Accepted Article
Postpartum plasma C-peptide and ghrelin concentrations are predictive of type 2 diabetes in women with previous gestational diabetes mellitus1 Running title: Metabolic hormones in GDM to predict diabetes
Martha LAPPAS 1,2, Debra JINKS 2, Antony UGONI 4, Connie CJ LOUIZOS 2,3,
1
Michael PERMEZEL 2,3, Harry M GEORGIOU 2,3
Obstetrics, Nutrition and Endocrinology Group, Department of Obstetrics and Gynaecology, University of Melbourne, Mercy Hospital for Women, Heidelberg, Victoria, Australia
2
3
Mercy Perinatal Research Centre, Mercy Hospital for Women, Heidelberg, Victoria, Australia
Department of Obstetrics and Gynaecology, University of Melbourne, Mercy Hospital for Women,
4
Heidelberg, Victoria, Australia
School of Physiotherapy, University of Melbourne, Melbourne, Victoria, Australia
Correspondence and Reprint Requests to: Associate Professor Martha Lappas
Department of Obstetrics and Gynaecology, University of Melbourne Mercy Hospital for Women, Level 4/163 Studley Road Heidelberg, 3084, Victoria, Australia Ph: 61-3-8458 4370; Fax: 61-3-8458 4380 E-mail: [email protected]
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1753-0407.12209 This article is protected by copyright. All rights reserved.
2
Accepted Article
ABSTRACT
Objectives: Women with previous gestational diabetes mellitus (pGDM) are at increased risk of
developing type 2 diabetes later in life. The aim of this study was to determine if circulating levels
of metabolic hormones 12 weeks following a GDM pregnancy are associated with an increased risk of type 2 diabetes 8-10 years later.
Study design: Fasting plasma concentrations of glucose, insulin, C-peptide, ghrelin, GIP, GLP-1, glucagon, leptin, PAI-1, resistin and visfatin were measured in 98 normal glucose tolerant women, 12 weeks following an index GDM pregnancy. Women were assessed every 2 years for up to 10 years for development of overt type 2 diabetes.
Results: After a median follow-up period of 8.7 years, 22.5% of women with a pGDM pregnancy developed type 2 diabetes. Significant risk factors for the development of type 2 diabetes were fasting plasma glucose levels > 5 mmol/L during pregnancy and at 12 weeks post pregnancy. In addition, higher C-peptide levels and lower ghrelin levels at 12 weeks post pregnancy were also significant risk factors for the development of type 2 diabetes.
Conclusions: Fasting plasma glucose during pregnancy and postpartum, and post-partum C-peptide
and ghrelin levels were significant risk factors for the development of type 2 diabetes in women with pGDM. This is the first report that identifies C-peptide and ghrelin as potential biomarkers for the prediction of type 2 diabetes in women with a history of GDM.
The significant findings of the study: High postpartum C-peptide, and low postpartum ghrelin levels are significant risk factors for the development of type 2 diabetes in women with a previous history of GDM. This article is protected by copyright. All rights reserved.
Accepted Article
3
This study adds: This study has identified two biochemical markers that can predict the
development of type 2 diabetes in women with previous GDM. The early detection of those at increased risk provides an opportunity for early treatment to prevent type 2 diabetes.
Keywords: GDM; type 2 diabetes; metabolic hormones
Abbreviations: pGDM, previous gestational diabetes mellitus; OGTT, oral glucose tolerance test; GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide-1; PAI-1, plasminogen activator inhibitor-1
This article is protected by copyright. All rights reserved.
4
Accepted Article
INTRODUCTION
Internationally the burden of disease that can be attributed to diabetes is increasing; with substantial societal and financial costs 1. Alarmingly, it is predicted that approximately 350 million people in the world will be diagnosed with type 2 diabetes by the year 2030; double the current number. Women with a previous history of gestational diabetes mellitus (pGDM) have a significantly increased risk of developing type 2 diabetes and cardiovascular disease later in life. Several studies performed early after pregnancy, have identified an association of glucose intolerance with increased risk of cardiovascular disease and type 2 diabetes 2. We have previously reported that the
risk of developing diabetes is 9.6 times greater for patients with pGDM, with the cumulative risk being 25% at 15 years post diagnosis 2. Thus, GDM may be an important predictor to the type 2 diabetes epidemic in many populations.
