CT-based measurement of visceral adipose tissue volume as a reliable tool for assessing metabolic risk factors in prediabetes across subtypes - Scientific Reports

Prediabetes, a high-risk factor for the development of type 2 diabetes mellitus (T2DM), is defined as a blood glucose level higher than normal but not yet at the diabetes diagnostic threshold, and is associated with an increased risk of cardiovascular disease (CVD)^1,2,3. Visceral obesity is an important risk factor for CVD and T2DM, and the metabolic results of visceral obesity are associated with insulin resistnace^4,5,6. Insulin resistance and a decline in pancreatic beta cell function are major pathophysiological factors for the development of diabetes in individuals with prediabetes⁷. The evaluation of visceral obesity may be a useful approach for predicting the risk of developing diabetes and CVD among individuals who have been diagnosed with prediabetes.

Based on the results of the oral glucose tolerance test (OGTT), prediabetes is classified as impaired fasting glucose (IFG), impaired glucose tolerance (IGT), or a combination of both (IFG and IGT). Different responses to the OGTT arise from differing pathophysiological mechanisms, including variations in the site of insulin resistance and insulin secretion patterns, among the different prediabetes subtypes⁸. Thus, the correlation between insulin resistance and visceral obesity may differ according to prediabetes subtype.

Anthropometric measurements such as body mass index (BMI), waist circumference (WC), and waist-to-hip ratio (W/H ratio) have been used in clinical settings to evaluate visceral obesity. However, given their indirect assessment of visceral adipose tissue (VAT) and racial/ethnic difference in the relationships between WC or BMI and VAT, these measurements have limitations⁹. Consequently, computed tomography (CT) is the gold standard for accurate quantitative evaluation of VAT¹⁰. VAT is an active endocrine organ that releases adipocytokines and is associated with inflammatory markers, hepatic steatosis, and metabolic risk factors, all of which may increase the risk of developing metabolic syndrome and T2DM^11,12,13. Accurate measurement of VAT is therefore crucial for assessing the risk of diabetes and CVD among individuals with prediabetes.

In this study, we compared VAT volume measured by CT to other adiposity indices, in individuals with prediabetes. Furthermore, we evaluated the associations between adiposity indices and insulin resistance among individuals with prediabetes, based on the subtypes.

Study design and population

The electronic records of individuals with prediabetes at Pusan National University Hospital (Busan, Republic of Korea) were retrospectively reviewed between December 2017 and December 2019. Prediabetes was defined according to the American Diabetes Association criteria; therefore, we included those subjects with a fasting plasma glucose (IFG) level between 100–125 mg/dL or with a 2-h glucose level (IGT) between 140 and 199 mg/dL during a 75 g OGTT, and had a hemoglobin A1c (HbA1c) levels of 6.4% or below¹. Individuals with prediabetes aged ≥ 20 years, who had abdominal CT screening for health check-up, suspected pancreatic disease, or follow-up imaging test for cancer, etc., were included in the study. Individuals were excluded from the study if they had undergone abdominal surgery or antidiabetic medications. The patients who are at the ongoing or progressive status of malignant disease, active fever, visible edema or active systemic comorbidity such as chronic renal disease, cirrhosis and heart failure which could significantly affect anthropometric and quantitative CT measurement were also excluded. Individuals with prediabetes were classified into subgroups according to prediabetes type (isolated IFG [I-IFG], isolated IGT [I-IGT], or combined IFG and IGT [C-IFG/IGT]). This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board named Pusan National University Hospital (IRB no. 2105-006-102, protocol code: 2021-NM-001). The requirement of written informed consent from enrolled subjects was waived by the Institutional Review Board of Pusan National University Hospital due to the retrospective study design.

