Optimal staffing for a good quality inpatient diabetes service.

INTRODUCTION: Increasing numbers of people admitted to hospital have diabetes and need specialist support. To date, there is no mechanism which can help teams estimate the number of health care professionals they need to provide optimal care for people with diabetes in hospitals. METHODS: The Joint British Diabetes Societies (JBDS) for Inpatient Care Group organised a survey of specialist inpatient diabetes teams in the UK for current staffing and the perception of optimal staffing using mailing lists available through their representative organisations. The results were verified and confirmed by one-to-one conversations with individual respondents and discussed in multiple expert-group meetings to agree on the results. RESULTS: Responses were received from 17 Trusts covering 30 hospital sites. Current diabetes specialist staffing level per 100 people with diabetes in hospital (Median, IQR) for consultants was 0.24 (0.22-0.37), diabetes inpatient specialist nurses was 1.94 (1.22-2.6), dieticians was 0.00 (0.00-0.00), podiatrists was 0.19 (0.00-0.62), pharmacists was 0.00 (0.00-0.37), psychologists was 0.00 (0.00-0.00). The teams also reported that for optimal care the total staff needed for each group (Median, IQR) was much higher; consultants 0.65 (0.50-0.88), specialist nurses 3.38 (2.78-4.59), dieticians 0.48 (0.33-0.72), podiatrists, 0.93 (0.65-1.24), pharmacists, 0.65 (0.40-0.79) and psychologists 0.33 (0.27-0.58). Based on the results of the survey, the JBDS expert group produced an Excel calculator to estimate staffing needs of any hospital site in question just by populating a few of the cells. CONCLUSION: Current inpatient diabetes staffing is much lower than needed in most Trusts who responded to the survey. The JBDS calculator can provide an estimate of the staffing needs of any hospital.

Improving the peri-operative pathway of people with diabetes undergoing elective surgery: the IP3D project.

AIM: To determine whether outcomes for people with diabetes undergoing elective surgery improve following the introduction of innovations in the peri-operative care pathway. METHODS: Following a baseline audit of 185 people with diabetes listed for elective surgery (July to December 2017) with a length of stay > 24 hours, a number of changes in practice were implemented. These included dissemination of a 'diabetes peri-operative passport' to participants preoperatively, formation of a diabetes surgery working group, recruitment of surgical diabetes champions and the roll-out of surgical diabetes study days. Crucial was recruitment of a diabetes peri-operative nurse, whose role included engaging and educating others and supporting individuals throughout their peri-operative diabetes care. Records of 166 individuals listed for surgery during the implementation period (July to December 2018) were then audited using the same methodology. RESULTS: The availability of a recent HbA1c measurement significantly increased (63% vs 92%; P ≤ 0.001). The mean HbA1c of those seen for optimizations by the diabetes peri-operative nurse significantly decreased [84 mmol/mol (9.8%) vs 62 mmol/mol (7.8%); P ≤ 0.001]. Recurrent hypoglycaemia significantly decreased (7.0% vs 0.6%; P = 0.002) and the mean number of hyperglycaemic events in people experiencing hyperglycaemia almost halved (3.0 vs 1.7; P=0.007). The mean length of hospital stay significantly decreased (4.8 vs 3.3 days; P=0.001) and, crucially, 30-day readmissions did not increase (12% vs 9%; P=0.307). Postoperative complications significantly decreased (28% vs 16%; P=0.008), including a composite of dysglycaemic complications, poor wound healing, wound infection and other infections (12% vs 5.4%; P=0.023). CONCLUSION: The new pathway improved important peri-operative outcomes for people with diabetes undergoing elective surgery with the potential for cost savings. These findings could have important implications for peri-operative care on a wider scale.

Impact of the Diabetes Inpatient Care and Education (DICE) project on length of stay and mortality.