Identifying women with GDM as a high-risk group for subsequent disease offers an opportunity to alter women’s future health. Indeed, lifestyle intervention and/or medications have been shown to prevent or delay the development of type 2 diabetes in women with pGDM 3, 4. Higher fasting
plasma glucose levels during pregnancy, after adjustments for BMI and age, have been shown to be predictors of the metabolic syndrome 40 months 5 and 8.5 years 6 after the index pregnancy. Genetic
variations have also recently been evaluated to determine if they can be used to predict postpartum diabetes in women with GDM 7, 8. However, there have been no studies that have examined biochemical markers in addition to the postnatal OGTT. Thus, the purpose of this study was to identify if known diabetes markers (insulin, C-peptide, ghrelin, gastric inhibitory polypeptide (GIP), glucagon-like peptide-1 (GLP-1), glucagon, leptin, plasminogen activator inhibitor-1 (PAI-1), resistin and visfatin) can be used as predictive factors for the subsequent development of type 2 diabetes in a cohort of women with a pGDM pregnancy.
This article is protected by copyright. All rights reserved.
5
Accepted Article
MATERIALS AND METHODS
Patient recruitment and sample collection The study was approved by the Mercy Health Research and Ethics Committee and written informed consent was obtained from all participants. Women were eligible for this study if they had a pGDM pregnancy, aged at least 18 years and English speaking. Women with pre-existing diabetes were excluded.
Between June 2003 and December 2005, 118 women were recruited for this study by a Research coordinator and for all participants, this was their first pregnancy diagnosed with GDM. The diagnosis of GDM was made by an oral glucose tolerance test (OGTT) at 24-28 weeks gestation, according to the Australasian Diabetes in Pregnancy Society (ADIPS) guidelines by either a fasting venous plasma concentration of ≥5.5 mmol/L glucose and/or ≥8.0 mmol/L glucose 2 h after a 75 g
oral glucose load. All women diagnosed with GDM met with the Hospital’s dietician and were advised to follow the recommended Standard of Care diet for controlling blood glucose (40% carbohydrate, 15% protein, and 45% fat).
The women who participated in this study had an OGTT performed at 12 weeks postnatally, and only those women that had a normal glucose tolerance were further evaluated for this study. The classification of non-pregnant glucose tolerance status was made using the American Diabetes Association criteria using the following venous plasma glucose concentrations: fasting ≥7.0 mmol/L
or ≥11.1 mmol/L after 2 h after a 75 g oral glucose load 9; the criteria remained consistent during the duration of the study. At the time of the postnatal OGTT, women had their weight and height measured and a blood sample was taken for biomarker evaluation. The participants were followed up every one to two years for 8-10 years. During these follow-up visits, the participants underwent
This article is protected by copyright. All rights reserved.
6
Accepted Article
further OGTT. Upon diagnosis of type 2 diabetes, the women were referred to their doctor for further specialist care.
Blood was collected postnatally at the time of the follow-up OGTT, at approximately 12 weeks after delivery. Blood samples were centrifuged at 1,000 g for 10 min, plasma was supplemented
with 0.1 mM phenylmethylsulfonyl fluoride (PMSF) protease inhibitor (USB, Cleveland, OH) and immediately stored at -80ºC until assayed by multiplex assay as detailed below.
Diabetes multiplex assay Metabolic hormone determinations were performed using the Bio-Plex suspension assay system and a 10-plex human plasma Diabetes Assay Kit (Bio-Rad Laboratories, Hercules, CA) according to the manufacturers’ instructions. The analytes measured were insulin, C-peptide, ghrelin, gastric inhibitory polypeptide (GIP), glucagon-like peptide-1 (GLP-1), glucagon, leptin, plasminogen activator inhibitor-1 (PAI-1), resistin and visfatin. The assay utilised 50 μL of plasma at 1/4 dilution. The limit of detection varied for each marker and ranged from 1.5 to 40 pg/mL. The calculated inter-assay and intra-assay coefficients of variation (CV) were all less than 8%.
