Association of serum phosphate levels with glucose tolerance, insulin sensitivity and insulin secretion in non-diabetic subjects

Transcription

1 (2006) 60, & 2006 Nature Publishing Group All rights reserved /06 $ ORIGINAL ARTICLE Association of serum phosphate levels with glucose tolerance, insulin sensitivity and insulin secretion in non-diabetic subjects M Haap 1, E Heller 1, C Thamer, O Tschritter, N Stefan and A Fritsche Department of Endocrinology, Metabolism, Pathobiochemistry, Vascular Medicine and Nephrology, University of Tübingen, Tübingen, Germany Background: Hypophosphatemia is associated with impaired glucose tolerance and insulin resistance in primary hyperparathyroidism. However, little is known about the association between serum phosphate and glucose metabolism in healthy subjects. Methods: We measured fasting serum phosphate levels (SP, normal range mg/dl) and serum calcium (S-Ca, normal range mmol/l) in 881 non-diabetic subjects (341 male/540 female, age: 3871 years, body mass index kg/m 2 (mean7standard error of the mean). An oral glucose tolerance test (OGTT) with determination of glucose and insulin every 30 min was performed in all subjects. Insulin secretion and insulin sensitivity (IS) were estimated from the OGTT using validated indices. Furthermore, we tested whether serum phosphate predicts glucose tolerance in 115 subjects during a lifestyle intervention program (LIP). Results: Serum phosphate was negatively correlated with 2-h blood glucose levels independent of age, gender and percent body fat (r ¼ 0.13, Po0.0001). This association remained significant after additional adjustment for S-Ca, creatinine and parathyroid hormone. Serum phosphate was positively correlated with IS (r ¼ 0.10, P ¼ ), but not with insulin secretion. This was independent of age, gender, percent body fat, S-Ca and serum creatinine. In the subjects taking part in the LIP low serum phosphate levels at baseline were associated with higher postprandial glucose levels. Conclusions: In non-diabetic subjects, low serum phosphate levels are associated with high 2-h blood glucose levels and reduced IS. Whether low serum phosphate levels are a cause or a consequence of low IS and impairment of glucose tolerance needs to be tested in further studies. (2006) 60, doi: /sj.ejcn ; published online 4 January 2006 Keywords: serum phosphate; glucose tolerance; insulin sensitivity; insulin secretion Introduction In patients with primary hyperparathyroidism (phpt) the prevalence of type 2 diabetes (T2DM) is three times higher compared to healthy subjects (Taylor and Khaleeli, 2001). phpt is also associated with impaired glucose tolerance (IGT) and reduced insulin sensitivity (IS) (Ljunghall et al., 1983). The hypophosphatemia present in phpt and other Correspondence: Dr A Fritsche, Department of Endocrinology, Metabolism, Pathobiochemistry, Vascular Medicine and Nephrology, Medizinische Universitätsklinik, Otfried-Müller-Str. 10, Tübingen D-72076, Germany. andreas.fritsche@med.uni-tuebingen.de 1 These authors contributed equally to this work. Received 25 July 2005; revised 28 October 2005; accepted 31 October 2005; published online 4 January 2006 conditions of various pathogenic origin is thought to be the cause of altered glucose metabolism and impaired responsiveness to insulin (DeFronzo and Lang, 1980). However, there is also evidence, that phosphate supplementation in glucose intolerant hypophosphatemic patients improves glucose tolerance significantly (Wittmann and Nagy, 1997). In addition, surgical treatment of phpt at least partly reverses IGT and insulin resistance in these subjects (Prager et al., 1990; Kautzky-Willer et al., 1992). There are several possible mechanisms by which phosphate levels influences glucose metabolism. Phosphate is involved in energy balance and ATP generation. In obese individuals it was shown, that an infusion of several electrolytes/minerals including phosphate was able to increase postprandial thermogenesis (Jaedig et al., 1994).

