Endocr. J./ R. YAMAMOTO-HONDA et al.: BLOOD GLUCOSE AND HbA1c LEVELS IN DIABETIC PATIENTS doi: /endocrj.K08E-071

Transcription

1 ORIGINAL Distribution of Blood Glucose and the Correlation between Blood Glucose and Hemoglobin A1c Levels in Diabetic Outpatients Ritsuko YAMAMOTO-HONDA 1, Hiroji KITAZATO 1, Shinji HASHIMOTO 2, Yoshihiko TAKAHASHI 3, Yoko YOSHIDA 1, Chiyoko HASEGAWA 1, Yasuo AKANUMA 1 and Mitsuhiko NODA 1,3 1 Department of Endocrinology and Metabolism, Marunouchi Hospital attached to the Institute for Adult Diseases, Asahi Life Foundation, 1-6-1, Marunouchi, Chiyoda-ku, Tokyo, , Japan 2 Research Management Department, Sanwa Kagaku Kennkyusho Co., Ltd., 363 Shiosaki, Hokusei, Inabe, Mie, , Japan 3 Department of Diabetes and Metabolic Medicine, International Medical Center of Japan, , Toyama, Shinjuku-ku, Tokyo, , Japan Received March 8, 2008; Accepted June 12, 2008; Released online July 9, 2008 Correspondence to: Mitsuhiko Noda, Department of Diabetes and Metabolic Medicine, International Medical Center of Japan, , Toyama, Shinjuku-ku, Tokyo, , Japan Abstract. Purpose of the study: Tight glycemic control is important for the prevention of microvascular complications in diabetic patients. We examined the reliability of using blood glucose levels measured at various time-points relative to a meal as an index of glycemic control in Japanese diabetic outpatients. Basic procedures followed: We examined the correlation between the fasting blood glucose (FBG) level; the one-hour (1-h), two-hour (2-h), and three-hour (3-h) post breakfast blood glucose (PBBG) levels, the 1h, 2h, and 3h post lunch blood glucose (PLBG) levels and the hemoglobin A1c levels in Japanese diabetic outpatients. A total of patient-visits to the Marunouchi Hospital between January 2002 and December 2002 were included in the study. The main findings: The blood glucose levels measured at all of the above time-points were significantly correlated with the HbA1c level. As calculated using local polynomial regression fitting, the FPG, 1-h, 2-h, and 3-h PBBG levels that corresponded to an HbA1c level of 6.5% were 132 mg/dl, 174 mg/dl, 170 mg/dl, and 143 mg/dl, respectively. The FPG and 2-h PBBG levels exhibited a good sensitivity and specificity for predicting a glycemic control corresponding to an HbA1c <5.8%, while the FPG and 3-h PBBG levels exhibited fair sensitivity and specificity for predicting glycemic control corresponding to an HbA1c <6.5%. The principal conclusions: The FBG, 2-hPBBG, and 3-hPBBG levels can be used as rough estimates of glycemic control in Japanese diabetic outpatients. Key words: diabetes, blood glucose, HbA1c TIGHT glycemic control is important for preventing microvascular complications in diabetic patients [1-3]. HbA1c is an established marker of long-term glycemic control in patients with diabetes. The UKPDS and DCCT demonstrated a significant association between HbA1c levels and the risk of occurrence of microvascular complications [3,4]. In addition, persistent postprandial hyperglycemia as well as an elevation of HbA1c has been linked to the risk of atherosclerotic diseases [5,6]. Numerous reports have suggested a good correlation between the HbA1c level and the fasting blood glucose level [7,8]. Nevertheless, the diagnostic value of postprandial blood glucose levels for estimating the HbA1c level in diabetic patients remains controversial [9]. In the present study, we analyzed the blood glucose levels (measured at various time-points relative to a meal) and the HbA1c levels in a large number of patients 1