Statistical analysis The significance of predictors of diabetes status was determined through the application of logistic regression using the median value to dichotomise each parameter. The median value was utilised as there was an unequal number of ‘normal’ and ‘diabetic’ subjects and data for each predictor were not normally distributed. Univariate models were fitted, where each predictor was considered in isolation to all other predictors. A multivariate model, where predictors were adjusted for age and BMI is included. Results are presented as odds ratios with corresponding confidence intervals (95%) and P-values. P 5 mmol/L at the time of GDM diagnosis, postnatal OGTT fasting glucose plasma concentrations > 5 mmol/L, postnatal C-peptide levels > 450 pg/ml and postnatal ghrelin levels ≤ 400 pg/mL were associated with significantly increased odds for the development of type 2 diabetes. In the multivariate analysis, after adjusting for age and BMI, these variables remained significant risk factors for the development of type 2 diabetes.
This article is protected by copyright. All rights reserved.
8
Accepted Article
DISCUSSION
This study revealed that 22/98 (22.5%) women with a pGDM pregnancy developed type 2 diabetes during a median follow-up period of 8.7 years. Fasting plasma glucose during pregnancy and postpartum, and post-partum C-peptide and ghrelin levels were significant risk factors for the development of type 2 diabetes in women with pGDM. This is the first report that identifies Cpeptide and ghrelin as potential biomarkers for the prediction of type 2 diabetes in women with a history of GDM.
Due to the limited sample size of this study (n=98), it is not practical to include a large number of confounding variables into the multivariate analysis. In this study, only maternal age and BMI were considered but these did not impact on the significance of the dependent variables measured. Of note, the study was sufficiently powered to identify a 2-fold difference between group means with a standard deviation equal to the mean (alpha error 0.05% and power 80%).
Women with a normal glucose tolerance and prior GDM have altered glucose homeostasis as characterised by insulin resistance and/or beta cell dysfunction. Higher fasting plasma glucose levels during pregnancy, after adjustments for BMI and age, have been shown to be predictors of the metabolic syndrome 40 months 5 and 8.5 years 6 after the index pregnancy. Furthermore, a high
fasting plasma glucose level at the diagnostic OGTT in pregnancy and hyperglycemia during the postpartum OGTT are predictive for subsequent development of overt diabetes 10. In agreement, after adjusting for BMI and age, our study indicates that fasting plasma glucose levels > 5 mmol/L during pregnancy and postpartum are significant risk factors for the development of diabetes 8.5
years post index pregnancy.
This article is protected by copyright. All rights reserved.
9
Accepted Article
High C-peptide and low ghrelin levels were found to be significant risk factors for developing type 2 diabetes. These relationships were maintained after adjustment for age and BMI. Interestingly, other studies have shown that C-peptide is a risk factor for other diseases, including cancer 11, and
all-cause cardiovascular-related disease and coronary artery–related mortality among adults without diabetes 12. C-peptide is a small peptide that is cleaved from proinsulin in the synthesis of active insulin. Although there was a significant correlation between insulin and C-peptide (r=0.7, P 30
51
18
26.1
≤ 25
41
7
14.6
> 25
35
15
30.0
No
46
10
17.9
Yes
30
12
28.6
≤5
56
8
12.5
>5
20
14
41.2
≤5
14
4
22.2
>5
62
18
22.5
≤5
64
12
15.8
>5
12
10
45.5
≤ 6.5
46
8
14.8
> 6.5
30
14
31.8
≤5
36
8
18.2
>5
40
14
25.9
≤ 450
44
6
12.0
> 450
32
16
33.3
This article is protected by copyright. All rights reserved.
