A fully convolutional neural network for comprehensive compartmentalization of abdominal adipose tissue compartments in MRI

BACKGROUND

Existing literature has highlighted structural, physiological, and pathological disparities among abdominal adipose tissue (AAT) sub-depots. Accurate separation and quantification of these sub-depots are crucial for advancing our understanding of obesity and its comorbidities. However, the absence of clear boundaries between the sub-depots in medical imaging data has challenged their separation, particularly for internal adipose tissue (IAT) sub-depots. To date, the quantification of AAT sub-depots remains challenging, marked by a time-consuming, costly, and complex process.

PURPOSE

To implement and evaluate a convolutional neural network to enable granular assessment of AAT by compartmentalization of subcutaneous adipose tissue (SAT) into superficial subcutaneous (SSAT) and deep subcutaneous (DSAT) adipose tissue, and IAT into intraperitoneal (IPAT), retroperitoneal (RPAT), and paraspinal (PSAT) adipose tissue.

MATERIAL AND METHODS

MRI datasets were retrospectively collected from Singapore Preconception Study for Long-Term Maternal and Child Outcomes (S-PRESTO: 389 women aged 31.4 ± 3.9 years) and Singapore Adult Metabolism Study (SAMS: 50 men aged 28.7 ± 5.7 years). For all datasets, ground truth segmentation masks were created through manual segmentation. A Res-Net based 3D-UNet was trained and evaluated via 5-fold cross-validation on S-PRESTO data (N = 300). The model's final performance was assessed on a hold-out (N = 89) and an external test set (N = 50, SAMS).

RESULTS

The proposed method enabled reliable segmentation of individual AAT sub-depots in 3D MRI volumes with high mean Dice similarity scores of 98.3%, 97.2%, 96.5%, 96.3%, and 95.9% for SSAT, DSAT, IPAT, RPAT, and PSAT respectively.

CONCLUSION

Convolutional neural networks can accurately sub-divide abdominal SAT into SSAT and DSAT, and abdominal IAT into IPAT, RPAT, and PSAT with high accuracy. The presented method has the potential to significantly contribute to advancements in the field of obesity imaging and precision medicine.

Long-term effects of bariatric surgery on cardiovascular risk factors in Singapore

INTRODUCTION

Bariatric surgery is considered an effective treatment for weight loss and for improving the metabolic profile of patients with obesity. Obesity-related comorbidities such as hyperlipidaemia and type 2 diabetes mellitus (DM) are significant cardiovascular risk factors. Additionally, prospective clinical trials have shown that statins increase the risk of development of DM, and many patients with obesity are on statins. We retrospectively examined the effect of bariatric surgery on lipid profile, DM control and weight loss at the five-year follow-up.

METHODS

In total, 104 patients undergoing bariatric surgery from 2008 to 2012 were retrospectively studied. 36 patients were on preoperative statins. Their lipid profile, DM control and weight loss were examined at the one-year and five-year follow-ups.

RESULTS

Both high-density lipoprotein and triglyceride levels showed significant improvement at the one-year and five-year follow-ups (p = 0.01). Total cholesterol showed significant improvement at the one-year follow-up (-0.30 mmol/dL, p = 0.0338); however, better control was not sustained at the five-year follow-up (-0.15 mmol/dL, p = 0.133). Low-density lipoprotein did not show any considerable improvement at the one- and five-year follow-ups (-0.27 mmol/dL, p = 0.150 and -0.24 mmol/dL, p = 0.138, respectively). A statistically significant improvement in DM control was observed in these patients and in those on preoperative statins. Weight loss was sustained at one and five years.

CONCLUSION

Bariatric surgery does not confer a uniform improvement in lipid profile in the long term. It does, however, induce efficient weight loss and improvement in diabetic profile, even in patients on preoperative statins.

A Pilot Study of External Counterpulsation on Reactive Hyperemia, Levels of Glycemia and Metabolic Parameters in Type 2 Diabetes Mellitus

Introduction: External counter-pulsation (ECP) involves cuff inflation over the lower extremities to generate sheer stress, thereby improving endothelial function and anginal symptoms in coronary artery disease. Endothelial dysfunction is also involved in the pathogenesis of T2DM. We hypothesized that 1) ECP will be associated with an improvement in endothelial function in T2DM as measured by peripheral artery tonometry, and 2) explored whether this would vary with different dose and frequency regimens. A shorter or less intensive regimen could potentially reduce cost and improve patient compliance if a similar therapeutic response is achieved.

Methods: This single-center prospective study in a tertiary institute in Singapore involving 46 adults with T2DM of HbA1c between 7 to 10%, who were randomly assigned to receive 35 sessions of ECP at different regimens and duration. Subjects in arm 1 received 1-hour daily sessions 5x per week for 7 consecutive weeks, subjects in arm 2 received 0.5-hour sessions 5x per week for 7 consecutive weeks, and subjects in arm 3 received 1-hour sessions 3x per week for 12 consecutive weeks. Endothelial function was evaluated by reactive hyperemia index (RHI) via peripheral arterial tonometry measured at the start, midpoint and end of study. Other secondary outcomes included fasting glucose, homeostatic model assessment of insulin resistance (HOMA-IR), HbA1c, blood pressure, lipid profile, weight and vibration sense.