2 Furthermore, in vitro studies have demonstrated that pancreatic islets of phosphate-depleted rats had low ATP levels, elevated cytosolic calcium and impaired glycolytic activity. These islets responded to the phosphate depletion with an impaired insulin secretion. These effects were normalized after phosphate supplementation (Castillo et al., 1982). So far, correlations between low serum phosphate and IGT and reduced IS have been shown mainly in states of pronounced phosphate depletion and IGT. However, little is known about these associations in non diabetic individuals with phosphate levels in the normal range. Therefore the objective of the present study was to determine the association of serum phosphate with glucose tolerance, IS and insulin secretion in a large healthy, nondiabetic population. In addition, we tested whether serum phosphate levels at baseline predict postprandial glucose levels in a longitudinal study. Materials and methods Subjects A total of 881 individuals who participated in the Tübingen Family Study for T2DM were investigated. In that study, primarily normal glucose-tolerant individuals with a family history of T2DM are recruited and metabolically characterized. Recruitment mechanisms include newspaper ads and word-of-mouth proposing diabetes screens. Subjects with T2DM and other severe diseases or with medication (especially drugs affecting glucose metabolism) were excluded from the current analysis. The local ethics committee had approved all protocols. All subjects gave informed written consent and underwent the standard procedures of the protocols including medical history, medication, physical examination, routine blood tests and an oral glucose tolerance test (OGTT) (data Table 1) 332 subjects underwent a hyperinsulinemic euglycemic clamp. In an additional prospective analysis, 115 subjects taking part in a lifestyle intervention program (TUebingen Lifestyle Intervention Program (TULIP)) were included. They were studied at baseline and after a mean follow-up of 9 months. The subjects were at increased risk for T2DM either because of being overweight (body mass index (BMI)427 kg/m 2 )or being a first-degree relative of a type 2 diabetic patient or having an IGT or females after gestational diabetes. This program includes the goals of the diabetes prevention study program (Tuomilehto et al., 2001): reduction of body weight by more than 5%, reduction of dietary intake of fat to less than 30% of calories, reduction of intake of saturated fatty acids to less than 10% of calories, increase of the daily amount of ingested fiber to more than 15 g per 100 kcal and increase of the amount of weekly exercise to more than 3 h per week. To achieve those goals, subjects were looked after by a team of dieticians. During the first 6 months, they had Table 1 Subjects characteristics (cross-sectional analysis) Mean7s.e.m. Range Normal range n 881 Male/female 341/540 NGT/IGT 770/111 Age (years) BMI (kg/m 2 ) Body fat (%) Fasting glucose (mmol/l) o6.0 2-h-glucose (mmol/l) o11.0 Fasting-insulin (mmol/l) h-insulin (mmol/l) Insulin sensitivity a (OGTT) st-phase insulin secretion (pm) HbA1c (%) Creatinine (mg/dl) o1.2 Potassium (mmol/l) Sodium (mmol/l) Calcium (mmol/l) Phosphate (mg/dl) Magnesium (mmol/l) Parathyroid hormone (pmol/l) b a Estimated according to the equation of Matsuda and de Fronzo (in arbitrary units) (Matsuda and DeFronzo, 1999). b Parathyroid hormone measured n ¼ 348. eight sessions with their lifestyle educators where they received individual advice. During each visit participants had to present a 2-day food diary and discussed the results with the dieticians. Diet composition was determined using a validated computer program (DGE-PC 3.0, Deutsche Gesellschaft für Ernährung, Bonn, Germany). In all, 90 subjects presented food diaries on a regular basis (1373 days of food diaries (mean7s.d.)) and were therefore included in the analysis of phosphate intake. OGTT A standard 75 g OGTT was performed after a 10-h overnight fast, with determination of glucose, insulin and C-peptide at 0, 30, 60, 90, and 120 min. Insulin sensitivity