diabetes research and clinical practice 106 (2014) 531–537
Contents available at ScienceDirect
Diabetes Research and Clinical Practice journ al h ome pa ge : www .elsevier.co m/lo cate/diabres
The effects of pitavastatin on glucose metabolism in patients with type 2 diabetes with hypercholesterolemia Hisashi Daido a,b, Yukio Horikawa b,*, Jun Takeda b a b
Hashima City Hospital, Gifu, Japan Department of Diabetes and Endocrinology, Gifu University Graduate School of Medicine, Gifu, Japan
article info
abstract
Article history:
Background: Although there have been several reports that statins cause insulin resistance
Received 5 April 2014
that leads to the occurrence of type 2 diabetes in Caucasians, there has been no Japanese
Received in revised form
prospective studies investigating the effects of statins on the glucose metabolism system.
4 June 2014
Materials and methods: Our subjects were 86 Japanese patients with type 2 diabetes with
Accepted 14 September 2014
hypercholesterolemia. Pitavastatin 2 mg/day was administered for 12 months and the lipid-
Available online 22 October 2014
related values, glucose metabolism values, and the presence/absence of side effects were investigated.
Keywords:
Results: None of these factors was found to differ between before and after administration of
Statin
pitavastatin in overall analysis of all subjects. In subgroup analysis, fasting blood glucose
Glucose metabolism
showed a decrease in the BMI 25 group and there was a significant difference between the
Insulin resistance
BMI < 25 and BMI 25 groups (P-values: 0.021 and 0.0036). Although HbA1c showed an
Insulin secretion
increase both in the group switched to pitavastatin and the BMI < 25 group (P-values: 0.035
Dyslipidemia
and 0.033) and HOMA-b showed a decrease in the BMI < 25 group (P-values: 0.044), there were no significant differences in changes between each divided group and their counterparts. Conclusion: In the Japanese obese group with BMI 25, pitavastatin elicited a significant decrease in fasting blood glucose. It is not clear whether or not this is due to improved insulin resistance as a direct effect of pitavastatin, but in contrast to findings in Caucasians pitavastatin does not worsen insulin resistance in Japanese patients with type 2 diabetes complicated by hypercholesterolemia. # 2014 Elsevier Ireland Ltd. All rights reserved.
1.
Introduction
Dyslipidemia in patients with type 2 diabetes is characterized by the increased production of VLDL, decreased production of HDL, and increased sdLDL (small dense LDL) levels. Free Fatty
Acids (FFAs) from visceral fat are excreted directly through the portal vein, causing high FFA concentrations in the liver and pancreas. Such chronically high FFA levels affect the ability to secrete insulin as well as the action of insulin [1]. Japanese patients with type 2 diabetes are characterized mainly by intrinsic impaired insulin secretion [2], which is well known to
* Corresponding author at: Department of Diabetes and Endocrinology, Gifu University Graduate School of Medicine, 1-1 Yanagido, Gifucity, Gifu 501-1194, Japan. Tel: +81 58 230 6564; fax: +81 58 230 6376. E-mail address: [email protected] (Y. Horikawa). http://dx.doi.org/10.1016/j.diabres.2014.09.048 0168-8227/# 2014 Elsevier Ireland Ltd. All rights reserved.
532
diabetes research and clinical practice 106 (2014) 531–537
be influenced by hyperglycemia and hyper FFA-emia, known as glucotoxicity and lipotoxicity, respectively. Both hyper-FFA-emia and hyper-LDL-emia are known to exert an effect on b cell function. LRP (LDL-receptor related protein), a specific LDL receptor, is expressed in the pancreatic islets, and it is known that pancreatic b cells show increased apoptosis under conditions of hyper-LDL-emia [3,4]. It has been shown in Tangier disease, which is associated with hyper-LDL-emia due to a mutation in the ABCA1 protein, that insulin secretion is decreased and hyperglycemia occurs [5,6]. Statins, which are inhibitors of HMG-CoA reductase, a ratelimiting enzyme of cholesterol synthesis, effectively lower LDL cholesterol (LDL-C). Cardiovascular morbidity, mainly due to dyslipidemia and hypertension, is a major burden in patients with type 2 diabetes. In Caucasians, the efficacy of statins in the prevention of cardiovascular disease has been well documented in long-term outcomes trials involving patients with varying degrees of cardiovascular risk, including patients with type 2 diabetes [7–11]. Thus, it might be expected that statins can both lower LDLC and preserve the insulin secretory function of b cells, but there have been several reports in Caucasians that statins have caused insulin resistance, although the mechanism is unknown, and that this has led to the occurrence of type 2 diabetes [12,13]. Approximately 80% of Japanese patients with type 2 diabetes also have dyslipidemia, so it is particularly important to elucidate the effect of statins on glucose metabolism with patients with type 2 diabetes. The main characteristic of type 2 diabetes in Caucasians is generally obesity-related insulin resistance, while that of Japanese is mainly impaired insulin secretion [2]. The response of glucose metabolism to statins might therefore differ in Caucasians and Japanese. In fact, there have been no prospective Japanese studies investigating whether statins exert any effects on the glucose metabolism system. We performed this study (Gifu Lipid Study) using pitavastatin to investigate the effect of statins on insulin resistance and insulin secretion in Japanese.