CT protocol and CT-based adiposity indices

Unenhanced spiral CT was performed using a Philips Brilliance 16-slice multidetector helical CT scanner (GEMINI TF CT; Philips, Eindhoven, Netherlands) at a voltage of 120 kVp, with a 3-mm slice thickness from the diaphragm to the mid-thigh. The VAT was manually defined as ranging from the diaphragm to the pelvic floor. The VAT volume was calculated by setting the attenuation values from -45 to -195 Hounsfield units (HU) using CT software (SIEMENS; Syngo CT basic evaluation). The WC and hip circumference (HC) were also measured using CT images. The perimeter of the free-drawn region of interest (ROI) at the umbilicus level was defined as the WC, and the ROI at the greater trochanter level of both femurs was defined as the HC (Fig. 1). The W/H ratio was calculated as the WC/HC ratio. The liver attenuation index (LAI) was calculated as follows: the average HU of five ROIs with 2–3 mm diameters within the liver was divided by the average HU of three ROIs within the spleen¹⁴.

An example of region of interest (ROI)s on computed tomography. ROIs are used to obtain the waist circumference (WC) and hip circumference, and are manually drawn on the transaxial non-enhanced computed tomography image at the umbilicus level for WC (A) and the greater trochanter level of the femur (B).

Full size image

Laboratory assessments and metabolic indices

Height (cm) and weight (kg) were measured using an automatic height-weight scale (GL-150; G-tech, Seoul, Korea). Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared.

Venous blood samples were collected from all participants in the morning after 12 h of overnight fasting. Biochemical assays were performed using an autoanalyzer (Hitachi 747; Hitachi Corp., Tokyo, Japan). Lipid profiles included total cholesterol (TC; mg/dL), triglycerides (TG; mg/dL), high-density lipoprotein (HDL) cholesterol (mg/dL), and low-density lipoprotein (LDL) cholesterol (mg/dL). Fasting plasma glucose (mg/dL) was measured by the glucose oxidase method using a Synchron LX20 (Beckman Coulter; Fullerton, CA, USA). Plasma insulin level (µU/mL) was determined using an enzyme immunoassay (Dainabot; Tokyo, Japan).

Insulin resistance was calculated using the homeostasis model – insulin resistance (HOMA-IR), by the following formula: fasting insulin (µU/mL) × plasma glucose (mg/dL)/405 and Matsuda insulin sensitivity index: calculated as 10,000/(fasting glucose (mg/dL) × fasting insulin (μU/mL) × mean glucose during OGTT (mg/dL) × mean insulin during OGTT (μU/mL))^0.5. Insulin secretion was determined using the homeostasis model–beta-cell function (HOMA-beta), by the following formula: HOMA-beta = 360 × fasting insulin (μU/mL)/(fasting glucose [mg/Dl]—63)^15,16.

The visceral adiposity index (VAI) was defined as follows: men: VAI = (WC/[39.68 + 1.88 × BMI]) × (TG /1.03) × (1.31/HDL); women: VAI = (WC/[36.58 + 1.89 × BMI]) × TG/ 0.81 × (1.52/HDL).

Statistical analyses

All non-normally distributed variables are expressed as medians and interquartile ranges (IQRs; 25–75%). The Mann–Whitney U test was used to compare continuous variables between the two groups. Categorical group data were compared using the Chi-squared test. Spearman's rank correlation was applied to analyze the degree of association between the variables. To evaluate the relationship between adiposity or metabolic indices and HOMA-IR, we conducted a stepwise multiple linear regression by considering a set of potential variables. The set of variables includes BMI, WC, W/H ratio, LAI, VAI, VAT area and VAT volume, which were revealed as significant variables from the univariate analysis. Statistical analyses were performed using MedCalc® (version 16.4.3; MedCalc, Mariakerke, Belgium) and GraphPad Prism (ver 8.0; GraphPad Software, Inc., San Diego, CA, USA). Statistical significance was set at p < 0.05.

Baseline characteristics and differences based on prediabetes subtypes

A total of 78 individuals diagnosed with prediabetes were enrolled (M:F = 14:64; median age, 60 years; IQR, 56–64 years). All study subjects were of Korean ethnicity. Among individuals with prediabetes, 32 (41.0%) had either I-IFG or I-IGT, while 46 (59.0%) had C-IFG/IGT. Approximately 30% (n = 24) of the individuals were treated for hypertension, and their systolic and diastolic blood pressures were relatively well controlled (Table 1).