2 and then examined the reliability of using these blood glucose levels measured at arbitrary time-points relative to a meal to estimate the HbA1c levels. Materials and Methods We retrospectively analyzed the charts of outpatients attending the Marunouchi Hospital attached to the Institute for Adult Disease, Asahi Life Foundation (located in the center of Tokyo) between January 2002 and December Patients with diabetes or impaired glucose tolerance were asked about the timing of their previous meal and their capillary blood was drawn to determine the blood glucose and HbA1c levels. The medical technologist judged the postprandial interval time and described it in 15-minute units. Among the records, we selected the blood glucose and HbA1c levels that were determined before breakfast, one hour after breakfast, two hours after breakfast, three hours after breakfast, one hour after lunch, two hours after lunch and three hours after lunch. A total of patient-visits made by 4120 patients (9334 patient-visits made by men and 2117 patient-visits made by women) were enrolled in the study. The mean (± standard deviation) age of the 4120 enrolled patients was 62.7 (±10.4) years, and the median, 25th percentile, and 75th percentile values of HbA1c were 6.6%, 5.9%, and 7.4%, respectively; and the duration of treatment for diabetes at the hospital was years (median, 8.9 years). Of the 4120 patients, 259 and 279 were diagnosed as having type 1 diabetes and impaired glucose tolerance, respectively. Treatment consisted of lifestyle intervention only in 21.1% of the patient-visits, sulfonylureas in 27.4%, a combination of sulfonylureas and biguanides in 12.7%, a combination of sulfonylureas and alpha-glucosidase inhibitors in 2.8%, combinations of other oral hypoglycemic agents in 4.4%, insulin only in 21.1%, a combination of insulin and sulfonylureas in 6.1%, a combination of insulin, sulfonylureas, and biguanides in 2.3%, and combinations of insulin and other oral hypoglycemic agents in 2.1%. Less than 0.2% of the patients were treated with the single use of nateglinide, biguanides, alpha-glucosidase inhibitors, or pioglitazone. The capillary blood glucose values were measured by the glucose-oxidase method (Fuji DRY-CHEM 300). The HbA1c values were measured using high-performance liquid chromatography (HLC-723 GHb V, Tosoh, Tokyo; the reference mean HbA1c value in non-diabetic subjects was 4.7%) [10]. Standard procedures were used to calculate the median, mean, SD and SEM. The relations between the HbA1c levels and the blood glucose levels were analyzed using linear regression according to the least squares method, non-parametric linear regression according to the method of Bablok et al. [11], and polynomial regression fitting according to locally weighted regression (loess {stats}, provided by R [12]). Correlations were calculated using Spearman s correlation coefficient. Statistical significance was considered to exist at the 5% level. Two-by-two contingency tables were formed for various cutoff points to examine the performance characteristics of the blood glucose levels for defining the sensitivity and specificity of the diagnostic test. The data were summarized using receiver-operating characteristic (ROC) curves. All the data was analyzed by R [12]. Glycemic control was categorized in accordance with the criteria of the Japan Diabetes Society as follows: HbA1c <5.8% (excellent), 5.8% HbA1c <6.5% (good), 6.5% HbA1c <8.0% (fair), and 8.0% HbA1c (poor) [13]. Results The distribution of the blood glucose levels is presented in Figure 1. Figure 2 shows the relations between the HbA1c and blood glucose levels at different time-points of measurement in the outpatients. There was a strong correlation between the HbA1c levels and the FBG levels. There were also good correlations between the HbA1c levels and the 1-hour postbreakfast blood glucose (1-h PBBG) level, the 2-hour postbreakfast blood 2