Results: 42 subjects completed the 35-session course of ECP. Mean age was 56.1±9.3 years, duration of diabetes 8.8±4.7 years, baseline RHI 2.0 (1.3–3.7) and baseline HOMA-IR was 3.1 (0.5–18.7). All regimes of ECP were well-tolerated. There was no change in RHI across all 3 regimens of ECP individually or collectively at the end of the study (ΔRHI +0.01%, p=0.458). Glycaemic markers of fasting glucose, HbA1c and HOMA-IR, as well as blood pressure, lipid profile, weight and vibration sense also remained unchanged at endpoint. Subgroup analysis showed a significant improvement in RHI (ΔRHI +20.6%, p=0.0178) in 7 subjects with more severe endothelial dysfunction (defined by RHI<1.67) at baseline who had a trend to having a longer duration of diabetes, however there was no improvement in fasting glucose, HbA1c, HOMA-IR or metabolic parameters in this group.

Conclusion: ECP did not show a beneficial effect on endothelial function, glycemic control or metabolic parameters in this South-East Asian population with T2DM at any of the three regimens. This may partly be explained by less severe endothelial dysfunction and less insulin resistance in our population at baseline. Future studies of ECP may investigate its potential benefits in a larger population of T2DM with more severe endothelial dysfunction, higher insulin resistance and/or longer duration of diabetes at baseline.

Evaluating the effectiveness of a multi-faceted inpatient diabetes management program among hospitalised patients with diabetes mellitus

BACKGROUND

Diabetes mellitus (DM) is one of the most common chronic diseases. Individuals with DM are more likely to be hospitalised and stay longer than those without DM. Inpatient hypoglycemia and hyperglycemia, which are associated with adverse outcomes, are common, but can be prevented through hospital quality improvement programs.

METHODS

We designed a multi-faceted intervention program with the aim of reducing inpatient hypoglycemia and hyperglycemia. This was implemented over seven phases between September 2013 to January 2016, and covered all the non-critical care wards in a tertiary hospital. The program represented a pragmatic approach that leveraged on existing resources and infrastructure within the hospital. We calculated glucometric outcomes in June to August 2016 and compared them with those in June to August 2013 to assess the overall effectiveness of the program. We used regression models with generalised estimating equations to adjust for potential confounders and account for correlations of repeated outcomes within patients and admissions.

RESULTS

We observed significant reductions in patient-days affected by hypoglycemia (any glucose reading < 4 mmol/L: OR = 0.71, 95% CI: 0.61 to 0.83, p <  0.001), and hyperglycemia (any glucose reading > 14 mmol/L: OR = 0.84, 95% CI: 0.71 to 0.99, p = 0.041). Similar findings were observed for admission-level hypoglycemia and hyperglycemia. Further analyses suggested that these reductions started to occur four to 6 months post-implementation.

CONCLUSIONS

Our program was associated with sustained improvements in clinically relevant outcomes. Our described intervention could be feasibly implemented by other secondary and tertiary care hospitals by leveraging on existing infrastructure and work force.

Effects of external counter-pulsation on endothelial function assessed by peripheral artery tonometry, levels of glycaemia and metabolic markers in individuals with type 2 diabetes mellitus

BACKGROUND AND AIMS

External counter-pulsation (ECP) generates sheer stress thereby improving endothelial function and anginal symptoms in coronary artery disease. Endothelial dysfunction is also involved in the pathogenesis of T2DM. The aim of this pilot study was to investigate the use of ECP at different doses in improving endothelial function and glycaemic markers in T2DM.

METHODS

This prospective study involved 46 subjects with T2DM randomly assigned to receive 35 sessions of ECP at different regimens (0.5 h versus 1 h) and duration (7 versus 12 weeks). Endothelial function was evaluated by reactive hyperaemia index (RHI) via peripheral arterial tonometry at the start, midpoint and end of study. Other secondary outcomes included fasting glucose, HOMA-IR, HbA1c, blood pressure, lipid profile, weight and vibration sense.

RESULTS

There was no change in RHI across all 3 regimens of ECP individually or collectively at the end of the study (ΔRHI +0.01%, p = 0.458). Glycaemic markers also remained unchanged at endpoint. Subgroup analysis showed an improvement in RHI (ΔRHI +20.6%, p = 0.0178) in subjects with more severe endothelial dysfunction at baseline.

CONCLUSION

ECP did not show a beneficial effect on endothelial function or glycemic control in this South-East Asian population with T2DM at any of the three regimens. This may partly be explained by less severe endothelial dysfunction and less insulin resistance in our population at baseline.

Robust estimation of the effect of an exposure on the change in a continuous outcome

Background

The change in two measurements of a continuous outcome can be modelled directly with a linear regression model, or indirectly with a random effects model (REM) of the individual measurements. These methods are susceptible to model misspecifications, which are commonly addressed by applying monotonic transformations (e.g., Box-Cox transformation) to the outcomes. However, transforming the outcomes complicates the data analysis, especially when variable selection is involved. We propose a robust alternative through a novel application of the conditional probit (cprobit) model.

Methods

The cprobit model analyzes the ordered outcomes within each subject, making the estimate invariant to monotonic transformation on the outcome. By scaling the estimate from the cprobit model, we obtain the exposure effect on the change in the observed or Box-Cox transformed outcome, pending the adequacy of the normality assumption on the raw or transformed scale.