Table 2 presents the individual characteristics based on subgroup analysis. Individuals with C-IFG/IGT showed significantly higher serum HbA1c levels and HOMA-IR values. Individuals with C-IFG/IGT shows a lower tendency in HOMA-beta and Matsuda insulin sensitivity index results compared to those with I-IFG or I-IGT, but the observed differences did not reach statistical significance. Compared to the individuals in the other prediabetes subtype groups, those with C-IFG/IGT had significantly higher WCs, W/H ratios, VAT areas, VAT volumes, and VAI values, as well as significantly lower LAI values.

Multiple regression analysis between adiposity indices and insulin resistance

In our regression analysis between CT-based adiposity indices and insulin resistance, all adiposity indices were associated with HOMA-IR and Matsuda insulin sensitivity index in univariate analyses. However, multiple regression analyses showed that the W/H ratio (β = 0.307, p = 0.014) and VAT volume (β = 0.307, p = 0.010) remained independent predictors of HOMA-IR. For Matsuda insulin sensitivity index, only VAT volume (β = − 1.210, p = 0.004) was the only significant factor in individuals with prediabetes (Table 3).

Heterogeneous associations in adiposity indicators according to prediabetes types

In all individuals with prediabetes, we evaluated the correlations among all indices studied, including CT-based adiposity indices, HOMA-IR, and others. The weakest correlation was observed for serum HbA1c levels (Fig. 2A). During the subgroup analysis, significant associations were observed between the indices studied in individuals with I-IFG or I-GT (Fig. 2B). However, in individuals with C-IFG/IGT, the association between most of the indices studied were found to be insignificant and heterogeneous (Fig. 2C). Notably, VAT volume and LAI showed significant correlations with the other indices.

The heat-maps of colored by the p value-based significance of associations between the adiposity and metabolic indices. The lightest green color indicates p value of 0 and darkest red indicates p value of 1, which are presented as vertical shaped rectangle on the right side of the maps. Among the individuals with prediabetes, all adiposity and metabolic indices were significantly correlated with one another. Hemoglobin A1c showed the weakest correlation with the other metabolic parameters (A). Among individuals with isolated (I)-impaired fasting glucose (IFG) or I-impaired glucose tolerance (IGT), all adiposity and metabolic indices were significantly correlated with one another, except hemoglobin A1c (B). Among individuals with combined IFG/IGT, the associations between most of the adiposity and metabolic indices were insignificant except visceral adipose tissue volume and liver attenuation index (C).

Full size image

This study evaluated the relationships between various obesity or metabolic indices, such as VAT volume and insulin resistance, according to prediabetes subtypes. Despite showing no differences in BMIs, individuals with the C-IFG/IGT prediabetes subtype showed higher VAT volumes, VAT areas, WCs, and W/H ratios compared to those with the I-IFG or I-IGT subtypes. Moreover, a significant difference was observed in the visceral adiposity functional indices (VAT volume and LAI) between I-IFG or I-IGT and C-IFG/IGT. In this study, VAT volume demonstrated a better correlation with insulin resistance, a crucial factor in obesity-related comorbidities, compared to other obesity indices, across all prediabetes subtypes. These results were consistent with previous reports that showed that excess visceral fat was associated with prediabetes and T2DM, and that VAT volume was a stronger predictor for it than other anthropometric obesity indices^17,18,19.