UNIVARIATE
MULTIVARIATE
95% Odds Confidence Ratio Interval
P value
Odds Ratio
95% Confidence Interval
P value
2.21
0.67-7.21
0.19
2.51
0.92-6.85
0.07
1.84
0.71-4.79
0.21
4.90
1.79-13.42
55
41
10
19.6
≤ 45
31
11
26.2
> 45
45
11
19.6
≤ 35
35
13
27.1
> 35
41
9
18.0
≤ 45
39
8
17.0
> 45
37
14
27.5
≤ 1,000
41
9
18.0
> 1,000
35
13
27.1
≤ 3,500
43
11
20.4
> 3,500
33
11
25.0
≤ 850
37
12
24.5
> 850
39
10
20.4
≤ 900
33
11
25.0
> 900
43
11
20.4
0.21
0.07-0.64
Postpartum plasma C-peptide and ghrelin concentrations are predictive of type 2 diabetes in women with previous gestational diabetes mellitus1 Running title: Metabolic hormones in GDM to predict diabetes
Martha LAPPAS 1,2, Debra JINKS 2, Antony UGONI 4, Connie CJ LOUIZOS 2,3,
1
Michael PERMEZEL 2,3, Harry M GEORGIOU 2,3
Obstetrics, Nutrition and Endocrinology Group, Department of Obstetrics and Gynaecology, University of Melbourne, Mercy Hospital for Women, Heidelberg, Victoria, Australia
2
3
Mercy Perinatal Research Centre, Mercy Hospital for Women, Heidelberg, Victoria, Australia
Department of Obstetrics and Gynaecology, University of Melbourne, Mercy Hospital for Women,
4
Heidelberg, Victoria, Australia
School of Physiotherapy, University of Melbourne, Melbourne, Victoria, Australia
Correspondence and Reprint Requests to: Associate Professor Martha Lappas
Department of Obstetrics and Gynaecology, University of Melbourne Mercy Hospital for Women, Level 4/163 Studley Road Heidelberg, 3084, Victoria, Australia Ph: 61-3-8458 4370; Fax: 61-3-8458 4380 E-mail: [email protected]
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1753-0407.12209 This article is protected by copyright. All rights reserved.
2
Accepted Article
ABSTRACT
Objectives: Women with previous gestational diabetes mellitus (pGDM) are at increased risk of
developing type 2 diabetes later in life. The aim of this study was to determine if circulating levels
of metabolic hormones 12 weeks following a GDM pregnancy are associated with an increased risk of type 2 diabetes 8-10 years later.
Study design: Fasting plasma concentrations of glucose, insulin, C-peptide, ghrelin, GIP, GLP-1, glucagon, leptin, PAI-1, resistin and visfatin were measured in 98 normal glucose tolerant women, 12 weeks following an index GDM pregnancy. Women were assessed every 2 years for up to 10 years for development of overt type 2 diabetes.
Results: After a median follow-up period of 8.7 years, 22.5% of women with a pGDM pregnancy developed type 2 diabetes. Significant risk factors for the development of type 2 diabetes were fasting plasma glucose levels > 5 mmol/L during pregnancy and at 12 weeks post pregnancy. In addition, higher C-peptide levels and lower ghrelin levels at 12 weeks post pregnancy were also significant risk factors for the development of type 2 diabetes.
Conclusions: Fasting plasma glucose during pregnancy and postpartum, and post-partum C-peptide
and ghrelin levels were significant risk factors for the development of type 2 diabetes in women with pGDM. This is the first report that identifies C-peptide and ghrelin as potential biomarkers for the prediction of type 2 diabetes in women with a history of GDM.
The significant findings of the study: High postpartum C-peptide, and low postpartum ghrelin levels are significant risk factors for the development of type 2 diabetes in women with a previous history of GDM. This article is protected by copyright. All rights reserved.
Accepted Article
3
This study adds: This study has identified two biochemical markers that can predict the
development of type 2 diabetes in women with previous GDM. The early detection of those at increased risk provides an opportunity for early treatment to prevent type 2 diabetes.
Keywords: GDM; type 2 diabetes; metabolic hormones
Abbreviations: pGDM, previous gestational diabetes mellitus; OGTT, oral glucose tolerance test; GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide-1; PAI-1, plasminogen activator inhibitor-1
This article is protected by copyright. All rights reserved.
4
Accepted Article
INTRODUCTION
Internationally the burden of disease that can be attributed to diabetes is increasing; with substantial societal and financial costs 1. Alarmingly, it is predicted that approximately 350 million people in the world will be diagnosed with type 2 diabetes by the year 2030; double the current number. Women with a previous history of gestational diabetes mellitus (pGDM) have a significantly increased risk of developing type 2 diabetes and cardiovascular disease later in life. Several studies performed early after pregnancy, have identified an association of glucose intolerance with increased risk of cardiovascular disease and type 2 diabetes 2. We have previously reported that the
risk of developing diabetes is 9.6 times greater for patients with pGDM, with the cumulative risk being 25% at 15 years post diagnosis 2. Thus, GDM may be an important predictor to the type 2 diabetes epidemic in many populations.