AIM: To determine whether the Diabetes Inpatient Care and Education (DICE) programme, a whole-systems approach to managing inpatient diabetes, reduces length of stay, in-hospital mortality and readmissions. RESEARCH DESIGN AND METHODS: Diabetes Inpatient Care and Education initiatives included identification of all diabetes admissions, a novel DICE care-pathway, an online system for prioritizing referrals, use of web-linked glucose meters, an enhanced diabetes team, and novel diabetes training for doctors. Patient administration system data were extracted for people admitted to Ipswich Hospital from January 2008 to June 2016. Logistic regression was used to compare binary outcomes (mortality, 30-day readmissions) 6 months before and after the intervention; generalized estimating equations were used to compare lengths of stay. Interrupted time series analysis was performed over the full 7.5-year period to account for secular trends. RESULTS: Before-and-after analysis revealed a significant reduction in lengths of stay for people with and without diabetes: relative ratios 0.89 (95% CI 0.83, 0.97) and 0.93 (95% CI 0.90, 0.96), respectively; however, in interrupted time series analysis the change in long-term trend for length of stay following the intervention was significant only for people with diabetes (P=0.017 vs P=0.48). Odds ratios for mortality were 0.63 (0.48, 0.82) and 0.81 (0.70, 0.93) in people with and without diabetes, respectively; however, the change in trend was not significant in people with diabetes, while there was an apparent increase in those without diabetes. There was no significant change in 30-day readmissions, but interrupted time series analysis showed a rising trend in both groups. CONCLUSION: The DICE programme was associated with a shorter length of stay in inpatients with diabetes beyond that observed in people without diabetes.

Guidelines for the management of diabetes services and patients during the COVID-19 pandemic.

The UK National Diabetes Inpatient COVID Response Group was formed at the end of March 2020 to support the provision of diabetes inpatient care during the COVID pandemic. It was formed in response to two emerging needs. First to ensure that basic diabetes services are secured and maintained at a time when there was a call for re-deployment to support the need for general medical expertise across secondary care services. The second was to provide simple safe diabetes guidelines for use by specialists and non-specialists treating inpatients with or suspected of COVID-19 infection. To date the group, comprising UK-based specialists in diabetes, pharmacy and psychology, have produced two sets of guidelines which will be continually revised as new evidence emerges. It is supported by Diabetes UK, the Association of British Clinical Diabetologists and NHS England.

The cost of diabetic foot ulcers and amputations to the National Health Service in England.

AIM: To estimate the healthcare costs of diabetic foot disease in England. METHODS: Patient-level data sets at a national and local level, and evidence from clinical studies, were used to estimate the annual cost of health care for foot ulceration and amputation in people with diabetes in England in 2014-2015. RESULTS: The cost of health care for ulceration and amputation in diabetes in 2014-2015 is estimated at between £837 million and £962 million; 0.8% to 0.9% of the National Health Service (NHS) budget for England. More than 90% of expenditure was related to ulceration, and 60% was for care in community, outpatient and primary settings. For inpatients, multiple regression analysis suggested that ulceration was associated with a length of stay 8.04 days longer (95% confidence interval 7.65 to 8.42) than that for diabetes admissions without ulceration. CONCLUSIONS: Diabetic foot care accounts for a substantial proportion of healthcare expenditure in England, more than the combined cost of breast, prostate and lung cancers. Much of this expenditure arises through prolonged and severe ulceration. If the NHS were to reduce the prevalence of diabetic foot ulcers in England by one-third, the gross annual saving would be more than £250 million. Diabetic foot ulceration is a large and growing problem globally, and it is likely that there is potential to improve outcomes and reduce expenditure in many countries.

Could FreeStyle Libre™ sensor glucose data support decisions for safe driving?