2.
Subjects and methods
This study is a multi-center, open-label, single-arm study. The registration period was within one year. The subjects were insulin-naı¨ve patients with type 2 diabetes and HbA1c values of 6.6% [NGSP] (48 mmol/mol [IFCC]) or more and less than 8.8% (72 mmol/mol), which did not vary more than 1% measured within the previous 3 months, who also had hypercholesterolemia with fasting serum total cholesterol (TC) of 5.7 mM (220 mg/dL) or LDL-C of 3.6 mM (140 mg/dL). Even when another statin had already been administered to a patient, if the patient satisfied these conditions, it was considered reasonable to switch drugs, and since this study was conducted within the limits of daily clinical practice, a wash-out period was not used. Patients who satisfied the above conditions and gave written informed consent when the purpose and the possible effects of the study were explained, were registered as subjects by a central registration system. The exclusion criteria as well as the inclusion criteria are shown in supplemental information. Pitavastatin 2 mg/day was administered to the registered patients for 12 months, and the lipid-related values, glucoserelated values, and presence/absence of side effects, as indicated in Table 1, were investigated. If LDL-C was not measured by a direct method, the Friedewald method was used for calculation (LDL-C value = TC value HDL-C value [TG value/5]). The blood specimens were taken mostly at fasting state. If patients were not fasting, only HbA1c, TChol, LDL-C, and HDL-c were used for analysis. The primary endpoint of this study was to examine the response of glucose metabolism to pitavastatin. The main analysis was performed in all subjects who met the inclusion criteria. Sub-analysis was performed in groups divided according HbA1c, lipid-lowering drug naivety, and BMI. Subjects were divided into two groups of glycemic control by HbA1c level at the start of administration: HbA1c < 7.0% (52 mmol/mol) and HbA1c 7.0% (52 mmol/mol) to evaluate lipid- and glucose-related values.
Table 1 – Study schedule. Observation period (months)
Time points of case reports Patient background Therapy Concomitant drugs Diet/exercise therapy HbA1c Fasting blood glucose level Blood insulin concentration Serum lipids (TC, LDL-C, TG, HDL-C) CK (CPK) AST, ALT, g-GTP Presence/absence of pregnancy Presence/absence of rhabdomyolysis Presence/absence of adverse events *, required; ~, optional.
Start of administration
1
3
6
9
12
Follow-up at the time of discontinuation
* *
*
*
*
*
*
*
* * * * * *
* * ~ * * *
* * ~ * * *
* * ~ * * *
* * ~ * * *
* * * * * *
* * ~ * * *
diabetes research and clinical practice 106 (2014) 531–537
In this study, NGSP values were used for HbA1c throughout. The data also were reviewed for statin administration in a ‘‘first statin group,’’ and a ‘‘switched statin group.’’ If any fibrate had already been administered to a patient, pitavastain was not allowed for concurrent administration. Accordingly, we stopped using fibrates and the cases were included in the ‘‘switched statin group.’’ The glucose-related values also were compared by dividing the subjects according to BMI < 25 and BMI 25. The study protocol was approved by the Institutional Review Board of Gifu University (No. 17-53). Informed consent was obtained from all participants.
2.1.