Results

Using simulated data, we demonstrated a similar good performance of the cprobit model and REM with and without transformation, except for some bias from both methods when the Box-Cox transformation was applied to scenarios with small sample size and strong effects. Only the cprobit model was robust to skewed subject-specific intercept terms when a Box-Cox transformation was used. Using two real datasets from the breast cancer and inpatient glycemic variability studies which utilize electronic medical records, we illustrated the application of our proposed robust approach as a seamless three-step workflow that facilitates the use of Box-Cox transformation to address non-normality with a common underlying model.

Conclusions

The cprobit model provides a seamless and robust inference on the change in continuous outcomes, and its three-step workflow is implemented in an R package for easy accessibility.

Using marginal standardisation to estimate relative risk without dichotomising continuous outcomes

Background

Although criticisms regarding the dichotomisation of continuous variables are well known, applying logit model to dichotomised outcomes is the convention because the odds ratios are easily obtained and they approximate the relative risks (RRs) for rare events.

Methods

To avoid dichotomisation when estimating RR, the marginal standardisation method that transforms estimates from logit or probit model to RR estimate is extended to include estimates from linear model in the transformation. We conducted a simulation study to compare the statistical properties of the estimates from: (i) marginal standardisation method between models for continuous (i.e., linear model) and dichotomised outcomes (i.e., logit or probit model), and (ii) marginal standardisation method and distributional approach (i.e., marginal mean method) applied to linear model. We also compared the diagnostic test for probit, logit and linear models. For the real dataset analysis, we applied these analytical approaches to assess the management of inpatient hyperglycaemia in a pilot intervention study.

Results

Although the RR estimates from the marginal standardisation method were generally unbiased for all models in the simulation study, the marginal standardisation method for linear model provided estimates with higher precision and power than logit or probit model, especially when the baseline risks were at the extremes. When comparing approaches that avoid dichotomisation, RR estimates from these approaches had comparable performance. Assessing the assumption of error distribution was less powerful for logit or probit model via link test when compared with diagnostic test for linear model. After accounting for multiple thresholds representing varying levels of severity in hyperglycaemia, marginal standardisation method for linear model provided stronger evidence of reduced hyperglycaemia risk after intervention in the real dataset analysis although the RR estimates were similar across various approaches.

Conclusions

When compared with approaches that do not avoid dichotomisation, the RR estimated from linear model is more precise and powerful, and the diagnostic test from linear model is more powerful in detecting mis-specified error distributional assumption than the diagnostic test from logit or probit model. Our work describes and assesses the methods available to analyse data involving studies of continuous outcomes with binary representations.

Electronic supplementary material

The online version of this article (10.1186/s12874-019-0778-9) contains supplementary material, which is available to authorized users.

Large-scale lipidomics identifies associations between plasma sphingolipids and T2DM incidence

BACKGROUND

Sphingolipids (SPs) are ubiquitous, structurally diverse molecules that include ceramides, sphingomyelins, and sphingosines. They are involved in various pathologies including obesity and type 2 diabetes mellitus (T2DM). Therefore, it is likely that perturbations in plasma concentrations of SPs are associated with disease. Identifying these associations may reveal useful biomarkers or provide insight into disease processes.

METHODS

We performed a lipidomics evaluation of molecularly-distinct SPs in the plasma of 2,302 ethnically-Chinese Singaporeans using electrospray ionization mass spectrometry coupled with liquid chromatography. SP profiles were compared to clinical and biochemical characteristics, and subjects were evaluated by follow-up visits for 11 years.

RESULTS

We found that ceramides correlate positively but hexosylceramides correlate negatively with body mass index (BMI) and homeostatic model assessment of insulin resistance (HOMA-IR). Furthermore, SPs with a d16:1 sphingoid backbone correlate more positively with BMI and HOMA-IR, while d18:2 SPs correlate less positively, relative to canonical d18:1 SPs. We also found that higher concentrations of two distinct sphingomyelins were associated with a higher risk of T2DM (HR 1.45, 95% CI 1.18-1.78 for SM d16:1/C18:0; and HR 1.40, 95% CI 1.17-1.68 for SM d18:1/C18:0).

CONCLUSION

We identified significant associations between SPs and obesity/T2DM characteristics, specifically, that of hexosylceramides, d16:1 SPs, and d18:2 SPs. This suggests that the balance of SP metabolism, rather than ceramide accumulation, is associated with the pathology of obesity. We further identified two specific SPs that may represent prognostic biomarkers for T2DM.

FUNDING

Funding sources are listed in the Acknowledgements section.

Plasma Sphingolipids and Subclinical Atherosclerosis – Novel Associations Uncovered by a Large-scale Lipidomic Analysis (P18-129-19)

Objectives

Sphingolipids (SP) are a diverse class of heterogenous lipids that includes ceramides, sphingomyelins, and glycosphingolipids. Many SPs have diverse roles in cell functions including cell growth, inflammation and angiogenesis, and previous studies suggest that SPs may also be involved in the pathogenesis of cardiovascular diseases. We aimed to identify plasma SPs and subsequently evaluate associations between plasma SPs and subclinical atherosclerosis in a subset of the Multiethnic Cohort of Singapore with the goal of uncovering novel biomarkers predictive of heart disease.

