Clinical care and other categories posters: Pregnancy

Quality improvement in the diabetes and pregnancy service: Improving folic acid uptake, undertaking third trimester HbA1c and increasing access to pumps

P King¹, L Bancroft², K Gale¹, E Marchant¹, K Dent²

¹Diabetes and Endocrinology, University Hospitals of Derby and Burton NHS FT, Derby, UK, ²Obstetrics and Gynaecology, University Hospitals of Derby and Burton NHS FT, Derby, UK

Aims As one of 19 centres that are part of the diabetes and pregnancy quality improvement collaborative, we undertook a quality improvement project locally focusing on areas of suboptimal performance in the 2016 national diabetes and pregnancy audit (NDIPA). Specifically, we aimed to:1Increase the percentage of women conceiving on Folic Acid 5 mg daily from 34 to 50%.2Increase the percentage of women with type 1 diabetes who are pregnant or preconceptual accessing pumps from 15 to 25%.3Increase the percentage of women having a third trimester HbA1c from 5 to 75%

Method Data from the 35 pregnancies submitted to the 2016 NDIPA was compared with the results from 53 pregnancies between April 2018 and March 2019. Root cause analysis and driver diagrams were used to inform strategies to improve outcomes. Run charts were used to assess improvement continuously and reviewed 3 monthly to inform plan do study act cycles.

Results

• 32 women (60%) were seen preconceptually.

• Run charts for Folic Acid and third trimester HbA1c demonstrated shifts three months after the interventions were implemented. The improvements were sustained.

• Overall, folic acid uptake increased from 34 to 58% and 3rd trimester Hba1c was undertaken in 83%, both above target.

• 7/29 (25%) women of women with type 1 diabetes accessed pumps before or during pregnancy, which is on target.

Conclusion The combination of run charts and conventional analysis confirm quality improvements in our diabetes and pregnancy service which were sustained over a 12-month period.

Follow-up screening for diabetes after gestational diabetes (GDM)

SJ McNulty, P Wilkinson, S Michaels, S Pankajar, H Davies, H Sulivan, S Bujawansa

Department of Diabetes, St Helens and Knowsley NHS Trusts, St Helens, UK

Aims NICE guidelines for screening state that all GDM patients should be screened for diabetes at 6–12 weeks postpartum. Previous published data have given a national rate of less than 20%¹. We reviewed our screening rate after the interventions of: discussion with the patient at first and last visit, written instructions to the patients’ GPs and copying these instructions to patients at first and last visit regarding follow-up screening.

Methods We reviewed our joint diabetes antenatal clinic (JANC) new patient lists from December 2018 through March 2019 collecting demographic data (including post codes), diagnosis, estimated date of delivery, HbA1c at diagnosis and follow-up screening (type, date and result).

Results 53 patients (aged 21–41 years) were identified of which 40 had GDM. Patients came from various post codes – six Liverpool, six Warrington, one Widnes and one Chester. All the OGTTs were diagnostic. HbA1c at first visit for our GDM population ranged between 30 and 61 mmol/mol. Eighteen patients had follow-up screening of which 16 were HbA1c. Twelve patients were at moderate risk, five at high risk and one confirmed a diagnosis of type 2 diabetes.

Conclusions With this extensive intervention we were able to demonstrate at least 45% follow-up screening of our GDM patient (over twice the national average). This figure maybe higher as we service patients from outside of our CCG and therefore some of the patients who appear not to have had a follow-up screen may have had them through a different laboratory.

References ¹McGovern A et al. Diabetes screening after gestational diabetes in England: a quantitative retrospective cohort study. Br J Gen Pract 2014;64(618):e17–23

Acknowledgement Diabetes Joint Antenatal Service

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

Pregnancy outcomes in a diverse London population of women with type 1 diabetes

SG Wijetilleka¹, J Burrows¹, E Grocholski¹, A Fayadh², D Gatmaytan¹, T Lewis¹, R Hamid², P Tripathi², E Hui¹, M Rahman¹, C Adjetey², A Vyas², T Zikhali²

¹Diabetes and Endocrinology, Northwick Park Hospital, Harrow, UK, ²Obstetrics and Gynaecology, Northwick Park Hospital, Harrow, UK

Aim To review the pregnancy outcomes in women with type 1 diabetes at a district general hospital with over 5,000 livebirths per annum.

Design A retrospective observational study of all women with type 1 diabetes who attended our antenatal service between January and December 2017. The impact of multidisciplinary care, based on NICE guideline NG3 (Diabetes in pregnancy: management from preconception to the postnatal period), on maternal and foetal outcomes was assessed.

Methods Women with pre-existing type 1 diabetes were selected using clinic outcome data. Data were collected on population demographics, interpreter use, maternal HbA1c at booking, gestational age, BM, mode of delivery and neonatal outcomes.

Results A total of 10 women with type 1 diabetes attended in 2017 with a mean age of 35.1 years (SD ± 3.8). 50% were of non-white ethnicity. All patients attended prior to 10 weeks gestation. Mean BMI at booking was 30 kg/m² (SD ± 4.8). Only one patient required an interpreter. The mean duration of diabetes was 17.1 years (SD ± 6.52). Mean HbA1c at first visit was 64 mmol/l (SD ± 10.26). 60% attended retinal screening in the first trimester and retinopathy outcomes were unchanged during pregnancy. Mean gestation at delivery was 37 weeks and 75% of deliveries were via caesarean section. No neonatal hypoglycaemia or macrosomia occurred; two newborns received special care.

Conclusion The results show that our antenatal diabetic services and pregnancy outcomes compare well with national data, despite the co-morbidity (advanced age, elevated BMI and 60% use of pre-pregnancy folic acid) and deprivation faced by our cohort.

Dietitian-led care is safe and effective in women with gestational diabetes: The dietitian-led secondary care multidisciplinary gestational diabetes service evaluation 2018 audit

A Bowes¹, J Beeson², M Conde², K Thomas², MG Masding¹

¹Diabetes Centre, Poole Hospital NHS FT, Poole, UK, ²Maternity Unit, Poole Hospital NHS FT, Poole, UK

Background In our weekly joint obstetric diabetic clinic dietitians initially review women with gestational diabetes (GDM) every one to two weeks with obstetric review after growth scans. The extended diabetes team see women when insulin is required.

Aims Audit of dietitian-led care for GDM to ensure same outcomes whether have just had dietetic review or have seen the wider diabetes team.

Methods A retrospective audit of routine clinic data collected from January to December 2018.

Results Of ninety-nine mothers with GDM (mean±SD; 32.7 ± 5.4 years; BMI 31.4 ± 6.3 kg/m²) who delivered in 2018, 41 women were treated with dietary therapy only (D), 47 were on only metformin (M) and 16 on insulin ± metformin (I+M). Gestational age at the delivery for the D group was 39.1 ± 1.4 weeks, this was 7.3 days longer then for M group (p < 0.05) and 10.3 days longer for I±M (p < 0.05). Baby's average weight was 3.30 ± 0.4 kg with no difference between treatment groups. Macrosomia rates were 1%, and caesarean section rates 46.5%; for both parameters, no significant difference between groups. Mothers’ weight gain was 6.4 ± 6.1 kg, which was higher for D compare to M and I±M groups (p < 0.05). Average number of dietetic appointments per patient were 4.8 ± 2.4, phone reviews 2.1 ± 1.6. Number of diabetes consultant or nurses’ appointments were 2.8 ± 3.9 and phone reviews 1.7 ± 2.8 per patient. Duration of post-natal stay was 2.2 ± 1.7 days with no difference between treatment groups.

Conclusion Women with GDM can safely be managed by specialist dietitians, without the involvement of the wider diabetes team until insulin therapy is required. This allows the diabetes MDT to use resources in the most efficient way.

A patient-lead rapid treatment intensification protocol for fasting glucose substantially improves glycaemic control in gestation diabetes

KD Hirwa, A Hashimi, R Thompson, V Creese, S Havill, T Sanders, B Vaidya, AT Hattersley, A McGovern

Macleod Diabetes and Endocrine Centre (MDEC), Royal Devon and Exeter Hospital, Exeter, UK

Aims Elevated fasting blood glucose in gestational diabetes (GDM), is the strongest predictor of higher birth weight babies and adverse neonatal outcomes but is hard to treat. We assessed the impact of implementing a rapid patient-lead insulin treatment protocol for fasting glycaemic control in GDM.

Methods We introduced a protocol for GDM where overnight basal insulin was introduced with fasting readings ≥5.3 mmol/l with subsequent insulin up-titration by patients of 4 units daily after every fasting glucose value ≥5.0 mmol/mol without further consultation. Women had at least weekly review of their glucose values by nursing staff. This automatic uptitration replaced patients reporting their glucose values to nursing staff and being advised on the appropriate insulin dose. We audited medication use and glucose values at 36 weeks gestation, and maternal and neonatal outcomes in 27 women completing pregnancy before and 30 women after implementing the protocol.

Results After implementation, insulin use increased (53% v 22%;p=0.033) and those on insulin were on a substantially higher dose at 36 weeks (45 vs 2 1units/day; 0.50 vs 0.20units/kg/day; p=0.027). Thirty-six week mean fasting glucose was lower overall (4.6 vs 5.7 mmol/l; p=0.005) and those on insulin (4.4 vs 5.6 mmol/l; p=0.002). Patient acceptance was high and no women experienced severe hypoglycaemia. There was no change in median delivery (38 + 2 v 38 + 1 weeks) or caesarean sections (23 vs 33%; p=0.59) and birth weight was not significantly reduced (mean reduction −120 g; 95%CI −369,129 g; p=0.325).

Conclusions A patient-lead rapid treatment intensification protocol for overnight insulin can result in rapid dose titration and substantially improve fasting glycaemic control in GDM.

An assessment of predictors for maternal glycaemic variability in pregnancies affected by type 1 diabetes and its effect on foetal birth weight

JA Allen^1,2, AD Dornhorst¹, RA Agha-Jaffar²

¹Department of Diabetes and Endocrinology, Hammersmith Hospital, Imperial College NHS Trust, London, UK, ²Division of Diabetes, Endocrinology and Metabolism, G3 Medical School Building, Imperial College London, London, UK

Aims To evaluate the relationship between maternal factors and glycaemic variability during pregnancies affected by type 1 diabetes, and the association between this variability and excess foetal growth- a major driver of both maternal and foetal complications.

Methods Thirteen pregnant women with type 1 diabetes who attended one of two inner-city antenatal clinics and received continuous glucose monitoring during their pregnancy were recruited. Baseline characteristics at booking, together with foetal birth weight and centile in those who delivered, were recorded. Indices of glycaemic variability, primarily assessed via coefficient of variation (CV%), were calculated from the glucose data. A separate data set of 14 similar pregnancies was also incorporated into the analyses.

Results CV% correlated well with other glycaemic variability metrics such as LBGI (r²=0.77, p < 0.001) and HBGI (r²=0.35, p=0.001). Neither maternal BMI (r²=0.0032, p=0.78) nor age (r²=0.64, p=0.69) correlated with CV%. Ethnicity was the only factor with a significant relationship with CV%: the Arab/North African group having a greater CV% compared to that of White Caucasians (54.44 vs 36.88, p=0.035). Adjusted foetal birth weight centile was associated with maternal booking HbA1c (r²=0.18, p=0.046), but not CV% during pregnancy (r²=0.032, p=0.424).

Conclusions CV% appears to be a suitable index of glycaemic variability during pregnancies complicated by type 1 diabetes, able to represent hyper- and hypoglycaemic excursions, but its prediction from maternal characteristics is challenging. Further research is required, but greater understanding of the influences upon, and effects of, glycaemic variability may aid risk prediction during these pregnancies.

Effect of Freestyle Libre use on HbA1c in patients with type 1 diabetes and pregnancy

JL Cowan, C Hogg, C Challapalli, V McAulay

Diabetes Department, University Hospital Crosshouse, Kilmarnock, UK

Refer to Oral number A9.

Steroid administration for late elective caesarean section is associated with neonatal hypoglycaemia

D Simmons^1,3, KJ Gupta¹, R Rajagopal^1,3, F King²

¹Macarthur Diabetes Service, Campbelltown Hospital, Campbelltown, Australia, ²Midwifery, Campbelltown Hospital, Campbelltown, Australia, ³Macarthur Clinical School, Western Sydney University, Campbelltown, Australia

Aims Women who undergo elective caesarean section <39 weeks gestation benefit from the administration of corticosteroids to reduce neonatal respiratory morbidity. However, it is unclear whether such benefits also exist for women with pregnancies complicated by diabetes. The aim of this study was to compare neonatal hypoglycaemia and respiratory morbidity rates in pregnancies complicated by diabetes following elective caesarean section (ECS) from ≥37 weeks gestation with and without maternal corticosteroid administration.

Methods Retrospective cohort study of women with any form of diabetes in pregnancy undergoing ECS at ≥37 weeks gestation between 1 May 2016 and 30 April 2018. Primary outcomes were admission rates to the neonatal intensive care unit (NICU) for respiratory distress syndrome (RDS)/transient tachypnoea of the newborn (TTN) and/or neonatal hypoglycaemia.

Results Of 102 women with pregnancies complicated by diabetes (94 with gestational diabetes), 32 received corticosteroids (31 with gestational diabetes). Women with and without corticosteroid administration were similar with regards to age (33 ± 5 years), body mass index (33 ± 7 kg/m²), smoking, type of diabetes, total daily insulin dose (29 IU) and second or third trimester HbA1c (36 mmol/mol). NICU admission rates for neonatal hypoglycaemia were significantly higher (5.8% vs 21.9%, p=0.016), and RDS/TTN were non-significantly higher (8.6% vs 12.5%, p=0.536) following corticosteroid administration.

Conclusions Corticosteroids were not beneficial among women at ≥37 weeks gestation with any form of diabetes in pregnancy undergoing ECS and indeed may be harmful. Larger studies are urgently needed to confirm these findings. We have ceased the use of steroids for women under these circumstances in our hospital.

Barriers to completing oral glucose tolerance testing in women at risk of gestational diabetes

EH Lachmann¹, RA Fox¹, RA Dennison², JA Usher-Smith², CL Meek^3,4,5, CE Aiken^6,7

¹School of Clinical Medicine, University of Cambridge, Cambridge, UK, ²Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK, ³Institute of Metabolic Science, University of Cambridge, Cambridge, UK, ⁴Department of Clinical Biochemistry, University of Cambridge, Cambridge, UK, ⁵Wolfson Diabetes and Endocrinology Clinic, Cambridge University Hospitals, Cambridge, UK, ⁶Department of Obstetrics and Gynaecology, Cambridge University Hospitals, Cambridge, UK, ⁷University Department of Obstetrics and Gynaecology, University of Cambridge, Cambridge, UK

Aims Complications of gestational diabetes can be mitigated if the diagnosis is recognised. However, some at-risk women do not complete antenatal diagnostic oral glucose tolerance testing (OGTT). We aimed to understand the reasons contributing to non-completion, particularly to identify modifiable factors.

Methods In total, 1,906 women at a single tertiary UK obstetrics centre (2018–2019) were invited for OGTT, based on risk-factor assessment. Demographic information, test results and reasons for non-completion were collected from the medical record. Logistic regression was used to analyse factors associated with non-completion.

Results 13% (n=242) of women failed to complete at least one OGTT, of whom 32% (n=78) never completed testing. In adjusted analysis, any non-completion was associated with younger maternal age (≤30 years; OR 2.3, CI: 1.6–3.4, p < 0.001), Black African ethnicity (OR 2.7, CI: 1.2–5.5, p < 0.05), and higher parity (≥2; OR 1.8, CI: 1.1–2.8, p < 0.01). Non-completion was more likely if testing indications included BMI ≥30 kg/m² (OR 1.7, CI: 1.1–2.4, p < 0.01) or family history of diabetes (OR 2.2, CI; 1.5–3.3, p < 0.001) and less likely if the indication was an ultrasound finding (OR 0.4, CI: 0.2–0.9, p < 0.05). The majority (54/78, 69%) of women who never completed testing cited multiple reasons. We identified a common over-lapping cluster of reasons, including inability to tolerate the testing protocol (20.5%), social or mental health issues (21.8%), and difficulty keeping track of multiple antenatal appointments (15.4%).

Conclusions There is a need to investigate methods of testing that are easier for high-risk groups to schedule and tolerate. Fuller explanation of test indications and additional support for vulnerable groups are also required.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

Pre-pregnancy and pregnancy care for women with type 1 and type 2 diabetes

PS Yap¹, KW Tan², K Shearer³, WA Watson^1,2

¹Department of Diabetes & Endocrinology, NHS Grampian, Aberdeen, UK, ²The School of Medicine, Medical Sciences and Nutrition, University of Aberdeen, Aberdeen, UK, ³Aberdeen Maternity Hospital, NHS Grampian, Aberdeen, UK

Aims This study aimed to assess pregnancy outcomes and impact of pre-pregnancy care in both type 1 and type 2 diabetes patients.

Methods Patients due to deliver between January 2017 and October 2018 were identified using BadgerNet database. We compared diabetes care and pregnancy outcomes.

Results Eighty-eight patients were identified. Sixty-four with type 1 diabetes and 24 type 2 diabetes of which 18.8% and 66.7% had HbA1c <53 mmol/mol respectively. Twenty-nine per cent type 1 diabetes delivered large for gestational age babies (p=0.017). Type 2 diabetes patients more likely to deliver less than 34 weeks (p=0.023) and small for gestational age babies (p=0.017).

Fifteen patients were on continuous subcutaneous insulin infusion (CSII) and 47 on multi daily injections (MDI). Eleven (73.3%) patients on CSII delivered >24 weeks and 4(26.7%) had miscarriages. In MDI group, 44(93.6%) delivered >24 weeks and 3(6.4%) had miscarriages. The difference was not significant(p=0.473). Large for gestational age babies were greater in MDI group (59.1% vs 36.4%, p=0.054).

Twenty-nine (33%) attended pre-pregnancy clinic. Type 2 diabetes patients had lower attendance (16.7%, p=0.047). Babies born <34 weeks and small for gestational age were delivered by mothers who did not attend pre-pregnancy clinic (p=0.015 and p=0.157 respectively).

Those who attended pre-pregnancy clinic were more likely to be on folic acid (p < 0.001) with less adverse outcomes (p=0.589).

Conclusion Pre-pregnancy care can improve pregnancy outcomes. Many type 2 diabetes women care was delivered in primary care pre-conception. We need to consider educational toolkit to highlight the importance of pregnancy planning to support optimisation of health and glycaemic control.

Glucose monitoring during pregnancy in type 1 diabetes: A comparison between flash and capillary glucose monitoring

N Galloway^1,2, S Ritchie², C Love¹, A Dover¹, C Alexander¹, M Strachan², N Aedla¹, N Zammitt¹

¹Royal Infirmary of Edinburgh, Edinburgh, UK, ²Western General Hospital, Edinburgh, UK

Aims Flash glucose monitoring (FGM) could improve glycaemic control in pregnancy, but the identification of previously unrecognised low glucose events could have a negative impact on average glucose. This study aimed to compare glycaemic control during pregnancy and obstetric outcomes between women with type 1 diabetes using FGM and self-monitoring of blood glucose (SMBG).

Methods This retrospective audit reviewed women with type 1 diabetes who were pregnant during 2017 in Edinburgh. This period incorporated the change-over in our pregnancy service from routine SMBG to routine FGM. Data on HbA1c during each trimester and obstetric outcomes were collected. Data are given as mean ±SD.

Results A total of 54 women with type 1 diabetes were studied. 59% (n=32) used FGM and 41% (n=22) used SMBG. There was no significant difference in HbA1c between the two groups: first trimester HbA1c was 56 ± 15 mmol/mol and 55 ± 12 mmol/mol respectively (p=0.96); second trimester HbA1c was 49 ± 10 mmol/mol and 47 ± 9 mmol/mol (p=0.52); third trimester HbA1c was 49 ± 8 mmol/mol and 48 ± 8 mmol/mol (p=0.81). Mean birth weight was 3168 ± 951 g and 3529 ± 735 g (p=0.11) respectively. Gestation at delivery was 37 weeks in both groups (p=0.62). Spontaneous vaginal deliveries occurred in 18% and 19% respectively (p=0.94).

Conclusions HbA1c levels during pregnancy and obstetric outcomes were similar between women using FGM and SMBG. Future studies should assess the impact of FGM on symptomatic low glucose and quality of life.

SUPPORTING INFORMATION The conference poster for this abstract is available online in the Supporting Information section at the end of this page.

How should we diagnose gestational diabetes in late pregnancy?

AJ Chakera^1,2

¹Department of Diabetes and Endocrinology, Brighton and Sussex University Hospitals NHS Trust, Brighton, UK, ²Department of Medicine, Brighton and Sussex Medical School, Brighton, UK

Introduction There is good evidence for the criteria to diagnose, and the benefits of treating Gestational diabetes (GDM) up to 34 weeks gestation. Polyhydramnios and glycosuria are indications to suspect GDM, if these occur after 34 weeks gestation, there is no evidence on how to best diagnose GDM.

Methods I asked antenatal services in Kent, Surrey and Sussex: do you use a different test to diagnose GDM in late pregnancy? If so, from when, with what test and using what glucose criteria?

Results 12 sites responded. In late pregnancy:

• Three sites use an OGTT (fasting ≥5.6 mmol/l; 2h ≥7.8 mmol/l) throughout pregnancy.

• Seven sites diagnose GDM only using pre-and post-meal capillary blood glucose testing (CBG). Among these sites, there are five different methods for diagnosing GDM. The methods are:

• Five use a week of testing, one uses five days and one two weeks.

• Five start CBG-testing from 34 weeks, one from 32 weeks and one 30 weeks.

• Five use fasting ≥5.3; 1h ≥7.8 (NICE treatment targets), two use fasting ≥5.6; 1h ≥7.8.

• Six require ≥three raised glucose values over the period of testing, one requires two raised values.

• One site uses a mixture of OGTT or CBG dependent on clinician.

• One site performs a single random CBG after 34 weeks; only if ≥7 mmol/l are women referred to the diabetes-antenatal team.

Conclusions The lack of consensus for diagnosing GDM in late pregnancy (eight methods in 12 sites) reflects the paucity of evidence. The next step is to create an area-wide policy, and measure indications and outcomes in women diagnosed with GDM after 34 weeks.

Acknowledgement Diabetes Antenatal Services in Kent, Surrey and Sussex

Neonatal outcomes are not affected by glycaemic variability in the 24h prior to delivery in women with type 1 diabetes

R Scott¹, J Allen², B Jones³, C Yu³, J Terry³, E Mullins³, A Dornhorst¹, S Robinson¹, R Agha-Jaffar¹

¹Department of Diabetes and Endocrinology, Imperial College Healthcare NHS Trust, London, UK, ²Department of Metabolism, Digestion and Reproduction, Imperial College, London, UK, ³Department of Obstetrics, Imperial College Healthcare NHS Trust, London, UK

Introduction Women with diabetes are encouraged to maintain tight glycaemic control during delivery, to reduce neonatal complications. This study aimed to determine if an association existed between glycaemic control in the 24-h preceding delivery and adverse neonatal outcomes.

Methods Twelve women with type 1 diabetes were retrospectively reviewed. Variable rate insulin infusions (VRII) were started before elective c-section, or once in active labour. Glycaemic control in the 24-h preceding delivery was assessed through real-time continuous glucose monitoring data and/or computerised records of capillary blood glucose values with the following definitions applied: ‘below range’ <4.0 mmol/l; ‘within range’ 4–8 mmol/l, and ‘above range’ >8.0 mmol/l.

Results Four babies (33%) had neonatal hypoglycaemia; six (50%) were admitted to the special care baby unit (SCBU); three (25%) had low initial APGARS and three (25%) had respiratory distress syndrome (RDS). Time below, within and above range did not differ in those with and without neonatal hypoglycaemia (p=0.38, 0.60 and 0.93 respectively); those admitted to SCBU (p=0.95, 0.41 and 0.43); initial low APGAR scores (p=0.87, 0.53 and 0.28); or RDS (p=0.93, 0.28 and 0.26). VRIIs were poor at maintaining blood glucose levels ‘within range’; women spent 24% of time within range, 7% of time below range and 69% of time above range.

Discussion Adverse neonatal outcomes were not associated with poor glycaemic control in the 24-h prior to delivery. Obligatory VRII may not be required in active labour, given poor efficacy at achieving control, and the lack of association with improved neonatal outcomes.

Seasonal variation and insulin treatment requirements in gestational diabetes

RA Fox¹, B Devoy², CL Meek^3,4,6, CE Aiken^2,5,6

¹School of Clinical Medicine, University of Cambridge, Cambridge, UK, ²Department of Obstetrics and Gynaecology, Rosie Hospital, Cambridge University Hospitals, Cambridge, UK, ³Institute of Metabolic Science, Addenbrooke's Hospital, Cambridge, UK, ⁴Wolfson Diabetes and Endocrinology Clinic, Cambridge University Hospitals, Addenbrooke's Hospital, Cambridge, UK, ⁵University Department of Obstetrics and Gynaecology, NIHR Cambridge Comprehensive Biomedical Research Centre, University of Cambridge, Cambridge, UK, ⁶Addenbrooke's Hospital, Cambridge University Hospitals NHS FT, Cambridge, UK

Aims Several studies demonstrate seasonal variation in incidence of gestational diabetes (GDM) with a greater number of diagnoses in summer compared to winter. We have recently shown that birth weight and caesarean section rates also vary seasonally in GDM. The aim of the current study was to identify if seasonal variation impacts upon treatment requirements in GDM pregnancies.

Methods The cohort included 791 women with GDM who delivered singleton infants at term at The Rosie hospital, Cambridge (January 15 to January 19). Data were extracted from the mother and babies’ electronic medical records. Maternal variables included treatment, obstetric history, intrapartum and demographic data. Neonatal variables included birth weight, neonatal capillary glucose and NICU admission. Data were analysed by day of diagnosis or birth for maternal and neonatal data respectively using logistic regression to adjust for maternal age, parity, ethnicity and BMI.

Results Likelihood of treatment with insulin varied seasonally (>40% above baseline November–January; >10% below baseline March–August; p < 0.05) while diet or metformin treatment did not. Total insulin requirement at 36 weeks varied by 19units/day from greatest (January) to lowest (July; p < 0.05). The risk of neonatal hypoglycaemia varied seasonally (40% above baseline December; 30% below baseline July; p < 0.05) but risk of admission to NICU did not.

Conclusion Increased risk of insulin treatment, greater daily units required and increased risk of neonatal hypoglycaemia in winter suggests seasonal variation has an impact upon the likelihood of achieving optimal glycaemic control. Findings of no seasonal variation in other maternal or neonatal outcomes suggests differential treatment adequately compensates.

Impact of early diagnosis in gestational diabetes

P Amin^1,2, R Fox^1,2, E Gurnell¹, C Aiken¹, C Meek¹

¹Wolfson Diabetes & Endocrine Clinic, Cambridge University Hospitals NHS FT, Cambridge, UK, ²School of Clinical Medicine, University of Cambridge, Cambridge, UK

Aims Gestational diabetes (GDM) is associated with adverse pregnancy outcomes. Current UK guidance recommends screening from 24 weeks gestation, but earlier intervention could improve outcomes. We aimed to assess whether early diagnosis (<24 weeks) affects pregnancy outcomes.

Methods This retrospective study included 826 women with GDM who delivered viable singleton infants (≥24 weeks) at The Rosie Hospital (October 14 to January 19). Diagnosis was based on risk-factor screening, 75 g OGTT and WHO diagnostic criteria (24–28 weeks) or earlier (<24 weeks) in women with previous GDM or symptoms of hyperglycaemia. Multivariate regression analysis was used to compare pregnancy outcomes according to gestational age at diagnosis.

Results The 87/826 (10.5%) women with early GDM were older (p < 0.05) with a similar BMI but lower gestational weight gain (p < 0.001) than women with late GDM. Women with early GDM had higher fasting OGTT glucose (p < 0.01) and HbA1c (p < 0.05) and higher rates of metformin (p < 0.01) and insulin (p < 0.001) prescription. Among women taking insulin (413/826), early GDM required higher daily doses (29.2 vs 8.9units/day; p < 0.001).

Women with early GDM gave birth earlier (38.3 weeks vs 38.6 weeks; p < 0.01), had smaller babies (average centile 51.9 vs 59.7; p < 0.01), increased rates of NICU admission (23.0% vs 14.2%; p < 0.05) but similar rates of C-sections and neonatal hypoglycaemia.

395 women had a postnatal OGTT (>6 weeks) and 327 had an HbA1c (>3 months). Women with early GDM were at highest risk of early postnatal diabetes (15.2 vs 2.5%; p=0.001).

Conclusions Women with early GDM required more intensive treatment which normalised infant birth weight but they remained at high risk of early postnatal diabetes.

Seasonal variations in incidence and neonatal outcomes of gestational diabetes: Relevant to service delivery?

LC Kusinski¹, CL Meek^1,2,3, B Devoy², D Simmons⁴, CJ Patient², HR Murphy^2,5,6, CE Aiken^1,2

¹Wellcome-Trust MRC Institute of Metabolic Science, University of Cambridge, Cambridge, UK, ²Department of Clinical Biochemistry, Cambridge University Hospitals, Cambridge, UK, ³Department of Chemistry, Peterborough City Hospital, Peterborough, UK, ⁴School of Medicine, University of Western Sydney, Sydney, Australia, ⁵Norwich Medical School, University of East Anglia, Norwich, UK, ⁶Department of Women's Health, King's College London, London, UK

Objectives Gestational diabetes (GDM) incidence reportedly demonstrates seasonal variation. We aimed to assess if season also affects maternal-fetal outcomes in GDM.

Methods 23,735 women who consecutively delivered singleton, live-born term infants in a single tertiary obstetrics centre (2004–2008) were included (1,094 GDM; 4.6%). Additive dynamic regression models (age, BMI, parity, ethnicity adjusted) compared GDM incidence and outcomes (GDM diagnosis, plasma glucose, birth-weight centile, delivery mode) over annual cycles.

Results GDM rates were low overall in our population. GDM incidence varied by over 30% from peak (October births) to nadir (March births; p < 0.05), resulting in variations in GDM incidence between 3.53 and 4.77%. A higher temperature at time of the 28-week oral glucose tolerance test (OGTT) was strongly associated with a positive diagnosis (p < 0.001). Birth weight in GDM-affected pregnancies also demonstrated seasonal variation (mean: 58th centile June – September; 67th centile December-March; p < 0.05) and was closely temporally correlated with rates of emergency Caesarean rates (50% variation) (p < 0.05).

Conclusions Substantial seasonal variation exists in GDM diagnoses and maternal-fetal outcomes in GDM-affected pregnancies (birth weight; Caesarean rates). OGTTs performed with higher ambient temperature resulted in more positive diagnoses but these women were at lower risk of suboptimal outcomes compared to women diagnosed in cool ambient temperatures. Although the absolute difference in GDM prevalence was only 1.24% in our low-risk population, this represents an increase in diagnoses of over 30% affecting October-born infants. A greater awareness of seasonal variation may allow improved service planning to optimise outcomes for all months of the year.

Optimal methods for assessment of glycaemia in type 1 diabetes pregnancy: How does HbA1c compare to continuous glucose monitoring metrics? An ancillary study of the CONCEPTT trial

CL Meek^1,2, LC Kusinski¹, D Tundidor³, JM Yamamoto⁴, D Ma⁵, EM Scott⁶, J Halperin⁵, DS Feig⁷, HR Murphy^2,8, R Corcoy⁹

¹Wellcome-Trust Institute of Metabolic Science, University of Cambridge, Cambridge, UK, ²Department of Clinical Biochemistry, Cambridge University Hospitals, Cambridge, UK, ³Institut de Recerca, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, ⁴Departments of Medicine and Obstetrics and Gynaecology, University of Calgary, Calgary, Canada, ⁵Laboratory for Translational Research, Harvard Medical School, Boston, Massachussetts, USA, ⁶Leeds Institute of Cardiovascular and Metabolic Medicine, University of Leeds, Leeds, UK, ⁷Mount Sinai Hospital, University of Toronto, Toronto, Canada, ⁸Norwich Medical School, University of East Anglia, Norwich, UK, ⁹Servei d'Endocrinologia i Nutrició, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain

Objectives Type 1 diabetes in pregnancy is associated with increased neonatal morbidity, which improves with optimal glycaemic control. However, measuring glycaemic control in pregnancy is challenging. We aimed to assess the ability of continuous glucose monitoring (CGM) metrics and HbA1c to identify pregnancies at risk of neonatal complications.

Methods 225 CONCEPTT participants had 6-day CGM and venesection at 12, 24 and 34 weeks. Pregnancy outcomes included preterm delivery (<37 weeks), large for gestational age (LGA), neonatal hypoglycaemia requiring IV dextrose (NH) and neonatal intensive care unit (NICU) admission for >24h.

Results Most women did not achieve glycaemic control targets for early pregnancy (9.1% achieving time-in-range (TIR; 3.5–7.8 mmol/l) >70%; 24.0% with HbA1c <48 mmol/mol, 12 weeks). At 12 weeks, CGM metrics were associated with preterm delivery and LGA, while HbA1c was only associated with LGA. However, both HbA1c and CGM metrics were strongly associated with neonatal complications at 24 and 34 weeks (p < 0.05). For HbA1c, each 1 mmol/mol increase at 24 weeks was significantly associated with increased rates of preterm delivery (5%), NICU admission (5%), NH (7%) and LGA (7%). Using CGM, a 5% decrease in TIR was associated with increased rates of preterm delivery (11%), NICU admission (10%), NH (14%) and LGA (17%). Time-above-range (>7.8 mmol/l) and mean CGM glucose were also significantly associated with outcomes.

Conclusions In women with type 1 diabetes, both CGM metrics and HbA1c are significantly associated with neonatal outcomes at 24 and 34 weeks. Early pregnancy CGM metrics show stronger associations with outcomes compared to HbA1c.

Variation in prevalence of gestational diabetes across North West London: An analysis of 10,307 deliveries in an inner city multi-ethnic cohort

S Gnanasambanthan¹, N Oliver², B Jones³, E Mullins³, J Terry³, C Yu³, A Dornhorst², S Robinson², R Agha-Jaffar²

¹Obstetrics and Gynaecology, Princess Royal University Hospital, Kings College NHS Trust, London, UK, ²Metabolic Medicine, Imperial College Healthcare NHS Trust, London, UK, ³Maternal Medicine, Imperial College Healthcare NHS Trust, London, UK

Refer to Oral number A17.

Gestational diabetes and type 2 diabetes in women in North West London

EWS Mullins^1,2, BA Jones², P Carnduff^3,4, R Delgardo⁵, A Dornhorst¹

¹Metabolism, Digestion and Reproduction, Imperial College, London, UK, ²Women and Children's Services, Imperial College Healthcare NHS Trust, London, UK, ³PRC Practice Consultancy Ltd, London, UK, ⁴CWHHE collaborative of CCGs, London, UK, ⁵Diabetes GP Lead, NHS Hounslow Clinical Commissioning Group, London, UK

Aims To estimate the prevalence of type 2 diabetes in women after gestational diabetes (GDM) and the quality of care post-GDM in NW London.

Methods Practice level data were obtained from SystmOne for Hounslow CCG in a snapshot of the population on 15/10/19. Searches for; women aged 18–50 (‘childbearing age’), those on diabetes register, type of diabetes, pregnant ever, GDM ever, fasting plasma glucose (FPG) or HbA1c in the last 12 months, BMI, ethnicity, were made, covering the period from 1/1/00 to 15/10/19.

Results Women of childbearing age (WOCA) comprised 83,198 of the 3,21,779 Hounslow population. 2.2% of WOCA are on the diabetes register; for 4/5 boroughs 90% of these had type 2 diabetes however in Chiswick (85% Caucasian cf. higher diversity in other boroughs) this was only 65%. 36,204 of the 83,198 WOCA had ever been pregnant, 1,573 (4.35%) of these had a diagnosis of GDM recorded. Of those with pregnancy and GDM recorded, 867 (55%) had a diagnosis of type 2 diabetes. Of those with pregnancy and GDM recorded, 766 (71%, range 33–100%) had a FPG or HbA1c recorded in the last 12 months.

Conclusions It seems likely that GDM, at 4.35% in a diverse population, is under-recorded. Cumulative progression from GDM to type 2 diabetes is estimated at 55%. There is variation between practices in uptake of subsequent diabetes screening, targeted improvement could improve quality of care. We will obtain data from all 8 CCGS in NW London and analyse a sub-group of women pregnant from 2015 when GDM recording is likely to have improved.