Statistical analyses
All values are indicated as mean standard deviation (SD). All data were examined regarding whether or not they were distributed normally. Wilcoxon signed-rank test was used for non-parametric analysis and paired Student-t test was used for parametric analysis of comparison between the before and after pitavastatin treatment groups. Mann–Whitney U test and unpaired Student t-test were performed for non-parametric and parametric analysis of comparison between two groups divided according to HbA1c, statin administration, and BMI. Statistical analysis was performed with JMP1 version 11 software (SAS Institute, Inc., Cary, NC).
3.
Results
3.1.
Study progress
Altogether, 114 subjects at nine institutions were registered in the study. Of these, two withdrew consent and one did not comply with the hospital visits; 111 subjects were analyzed (Fig. 1). Of the 111 registered subjects, six had adverse effects
533
with a percentage of incidence of 5.4%. There was one case of palsy, two of increased serum CPK level, one of liver dysfunction, four of muscle pain, and one of rash; some patients had multiple complications. We stopped using pitavastatin as soon as possible when these adverse effects occurred, after which all patients recovered immediately. We did not observe any side effects in the 86 subjects who finished the study. Table 2 shows the background of subjects whose glucose metabolism was analyzed. There were 39 males and 47 females. In the groups divided by HbA1c as an index of glycemic control at the time of registration, 40 had HbA1c < 7% (52 mmol/mol) and 46 had HbA1c 7% (52 mmol/mol). There were 43 subjects in the first statin group and 43 subjects in the switched statin group. The main complications were hypertension (45.3%), stroke (5.8%), renal disorders (2.3%), arteriosclerosis obliterans (2.3%), and liver disorders (1.2%); 50 of the subjects (58.1%) had at least one complication (Table 3).
3.2.
Effect of statin on glucose metabolism
HbA1c, fasting plasma glucose (FPG), insulin concentration (IRI), HOMA-R, HOMA-b, and BMI were examined. There was no difference in any of these factors between before and after administration of pitavastatin in overall analysis of all subjects (Table 4). We analyzed the correlation between change of HbA1c values and changes of various other parameters by univariate regression analysis, and found a significant correlation only between change of HbA1c and that of TG (P-values: 0.021). When the data were analyzed by dividing subjects by glycemic control into HbA1c < 7% (52 mmol/mol) and HbA1c 7% (52 mmol/mol), by statin treatment into a first statin group and a switched statin group, and by BMI to
Contents available at ScienceDirect
Diabetes Research and Clinical Practice journ al h ome pa ge : www .elsevier.co m/lo cate/diabres
The effects of pitavastatin on glucose metabolism in patients with type 2 diabetes with hypercholesterolemia Hisashi Daido a,b, Yukio Horikawa b,*, Jun Takeda b a b
Hashima City Hospital, Gifu, Japan Department of Diabetes and Endocrinology, Gifu University Graduate School of Medicine, Gifu, Japan
article info
abstract
Article history:
Background: Although there have been several reports that statins cause insulin resistance
Received 5 April 2014
that leads to the occurrence of type 2 diabetes in Caucasians, there has been no Japanese
Received in revised form
prospective studies investigating the effects of statins on the glucose metabolism system.
4 June 2014
Materials and methods: Our subjects were 86 Japanese patients with type 2 diabetes with
Accepted 14 September 2014
hypercholesterolemia. Pitavastatin 2 mg/day was administered for 12 months and the lipid-
Available online 22 October 2014
related values, glucose metabolism values, and the presence/absence of side effects were investigated.
Keywords:
Results: None of these factors was found to differ between before and after administration of
Statin
pitavastatin in overall analysis of all subjects. In subgroup analysis, fasting blood glucose
Glucose metabolism
showed a decrease in the BMI 25 group and there was a significant difference between the
Insulin resistance
BMI < 25 and BMI 25 groups (P-values: 0.021 and 0.0036). Although HbA1c showed an
Insulin secretion
increase both in the group switched to pitavastatin and the BMI < 25 group (P-values: 0.035
Dyslipidemia
and 0.033) and HOMA-b showed a decrease in the BMI < 25 group (P-values: 0.044), there were no significant differences in changes between each divided group and their counterparts. Conclusion: In the Japanese obese group with BMI 25, pitavastatin elicited a significant decrease in fasting blood glucose. It is not clear whether or not this is due to improved insulin resistance as a direct effect of pitavastatin, but in contrast to findings in Caucasians pitavastatin does not worsen insulin resistance in Japanese patients with type 2 diabetes complicated by hypercholesterolemia. # 2014 Elsevier Ireland Ltd. All rights reserved.
1.
Introduction
Dyslipidemia in patients with type 2 diabetes is characterized by the increased production of VLDL, decreased production of HDL, and increased sdLDL (small dense LDL) levels. Free Fatty
Acids (FFAs) from visceral fat are excreted directly through the portal vein, causing high FFA concentrations in the liver and pancreas. Such chronically high FFA levels affect the ability to secrete insulin as well as the action of insulin [1]. Japanese patients with type 2 diabetes are characterized mainly by intrinsic impaired insulin secretion [2], which is well known to
* Corresponding author at: Department of Diabetes and Endocrinology, Gifu University Graduate School of Medicine, 1-1 Yanagido, Gifucity, Gifu 501-1194, Japan. Tel: +81 58 230 6564; fax: +81 58 230 6376. E-mail address: [email protected] (Y. Horikawa). http://dx.doi.org/10.1016/j.diabres.2014.09.048 0168-8227/# 2014 Elsevier Ireland Ltd. All rights reserved.
532
diabetes research and clinical practice 106 (2014) 531–537
be influenced by hyperglycemia and hyper FFA-emia, known as glucotoxicity and lipotoxicity, respectively. Both hyper-FFA-emia and hyper-LDL-emia are known to exert an effect on b cell function. LRP (LDL-receptor related protein), a specific LDL receptor, is expressed in the pancreatic islets, and it is known that pancreatic b cells show increased apoptosis under conditions of hyper-LDL-emia [3,4]. It has been shown in Tangier disease, which is associated with hyper-LDL-emia due to a mutation in the ABCA1 protein, that insulin secretion is decreased and hyperglycemia occurs [5,6]. Statins, which are inhibitors of HMG-CoA reductase, a ratelimiting enzyme of cholesterol synthesis, effectively lower LDL cholesterol (LDL-C). Cardiovascular morbidity, mainly due to dyslipidemia and hypertension, is a major burden in patients with type 2 diabetes. In Caucasians, the efficacy of statins in the prevention of cardiovascular disease has been well documented in long-term outcomes trials involving patients with varying degrees of cardiovascular risk, including patients with type 2 diabetes [7–11]. Thus, it might be expected that statins can both lower LDLC and preserve the insulin secretory function of b cells, but there have been several reports in Caucasians that statins have caused insulin resistance, although the mechanism is unknown, and that this has led to the occurrence of type 2 diabetes [12,13]. Approximately 80% of Japanese patients with type 2 diabetes also have dyslipidemia, so it is particularly important to elucidate the effect of statins on glucose metabolism with patients with type 2 diabetes. The main characteristic of type 2 diabetes in Caucasians is generally obesity-related insulin resistance, while that of Japanese is mainly impaired insulin secretion [2]. The response of glucose metabolism to statins might therefore differ in Caucasians and Japanese. In fact, there have been no prospective Japanese studies investigating whether statins exert any effects on the glucose metabolism system. We performed this study (Gifu Lipid Study) using pitavastatin to investigate the effect of statins on insulin resistance and insulin secretion in Japanese.
2.
Subjects and methods
This study is a multi-center, open-label, single-arm study. The registration period was within one year. The subjects were insulin-naı¨ve patients with type 2 diabetes and HbA1c values of 6.6% [NGSP] (48 mmol/mol [IFCC]) or more and less than 8.8% (72 mmol/mol), which did not vary more than 1% measured within the previous 3 months, who also had hypercholesterolemia with fasting serum total cholesterol (TC) of 5.7 mM (220 mg/dL) or LDL-C of 3.6 mM (140 mg/dL). Even when another statin had already been administered to a patient, if the patient satisfied these conditions, it was considered reasonable to switch drugs, and since this study was conducted within the limits of daily clinical practice, a wash-out period was not used. Patients who satisfied the above conditions and gave written informed consent when the purpose and the possible effects of the study were explained, were registered as subjects by a central registration system. The exclusion criteria as well as the inclusion criteria are shown in supplemental information. Pitavastatin 2 mg/day was administered to the registered patients for 12 months, and the lipid-related values, glucoserelated values, and presence/absence of side effects, as indicated in Table 1, were investigated. If LDL-C was not measured by a direct method, the Friedewald method was used for calculation (LDL-C value = TC value HDL-C value [TG value/5]). The blood specimens were taken mostly at fasting state. If patients were not fasting, only HbA1c, TChol, LDL-C, and HDL-c were used for analysis. The primary endpoint of this study was to examine the response of glucose metabolism to pitavastatin. The main analysis was performed in all subjects who met the inclusion criteria. Sub-analysis was performed in groups divided according HbA1c, lipid-lowering drug naivety, and BMI. Subjects were divided into two groups of glycemic control by HbA1c level at the start of administration: HbA1c < 7.0% (52 mmol/mol) and HbA1c 7.0% (52 mmol/mol) to evaluate lipid- and glucose-related values.
Table 1 – Study schedule. Observation period (months)
Time points of case reports Patient background Therapy Concomitant drugs Diet/exercise therapy HbA1c Fasting blood glucose level Blood insulin concentration Serum lipids (TC, LDL-C, TG, HDL-C) CK (CPK) AST, ALT, g-GTP Presence/absence of pregnancy Presence/absence of rhabdomyolysis Presence/absence of adverse events *, required; ~, optional.
Start of administration
1
3
6
9
12
Follow-up at the time of discontinuation
* *
*
*
*
*
*
*
* * * * * *
* * ~ * * *
* * ~ * * *
* * ~ * * *
* * ~ * * *
* * * * * *
* * ~ * * *
diabetes research and clinical practice 106 (2014) 531–537
In this study, NGSP values were used for HbA1c throughout. The data also were reviewed for statin administration in a ‘‘first statin group,’’ and a ‘‘switched statin group.’’ If any fibrate had already been administered to a patient, pitavastain was not allowed for concurrent administration. Accordingly, we stopped using fibrates and the cases were included in the ‘‘switched statin group.’’ The glucose-related values also were compared by dividing the subjects according to BMI < 25 and BMI 25. The study protocol was approved by the Institutional Review Board of Gifu University (No. 17-53). Informed consent was obtained from all participants.
2.1.
Statistical analyses
All values are indicated as mean standard deviation (SD). All data were examined regarding whether or not they were distributed normally. Wilcoxon signed-rank test was used for non-parametric analysis and paired Student-t test was used for parametric analysis of comparison between the before and after pitavastatin treatment groups. Mann–Whitney U test and unpaired Student t-test were performed for non-parametric and parametric analysis of comparison between two groups divided according to HbA1c, statin administration, and BMI. Statistical analysis was performed with JMP1 version 11 software (SAS Institute, Inc., Cary, NC).
3.
Results
3.1.
Study progress
Altogether, 114 subjects at nine institutions were registered in the study. Of these, two withdrew consent and one did not comply with the hospital visits; 111 subjects were analyzed (Fig. 1). Of the 111 registered subjects, six had adverse effects
533
with a percentage of incidence of 5.4%. There was one case of palsy, two of increased serum CPK level, one of liver dysfunction, four of muscle pain, and one of rash; some patients had multiple complications. We stopped using pitavastatin as soon as possible when these adverse effects occurred, after which all patients recovered immediately. We did not observe any side effects in the 86 subjects who finished the study. Table 2 shows the background of subjects whose glucose metabolism was analyzed. There were 39 males and 47 females. In the groups divided by HbA1c as an index of glycemic control at the time of registration, 40 had HbA1c < 7% (52 mmol/mol) and 46 had HbA1c 7% (52 mmol/mol). There were 43 subjects in the first statin group and 43 subjects in the switched statin group. The main complications were hypertension (45.3%), stroke (5.8%), renal disorders (2.3%), arteriosclerosis obliterans (2.3%), and liver disorders (1.2%); 50 of the subjects (58.1%) had at least one complication (Table 3).
3.2.
Effect of statin on glucose metabolism
HbA1c, fasting plasma glucose (FPG), insulin concentration (IRI), HOMA-R, HOMA-b, and BMI were examined. There was no difference in any of these factors between before and after administration of pitavastatin in overall analysis of all subjects (Table 4). We analyzed the correlation between change of HbA1c values and changes of various other parameters by univariate regression analysis, and found a significant correlation only between change of HbA1c and that of TG (P-values: 0.021). When the data were analyzed by dividing subjects by glycemic control into HbA1c < 7% (52 mmol/mol) and HbA1c 7% (52 mmol/mol), by statin treatment into a first statin group and a switched statin group, and by BMI to