Methods

We conducted a lipidomics evaluation of 103 molecularly-distinct SPs in the plasma of 559 Singaporeans aged 50 years and above using electrospray ionization mass spectrometry coupled with liquid chromatography. All participants did not have a history of diabetes, heart disease, stroke or cancer at baseline, and completed a detailed health screening that evaluated risk factors for cardiovascular disease including computed tomography scans of the abdomen and coronary arteries. Multivariable logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for associations between SPs and subclinical atherosclerosis (defined as coronary artery calcium score ≥ 100).

Results

Ceramides (Cer), particularly those with a d16:1 or d18:1 backbone, were directly associated with higher risk of subclinical atherosclerosis whereas a small number of monohexosylceramides (MHCer), dihexosylceramides (DHCer) and sphingomyelins (SM) with a d18:2 sphingoid backbone were inversely associated. However most associations were attenuated after adjusting for conventional cardiovascular risk factors, including blood lipid (low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides) concentrations and glycemic markers, suggesting that the associations may be mediated by these risk factors.

Conclusions

High-throughput lipidomics may uncover novel sphingolipids predictive of heart disease. Characterization of these lipid species could provide insights into disease etiology.

Funding Sources

This work was supported by the National University Health System and the National Research Foundation Investigatorship grant.

Supporting Tables, Images and/or Graphs

Ethnicity-Specific Skeletal Muscle Transcriptional Signatures and Their Relevance to Insulin Resistance in Singapore

CONTEXT

Insulin resistance (IR) and obesity differ among ethnic groups in Singapore, with the Malays more obese yet less IR than Asian-Indians. However, the molecular basis underlying these differences is not clear.

OBJECTIVE

As the skeletal muscle (SM) is metabolically relevant to IR, we investigated molecular pathways in SM that are associated with ethnic differences in IR, obesity, and related traits.

DESIGN, SETTING, AND MAIN OUTCOME MEASURES

We integrated transcriptomic, genomic, and phenotypic analyses in 156 healthy subjects representing three major ethnicities in the Singapore Adult Metabolism Study.

PATIENTS

This study contains Chinese (n = 63), Malay (n = 51), and Asian-Indian (n = 42) men, aged 21 to 40 years, without systemic diseases.

RESULTS

We found remarkable diversity in the SM transcriptome among the three ethnicities, with >8000 differentially expressed genes (40% of all genes expressed in SM). Comparison with blood transcriptome from a separate Singaporean cohort showed that >95% of SM expression differences among ethnicities were unique to SM. We identified a network of 46 genes that were specifically downregulated in Malays, suggesting dysregulation of components of cellular respiration in SM of Malay individuals. We also report 28 differentially expressed gene clusters, four of which were also enriched for genes that were found in genome-wide association studies of metabolic traits and disease and correlated with variation in IR, obesity, and related traits.

CONCLUSION

We identified extensive gene-expression changes in SM among the three Singaporean ethnicities and report specific genes and molecular pathways that might underpin and explain the differences in IR among these ethnic groups.

Feasibility of representing adherence to blood glucose monitoring through visualizations: A pilot survey study among healthcare workers

BACKGROUND

Measuring adherence to processes is one of the established ways to quantify the quality of healthcare. Providing timely feedback to healthcare workers on the level of adherence can improve process measures. However, it is challenging to present data on adherence to repetitive time-sensitive tasks in a clear manner.

OBJECTIVES

We used inpatient glucose monitoring as a test case to explore the feasibility of using visualizations to communicate adherence to repetitive scheduled tasks to healthcare workers.

METHODS

We selected four candidate plots that represented distribution across time: histogram, probability density function plot (pdf plot), violin plot and cumulative density function plot (cdf plot). Doctors and nurses involved in inpatient diabetes care in a tertiary hospital were invited to complete a self-administered questionnaire that measured self-reported baseline knowledge, performance, and perception towards the visualizations. Performance was assessed by determining if a participant was able to correctly identify visualizations representing protocol adherence. We also assessed the perception of usability of these visualizations for monitoring protocol adherence. Binomial regression models were used to identify factors associated with overall performance and perception. Logistic regression models with generalized estimating equation were used to compare performance and perception between visualizations, and identify effect modifiers.

RESULTS

A total of 57 doctors and nurses completed the questionnaire. Participants were most familiar with histogram (87.7%), followed by cdf plot (61.4%), pdf plot (40.4%), and violin plot (7%). However, the percentages of participants who identified non-adherence using these plots were generally lower, ranging from 29.8% to 40.4%. Participants' perception of usability ranged from 14% to 17.5% across these visualizations. More favorable perceptions were found among participants with baseline knowledge for two or more visualizations (adjusted odds ratio: 3.21; 95%CI: 1.29, 7.96; p-value: 0.012) and having identified two or more non-adherent visualizations (adjusted odds ratio: 4.23; 95%CI: 1.95, 9.16; p-value: < 0.001).

CONCLUSIONS

Adherence to repetitive time-sensitive tasks can be presented in the form of visualizations. However, nurses' and doctors' knowledge and understanding of these visualizations are generally poor. This may influence their perception of usability of these plots. Therefore, these visualizations need to be implemented in tandem with training on their interpretation, to enhance the usefulness of these plots in motivating quality improvement.

A stratification approach using logit-based models for confounder adjustment in the study of continuous outcomes

The control of confounding is an area of extensive epidemiological research, especially in the field of causal inference for observational studies. Matched cohort and case-control study designs are commonly implemented to control for confounding effects without specifying the functional form of the relationship between the outcome and confounders. This paper extends the commonly used regression models in matched designs for binary and survival outcomes (i.e. conditional logistic and stratified Cox proportional hazards) to studies of continuous outcomes through a novel interpretation and application of logit-based regression models from the econometrics and marketing research literature. We compare the performance of the maximum likelihood estimators using simulated data and propose a heuristic argument for obtaining the residuals for model diagnostics. We illustrate our proposed approach with two real data applications. Our simulation studies demonstrate that our stratification approach is robust to model misspecification and that the distribution of the estimated residuals provides a useful diagnostic when the strata are of moderate size. In our applications to real data, we demonstrate that parity and menopausal status are associated with percent mammographic density, and that the mean level and variability of inpatient blood glucose readings vary between medical and surgical wards within a national tertiary hospital. Our work highlights how the same class of regression models, available in most statistical software, can be used to adjust for confounding in the study of binary, time-to-event and continuous outcomes.

The Utility of Brief Cognitive Tests for Patients With Type 2 Diabetes Mellitus: A Systematic Review

BACKGROUND

Type 2 diabetes mellitus (T2DM) is associated with an increased risk for mild cognitive impairment and dementia in both middle-aged and older individuals. Brief cognitive tests can potentially serve as a reliable and cost effective approach to detect for cognitive decrements in clinical practice.

OBJECTIVE

This systematic review examined the utility of brief cognitive tests in studies with patients with T2DM.

METHOD

This systematic review was conducted according to guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. "PubMed," "PsychINFO," "ScienceDirect," and "ProQuest" electronic databases were searched to identify articles published from January 1, 2005 to December 31, 2015.

RESULTS

The search yielded 22 studies, with only 8 using brief tests as a cognitive screening tool, whereas the majority using these tests as a measure of global cognitive functions. In regard to cognitive screening studies, most had failed to fulfil the standard reporting of diagnostic test accuracy criteria such as Standards for Reporting of Diagnostic Accuracy for dementia and cognitive impairment. Moreover, few studies reported discriminant indices such as sensitivity, specificity, and positive and negative predictive values of brief cognitive tests in detecting cognitive impairment in patients with T2DM. Among studies which used brief cognitive tests as a measure of global cognitive function, patients with diabetes tended to perform worse than patients without diabetes. Processing speed appeared to be particularly impaired among patients with diabetes, therefore, measures of processing speed such as the Digit Symbol Substitution Test may add value to brief cognitive tests such as the Montreal Cognitive Assessment.

CONCLUSIONS

The Montreal Cognitive Assessment supplemented by the Digit Symbol Substitution Test indicate initial promise in screening for cognitive impairment in T2DM.

Utilizing distributional analytics and electronic records to assess timeliness of inpatient blood glucose monitoring in non-critical care wards

Background

Regular and timely monitoring of blood glucose (BG) levels in hospitalized patients with diabetes mellitus is crucial to optimizing inpatient glycaemic control. However, methods to quantify timeliness as a measurement of quality of care are lacking. We propose an analytical approach that utilizes BG measurements from electronic records to assess adherence to an inpatient BG monitoring protocol in hospital wards.

Methods

We applied our proposed analytical approach to electronic records obtained from 24 non-critical care wards in November and December 2013 from a tertiary care hospital in Singapore. We applied distributional analytics to evaluate daily adherence to BG monitoring timings. A one-sample Kolmogorov-Smirnov (1S-KS) test was performed to test daily BG timings against non-adherence represented by the uniform distribution. This test was performed among wards with high power, determined through simulation. The 1S-KS test was coupled with visualization via the cumulative distribution function (cdf) plot and a two-sample Kolmogorov-Smirnov (2S-KS) test, enabling comparison of the BG timing distributions between two consecutive days. We also applied mixture modelling to identify the key features in daily BG timings.

Results

We found that 11 out of the 24 wards had high power. Among these wards, 1S-KS test with cdf plots indicated adherence to BG monitoring protocols. Integrating both 1S-KS and 2S-KS information within a moving window consisting of two consecutive days did not suggest frequent potential change from or towards non-adherence to protocol. From mixture modelling among wards with high power, we consistently identified four components with high concentration of BG measurements taken before mealtimes and around bedtime. This agnostic analysis provided additional evidence that the wards were adherent to BG monitoring protocols.

Conclusions

We demonstrated the utility of our proposed analytical approach as a monitoring tool. It provided information to healthcare providers regarding the timeliness of daily BG measurements. From the real data application, there were empirical evidences suggesting adherence of BG timings to protocol among wards with adequate power for assessing timeliness. Our approach is extendable to other areas of healthcare where timeliness of patient care processes is important.

Smaller size of high metabolic rate organs explains lower resting energy expenditure in Asian-Indian Than Chinese men

BACKGROUND

In Singapore, the obesity prevalence is disproportionately higher in the Asian-Indians and Malays than the Chinese. Lower resting energy expenditure (REE) may be a contributory factor.

OBJECTIVE

We explored the association between ethnicity and REE in Chinese, Asian-Indian and Malay men living in Singapore and determined the influence of body composition, mass/volume of high metabolic rate organs, represented by brain volume and trunk fat-free mass (FFM), and physical activity on ethnic differences.