AIM: Many countries require individuals with diabetes to adhere to standards regarding blood glucose testing in order to be granted or retain a driving licence. Currently, interstitial glucose results may not be used. The aim of this study was to determine whether interstitial glucose measurements using flash glucose-sensing technology can provide additional information to augment safe driving. METHODS: Sensor data from two European studies (NCT02232698 and NCT02082184) of the FreeStyle Libre Glucose Monitoring System™ in insulin-treated Type 1 and Type 2 diabetes, 241 and 224 participants respectively, were used to determine the frequency of a low interstitial sensor glucose result (< 3.9 mmol/l) up to 4 h subsequent to a daytime (07:00-21:00 h) capillary blood glucose result ≥ 5 mmol/l. RESULTS: Within 4 h of a capillary blood glucose result ≥ 5 mmol/l a sensor glucose result of < 3.9 mmol/l occurred on 22.0% of occasions (2573 of 11 706 blood glucose readings) for those with Type 1 diabetes, and 8.4% of occasions (699/8352) for those with Type 2 diabetes; 13.8% (1610/11 628) and 4.4% (365/8203) within 2 h, and 10.0% (1160/11 601) and 3.1% (254/8152) within 1.5 h. Analysis of sensor glucose results 5-7 mmol/l demonstrated the glucose trend arrow descending on 14.7% (1163/7894, Type 1 diabetes) and 9.4% (305/3233, Type 2 diabetes) of occasions. CONCLUSIONS: Sensor-based glucose information with directional arrows has the potential to support assessment of safe glucose levels associated with driving and offers distinct advantages over blood glucose testing for individuals with Type 1 and Type 2 diabetes to concord with driving safety standards.

Temporal and external validation of a prediction model for adverse outcomes among inpatients with diabetes.

AIM: To temporally and externally validate our previously developed prediction model, which used data from University Hospitals Birmingham to identify inpatients with diabetes at high risk of adverse outcome (mortality or excessive length of stay), in order to demonstrate its applicability to other hospital populations within the UK. METHODS: Temporal validation was performed using data from University Hospitals Birmingham and external validation was performed using data from both the Heart of England NHS Foundation Trust and Ipswich Hospital. All adult inpatients with diabetes were included. Variables included in the model were age, gender, ethnicity, admission type, intensive therapy unit admission, insulin therapy, albumin, sodium, potassium, haemoglobin, C-reactive protein, estimated GFR and neutrophil count. Adverse outcome was defined as excessive length of stay or death. RESULTS: Model discrimination in the temporal and external validation datasets was good. In temporal validation using data from University Hospitals Birmingham, the area under the curve was 0.797 (95% CI 0.785-0.810), sensitivity was 70% (95% CI 67-72) and specificity was 75% (95% CI 74-76). In external validation using data from Heart of England NHS Foundation Trust, the area under the curve was 0.758 (95% CI 0.747-0.768), sensitivity was 73% (95% CI 71-74) and specificity was 66% (95% CI 65-67). In external validation using data from Ipswich, the area under the curve was 0.736 (95% CI 0.711-0.761), sensitivity was 63% (95% CI 59-68) and specificity was 69% (95% CI 67-72). These results were similar to those for the internally validated model derived from University Hospitals Birmingham. CONCLUSIONS: The prediction model to identify patients with diabetes at high risk of developing an adverse event while in hospital performed well in temporal and external validation. The externally validated prediction model is a novel tool that can be used to improve care pathways for inpatients with diabetes. Further research to assess clinical utility is needed.

Perioperative passport: empowering people with diabetes along their surgical journey.