3 glucose (2-h PBBG) level, the 3-hour postbreakfast blood glucose (3-h PBBG), the 1-hour postlunch blood glucose (1-h PLBG) level, the 2-hour postlunch blood glucose (2-h PLBG) level, and the 3-hour postlunch blood glucose (3-h PLBG) level; all of the correlations were highly significant. Regression analyses (linear regression by the least squares method, linear regression by Bablok et al. [11], and polynomial regression fitting (loess)) produced very similar results (Fig. 2). Median blood glucose levels at each time-point corresponding to the various HbA1c levels, as determined by loess are presented in Table 1. The fasting blood glucose, 1-h PBBG, 2-h PBBG, 3-h PBBG, 1-h PLBG, 2-h PLBG and 3-h PLBG levels corresponding to an HbA1c level of 6.5% (associated with a minimal risk of the development of microvascular complications [13,14]) were 132 mg/dl, 174 mg/dl, 170 mg/dl, 143 mg/dl, 169 mg/dl, 170 mg/dl and 157 mg/dl, respectively. The difference between 2-h postprandial blood glucose levels and 1-h postprandial blood glucose levels corresponding to a given HbA1c value were very small when the HbA1c value was above 6.5%. The differences among postprandial blood glucose values corresponding to an HbA1c level of 5.8% were larger than those among postprandial blood glucose values corresponding to an HbA1c level of 8.0%. The 1-h PBBG, 1-h PLBG, 2-h PBBG, 2-h PLBG and 3-h PLBG levels corresponding to an HbA1c level of 8.0% were indistinguishable. Blood glucose levels corresponding to an HbA1c level of 6.5% as obtained using the loess tended to be higher in men than in women (Table 2) and tended to be slightly higher after lunch in patients younger than 60 years old (Table 3). The sensitivity and specificity of the blood glucose levels examined at various time-points to predict glycemic control corresponding to HbA1c levels of <5.8%, <6.5% and <8.0% were calculated and are plotted in Figure 3. The sensitivity and specificity to predict an HbA1c level of <5.8% and an HbA1c level of <8.0% were superior to those for predicting an HbA1c level of <6.5% (Table 5). The sensitivity and specificity of blood glucose levels measured before breakfast were good, while the specificity of 1-h PLBG levels were poor. To predict glycemic control corresponding to an HbA1c level of <5.8%, the sensitivity and specificity of the 2-h PBBG value were superior to those of the 1-h PBBG, 3-h PBBG and 3-h PLBG values. To predict glycemic control corresponding to an HbA1c level of <8.0%, the sensitivity and specificity of the 1-h PBBG, 3-h PBBG and 3-h PLBG values were superior to those of the 2-h PBBG. To predict glycemic control corresponding to an HbA1c level of <6.5%, the sensitivity, specificity and positive predictive value of FPG <130 mg/dl were 71.5%, 72.4% and 66.2%, whereas those of 2-h PBBG <170 mg/dl were 67.2%, 70.0% and 61.0%, respectively. When the data were categorized according to the treatment regimen (with neither oral hypoglycemic agents nor insulin, with oral hypoglycemic agents only, or with insulin), the correlation between the HbA1c and blood glucose levels was the strongest in patients not on medication and weakest in patients treated with insulin (Table 4). Two-hour post-breakfast glucose levels and HbA1c levels obtained from patients treated with various treatment regimens are presented in Figure 4. When the HbA1c levels was larger than 7.0%, the 2-h PBBG levels obtained from patients treated with sulfonylureas or sulfonylureas plus biguanides were lower than those obtained from patients treated with lifestyle intervention only. Data obtained from patients treated with insulin lispro or insulin aspart plus NPH insulin for a given HbA1c level exhibited lower 2-h PBBG values than those of patients treated with regular insulin plus NPH insulin. Discussion In the present study, by analyzing a large number of values of blood glucose and HbA1c, we attempted to determine the degree of reliability with which blood glucose 3

4 levels measured at various time-points relative to a meal could be used as an index of glycemic control in Japanese diabetic outpatients. Many reports have discussed the correlation between blood glucose and HbA1c levels. In type 2 diabetes, studies have shown that postprandial glucose levels measured at 1 h and 2 h after an oral glucose load or a test meal were correlated with the HbA 1c as closely as, or sometimes even more closely than, the fasting plasma glucose levels [15,16]. Kosaka reported that blood glucose levels measured 90 to 150 min after breakfast were well correlated with the HbA 1c [17]. In the present study, we confirmed the correlations between blood glucose values measured at various time-points relative to a meal and the HbA1c levels. We present the reference blood glucose levels measured at various time-points relative to a meal for a given HbA1c level. The estimated levels were very similar to the median blood glucose values shown in Figure 1, and the mean levels calculated from a primary approximation calculated using the least squares method or a linear regression analysis according to the method of Bablok et al. (Fig. 2). Blood glucose levels corresponding to an HbA1c level of 6.5% as obtained from a locally weighted regression tended to be slightly higher after lunch in patients younger than 60 years old (Table 3). The difference between <60 and 60 might be dependent on lifestyle-related differences between relatively younger and elder generations. Blood glucose levels corresponding to an HbA1c level of 6.5% as obtained using a locally weighed regression tended to be higher in men than in women (Table 2). Only a few reports have discussed differences in blood glucose levels between men and women. The gender differences have been reported not to affect the relation between blood glucose and HbA1c levels in the population of the United States [16]. The blood glucose profile after oral glucose load has been reported to vary according to gender in Chinese and Japanese subjects but not in Indian subjects [18]. Thus our present observation needs to be reexamined in other cohorts in Japan. Recently, persistent postprandial hyperglycemia has been suggested to be associated with a risk of atherosclerotic diseases [5,6]. We therefore attempted to examine the difference in 2-hPBBG values for a given HbA1c level according to the various treatment regimens. In the present study, the number of patients receiving monotherapy of alpha-glucosidase inhibitors, nateglinide, or pioglitazone or combination therapy with sulfonylureas and alpha-glucosidase inhibitors was too small for the effect of these drugs on postprandial blood glucose patterns to be analyzed. Treatment with insulin aspart or insulin lispro with NPH insulin is known to lower postprandial blood glucose more than combination therapy with regular insulin and NPH insulin [19,20]. This result was reproducible with our present analysis (Fig. 4). The correlation between the value of blood glucose and HbA1c level is well known and an excellent relation between the HbA1c level and the mean glucose levels measured by continuous glucose monitoring was recently reported by Nathan et al. [21]. In clinical practice, whether the fasting blood glucose value or the postprandial blood glucose value is a better predictor of HbA1c level remains controversial. Avignon et al. reported the prebreakfast (8:00 A.M.), prelunch (11:00 A.M.), postlunch (2:00 P.M.) and extended postlunch (5:00 P.M.) plasma glucose levels in type 2 diabetic patients were all significantly correlated with the HbA1c level. They pointed out that the postlunch and extended postlunch plasma glucose levels exhibited a better sensitivity and specificity for predicting an HbA1c value of 8.5% than the prebreakfast plasma glucose level [9]. In contrast, Bonora et al. reported that HbA 1c level was more strongly correlated with the preprandial blood glucose level than the postprandial (2-3 h after a meal) blood glucose level [8]. Pistrosch et al. examined the relationship between HbA1c and blood glucose levels measured at various times. They reported that the pre-breakfast blood glucose level showed the strongest correlation with the HbA1c level [22]. 4

5 As shown in Figure 1, a given value of blood glucose was distributed over such a wide range of HbA1c levels that the prediction of HbA1c from a given blood glucose level was done at a cost of more than 24% false positives and more than 24% false negatives (Fig. 3). At best, the FBG levels exhibited a fair sensitivity and specificity throughout the range of HbA1c levels that were examined. The sensitivity and specificity of various postprandial levels for predicting a given HbA1c level varied with the target HbA1c. To predict an HbA1c level of <5.8%, the 2-h PBBG level, but not the 2-h PLBG level, appeared to have an equal specificity and sensitivity to those of the FBG level. This situation resembles the good specificity and sensitivity of the 2-h post glucose load value for diagnosing diabetes mellitus. However, the 2-h PBBG value did not show a satisfactory specificity or sensitivity for predicting an HbA1c level of <6.5% or <8.0%. In summary, by examining the blood glucose and HbA1c levels of diabetic patients during routine visits to the diabetic clinic, we attempted to ascertain the correlation between HbA1c and the blood glucose levels measured at various time-points relative to a meal, and presented the resulting data to help estimate the HbA1c using the clinical examination of blood glucose levels. The FBG and 2-h PBBG levels exhibited a fair sensitivity and specificity for predicting an excellent glycemic control. The values of FBG and those of 3-h PBBG exhibited fair sensitivity and specificity in predicting fair glycemic control. These data enable a basic understanding of the relation between postprandial blood glucose levels and HbA1c in diabetic outpatients. Acknowledgments We thank Ms. Kumiko Kimura and Ms. Tomoyo Sato for their excellent technical assistance. References 1. Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M (1995) Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: A randomized prospective 6-year study. Diabetes Res Clin Pract 28: The Diabetes Control and Complications Trial (DCCT) Research Group (1993) The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329: UK Prospective Diabetes Study (UKPDS) Group (1998) Intensive blood-glucose control with sulfonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes: UKPDS33. Lancet 352: Lachin JM, Genuth S, Nathan DM, Zinman B, Rutledge BN; DCCT/EDIC Research Group (2008). The effect of glycemic exposure on the risk of microvascular complications in the diabetes control and complications trial revisited. Diabetes 57: DECODE Study Group, the European Diabetes Epidemiology Group (2001). Glucose tolerance and cardiovascular mortality: comparison of fasting and 2-hour diagnostic criteria. Arch Intern Med 161: Nakagami T; DECODA Study Group (2004) Hyperglycemia and mortality from all causes and from cardiovascular disease in five populations of Asian origin. Diabetologia 47: Bouma M, Dekker JH, De Sonnaville JJ, van der Does FE, de Vries H, Kriegsman DM, Kostense PJ, Heine RJ, van Eijk JT (1999) How valid is fasting plasma glucose 5