After stratifying for prediabetes subtypes according to I-IFG/I-IGT and C-IFG/IGT, the C-IFG/IGT group was found to exhibit significantly higher HOMA-IR and adiposity indices such as W/H ratio, VAT volume, and VAT area, when compared to the I-IFG and I-IGT groups. Prediabetes is a heterogeneous disease according to each patient’s response to the OGTT, and the pathophysiologic characteristics of IFG and IGT are considerably different in terms of insulin resistance and insulin secretion. IFG has a higher hepatic insulin resistance and reduced early phase (first 30 min) insulin response to oral glucose, whereas IGT has a higher muscle insulin resistance and reduced early and late phase (60–120 min) insulin secretion^20,21,22. C-IFG/IGT have characteristics of both conditions, resulting in a higher risk of developing DM compared to I-IFG or I-IGT alone⁸. A previous study demonstrated that individuals with C-IFG/IGT had a significantly higher HOMA-IR compared to those with either I-IFG or I-IGT subtypes, whereas no significant differences were observed in BMI or W/H ratio²³. In this study, individuals with C-IFG/IGT had significantly higher WCs, W/H ratios, and VAT volumes, compared to those with either the I-IFG or I-IGT subtypes. Previous studies have demonstrated variations in the association between anthropometric measures and risk factors in prediabetes, which is attributed to differences in measurement protocols across studies and ethnic variations. Specifically, the higher percentage of body fat observed in Asians at lower WC and BMI levels may contribute to these differences^24,25. Notably, in this study, the correlations between most of the evaluated indices were diminished among individuals with C-IFG/IGT. However, CT-based measurement of VAT volume and LAI remained significantly associated with the others. Therefore, assessment of VAT volume and LAI may be valuable for evaluating metabolic risk in individuals with prediabetes, irrespective of their prediabetes subtypes.

CT-based VAT measurement is a potentially more precise and more objective method for VAT assessment than WC or W/H ratio but also has its limitations, including radiation exposure and high costs. However, volumetric quantification of VAT is highly reproducible, and there are significant differences between volumetric and single-slice measurements²⁶. Abdominal CT scans have been frequently performed for health check-up, abnormal laboratory test suspected pancreatic disease in prediabetes, follow-up imaging test for cancer, and other medical conditions for the diagnosis and follow-up of diseases that affect abdominal organs in individuals with prediabetes. Several studies have already reported that the abdominal CT is useful for the patients with diabetes to assess the pancreatic tumor or pancreatic endocrine function^27,28,29,30. Therefore, utilizing these abdominal CT images for VAT measurements can provide additional cardiometabolic information without additional radiation exposure as the enrolled participants in the current study could be evaluated for their CT-based VAT measurement using already acquired their abdominal CT.

In addition, NAFLD is highly prevalent in T2DM patients and is an early predictor of T2DM development³¹, as demonstrated by a recent meta-analysis that showed an increased risk of T2DM and metabolic syndrome in subjects with NAFLD³². Therefore, predisposed abdominal CT scans also provide information on non-alcoholic fatty liver disease (NAFLD)³³. In the present study, there was no patient showed significant decrease in hepatic density suggesting hepatosteatosis. LAI results from abdominal CT scan showed an important difference among prediabetes subtypes, and was associated with insulin resistance and other adiposity indicators in individuals with C-IFG/IGT.

This study has some limitations. First, it was conducted retrospectively in a single medical center with a relatively small sample size (mainly women), especially in terms of the subgroup analysis. However, several previous reports have also shown that C-IFG/IGT had a higher correlation with central obesity and insulin resistance, agreeing with the results of our study^8,23. Thus, the consistent correlation observed between VAT volume and insulin resistance across different subtypes of prediabetes indicates that VAT volume plays a crucial role in the development of DM in individuals with prediabetes. Second, there was a lower prevalence of I-IGT and an absence of normal glucose tolerance in this study population. However, our study demonstrated a significant difference in several clinical indices between individuals with a combined type of prediabetes subtypes and those with a single type of prediabetes subtypes. Further a large-population study that includes individuals with normal glucose tolerance would be needed in the future.

• 18 December 2023

  A Correction to this paper has been published: https://doi.org/10.1038/s41598-023-49371-z

1. American Diabetes, A. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2020. Diabetes Care 43, S14-S31 (2020).

2. Gerstein, H. C. et al. Annual incidence and relative risk of diabetes in people with various categories of dysglycemia: A systematic overview and meta-analysis of prospective studies. Diabetes Res. Clin. Pract. 78, 305–312 (2007).