DESIGN

Two hundred and forty-four men from Singapore (n=100 Chinese, 70 Asian-Indians and 74 Malays), aged 21-40 years and body mass index of 18.5-30.0 kg m(-2), were recruited in this cross-sectional study. REE was assessed by indirect calorimetry and body composition by dual-energy X-ray absorptiometry. Brain volume was measured by magnetic resonance imaging. Physical activity was assessed by the Singapore Prospective Study Program Physical Activity Questionnaire.

RESULTS

REE was significantly lower in Asian-Indians compared with that in Chinese after adjusting for body weight. FFM (total, trunk and limb) and total fat mass were important predictors of REE across all ethnic groups. Brain volume was positively associated with REE only in Malays. Moderate and vigorous physical activity was positively associated with REE only in Asian-Indians and Malays. The difference in REE between Asian-Indians and Chinese was attenuated but remained statistically significant after adjustment for total FFM (59±20 kcal per day), fat mass (67±20 kcal per day) and brain volume (54±22 kcal per day). The association between REE and ethnicity was no longer statistically significant after total FFM was replaced by trunk FFM (which includes heart, liver, kidney and spleen) but not when it was replaced by limb FFM (skeletal muscle).

CONCLUSIONS

We have demonstrated a lower REE in Asian-Indians compared with Chinese who may contribute to the higher rates of obesity in the former. This difference could be accounted for by differences in metabolically active organs.

Ethnic Differences in the Role of Adipocytokines Linking Abdominal Adiposity and Insulin Sensitivity Among Asians

CONTEXT

Among Asian ethnic groups, Chinese or Malays are more insulin sensitive than South Asians, in particular in lean individuals. We have further reported that body fat partitioning did not explain this ethnic difference in insulin sensitivity.

OBJECTIVE

We examined whether adipocytokines might explain the ethnic differences in the relationship between obesity and insulin resistance among the three major ethnic groups in Singapore.

DESIGN AND PARTICIPANTS

This was a cross-sectional study of 101 Chinese, 82 Malays, and 81 South Asian men. Insulin sensitivity index (ISI) was measured using hyperinsulinemic euglycemic clamp. Visceral (VAT) and subcutaneous adipose tissue (SAT) volumes were quantified using magnetic resonance imaging.

MAIN OUTCOME MEASURES

Plasma total and high-molecular-weight adiponectin, leptin, visfatin, apelin, IL-6, fibroblast growth factor 21 (FGF21), retinol binding protein-4 (RBP 4), and resistin were measured using enzyme-linked immunoassays.

RESULTS

Principle component (PC) analysis on the adipocytokines identified three PCs, which explained 49.5% of the total variance. Adiponectin loaded negatively, and leptin and FGF21 loaded positively onto PC1. Visfatin, resistin, and apelin all loaded positively onto PC2. IL-6 loaded positively and RBP-4 negatively onto PC3. Only PC1 was negatively associated with ISI in all ethnic groups. In the path analysis, SAT and VAT were negatively associated with ISI in Chinese and Malays without significant mediatory role of PC1. In South Asians, the relationship between VAT and ISI was mediated partly through PC1, whereas the relationship between SAT and ISI was mediated mainly through PC1.

CONCLUSIONS

The relationships between abdominal obesity, adipocytokines and insulin sensitivity differ between ethnic groups. Adiponectin, leptin, and FGF21 play a mediating role in the relationship between abdominal adiposity and insulin resistance in South Asians, but not in Malays or Chinese.

Psychometric Properties of the Problem Areas in Diabetes (PAID) Instrument in Singapore

BACKGROUND

Emotional distress is an important dimension in diabetes, and several instruments have been developed to measure this aspect. The Problem Areas in Diabetes (PAID) scale is one such instrument which has demonstrated validity and reliability in Western populations, but its psychometric properties in Asian populations have not been examined.

METHODS

This was a secondary analysis of data from patients with Type 2 diabetes mellitus recruited through convenience sampling from a diabetes specialist outpatient clinic in Singapore. The following psychometric properties were assessed: Construct validity through confirmatory factor analysis (CFA) and Rasch analysis, concurrent validity through correlation with related scales (Kessler Psychological Distress Scale, Diabetes Health Profile-psychological distress, Audit of Diabetes Dependent Quality of Life), reliability through assessment of internal consistency and floor and ceiling effects, and sensitivity by estimating effect sizes for known clinical and social functioning groups.

RESULTS

203 patients with mean age of 45±12 years were analysed. None of the previously published model structures achieved a good fit on CFA. On Rasch analysis, four items showed poor fit and were removed. The abridged 16-item PAID mapped to a single latent trait, with a high degree of internal consistency (Cronbach ɑ 0.95), but significant floor effect (24.6% scoring at floor). Both 20-item and 16-item PAID scores were moderately correlated with scores of related scales, and sensitive to differences in clinical and social functioning groups, with large effect sizes for glycemic control and diabetes related complications, nephropathy and neuropathy.

CONCLUSION

The abridged 16-item PAID measures a single latent trait of emotional distress due to diabetes whereas the 20-item PAID appears to measures more than one latent trait. However, both the 16-item and 20-item PAID versions are valid, reliable and sensitive for use among Singaporean patients with diabetes.

Ethnic differences in insulin sensitivity and beta-cell function among Asian men

BACKGROUND AND OBJECTIVES

Lean Asian Indians are less insulin sensitive compared with Chinese and Malays, but the pancreatic beta-cell function among these ethnic groups has yet to be studied in depth. We aimed to study beta-cell function in relation to insulin sensitivity among individuals of Chinese, Malay and Asian-Indian ethnicity living in Singapore.