AIM: To determine whether a handheld 'perioperative passport' could improve the experience of perioperative care for people with diabetes and overcome some of the communication issues commonly identified in inpatient extracts. METHODS: Individuals with diabetes undergoing elective surgery requiring at least an overnight stay were identified via a customized information technology system. Those allocated to the passport group were given the perioperative passport before their hospital admission. A 26-item questionnaire was completed after surgery by 50 participants in the passport group (mean age 69 years) and by 35 participants with diabetes who followed the usual surgical pathway (mean age 70 years). In addition, the former group had a structured interview about their experience of the passport. RESULTS: The prevalence of those who reported having received prior information about their expected diabetes care was 35% in the control group vs 92% in the passport group (P<0.001). The passport group found the information given significantly more helpful (P<0.001), including the advice on medication adjustment (P=0.008). Furthermore, those with the passport were more involved in planning their diabetes care (P <0.001), less anxious whilst in hospital (P<0.044) and better prepared to manage their diabetes on discharge (P≤0.001). The mean length of hospital stay was shorter in the passport group, although the difference did not reach significance (4.4 vs 6.5 days; P<0.058). Content analysis indicated that the passport was well liked and innovative. CONCLUSION: Our data indicate that the perioperative passport is effective in both informing and involving people in their diabetes care throughout the perioperative period.

Predicting inpatient hypoglycaemia in hospitalized patients with diabetes: a retrospective analysis of 9584 admissions with diabetes.

AIMS: To explore whether a quantitative approach to identifying hospitalized patients with diabetes at risk of hypoglycaemia would be feasible through incorporation of routine biochemical, haematological and prescription data. METHODS: A retrospective cross-sectional analysis of all diabetic admissions (n=9584) from 1 January 2014 to 31 December 2014 was performed. Hypoglycaemia was defined as a blood glucose level of <4 mmol/l. The prediction model was constructed using multivariable logistic regression, populated by clinically important variables and routine laboratory data. RESULTS: Using a prespecified variable selection strategy, it was shown that the occurrence of inpatient hypoglycaemia could be predicted by a combined model taking into account background medication (type of insulin, use of sulfonylureas), ethnicity (black and Asian), age (≥75 years), type of admission (emergency) and laboratory measurements (estimated GFR, C-reactive protein, sodium and albumin). Receiver-operating curve analysis showed that the area under the curve was 0.733 (95% CI 0.719 to 0.747). The threshold chosen to maximize both sensitivity and specificity was 0.15. The area under the curve obtained from internal validation did not differ from the primary model [0.731 (95% CI 0.717 to 0.746)]. CONCLUSIONS: The inclusion of routine biochemical data, available at the time of admission, can add prognostic value to demographic and medication history. The predictive performance of the constructed model indicates potential clinical utility for the identification of patients at risk of hypoglycaemia during their inpatient stay.

Impact of the Diabetes Inpatient Care and Education (DICE) project and the DICE Care Pathway on patient outcomes and trainee doctor's knowledge and confidence.

AIM: To evaluate the impact of the Diabetes Inpatient Care and Education project and a comprehensive diabetes care pathway, the Diabetes Inpatient Care and Education Care Pathway, on patient outcomes and on the knowledge and confidence of trainee doctors. METHODS: The effect on patient outcomes was evaluated by comparing the National Diabetes Inpatient Audit data before (2012) and after (2013) implementing the Diabetes Inpatient Care and Education project. The impact on trainee doctors was evaluated using the Modified Kirkpatrick model. Just before the project began and again 3 months later, trainee doctors were surveyed to assess their knowledge and confidence in inpatient diabetes care. RESULTS: Patient harm was found to have been reduced significantly when National Diabetes Inpatient Audit data for 2012 and 2013 were compared. Severe hypoglycaemia decreased from 15.4 to 9.7%, medication errors from 56.9 to 21.1% and insulin errors from 31 to 7%. Across the 96 trainee doctors surveyed, the mean (sd) knowledge and confidence scores increased significantly (P < 0.001 for both) from 57.1 (16.8) and 61.8 (14.9) to 68.4 (13.3) and 74.3 (11.7), respectively. CONCLUSION: The Diabetes Inpatient Care and Education project and the Diabetes Inpatient Care and Education Care Pathway improved patient outcomes and the knowledge and confidence of trainee doctors in this hospital. The impact of a similar project in other hospitals needs to be evaluated.

Serious harm from inpatient hypoglycaemia: a survey of hospitals in the UK.