   Article  PubMed  Google Scholar 

3. DeFronzo, R. A. & Abdul-Ghani, M. Assessment and treatment of cardiovascular risk in prediabetes: Impaired glucose tolerance and impaired fasting glucose. Am. J. Cardiol. 108, 3B-24B (2011).

   Article  PubMed  CAS  Google Scholar 

4. Reaven, G. M. Banting lecture 1988. Role of insulin resistance in human disease. Diabetes 37, 1595–1607 (1988).

   Article  PubMed  CAS  Google Scholar 

5. Despres, J. P. et al. Abdominal obesity and the metabolic syndrome: Contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol 28, 1039–1049 (2008).

   Article  PubMed  CAS  Google Scholar 

6. Fujimoto, W. Y. et al. Visceral adiposity and incident coronary heart disease in Japanese-American men. The 10-year follow-up results of the Seattle Japanese-American community diabetes study. Diabetes Care 22, 1808–1812 (1999).

   Article  PubMed  CAS  Google Scholar 

7. Moller, D. E. & Flier, J. S. Insulin resistance–mechanisms, syndromes, and implications. N. Engl. J. Med. 325, 938–948 (1991).

   Article  PubMed  CAS  Google Scholar 

8. Nathan, D. M. et al. Impaired fasting glucose and impaired glucose tolerance: implications for care. Diabetes Care 30, 753–759 (2007).

   Article  PubMed  CAS  Google Scholar 

9. Carroll, J. F. et al. Visceral fat, waist circumference, and BMI: Impact of race/ethnicity. Obesity 16, 600–607 (2008).

   Article  PubMed  Google Scholar 

10. Shuster, A., Patlas, M., Pinthus, J. H. & Mourtzakis, M. The clinical importance of visceral adiposity: A critical review of methods for visceral adipose tissue analysis. Br. J. Radiol. 85, 1–10 (2012).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

11. Despres, J. P. & Lemieux, I. Abdominal obesity and metabolic syndrome. Nature 444, 881–887 (2006).

    Article  ADS  PubMed  CAS  Google Scholar 

12. Neeland, I. J. et al. Associations of visceral and abdominal subcutaneous adipose tissue with markers of cardiac and metabolic risk in obese adults. Obesity (Silver Spring) 21, E439-447 (2013).

    Article  PubMed  CAS  Google Scholar 

13. Tchkonia, T. et al. Mechanisms and metabolic implications of regional differences among fat depots. Cell Metab. 17, 644–656 (2013).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

14. Lee, D. H. Imaging evaluation of non-alcoholic fatty liver disease: focused on quantification. Clin. Mol. Hepatol. 23, 290–301 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

15. Haffner, S. M., Kennedy, E., Gonzalez, C., Stern, M. P. & Miettinen, H. A prospective analysis of the HOMA model. The Mexico city diabetes study. Diabetes Care 19, 1138–1141 (1996).

    Article  PubMed  CAS  Google Scholar 

16. Matthews, D. R. et al. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28, 412–419 (1985).

    Article  PubMed  CAS  Google Scholar 

17. Neeland, I. J. et al. Dysfunctional adiposity and the risk of prediabetes and type 2 diabetes in obese adults. JAMA 308, 1150–1159 (2012).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

18. Jung, S. H., Ha, K. H. & Kim, D. J. Visceral fat mass has stronger associations with diabetes and prediabetes than other anthropometric obesity indicators among Korean adults. Yonsei. Med. J. 57, 674–680 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

19. Brahimaj, A. et al. Novel metabolic indices and incident type 2 diabetes among women and men: the Rotterdam Study. Diabetologia 62, 1581–1590 (2019).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

20. Abdul-Ghani, M. A., Tripathy, D. & DeFronzo, R. A. Contributions of beta-cell dysfunction and insulin resistance to the pathogenesis of impaired glucose tolerance and impaired fasting glucose. Diabetes Care 29, 1130–1139 (2006).

    Article  PubMed  CAS  Google Scholar 

21. Meyer, C. et al. Different mechanisms for impaired fasting glucose and impaired postprandial glucose tolerance in humans. Diabetes Care 29, 1909–1914 (2006).