SUBJECTS AND METHODS

This is a sub-group analysis of 59 normoglycemic lean (body mass index (BMI) <23 kg m(-)(2)) adult males (14 Chinese, 21 Malays and 24 Asian Indians) from the Singapore Adults Metabolism Study. Insulin sensitivity was determined using fasting state indices (homeostatic model assessment-insulin resistance), the euglycemic-hyperinsulinemic clamp (ISI-clamp) and a liquid mixed-meal tolerance test (LMMTT) (Matsuda insulin sensitivity index (ISI-Mat)). Beta-cell function was assessed using fasting state indices (homeostatic model assessment-beta-cell function) and from the LMMTT (insulinogenic index and insulin secretion index). The oral disposition index (DI), a measure of beta-cell function relative to insulin sensitivity during the LMMTT, was calculated as a product of ISI-Mat and insulin secretion index.

RESULTS

Asian Indians had higher waist circumference and percent body fat than Chinese and Malays despite similar BMI. Overall, Asian Indians were the least insulin sensitive whereas the Chinese were most insulin sensitive. Asian Indians had higher beta-cell function compared with Chinese or Malays but these were not statistically different. Malays had the highest incremental area under the curve for glucose during LMMTT compared with Asian Indians and Chinese. However, there were no significant ethnic differences in the incremental insulin area under the curve. The oral DI was the lowest in Malays, followed by Asian Indians and Chinese.

CONCLUSION

Among lean Asians, Chinese are the most insulin sensitive whereas Asian Indians are the least insulin sensitive. However, Malays demonstrate higher postprandial glucose excursion with lower beta-cell response compare with Chinese or Asian Indians. The paths leading to type 2 diabetes mellitus might differ between these Asian ethnic groups.

Body fat partitioning does not explain the interethnic variation in insulin sensitivity among Asian ethnicity: the Singapore adults metabolism study

We previously showed that ethnicity modifies the association between adiposity and insulin resistance. We sought to determine whether differential body fat partitioning or abnormalities in muscle insulin signaling associated with higher levels of adiposity might underlie this observation. We measured the insulin sensitivity index (ISI), percentage of body fat (%body fat), visceral (VAT) and subcutaneous (SAT) adipose tissue, liver fat, and intramyocellular lipids (IMCL) in 101 Chinese, 82 Malays, and 81 South Asians, as well as phosphorylated (p)-Akt levels in cultured myoblasts from Chinese and South Asians. Lean Chinese and Malays had higher ISI than South Asians. Although the ISI was lower in all ethnic groups when %body fat was higher, this association was stronger in Chinese and Malays, such that no ethnic differences were observed in overweight individuals. These ethnic differences were observed even when %body fat was replaced with fat in other depots. Myoblasts obtained from lean South Asians had lower p-Akt levels than those from lean Chinese. Higher adiposity was associated with lower p-Akt levels in Chinese but not in South Asians, and no ethnic differences were observed in overweight individuals. With higher %body fat, Chinese exhibited smaller increases in deep SAT and IMCL compared with Malays and South Asians, which did not explain the ethnic differences observed. Our study suggests that body fat partitioning does not explain interethnic differences in insulin sensitivity among Asian ethnic groups. Although higher adiposity had greater effect on skeletal muscle insulin sensitivity among Chinese, obesity-independent pathways may be more relevant in South Asians.

Variability in fasting lipid and glycogen contents in hepatic and skeletal muscle tissue in subjects with and without type 2 diabetes: a 1H and 13C MRS study

The measurement of tissue lipid and glycogen contents and the establishment of normal levels of variability are important when assessing changes caused by pathology or treatment. We measured hepatic and skeletal muscle lipid and glycogen levels using (1)H and (13)C MRS at 3 T in groups of subjects with and without type 2 diabetes. Within-visit reproducibility, due to repositioning and instrument errors was determined from repeat measurements made over 1 h. Natural variability was assessed from separate measurements made on three occasions over 1 month. Hepatic lipid content was greater in subjects with diabetes relative to healthy subjects (p = 0.03), whereas levels of hepatic and skeletal muscle glycogen, and of intra- and extra-myocellular lipid, were similar. The single-session reproducibility values (coefficient of variation, CV) for hepatic lipid content were 12% and 7% in groups of subjects with and without diabetes, respectively. The variability of hepatic lipid content over 1 month was greater than the reproducibility, with CV = 22% (p = 0.08) and CV = 44% (p = 0.004) in subjects with and without diabetes, respectively. Similarly, levels of variation in basal hepatic glycogen concentrations (subjects with diabetes, CV = 38%; healthy volunteers, CV = 35%) were significantly larger than single-session reproducibility values (CV = 17%, p = 0.02 and CV = 13%, p = 0.05, respectively), indicating substantial biological changes in basal concentrations over 1 month. There was a decreasing correlation in measurements of both hepatic lipid and glycogen content with increasing time between scans. Levels of variability in intra- and extra-myocellular lipid in the soleus muscle, and glycogen concentrations in the gastrocnemius muscle, tended to be larger than expected from single-session reproducibility, although these did not reach significance.