AIM: To estimate the incidence of serious harm to inpatients with diabetes from hypoglycaemia. METHOD: An anonymised questionnaire was e-mailed to lead organisers at the 142 acute NHS Trusts that contributed to the National Diabetes Inpatient Audit 2012. Each diabetes team was asked collectively to recall and report any serious adverse events from inpatient hypoglycaemia in the previous year. A total of 83 Trusts agreed to participate. Serious harm was defined as death, a cardiac or cerebral event or a fall resulting in permanent physical injury or fracture. RESULTS: A total of 41 Trusts returned the survey. Of these, only 28 (68.3%) were confident that robust methods existed in their Trust to ensure all such events were reported, and only 23 (56.1%) were confident that all such events were reported to the diabetes team. Despite these reporting concerns, the retrospective nature of the survey and the reliance on recall, 12 serious adverse events were reported from nine trusts: three deaths; two cases of permanent cerebral damage; two successfully resuscitated cardiac arrests; three seizures; and two undefined events. Insulin therapy was implicated in 10 events. Importantly, three events with two deaths occurred in patients who had received insulin/dextrose to correct hyperkalaemia; only one of whom had diabetes. CONCLUSIONS: An alarming number of serious adverse events was reported: 12 serious adverse events with three deaths over a 1-year period in 41 Trusts. This may be the tip of the iceberg, considering the potential under-reporting. Robust reporting mechanisms are required to determine the full extent of this serious preventable harm.

Cost of diabetic foot disease to the National Health Service in England.

AIM: To estimate the annual cost of diabetic foot care in a universal healthcare system. METHODS: National datasets and economic modelling were used to estimate the cost of diabetic foot disease to the National Health Service in England in 2010-2011. The cost of hospital admissions specific to foot disease or amputation was estimated from Hospital Episode Statistics and national tariffs. Multivariate regression analysis was used to estimate the impact of foot disease on length of stay in admissions that were not specific to foot disease or amputation. Costs in other areas were estimated from published studies and data from individual hospitals. RESULTS: The cost of diabetic foot care in 2010-2011 is estimated at £580 m, almost 0.6% of National Health Service expenditure in England. We estimate that more than half this sum (£307 m) was spent on care for ulceration in primary and community settings. Of hospital admissions with recorded diabetes, 8.8% included ulcer care or amputation. Regression analysis suggests that foot disease was associated with a 2.51-fold (95% CI 2.43-2.59) increase in length of stay.The cost of inpatient ulcer care is estimated at £219 m, and that of amputation care at £55 m. CONCLUSIONS: The cost of diabetic foot disease is substantial. Ignorance of the cost of current care may hinder commissioning of effective services for prevention and management in both community and secondary care.

The Ipswich Touch Test: a simple and novel method to screen patients with diabetes at home for increased risk of foot ulceration.

AIMS: The Ipswich Touch Test is a novel method to detect subjects with diabetes with loss of foot sensation and is simple, safe, quick, and easy to perform and teach. This study determines whether it can be used by relatives and/or carers to detect reduced foot sensation in the setting of the patient's home. METHODS: The test involves lightly and briefly (1-2 s) touching the tips of the first, third and fifth toes of both feet with the index finger. Reduced foot sensation was defined as ≥ 2 insensate areas. Patients due to attend clinic over a 4-week period were invited by post. The invitation contained detailed instructions and a sheet for recording the results. The findings were compared with those obtained in clinic using the 10-g monofilament at the same six sites. RESULTS: Of 331 patients (174 males), 25.1% (n = 83) had ≥ 2 insensate areas to 10-g monofilament testing. Compared with this, the Ipswich Touch Test at home had a sensitivity of 78.3% and a specificity of 93.9%. The predictive values of detecting 'at-risk' feet were positive at 81.2% and negative at 92.8%. The likelihood ratios were positive at 12.9 and negative at 0.23. CONCLUSIONS: With clearly written instructions, this simple test can be used by non-professionals to accurately assess for loss of protective sensation. We believe that the Ipswich Touch Test may also be a useful educational adjunct to improve awareness of diabetes foot disease in patients and relatives alike and empower them to seek appropriate care if sensation was found to be abnormal.

Diurnal temporal patterns of hypoglycaemia in hospitalized people with diabetes may reveal potentially correctable factors.

AIM: To determine whether diurnal temporal variations in hypoglycaemic frequency occur in hospitalized patients. METHODS: Hypoglycaemic events were identified in a snapshot bedside audit of capillary blood glucose results from diabetes charts of all inpatients receiving insulin or a sulphonylurea (with or without insulin) on 2 days separated by 6 weeks. Additionally, capillary blood glucose measurements were remotely captured over 2 months, in the same category of patients, and analysed for temporal patterns. Hypoglycaemia was defined as 'severe' when the capillary blood glucose was < 3.0 mmol/l and 'mild' when the capillary blood glucose was between 3.0 and 3.9 mmol/l. RESULTS: The bedside audit found that 74% of those audited experienced a hypoglycaemia event. Eighty-three per cent of all hypoglycaemic events and 70% of severe events were recorded between 21.00 and 09.00 h. This was confirmed in the longer duration remote monitoring study where 70% of all hypoglycaemic events and 66% of severe events occurred between 21.00 and 09.00 h. CONCLUSION: Hypoglycaemia occurs more frequently between 21.00 and 09.00 h in hospitalized patients receiving treatments that can cause hypoglycaemia. This may be related to insufficient carbohydrate intake during this period, and is potentially preventable by changes in catering practice.

Small fibre dysfunction, microvascular complications and glycaemic control in type 1 diabetes: a case-control study.

AIMS/HYPOTHESIS: The aim of this study was to determine the influence of microvascular disease on C-fibre function in patients with type 1 diabetes of moderate duration. METHODS: The axon-reflex flare area induced on the dorsum of the foot by local skin heating to 47 °C was measured with a laser Doppler imager (LDI) in sex-, age- and height-matched groups with type 1 diabetes, with and without microvascular disease (MV+ and MV-, respectively) and in healthy controls (HC). Each group consisted of 24 individuals and all were free from clinical neuropathy (neuropathy disability score <3 and Toronto clinical neuropathy score <5). RESULTS: LDI flare (LDIflare) was reduced in MV+ compared with HC (5.1 ± 1.8 vs 10.0 ± 3.1 cm², p < 0.0001) and MV- groups (9.9 ± 2.9 cm², p < 0.0001). MV- and HC groups did not differ. There was no difference in diabetes duration between MV- and MV+ groups (17.5 ± 5.7 and 20.1 ± 5.2 years, p = 0.21) nor current HbA(1c) (MV- 8.0 ± 1.2% [64 ± 10 mmol/mol]; MV+ 8.0 ± 0.9% [64 ± 9 mmol/mol], p = 0.53); neither variable correlated with flare size. In contrast, duration-averaged HbA(1c) was higher in the MV+ group (8.6 ± 0.9% [70 ± 9 mmol/mol] vs 7.6 ± 0.6% [60 ± 7 mmol/mol], p < 0.001) and correlated with LDIflare size (r = -0.50, p < 0.001). Triacylglycerols were higher in MV+ compared with MV- (1.23 ± 0.121 vs 0.93 ± 0.7 mmol/l, p = 0.04), but other metabolic variables did not differ between the groups. CONCLUSIONS/INTERPRETATION: We have shown that glycaemic burden and the presence of microvascular complications are associated with small fibre dysfunction in type 1 diabetes.

Audit of acute Charcot's disease in the UK: the CDUK study.

AIMS/HYPOTHESIS: We studied factors associated with the development and resolution of acute Charcot foot using a web-based observational study. METHODS: Clinicians managing cases of acute Charcot foot in the UK and Ireland between June 2005 and February 2007 were invited to register anonymised details on a secure website. RESULTS: A total of 288 cases (age 57.0 ± 11.3 years [mean ± SD]; 71.2% male) were registered from 76 centres. Of these, 36% of patients recalled an episode of relevant trauma in the preceding 6 months, while 12% had had surgery to the affected foot. In 101 (35%) cases, ulceration was present at registration and 20% of these had osteomyelitis. Non-removable off-loading devices were used at presentation in 35.4% of cases, with removable off-loading used in 50%. Data on resolution were available for 219 patients. The median time to resolution was 9 months in patients whose initial management included the use of non-removable off-loading, compared with 12 months in the remainder (p = 0.001). Bisphosphonates were administered intravenously in 25.4% and orally in 19.4% of cases. The median time to resolution in patients who received bisphosphonates was 12 months and was longer than in those who did not (10 months, p = 0.005). CONCLUSIONS/INTERPRETATION: The median time to resolution was longer than in earlier series. Although limited by being observational and non-randomised, these data suggest that the use of non-removable off-loading at presentation may shorten the time to resolution. They provide no evidence to indicate that the use of bisphosphonates is beneficial.

Pathophysiology of postprandial hyperglycaemia in women with type 1 diabetes during pregnancy.

AIMS/HYPOTHESIS: Although maternal hyperglycaemia is associated with increased risk of adverse pregnancy outcome, the mechanisms of postprandial hyperglycaemia during pregnancy are poorly understood. We aimed to describe glucose turnover in pregnant women with type 1 diabetes, according to stage of gestation (early vs late gestation). METHODS: The rates of systemic glucose appearance (R(a)) and glucose disposal (R(d)) were measured in ten pregnant women with type 1 diabetes during early (12-16 weeks) and late (28-32 weeks) gestation. Women ate standardised meals--a starch-rich 80 g carbohydrate dinner and a sugar-rich 60 g carbohydrate breakfast--and fasted between meals and overnight. Stable-label isotope tracers ([6,6-(2)H(2)]glucose and [U-(13)C]glucose) were used to determine R(a), R(d) and glucose bioavailability. Closed-loop insulin delivery maintained stable glycaemic conditions. RESULTS: There were no changes in fasting R(a) (10 ± 2 vs 11 ± 2 µmol kg(-1) min(-1); p = 0.32) or fasting R(d) (11 ± 2 vs 11 ± 1 µmol kg(-1) min(-1); p = 0.77) in early vs late gestation. There was increased hepatic insulin resistance (381 ± 237 vs 540 ± 242 µmol kg(-1) min(-1) × pmol/l; p = 0.04) and decreased peripheral insulin sensitivity (0.09 ± 0.04 vs 0.05 ± 0.02 µmol kg(-1) min(-1) per pmol/l dinner, 0.11 ± 0.05 vs 0.07 ± 0.03 µmol kg(-1) min(-1) per pmol/l breakfast; p = 0.002) in late gestation. It also took longer for insulin levels to reach maximal concentrations (49 [37-55] vs 71 [52-108] min; p = 0.004) with significantly delayed glucose disposal (108 [87-125] vs 135 [110-158] min; p = 0.005) in late gestation. CONCLUSIONS/INTERPRETATION: Postprandial glucose control is impaired by significantly slower glucose disposal in late gestation. Early prandial insulin dosing may help to accelerate glucose disposal and potentially ameliorate postprandial hyperglycaemia in late pregnancy. TRIAL REGISTRATION: ISRCTN 62568875 FUNDING: Diabetes UK Project Grant BDA 07/003551. H.R. Murphy is funded by a National Institute for Health Research (NIHR) research fellowship (PDF/08/01/036). Supported also by the Juvenile Diabetes Research Foundation (JDRF), Abbott Diabetes Care (Freestyle Navigator CGM and sensors free of charge), Medical Research Council Centre for Obesity and Related Metabolic Diseases and NIHR Cambridge Biomedical Research Centre.