was calculated as proposed by Matsuda and DeFronzo (1999). First phase insulin secretion was calculated using a validated index (Stumvoll et al., 2000) (1st PH ¼ 1283 þ Ins Gluc 30 þ Ins 0 ). Percentage of body fat was determined using bioelectrical impedance (RJL Systems, Detroit, MI, USA). In the prospective study, the OGTT was repeated after a mean follow-up time of 9 months. Hyperinsulinemic euglycemic clamp The hyperinsulinemic euglycemic clamp was performed as previously described (Thamer et al., 2003). The insulin sensitivity index (ISI) (in mmol kg 1 min 1 pm 1 ) for systemic glucose uptake was calculated as the mean infusion rate of glucose (GIR) (in mmol kg 1 min 1 ) necessary to 735

3 736 maintain euglycemia during the last 60 min of the euglycemic hyperinsulinemic clamp divided by the steady-state serum insulin concentration. a Analytical procedures Plasma glucose was measured by the glucose-oxidase method (YSI, Yellow Springs Instruments, Yellow Springs, USA). Plasma insulin was determined by a microparticle enzyme immunoassay (Abbott Laboratories, Tokyo, Japan). 2-h-glucose (mmol/l) log scale r= p = Statistical analysis All data are given as mean7standard error of the mean unless otherwise stated. For variables which were not normally distributed data were logarithmically transformed for further analysis. We used multivariate regression analysis to adjust for the independent effects of covariates. A P-value o0.05 was considered to be statistically significant. All calculations and statistical analyses were performed using the Statistical Package JMP Version 4.0 (SAS Institute Inc., Cary, NC, USA). b Insulin sensitivity (OGTT) log scale r = 0.10 p = Results c Anthropometric data of all subjects are shown in Table 1. The participants had a mean age of 3871 years and a mean BMI of kg/m 2. Out of 881 individuals 770 were normal glucose tolerant and 111 subjects were impaired glucose tolerant. Association of serum phosphate with BMI and glucose tolerance Serum phosphate was negatively correlated with BMI (r ¼ 0.22, Po0.0001). This was still significant after adjusting for the covariates age and gender (r ¼ 0.17, Po0.0001) (Figure 2). Serum phosphate was also negatively correlated with 2-h blood glucose concentration (r ¼ 0.21, Po0.0001). This effect remained significant after adjusting for the covariates gender, age and percent of body fat (Figure 1a) (r ¼ 0.13, P ¼ , Table 2). The additional inclusion of serum calcium (S-Ca), creatinine and parathyroid hormone into the model did not change the significant independent association of serum phosphate with 2 h blood glucose. Association of serum phosphate with IS and secretion Serum phosphate was positively correlated with IS calculated from the OGTT (r ¼ 0.19, Po This effect remained significant after adjusting for the covariates gender, age and percentage of body fat (Figure 1b) (r ¼ 0.10, Po0.006, Table 3). The additional inclusion of S-Ca and creatinine into the model did not change the significant independent association of serum phosphate with IS. Insulin secretion pm log scale r = 0.06 p = phosphate (mg/dl) Figure 1 (a) Correlation between serum phosphate and glucose tolerance (2 h blood glucose), r and P-values are expressed after adjustment for gender, age and percent body fat. (b) Correlation between serum phosphate and insulin sensitivity (Matsuda and DeFronzo, 1999). r and P-values are expressed after adjustment for gender, age and percent body fat. (c) Correlation between serum phosphate and insulin secretion (validated index (Stumvoll et al., 2000)) r and P-values are expressed after adjustment for gender, age and percent body fat. In a subset of 332 subjects which underwent a hyperinsulinemic euglycemic clamp, serum phosphate was also positively associated with IS independent of gender, age, and percentage of body fat (r ¼ 0.12, Po0.03). There was a significant correlation between serum phosphate and insulin secretion, (r ¼ 0.12, Po0.001) but this correlation was not significant after adjustment for the covariates gender, age and percentage of body fat (Figure 1c) (r ¼ 0.06, Po0.09).