    Article  PubMed  CAS  Google Scholar 

22. Rhee, S. Y. et al. Characteristics of insulin resistance and insulin secretory capacity in Korean subjects with IFG and IGT. Diabetes Res. Clin. Pract. 89, 250–255 (2010).

    Article  PubMed  CAS  Google Scholar 

23. Oh, M. R., Jung, S. J., Bae, E. J., Park, B. H. & Chae, S. W. Clinical characteristics and associated risk factors of prediabetes in the Southwestern Region of Korea from 2010–2019. J. Clin. Med. 9, 1114 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

24. Deurenberg-Yap, M. et al. Relationships between indices of obesity and its co-morbidities in multi-ethnic Singapore. Int. J. Obes. Relat. Metab. Disord. 25, 1554–1562 (2001).

    Article  PubMed  CAS  Google Scholar 

25. Lear, S. A. et al. Visceral adipose tissue accumulation differs according to ethnic background: Results of the multicultural community health assessment trial (M-CHAT). Am. J. Clin. Nutr. 86, 353–359 (2007).

    Article  PubMed  CAS  Google Scholar 

26. Maurovich-Horvat, P. et al. Comparison of anthropometric, area- and volume-based assessment of abdominal subcutaneous and visceral adipose tissue volumes using multi-detector computed tomography. Int. J. Obes. (Lond) 31, 500–506 (2007).

    Article  PubMed  CAS  Google Scholar 

27. Gritzmann, N., Macheiner, P., Hollerweger, A. & Hubner, E. CT in the differentiation of pancreatic neoplasms–progress report. Dig. Dis. 22, 6–17 (2004).

    Article  PubMed  Google Scholar 

28. Saisho, Y. et al. Pancreas volumes in humans from birth to age one hundred taking into account sex, obesity, and presence of type-2 diabetes. Clin. Anat. 20, 933–942 (2007).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

29. Sakata, N. et al. Computed tomography reflected endocrine function of the pancreas. J. Gastrointest. Surg. 15, 525–532 (2011).

    Article  PubMed  Google Scholar 

30. Yokota, K., Fukushima, M., Takahashi, Y., Igaki, N. & Seino, S. Insulin secretion and computed tomography values of the pancreas in the early stage of the development of diabetes. J. Diabetes Investig. 3, 371–376 (2012).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

31. Lee, Y. H. et al. Nonalcoholic fatty liver disease in diabetes. Part I: Epidemiology and diagnosis. Diabetes. Metab. J. 43, 31–45 (2019).

    Article  PubMed  Google Scholar 

32. Ballestri, S. et al. Nonalcoholic fatty liver disease is associated with an almost twofold increased risk of incident type 2 diabetes and metabolic syndrome. Evidence from a systematic review and meta-analysis. J. Gastroenterol. Hepatol. 31, 936–944 (2016).

    Article  PubMed  CAS  Google Scholar 

33. Zhang, Y. N. et al. Liver fat imaging-a clinical overview of ultrasound, CT, and MR imaging. Br. J. Radiol. 91, 20170959 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Authors and Affiliations

1. Department of Nuclear Medicine and Biomedical Research Institute, Pusan National University Hospital, 179, Gudeok-Ro, Seo-Gu, Busan, Republic of Korea

   Jihyun Kim & Keunyoung Kim

2. Department of Nuclear Medicine, College of Medicine, Pusan National University, Yangsan, 50612, Republic of Korea

   Keunyoung Kim

Contributions

Guarantor of integristy of the entire study: K.K, Study Conception or design: K.K, Literature research: J.K. and K.K, Data and statistical analysis: K.K., Manuscript preparation: J.K. and K.K., Manuscript editing: J.K. and K.K.

Corresponding author

Correspondence to Keunyoung Kim.

Competing interests

The authors declare no competing interests.

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The original online version of this Article was revised: The original version of this Article contained an error in the Acknowledgements section, where the grant number was incorrect. Full information regarding the correction made can be found in the correction for this article.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions