurements were similar and not significantly different from the main group of volunteers. Results are reported for the 29 volunteers who completed all measurements. The 24 h average blood glucose decreased significantly with time (p Z 0.001) by 1.6 0.5 and 0.7 0.4 mmol/L in HPD and SPD respectively at 4 months. Between 4 and 12 months the 24 h BG increased and was not different from baseline in either group. Adjusting for baseline values revealed a treatment effect at 4 months (p Z 0.023) but not at 12 months (p Z 0.12). %T > 10 decreased significantly between baseline and 4 months (p Z 0.003). The HPD group decreased from 22% to 6% (relative change 73%) and the SPD group decreased from 19% to 10% (relative change 47%) with no significant treatment effect. At the end of the study %T > 10 was not
significantly different from baseline (HPD 16% and SPD 20%). At 4 months Gmax had decreased by 2.2 1.1 HPD and 1.1 1.0 mmol/L SPD (p Z NS adjusted for baseline). Gmax increased between 4 and 12 months and did not differ significantly from baseline. AUC decreased in HPD by 2155 and in SPD by 1004 (mmol/L/min) p Z 0.001 for time. AUC increased in both groups between 4 and 12 months and was no longer significantly different from baseline. Adjusting for baseline values there was no significant effect of treatment at 12 months. The change in AER was significantly correlated with the change in Gmax at 12 months (r Z 0.389; p Z 0.037) but after adjusting for baseline AER and weight loss (r2 Z 0.57; p Z 0.028); change in Gmax was not significant (p Z 0.14). All CGMS variables were strongly correlated with HbA1c at both 4 and 12 months (Gmax r Z 0.574, p Z 0.001; % T > 10 r Z 0.695, p < 0.00; AUC r Z 0.692, p < 0.00; mean 24 h BG r Z 0.651, p < 0.00 at 12 months). SBP was significantly correlated with 24 h BG (r Z 0.38; p Z 0.04) at 4 months but not at 12 months.
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
Serum lipids At baseline all volunteers were well treated for dyslipidaemia. Total cholesterol (TC), low density lipoproteins (LDL), high density lipoproteins (HDL) and triglycerides (TG) were all within the recommended range (TC < 4.0, TG < 2.0, LDL < 2.0 and HDL >0.9; all measured as mmol/L) with no significant difference between groups at baseline. Total cholesterol did not change significantly between baseline and 12 months or between groups but showed significant changes in response to caloric restriction at 1 month (Supplementary data figure S2). TG decreased significantly with time (p < 0.001) with no treatment effect. HDL increased significantly in both groups with time (p Z 0.001) with no significant treatment effect. There were no significant changes in LDL. Adjusting for baseline values, there was a borderline significant difference in LDL cholesterol between the randomized groups at 12 months (2.02 in the HPD and 1.68 in the SPD group p Z 0.049). Statin use decreased in the HPD and increased in the SPD. When the analysis was confined to those who did not change medication no effect of diet was seen in LDL cholesterol. Non-HDL cholesterol differed significantly between groups at baseline (p Z 0.041). There was a significant change in non-HDL cholesterol with time (p < 0.005), but no treatment effect when adjusting for baseline (Fig. 3). Blood pressure The change in SBP and DBP between baseline and 12 months was not significant. However, when adjusting DBP for baseline values there was a significant difference between treatments, with lower DBP in the HPD at twelve months (Fig. 4, p Z 0.024).
559
Adding weight loss as covariate did not change the significance (p Z 0.03). Nine volunteers decreased their blood pressure medication. When the data was analysed excluding these participants both SBP (p Z 0.036) and DBP (p Z 0.054) were lower in HPD compared with SPD at 12 months.
Discussion This study showed a significant improvement in weight which reached a plateau after 6 months with no differences between treatments (as planned). The 13% difference in urinary urea at 12 months confirms that the planned dietary protein separation was maintained. It has been suggested that an HPD will have a negative effect on renal function in T2DM participants with renal impairment; but the evidence is scant and there is not sufficient evidence to recommend a protein intake lower than usual intake in T2DM with mild renal impairment nor is there evidence of harm with a higher intake [9,10]; however in this study renal function showed a trend towards a beneficial effect of weight loss in both those patients with hyperfiltration and those with slightly impaired GFR regardless of diet allocation. Some researchers have shown that weight loss is greater in subjects using a high protein diet [11,12] although others have reported similar weight loss in HPD and SPD when using controlled energy restricted diets [13e15]. By design we achieved matched weight loss in both groups so that we could tease out the effect of the different macronutrient prescription on metabolic parameters. Even a modest weight loss has been shown to improve metabolic control [16]. Müller et al. in their structured diabetes treatment and teaching program were
Figure 3 Changes in serum lipids (mmol/L). Data are means (SEM). p Values at 6 months (6) and 12 months (12) are changes between baseline and 6 months and changes between baseline and 12 months. Time effect was analysed using paired samples T-test (time *p < 0.001; yp < 0.005). Significant differences between treatments were analysed using analysis of covariance (ANCOVA) with baseline values as covariates. Only LDL was significantly different between treatments at 12 month (treatment #p Z 0.049). Abbreviations: HPD high protein diet and SPD standard protein diet. TC total cholesterol, HDL high density lipoprotein, LDL low density lipoprotein, Non-HDL-c non-HDL cholesterol, TG triglyceride. All values are in mmol/L.
560
E. Pedersen et al.
Figure 4 Changes in blood pressure over time. Data are means (SEM). p Values are 6 and 12 months values adjusted for baseline using a repeated measures analysis of covariance (ANCOVA). DBP was significantly lower with HPD at 12 months #p Z 0.024. Abbreviations: HPD high protein diet (dark grey), SPD standard protein diet (light grey), SBP systolic blood pressure, DBP diastolic blood pressure.
able to show continued benefit on HbA1c over 24 months with a modest weight loss of less than 2 kg [17]. While FBG decreased significantly from baseline and did not increase significantly between 6 and 12 months, the weight loss of more than 6 kg did not result in a sustained improvement in HbA1c. There are several explanations. Firstly, the initial optimizing of glycaemic control the volunteers went through in the run in period would have limited the scope for further reductions in HbA1c. Müller et al. showed a significant decrease in HbA1c only in the group with elevated HbA1c (>7.5%) [17]. Major improvements in HbA1c would demand a very high level of compliance to the planned energy prescription in which this study was not strictly maintained beyond 4 months. Secondly, some subjects reduced the number or dosage of diabetes medications as they lost weight. This will have attenuated the improvements seen in glycaemic control. Although we did not find any evidence of hypoglycaemia as control improved, those subjects who did not reduce medications may have been more cautious with dietary restriction so as to prevent hypoglycaemia. It has been suggested that postprandial blood glucose is an independent predictor of decline in GFR [18]. In this study we found a significant correlation between changes in albumin excretion and changes in Gmax but the linear regression did not show a significant independent contribution of Gmax. This may be explained by the very strict control of HbA1c and BP (7% and 125/74 mmHg respectively) in this group. It has been reported in a group of T2DM with HbA1c < 7.5% and a BP < 140/90 mmHg that postprandial blood glucose did not predict a change in GFR whereas the group with HbA1c > 7.5e8% had a marked decline in GFR with postprandial levels above 10 mmol/L [18]. Exchanging carbohydrate with protein or fat has been shown to be beneficial in lowering triglycerides, and increasing HDL [19,20] in clinical trials. However the
Nurses’ Health study found no association with low carbohydrate diet scores and total, HDL or LDL cholesterol. They did however report a beneficial effect on triglyceride levels [21]; but we failed to observe this in this study probably because the carbohydrate changes were too small at 12 months and the triglyceride levels were very normal. Higher HDL has been proposed to decrease the risk of incident chronic kidney disease in type 2 diabetes [22]. In the present study HDL increased throughout and was significantly higher than baseline in both groups at the end of the study. It was surprising that HDL increased from the start of the study and no further decrease was seen in response to the rather significant weight loss which is what is usually observed [23]. It has been proposed that non-HDL cholesterol is a better predictor of CVD risk as it may include a wider array of atherogenic particles such as intermediate density lipoproteins (IDL), very low density lipoproteins (VLDL), lipoprotein (a) and LDL and it can be calculated in a nonfasting sample [24,25]. In a population based study Lu et al. [25] found a strong association between non-HDL cholesterol and CVD in both men and women. The hazard ratio in the highest quartile of non-HDL cholesterol was 2.23 for men and 1.80 for women, higher than either LDL cholesterol alone or total cholesterol-to-HDL ratio. In our study, non-HDL cholesterol decreased significantly with time in both groups (p < 0.05) with weight loss alone with no effect of diet. The small increase in LDL cholesterol in the HPD was due to greater reductions in statin prescribing in this group. BP was not significantly different from baseline at the end of the study; however DBP was significantly lower in the HPD compared to the SPD at the end of the study after adjustment for baseline values. It has been shown that partially exchanging carbohydrate with protein may reduce blood pressure [26]. In a meta-analysis it was
High protein weight loss diets in obese subjects with type 2 diabetes mellitus
shown that SBP decreased by 1.76 mmHg (95% CI: 2.33, 1.20) and DBP decreased by 1.15 mmHg (95% CI: 1.59, 0.71) in the HPD compared to the SPD diets [26]. The median energy intake from protein in these trials was 27% TE which is comparable to the 26.6% of total energy ingested from protein in our trial. In the weight stable state, as well as in weight reduction, protein rich diets have a well described effect on both SBP and DBP [16,27e29]. It is not clear why in this study protein and weight loss had no effect on SBP although the fact that subjects had excellent blood pressure control to begin with led to dose reductions in antihypertensive agents and when analysing the data excluding participants who reduced their antihypertensive medication both SBP and DBP were lower in HPD compared to SPD at 12 months. Body composition may be a better predictor of risk for cardiovascular complications compared to weight loss per se in T2DM. In this study we saw a significant decrease in total weight, %FM and significant increases in %LBM over time with no significant difference between diets. This is in contrast to results reported in a recent meta-analysis [30] of 24 short term RCTs comparing an energy restricted SPD to an energy restricted HPD which showed greater reduction in weight and FM together with a lower reduction in FFM resulting in decreased triglycerides and sustained resting energy metabolism. It could be speculated that the length of the study would have an effect and that longer duration would show declining effect with declining compliance. However the meta analysis showed no difference between studies of 12weeks however only one study in the meta analysis was long term (52 weeks) and was excluded from the analysis due to lack of data. Our study had several strengths. Subjects had their cardiovascular risk factors optimized before study start therefore any change from baseline was likely due to different dietary composition between groups rather than a dramatic response to energy restriction in poorly controlled subjects. We achieved similar weight loss in both groups and had biochemical evidence that we achieved differences in protein intake. The study was designed to assess the effect of the different diets on the full array of cardiovascular risk factors and we were able to look at both short and longer term effects of our interventions on glucose by measuring HbA1c as well as using CGMS and measuring body composition. Our study would have been strengthened by more subjects. Because of concerns about using high protein diets in renal dysfunction, we were only permitted to recruit subjects with early renal dysfunction and did not obtain as many subjects as we would have liked. Although the benefits were modest on an individual level, they are clinically significant at a population level of obese subjects with diabetes and renal dysfunction. Conclusion This study shows that weight loss is the driving force in the improvement in metabolic control in this group of
561
participants with type 2 diabetes. Overall the HPD had a beneficial effect on metabolic control comparable to the benefits achieved with the SPD in this weight loss study with no adverse effects demonstrated. Whilst the metabolic improvements were modest further research could focus on less well controlled subjects where weight loss and dietary interventions may show a greater benefit. Disclosure PMC is the co-author of The CSIRO Total Wellbeing Diet Book. No other conflict of interest reported. Acknowledgements Thank you to all the volunteers and clinic staff for their hard work. Special thanks to Julia Weaver, Lindy Lawson, Kathryn Bastiaans, Mark Mano, Candita Sullivan, Rosemary McArthur, Xenia Cleanthous, Anne McGuffin, Vanessa Russell, Pennie Taylor and Jennifer Keogh. Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.numecd.2013.11.003. References [1] Hu FB. Protein, body weight, and cardiovascular health. Am J Clin Nutr 2005;82:242Se7S. [2] Gannon MC, Nuttall FQ, Saeed A, Jordan K, Hoover H. An increase in dietary protein improves the blood glucose response in persons with type 2 diabetes. Am J Clin Nutr 2003;78:734e41. [3] Hu T, Mills KT, Yao L, Demanelis K, Eloustaz M, Yancy Jr WS, et al. Effects of low-carbohydrate diets versus low-fat diets on metabolic risk factors: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol 2012;176(Suppl. 7):S44e54. [4] Cerasola G, Cottone S, Mule G. The progressive pathway of microalbuminuria: from early marker of renal damage to strong cardiovascular risk predictor. J Hypertens 2010;28:2357e69. [5] Schmiedel O, Schroeter ML, Harvey JN. Microalbuminuria in type 2 diabetes indicates impaired microvascular vasomotion and perfusion. Am J Physiol Heart Circ Physiol 2007;293:H3424e31. [6] Freedman BI, Langefeld CD, Lohman KK, Bowden DW, Carr JJ, Rich SS, et al. Relationship between albuminuria and cardiovascular disease in type 2 diabetes. J Am Soc Nephrol 2005;16: 2156e61. [7] American Diabetes A, Bantle JP, Wylie-Rosett J, Albright AL, Apovian CM, Clark NG, et al. Nutrition recommendations and interventions for diabetes: a position statement of the American Diabetes Association. Diabetes Care 2008;31(Suppl. 1):S61e78. [8] Jesudason DR, Pedersen E, Clifton PM. Weight-loss diets in people with type 2 diabetes and renal disease: a randomized controlled trial of the effect of different dietary protein amounts. Am J Clin Nutr 2013;98(2):494e501. [9] Martin WF, Armstrong LE, Rodriguez NR. Dietary protein intake and renal function. Nutr Metab (Lond) 2005;2:25. [10] Mann JI, De Leeuw I, Hermansen K, Karamanos B, Karlstrom B, Katsilambros N, et al. Evidence-based nutritional approaches to the treatment and prevention of diabetes mellitus. Nutr Metab Cardiovasc Dis 2004;14:373e94. [11] Skov AR, Toubro S, Ronn B, Holm L, Astrup A. Randomized trial on protein vs carbohydrate in ad libitum fat reduced diet for the treatment of obesity. Int J Obes Relat Metab Disord 1999;23: 528e36. [12] Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA. A randomized trial comparing a very low carbohydrate diet and a calorie-restricted
562
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
E. Pedersen et al. low fat diet on body weight and cardiovascular risk factors in healthy women. J Clin Endocrinol Metab 2003;88:1617e23. Clifton PM, Bastiaans K, Keogh JB. High protein diets decrease total and abdominal fat and improve CVD risk profile in overweight and obese men and women with elevated triacylglycerol. Nutr Metab Cardiovasc Dis 2009;19:548e54. Johnston CS, Tjonn SL, Swan PD. High-protein, low-fat diets are effective for weight loss and favorably alter biomarkers in healthy adults. J Nutr 2004;134:586e91. Krebs JD, Elley CR, Parry-Strong A, Lunt H, Drury PL, Bell DA, et al. The Diabetes Excess Weight Loss (DEWL) Trial: a randomised controlled trial of high-protein versus high-carbohydrate diets over 2 years in type 2 diabetes. Diabetologia 2012;55:905e14. Wing RR, Lang W, Wadden TA, Safford M, Knowler WC, Bertoni AG, et al. Benefits of modest weight loss in improving cardiovascular risk factors in overweight and obese individuals with type 2 diabetes. Diabetes Care 2011;34:1481e6. Müller N, Stengel D, Kloos C, Ristow M, Wolf G, Muller UA. Improvement of HbA(1c) and stable weight loss 2 years after an outpatient treatment and teaching program for patients with type 2 diabetes without insulin therapy based on urine glucose selfmonitoring. Int J Gen Med 2012;5:241e7. Nosadini R, Tonolo G. Relationship between blood glucose control, pathogenesis and progression of diabetic nephropathy. J Am Soc Nephrol 2004;15(Suppl. 1):S1e5. Nordmann AJ, Nordmann A, Briel M, Keller U, Yancy Jr WS, Brehm BJ, et al. Effects of low-carbohydrate vs low-fat diets on weight loss and cardiovascular risk factors: a meta-analysis of randomized controlled trials. Arch Intern Med 2006;166:285e93. Lichtenstein AH. Thematic review series: patient-oriented research. Dietary fat, carbohydrate, and protein: effects on plasma lipoprotein patterns. J Lipid Res 2006;47:1661e7. Halton TL, Willett WC, Liu S, Manson JE, Albert CM, Rexrode K, et al. Low-carbohydrate-diet score and the risk of coronary heart disease in women. N Engl J Med 2006;355:1991e2002.
[22] Zoppini G, Targher G, Chonchol M, Perrone F, Lippi G, Muggeo M. Higher HDL cholesterol levels are associated with a lower incidence of chronic kidney disease in patients with type 2 diabetes. Nutr Metab Cardiovasc Dis 2009;19:580e6. [23] Noakes M, Clifton PM. Changes in plasma lipids and other cardiovascular risk factors during 3 energy-restricted diets differing in total fat and fatty acid composition. Am J Clin Nutr 2000;71: 706e12. [24] Saenger A. Cardiovascular risk assessment beyond LDL cholesterol: non-HDL cholesterol, LDL particle number, and apolipoprotein B. Communiquet Mayo Clinik; 2011. pp. 1e9. [25] Lu W, Resnick HE, Jablonski KA, Jones KL, Jain AK, Howard WJ, et al. Non-HDL cholesterol as a predictor of cardiovascular disease in type 2 diabetes: the strong heart study. Diabetes Care 2003;26: 16e23. [26] Rebholz CM, Friedman EE, Powers LJ, Arroyave WD, He J, Kelly TN. Dietary protein intake and blood pressure: a meta-analysis of randomized controlled trials. Am J Epidemiol 2012;176(Suppl. 7): S27e43. [27] Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med 1997;336:1117e24. [28] Appel LJ, Sacks FM, Carey VJ, Obarzanek E, Swain JF, Miller 3rd ER, et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. J Am Med Assoc 2005;294:2455e64. [29] Santesso N, Akl EA, Bianchi M, Mente A, Mustafa R, HeelsAnsdell D, et al. Effects of higher- versus lower-protein diets on health outcomes: a systematic review and meta-analysis. Eur J Clin Nutr 2012;66:780e8. [30] Wycherley TP, Moran LJ, Clifton PM, Noakes M, Brinkworth GD. Effects of energy-restricted high-protein, low-fat compared with standard-protein, low-fat diets: a meta-analysis of randomized controlled trials. Am J Clin Nutr 2012;96:1281e98.