4 Association of other serum electrolytes with glucose tolerance, IS and secretion Serum potassium was associated with 2 h blood glucose (r ¼ 0.14, P ¼ ).This association remained significant BMI (kg/m 2 ) log-scale r = p < phosphate (mg/dl) Figure 2 Correlation between serum phosphate and body mass index (BMI). r and P-values are expressed after adjustment for gender, age. after adjustment for gender, age and percentage of body fat (r ¼ 0.13, Po0.0001). Serum potassium also showed an association with IS after adjusting for gender, age and percent body fat (r ¼ 0.08, Po0.02). S-Ca, serum sodium and serum magnesium were not associated with 2 h blood glucose levels or IS after adjusting for gender, age and body fatness (all P40.1). Prospective analysis The 115 (46 male/69 female) subjects included in the prospective analysis had a mean age of 4671 years. During lifestyle intervention (mean follow-up 971 months) their BMI decreased from to kg/m 2 (Po0.0001), their plasma glucose decreased from to mmol/l (Po0.0001) and their serum phosphate increased from to mg/dl (Po0.001). The estimated average daily phosphate intake during 9 months was mg per day. The average phosphate intake did not change during the 9 months follow-up (P ¼ 0.9). 737 Table 2 Effect of serum phosphate on glucose tolerance in multivariate regression models Independent variable Model 1 (r 2 ¼ 0.14, Po0.0001) Model 2 (r 2 ¼ 0.14, Po0.0001) Model 3 (r 2 ¼ 0.17, Po0.0001) Model 4 (r 2 ¼ 0.11 Po0.0001) Estimate s.e. P-value Estimate s.e. P-value Estimate s.e. P-value Estimate s.e. P-value Intercept o o o o Sex Age o o o % Body fat o o o o Phosphate o Calcium Creatinine Parathyroid homone Model 4 was performed in n ¼ 348 subjects were parathyroid hormone levels were available. In all models, 2-h glucose is the dependent variable. Bold characters display significant associations of phosphate with glucose tolerance. Table 3 Effect of serum phosphate on insulin sensitivity in multivariate regression models Independent variable Model 1 (r 2 ¼ 0.33, Po0.0001) Model 2 (r 2 ¼ 0.33, Po0.0001) Model 3 (r 2 ¼ 0.34, Po0.0001) Model 4 (r 2 ¼ 0.34, Po0.0001) Estimate s.e. P-value Estimate s.e. P-value Estimate s.e. P-value Estimate s.e. P-value Intercept o o o o Sex o o o o Age % Body fat o o o o Phosphate Calcium Creatinine Parathyroid homone Model 4 was performed in n ¼ 348 subjects were parathyroid hormone levels were available. In all models, insulin sensitivity is the dependent variable. Bold characters display significant associations of phosphate with insulin sensitivity.

5 738 Table 4 Effect of serum phosphate on glucose tolerance in a longitudinal study Independent variable In a multivariate model (Table 4) including gender, age, plasma glucose at baseline and change in body fat we found a significant predictive value of serum phosphate at baseline for postprandial plasma glucose levels at follow-up (Po0.01). Low serum phosphate levels at baseline are associated with higher postprandial plasma glucose levels after a lifestyle intervention. The change in serum phosphate during lifestyle intervention had no significant influence on postprandial plasma glucose. Discussion Dependent variable: 120 min plasma glucose at follow-up (OGTT) Model (r 2 ¼ 0.48, Po0.0001) Estimate s.e. P-value Intercept Sex Age Plasma glucose 120 min (baseline) o % body fat (baseline) % body fat (follow up) Serum phosphate (baseline) The objective of the present study was to investigate whether serum phosphate levels are associated with glucose tolerance, IS and insulin secretion in healthy non-diabetic subjects. There was a significant association of low serum phosphate concentrations with high 2-h blood glucose levels independent of anthropometric parameters like percent body fat, age and gender. These findings are in line with several other studies which demonstrated a significant correlation between serum phosphate, glucose tolerance (DeFronzo and Lang, 1980; Campillo et al., 1982) and adiposity (Lindgarde and Trell, 1977; Haglin et al., 2001). Especially the inverse correlation between BMI and low serum phosphate concentrations has been described earlier (Lind et al., 1993). The mechanism is thought to be attributed by an overconsumption of energy (diet with low nutrient density), in particular an overconsumption of carbohydrates accompanied by low protein intake which is a main source of phosphate (Haglin, 2001). The mechanism how hypophosphatemia could influence glucose tolerance is under investigation. Two main factors influencing glucose homeostasis are insulin secretion and IS (Gerich, 1998). The present study could demonstrate that low serum phosphate levels were associated with a reduced IS estimated from the OGTT. This association was independent of gender, age, percent body fat, S-Ca or serum creatinine. This finding could be confirmed in a smaller group of subjects using the gold standard for measurement of IS, the hyperinsulinemic euglycemic clamp. Moreover, these results are in line with findings in patients with hypophospatemia caused by phpt (DeFronzo and Lang, 1980; Sowers and Draznin, 1998). The results further indicate that there was no significant independent correlation between serum phosphate levels and insulin secretion. These findings are consistent with the results in humans by Paula et al. (1998). In contrast, Zhou et al. (1991) found a reduced insulin secretion capacity in phosphorus-depleted rats. The different results may be due to the extreme phenotype in the animal model. Taken together, one explanation for the association of hypophosphatemia with IGT is that hypophophatemia may cause insulin resistance in subjects with disturbed electrolytes metabolism as well as in a healthy population. Phosphate is needed for ATP generation and therefore seems to be an important component of energy metabolism (Massry et al., 1992; Thompson and Kemp, 1995). Reduction of serum phosphate could theoretically lead to disturbances of energy metabolism resulting in insulin resistance and IGT. For example, low serum phosphate levels inhibit the phosphorylation of carbohydrate intermediates in glycolysis and gluconeogenesis (Xie et al., 2000). In addition, limitations of glucose transport across the cell membrane by altering transmembraneous phospholipids are under investigation (Bouche et al., 2004). Another explanation for the present findings could be that low serum phosphate levels are not the cause but rather the consequence of increased postprandial blood glucose levels. It is known that osmotic diuresis present in hyperglycemic and acidemic states may cause an increased phosphate excretion due to competition between phosphate and glucose in the proximal tubular system (Simonson and DeFronzo, 1982). However, since the analysis could confirm the present results also in the subgroup of normal glucosetolerant subjects, the possibility that postprandial hyperglycemia may lead to hypophosphatemia seems to be unlikely. Furthermore, the results demonstrate that low serum phosphate levels in subjects entering a lifestyle intervention are associated with higher postprandial plasma glucose levels at follow-up. Serum phosphate levels increased during lifestyle intervention, however, this increase was not associated with changes in glucose tolerance. Therefore, it seems unlikely that small changes in phosphate intake as part of the lifestyle intervention are responsible for the changes in postprandial plasma glucose levels. In summary, the present analysis could demonstrate an association of low serum phosphate with high postprandial blood glucose and impaired IS. It can be speculated that insulin resistance and glucose intolerance can be at least worsened by low serum phosphate levels, even in otherwise healthy subjects. However, studies with acute and chronic phosphate supplementation and careful monitoring of glucose homeostasis are needed to finally determine whether hypophosphatemia is causative for hyperglycemia or is rather the consequence of IGT in otherwise healthy subjects.

6 Acknowledgements We thank the volunteers for their participation. We gratefully acknowledge the superb technical assistance of Anna Teigeler, Marjo Graf and Heike Luz. The studies were in part supported by grants from the Deutsche Forschungsgemeinschaft (KFO 114/1) and the Interdisciplinary Center of Clinical Research Tübingen (IZKF). References Bouche C, Serdy S, Kahn CR, Goldfine AB (2004). The cellular fate of glucose and its relevance in type 2 diabetes. Endocr Rev 25, Campillo JE, Aguayo J, Pages I, Castillo M, Osorio C (1982). Inorganic phosphate insulin relationships in normal subjects and in patients with moderate glucose intolerance. Diabet Metab 8, Castillo M, Campillo JE, Martinez VM, Osorio C (1982). Effect of phosphate omission on the glucose-induced insulin release in vitro in fed and fasted rats. Acta Diabetol Lat 19, DeFronzo RA, Lang R (1980). Hypophosphatemia and glucose intolerance: evidence for tissue insensitivity to insulin. N Engl J Med 303, Gerich JE (1998). The genetic basis of type 2 diabetes mellitus: impaired insulin secretion versus impaired insulin sensitivity. Endocr Rev 19, Haglin L (2001). Hypophosphataemia: cause of the disturbed metabolism in the metabolic syndrome. Med Hypotheses 56, Haglin L, Lindblad A, Bygren LO (2001). Hypophosphataemia in the metabolic syndrome. Gender differences in body weight and blood glucose. Eur J Clin Nutr 55, Jaedig S, Lindgarde F, Arborelius M (1994). Increased postprandial energy expenditure in obese women after peroral K- and Mg-phosphate. Miner Electrolyte Metab 20, Kautzky-Willer A, Pacini G, Niederle B, Schernthaner G, Prager R (1992). Insulin secretion, insulin sensitivity and hepatic insulin extraction in primary hyperparathyroidism before and after surgery. Clin Endocrinol (Oxford) 37, Lind L, Lithell H, Hvarfner A, Pollare T, Ljunghall S (1993). On the relationships between mineral metabolism, obesity and fat distribution. Eur J Clin Invest 23, Lindgarde F, Trell E (1977). Serum inorganic phosphate in middleaged men. I. Inverse relation to body weight. Acta Med Scand 202, Ljunghall S, Palmer M, Akerstrom G, Wide L (1983). Diabetes mellitus, glucose tolerance and insulin response to glucose in patients with primary hyperparathyroidism before and after parathyroidectomy. Eur J Clin Invest 13, Massry SG, Fadda GZ, Perna AF, Kiersztejn M, Smogorzewski M (1992). Mechanism of organ dysfunction in phosphate depletion: a critical role for a rise in cytosolic calcium. Miner Electrolyte Metab 18, Matsuda M, DeFronzo RA (1999). Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabet Care 22, Paula FJ, Plens AE, Foss MC (1998). Effects of hypophosphatemia on glucose tolerance and insulin secretion. Horm Metab Res 30, Prager R, Schernthaner G, Niederle B, Roka R (1990). Evaluation of glucose tolerance, insulin secretion, and insulin action in patients with primary hyperparathyroidism before and after surgery. Calcif Tissue Int 46, 1 4. Simonson D, DeFronzo RA (1982). Hypophosphatemia and glucose intolerance. Adv Exp Med Biol 151, Sowers JR, Draznin B (1998). Insulin, cation metabolism and insulin resistance. J Basic Clin Physiol Pharmacol 9, Stumvoll M, Mitrakou A, Pimenta W, Jenssen T, Yki-Jarvinen H, Van Haeften T et al. (2000). Assessment of insulin secretion from the oral glucose tolerance test in white patients with type 2 diabetes. Diabet Care 23, Taylor WH, Khaleeli AA (2001). Coincident diabetes mellitus and primary hyperparathyroidism. Diabet Metab Res Rev 17, Thamer C, Machann J, Bachmann O, Haap M, Dahl D, Wietek B et al. (2003). Intramyocellular lipids: anthropometric determinants and relationships with maximal aerobic capacity and insulin sensitivity. J Clin Endocrinol Metab 88, Thompson CH, Kemp GJ (1995). Reduced muscle cell phosphate (Pi) without hypophosphatemia in mild dietary Pi deprivation. Clin Chem 41, Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P et al. (2001). Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 344, Wittmann I, Nagy J (1997). Effectiveness of phosphate supplementation in glucose intolerant, hypophosphatemic patients. Miner Electrolyte Metab 23, Xie W, Tran TL, Finegood DT, van de WG (2000). Dietary P(i) deprivation in rats affects liver camp, glycogen, key steps of gluconeogenesis and glucose production. Biochem J 352 (Part 1), Zhou XJ, Fadda GZ, Perna AF, Massry SG (1991). Phosphate depletion impairs insulin secretion by pancreatic islets. Kidney Int 39,