Identifying women with GDM as a high-risk group for subsequent disease offers an opportunity to alter women’s future health. Indeed, lifestyle intervention and/or medications have been shown to prevent or delay the development of type 2 diabetes in women with pGDM 3, 4. Higher fasting
plasma glucose levels during pregnancy, after adjustments for BMI and age, have been shown to be predictors of the metabolic syndrome 40 months 5 and 8.5 years 6 after the index pregnancy. Genetic
variations have also recently been evaluated to determine if they can be used to predict postpartum diabetes in women with GDM 7, 8. However, there have been no studies that have examined biochemical markers in addition to the postnatal OGTT. Thus, the purpose of this study was to identify if known diabetes markers (insulin, C-peptide, ghrelin, gastric inhibitory polypeptide (GIP), glucagon-like peptide-1 (GLP-1), glucagon, leptin, plasminogen activator inhibitor-1 (PAI-1), resistin and visfatin) can be used as predictive factors for the subsequent development of type 2 diabetes in a cohort of women with a pGDM pregnancy.
This article is protected by copyright. All rights reserved.
5
Accepted Article
MATERIALS AND METHODS
Patient recruitment and sample collection The study was approved by the Mercy Health Research and Ethics Committee and written informed consent was obtained from all participants. Women were eligible for this study if they had a pGDM pregnancy, aged at least 18 years and English speaking. Women with pre-existing diabetes were excluded.
Between June 2003 and December 2005, 118 women were recruited for this study by a Research coordinator and for all participants, this was their first pregnancy diagnosed with GDM. The diagnosis of GDM was made by an oral glucose tolerance test (OGTT) at 24-28 weeks gestation, according to the Australasian Diabetes in Pregnancy Society (ADIPS) guidelines by either a fasting venous plasma concentration of ≥5.5 mmol/L glucose and/or ≥8.0 mmol/L glucose 2 h after a 75 g
oral glucose load. All women diagnosed with GDM met with the Hospital’s dietician and were advised to follow the recommended Standard of Care diet for controlling blood glucose (40% carbohydrate, 15% protein, and 45% fat).
The women who participated in this study had an OGTT performed at 12 weeks postnatally, and only those women that had a normal glucose tolerance were further evaluated for this study. The classification of non-pregnant glucose tolerance status was made using the American Diabetes Association criteria using the following venous plasma glucose concentrations: fasting ≥7.0 mmol/L
or ≥11.1 mmol/L after 2 h after a 75 g oral glucose load 9; the criteria remained consistent during the duration of the study. At the time of the postnatal OGTT, women had their weight and height measured and a blood sample was taken for biomarker evaluation. The participants were followed up every one to two years for 8-10 years. During these follow-up visits, the participants underwent
This article is protected by copyright. All rights reserved.
6
Accepted Article
further OGTT. Upon diagnosis of type 2 diabetes, the women were referred to their doctor for further specialist care.
Blood was collected postnatally at the time of the follow-up OGTT, at approximately 12 weeks after delivery. Blood samples were centrifuged at 1,000 g for 10 min, plasma was supplemented
with 0.1 mM phenylmethylsulfonyl fluoride (PMSF) protease inhibitor (USB, Cleveland, OH) and immediately stored at -80ºC until assayed by multiplex assay as detailed below.
Diabetes multiplex assay Metabolic hormone determinations were performed using the Bio-Plex suspension assay system and a 10-plex human plasma Diabetes Assay Kit (Bio-Rad Laboratories, Hercules, CA) according to the manufacturers’ instructions. The analytes measured were insulin, C-peptide, ghrelin, gastric inhibitory polypeptide (GIP), glucagon-like peptide-1 (GLP-1), glucagon, leptin, plasminogen activator inhibitor-1 (PAI-1), resistin and visfatin. The assay utilised 50 μL of plasma at 1/4 dilution. The limit of detection varied for each marker and ranged from 1.5 to 40 pg/mL. The calculated inter-assay and intra-assay coefficients of variation (CV) were all less than 8%.