Associations between ethnicity, body composition, and bone mineral density in a Southeast Asian population

CONTEXT AND OBJECTIVE

Chinese men in Singapore have a higher incidence of hip fractures than Malay and Indian men. We investigated whether there were corresponding ethnic differences in peak bone mineral density (BMD) in young men and whether differences in body composition influenced peak BMD.

DESIGN AND SETTING

This was a cross-sectional study of healthy volunteers in a tertiary medical center.

PARTICIPANTS

A total of 100 Chinese, 82 Malay, and 80 Indian men aged 21 to 40 years, with body mass index between 18 and 30 kg/m(2) underwent dual-energy x-ray absorptiometry to assess BMD, lean mass (LM) and fat mass (FM), and magnetic resonance imaging to quantify abdominal subcutaneous and visceral adipose tissue. Multiple linear regression models, with adjustment for age and height (as a proxy for skeletal size), were used.

RESULTS

Malay and Indian men had significantly higher BMD than Chinese men at the lumbar spine (Malay: B, 0.06 ± 0.02, P = .001; Indian: B, 0.03 ± 0.02, P = .049), femoral neck (Malay: B 0.04 ± 0.02, P = .034; Indian: B, 0.04 ± 0.02, P = .041), hip (Malay: B, 0.05 ± 0.02, P = .016; Indian: B, 0.06 ± 0.02, P = .001), and ultradistal radius (Malay: B, 0.03 ± 0.01, P < .001; Indian: B, 0.02 ± 0.01, P = .029), and this difference was retained after adjustment for LM and FM, except in Malay men at the femoral neck and in Indian men at the ultradistal radius. LM was an important independent determinant of BMD at all sites, whereas FM, subcutaneous adipose tissue, and visceral adipose tissue were not significantly associated with BMD at any site.

CONCLUSIONS

Lower peak BMD in Chinese men may partly explain the higher fracture incidence in this ethnic group. Further studies are needed to elucidate the reasons for these ethnic differences in bone accumulation.

Ethnicity modifies the relation between fasting plasma glucose and HbA1c in Indians, Malays and Chinese

AIMS

To study whether HbA(1c) , and its relationship with fasting plasma glucose, was significantly different among Chinese, Malays and Indians in Singapore.

METHODS

A sample of 3895 individuals without known diabetes underwent detailed interview and health examination, including anthropometric and biochemical evaluation, between 2004 and 2007. Pearson's correlation, analysis of variance and multiple linear regression analyses were used to examine the influence of ethnicity on HbA(1c) .

RESULTS

As fasting plasma glucose increased, HbA(1c) increased more in Malays and Indians compared with Chinese after adjustment for age, gender, waist circumference, serum cholesterol, serum triglyceride and homeostasis model assessment of insulin resistance (P-interaction < 0.001). This translates to an HbA(1c) difference of 1.1 mmol/mol (0.1%, Indians vs. Chinese), and 0.9 mmol/mol (0.08%, Malays vs. Chinese) at fasting plasma glucose 5.6 mmol/l (the American Diabetes Association criterion for impaired fasting glycaemia); and 2.1 mmol/mol (0.19%, Indians vs. Chinese) and 2.6 mmol/mol (0.24%, Malays vs. Chinese) at fasting plasma glucose 7.0 mmol/l, the diagnostic criterion for diabetes mellitus.

CONCLUSIONS

Using HbA(1c) in place of fasting plasma glucose will reclassify different proportions of the population in different ethnic groups. This may have implications in interpretation of HbA(1c) results across ethnic groups and the use of HbA(1c) for diagnosing diabetes mellitus.

Effects of exenatide on circulating glucose, insulin, glucagon, cortisol and catecholamines in healthy volunteers during exercise

AIMS/HYPOTHESIS

Exenatide, a glucagon like peptide-1 agonist, is a treatment for type 2 diabetes mellitus that stimulates insulin and suppresses glucagon secretion in a glucose-dependent manner. By contrast, during aerobic exercise, the serum insulin concentration normally falls, with a rise in plasma glucagon. We therefore assessed whether exenatide might predispose to hypoglycaemia during exercise.

METHODS

We studied eight non-diabetic men, who were 35.3 +/- 6.3 years of age with BMI of 24.7 +/- 1.7 kg/m(2) (mean +/- SD), using a randomised, crossover, double-blind design investigation. After an overnight fast, participants received 5 microg of subcutaneous exenatide or placebo and rested for 105 min before cycling at 60% of their maximal oxygen uptake (VO(2max)) for 75 min and then recovering for a further 60 min.

RESULTS

The insulin/glucagon molar ratio rose with exenatide at rest (p < 0.01), then fell during exercise with placebo and with exenatide. At rest, fasting blood glucose fell by approximately 1 mmol/l with exenatide to a nadir of 3.4 +/- 0.1 mmol/l (p < 0.01). During exercise, blood glucose fell with placebo but, unexpectedly, rose with exenatide. Plasma adrenaline (epinephrine) and noradrenaline (norepinephrine), but not cortisol concentrations increased to a greater extent during exercise after exenatide. No participant developed symptomatic hypoglycaemia and the lowest individual blood glucose recorded was 2.8 mmol/l with exenatide at 50 min in the pre-exercise period.

CONCLUSIONS/INTERPRETATION

In non-diabetic participants given exenatide, blood glucose concentrations rise rather than fall during aerobic exercise with an associated greater catecholamine response.