6 as a parameter of glycemic control in non-insulin-using patients with type 2 diabetes? Diabetes Care 22: Bonora E, Calcaterra F, Lombardi S, Bonfante N, Formentini G, Bonadonna RC, Muggeo M (2001) Plasma glucose levels throughout the day and HbA(1c) interrelationships in type 2 diabetes: implications for treatment and monitoring of metabolic control. Diabetes Care 24: Avignon A, Radauceanu A, Monnier L (1997) Nonfasting plasma glucose is a better marker of diabetic control than fasting plasma glucose in type 2 diabetes. Diabetes Care 20: Tominaga M, Makino H, Yoshino G, Kuwa K, Takei I, Aono Y, Hoshino T, Umemoto M, Shimatsu A, Sanke T, Kuwashima M, Taminato T, Ono J (2005) Japanese standard reference material for JDS Lot 2 haemoglobin A1c. I: Comparison of Japan Diabetes Society-assigned values to those obtained by the Japanese and USA domestic standardization programmes and by the International Federation of Clinical Chemistry reference laboratories. Ann Clin Biochem. 42 (Pt 1): Bablok W, Passing H, Bender R, Schneider B (1988) A general regression procedure for method transformation. Application of linear regression procedures for method comparison studies in clinical chemistry, Part III. J. Clin. Chem. Clin. Biochem. 26: Ihaka R and Gentleman R (1996) R: a language for data analysis and graphics. J. Comp. Graph. Stat. 5: Available via Japan Diabetes Society (eds.) (2007) Glycemic control indicators and assessment. In Treatment guide for diabetes. Bunkodo: Tokyo, p International Diabetes Federation (eds) (2005) Glucose control levels. In Global guideline for type 2 diabetes. International Diabetes Federation, p Monnier L, Lapinski H, Colette C (2003) Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: variations with increasing levels of HbA1 (c). Diabetes Care 26: Woerle HJ, Pimenta WP, Meyer C, Gosmanov NR, Szoke E, Szombathy T, Mitrakou A, Gerichc JE (2004) Diagnostic and therapeutic implications of relationships between fasting, 2-hour postchallenge plasma glucose and hemoglobin A1C values. Arch Intern Med 164: Kosaka K (1997) Parameters used for the diagnosis and screening of diabetes mellitus and their mutual relationships. Nippon Rinsho 55 (725): [Article in Japanese] 18. The DECODA Study Group (2003) Age- and sex- specific prevalence of diabetes and impaired glucose regulation in 11 Asian cohorts. Diabetes Care 26: Pampanelli S, Torlone E, Ialli C, Del Sindaco P, Ciofetta M, Lepore M, Bartocci L, Brunetti P, Bolli GB (1995) Improved postprandial metabolic control after subcutaneous injection of a short-acting insulin analog in IDDM of short duration with residual pancreatic beta-cell function. Diabetes Care 18: Rosenfalck AM, Thorby P, Kjems L, Dejgaard A, Hanssen KF, Madbad S (2000) Improved postprandial gylcemic control with insulin Aspart in type 2 diabetic patients treated with insulin. Acta Diabetol. 37: Nathan DM, Turgeon H, Regan S (2007) Relationship between glycated haemoglobin levels and mean glucose levels over time. Diabetologia 50: Pistrosch F, Koehler C, Wildbrett J, Hanefeld M (2006). Relationship between diurnal glucose levels and HbA1c in type 2 diabetes. Horm Metab Res 38:

7 Fig. 1. Distribution of fasting and postprandial blood glucose levels according to HbA1c levels. 7

8 Fig. 2. Relation between the HbA1c values and blood glucose values. Red line indicates the linear regression according to the least squares method. Yellow line indicates the linear regression according to the method of Bablok et al. [11]. Green line indicates the locally weighted regression line (loess). 8

9 Fig. 3. Receiver-operating characteristic (ROC) curves for predicting glycemic control corresponding to HbA1c levels of <5.8%, <6.5% and <8.0% based on the blood glucose measurement. 9

10 Fig. 4. Relation between HbA1c values and blood glucose values measured 2 h after breakfast; categorized according to diabetic treatment regimen. Samples were taken from patients with the fixed treatment regimen throughout the examination period. Lines indicate the locally weighted regression line (loess). 10

11 Table 1. Blood glucose values corresponding to HbA1c values of 5.8%, 6.5%, and 8.0% Time of examination Values corresponding to an HbA1c value of 5.8% Values corresponding to an HbA1c value of 6.5% Values corresponding to an HbA1c value of 8.0% Fasting 120 mg/dl 132 mg/dl 159 mg/dl One hour after breakfast 160 mg/dl 174mg/dL 208 mg/dl Two hours after breakfast 146 mg/dl 170 mg/dl 201 mg/dl Three hour after breakfast 123 mg/dl 143 mg/dl 182 mg/dl One hour after lunch 159 mg/dl 169 mg/dl 197 mg/dl Two hours after lunch 151 mg/dl 170 mg/dl 199 mg/dl Three hours after lunch 138 mg/dl 157 mg/dl 196 mg/dl Table 2. Relationship between the HbA1c values and blood glucose values according to gender Time of examination Fasting One hour after breakfast Two hours after breakfast Three hours after breakfast One hour after lunch Two hours after lunch Three hours after lunch Gender of the patients Number of patient-visits Median value of HbA1c (%) R Blood glucose values (mg/dl) corresponding to an HbA1c level of 6.5% men women * 122 men women men women ** 153 men women men women men women men women * 135 ** p < 0.01, * p < 0.05 versus R value of men. 11

12 Table 3. Relationship between the HbA1c values and blood glucose values according to patient age Time of examination Fasting Age of patients (years) Number of patient-visits Median value of HbA1c (%) R Blood glucose values (mg/dl) corresponding to an HbA1c level of 6.5% < ** 130 One hour after < breakfast Two hours after < breakfast Three hours < after breakfast ** 144 One hour after < lunch Two hours after < lunch Three hours < after lunch ** p < 0.01 versus R value of patients under 60 years old. 12

13 Table 4. Correlation between the HbA1c values and blood glucose values according to diabetic treatment regimen Timing of examination Medication for hyperglycemia Number of patient-visits Median value of HbA1c (%) R Blood glucose values (mg/dl) corresponding to an HbA1c level of 6.5% Fasting One hour after breakfast Two hours after breakfast Three hours after breakfast One hour after lunch Two hours after lunch Three hours after lunch None OHA *** 138 Insulin *** 125 None OHA Insulin None OHA *** 172 Insulin *** 169 None OHA Insulin *** 149 None OHA Insulin None OHA Insulin ** 166 None OHA Insulin *** p < 0.001, ** p < 0.01 versus R value of patients without medication for hyperglycemia. OHA: oral hypoglycemic agents. 13

14 Table 5. Values obtained from the receiver-operating characteristic curves for estimating the HbA1c levels from the blood glucose values Target HbA1c Level (%) < 5.8% < 6.5% < 8.0% Timing of examination PB: post breakfast PL: post lunch Value of blood glucose (mg/dl) with equal sensitivity and specificity (a) Sensitivity (=specificity) values corresponding to blood glucose values of (a) Value of blood glucose (mg/dl) with the maximum sensitivity + specificity (b) Sensitivity + specificity values corresponding to blood glucose values of (b) Fasting One hour PB Two hours PB Three hours PB One hour PL Two hours PL Three hours PL Fasting One hour PB Two hours PB Three hours PB One hour PL Two hours PL Three hours PL Fasting One hour PB Two hours PB Three hours PB One hour PL Two hours PL Three hours PL PB: post breakfast, PL: post lunch. 14