Statistical analysis The significance of predictors of diabetes status was determined through the application of logistic regression using the median value to dichotomise each parameter. The median value was utilised as there was an unequal number of ‘normal’ and ‘diabetic’ subjects and data for each predictor were not normally distributed. Univariate models were fitted, where each predictor was considered in isolation to all other predictors. A multivariate model, where predictors were adjusted for age and BMI is included. Results are presented as odds ratios with corresponding confidence intervals (95%) and P-values. P 5 mmol/L at the time of GDM diagnosis, postnatal OGTT fasting glucose plasma concentrations > 5 mmol/L, postnatal C-peptide levels > 450 pg/ml and postnatal ghrelin levels ≤ 400 pg/mL were associated with significantly increased odds for the development of type 2 diabetes. In the multivariate analysis, after adjusting for age and BMI, these variables remained significant risk factors for the development of type 2 diabetes.
This article is protected by copyright. All rights reserved.
8
Accepted Article
DISCUSSION
This study revealed that 22/98 (22.5%) women with a pGDM pregnancy developed type 2 diabetes during a median follow-up period of 8.7 years. Fasting plasma glucose during pregnancy and postpartum, and post-partum C-peptide and ghrelin levels were significant risk factors for the development of type 2 diabetes in women with pGDM. This is the first report that identifies Cpeptide and ghrelin as potential biomarkers for the prediction of type 2 diabetes in women with a history of GDM.
Due to the limited sample size of this study (n=98), it is not practical to include a large number of confounding variables into the multivariate analysis. In this study, only maternal age and BMI were considered but these did not impact on the significance of the dependent variables measured. Of note, the study was sufficiently powered to identify a 2-fold difference between group means with a standard deviation equal to the mean (alpha error 0.05% and power 80%).
Women with a normal glucose tolerance and prior GDM have altered glucose homeostasis as characterised by insulin resistance and/or beta cell dysfunction. Higher fasting plasma glucose levels during pregnancy, after adjustments for BMI and age, have been shown to be predictors of the metabolic syndrome 40 months 5 and 8.5 years 6 after the index pregnancy. Furthermore, a high
fasting plasma glucose level at the diagnostic OGTT in pregnancy and hyperglycemia during the postpartum OGTT are predictive for subsequent development of overt diabetes 10. In agreement, after adjusting for BMI and age, our study indicates that fasting plasma glucose levels > 5 mmol/L during pregnancy and postpartum are significant risk factors for the development of diabetes 8.5
years post index pregnancy.
This article is protected by copyright. All rights reserved.
9
Accepted Article
High C-peptide and low ghrelin levels were found to be significant risk factors for developing type 2 diabetes. These relationships were maintained after adjustment for age and BMI. Interestingly, other studies have shown that C-peptide is a risk factor for other diseases, including cancer 11, and
all-cause cardiovascular-related disease and coronary artery–related mortality among adults without diabetes 12. C-peptide is a small peptide that is cleaved from proinsulin in the synthesis of active insulin. Although there was a significant correlation between insulin and C-peptide (r=0.7, P 30
51
18
26.1
≤ 25
41
7
14.6
> 25
35
15
30.0
No
46
10
17.9
Yes
30
12
28.6
≤5
56
8
12.5
>5
20
14
41.2
≤5
14
4
22.2
>5
62
18
22.5
≤5
64
12
15.8
>5
12
10
45.5
≤ 6.5
46
8
14.8
> 6.5
30
14
31.8
≤5
36
8
18.2
>5
40
14
25.9
≤ 450
44
6
12.0
> 450
32
16
33.3
This article is protected by copyright. All rights reserved.
UNIVARIATE
MULTIVARIATE
95% Odds Confidence Ratio Interval
P value
Odds Ratio
95% Confidence Interval
P value
2.21
0.67-7.21
0.19
2.51
0.92-6.85
0.07
1.84
0.71-4.79
0.21
4.90
1.79-13.42
55
41
10
19.6
≤ 45
31
11
26.2
> 45
45
11
19.6
≤ 35
35
13
27.1
> 35
41
9
18.0
≤ 45
39
8
17.0
> 45
37
14
27.5
≤ 1,000
41
9
18.0
> 1,000
35
13
27.1
≤ 3,500
43
11
20.4
> 3,500
33
11
25.0
≤ 850
37
12
24.5
> 850
39
10
20.4
≤ 900
33
11
25.0
> 900
43
11
20.4
0.21
0.07-0.64
