Longitudinal audit of assessment and pharmaceutical intervention for cardiovascular risk in the Australasian Diabetes Data Network

Cardiovascular disease is the primary cause of increased morbidity and reduced longevity in type 1 diabetes (TID).1, 2 Children who develop T1D before the age of 10 years are particularly at risk of cardiovascular disease in adult life.3 Monitoring of cardiovascular risk factors during childhood and adolescence is recognized as an essential part of care. The International Society of Paediatric and Adolescent Diabetes (ISPAD),4 American Diabetes Association (ADA),5 and National Institute for Health and Care Excellence (NICE)6 provide recent guidelines for screening and treatment of elevated blood pressure (BP), dyslipidaemia, and albuminuria. Achievement of targets early in the course of T1D may provide early cardiorenal protection, as suggested in a 2-year follow-up of children with T1D who were screened according to international guidelines.7 The achievement of target metabolic control is also crucial to improve cardiovascular risk and specific risk factors such as dyslipidaemia. A recent cross-sectional review identifies the ongoing problems with undertreatment of cardiovascular risk factors and therapeutic inertia in young people with T1D in Europe and the United States.8 Factors contributing to undertreatment may include adherence issues,9 lack of provider familiarity with prescribing antihypertensive and lipid-lowering agents in adolescents,10 the perspective of patients and their families,11 and delays in pharmaceutical treatment after lifestyle intervention has proven ineffective.10 We recently reported the determinants of cardiovascular risk in young people aged 2 to 25 years with T1D who were enrolled in the Australasian Diabetes Data Network (ADDN).12 This audit aims to extend this work, to measure first the frequency of assessments of BP, lipid profile, and urinary albumin-creatinine ratio (ACR) in young people aged 2 to 25 years with T1D who are enrolled in the ADDN, and second the number of participants who received pharmaceutical treatment for raised BP, abnormal lipid profile, or raised ACR according to current ISPAD4 and ADA guidelines.5 The ADDN model involves the transfer of deidentified, prospectively collected participant data from the databases or electronic medical record systems of participating ADDN centres to a web-based server hosted by the University of Melbourne.13, 14 Participating centres collect data using a common data dictionary. Data are transferred every 6 months to the registry.14 Participant data were entered prospectively at each site since January 2012. Participants attended T1D multidisciplinary clinics in 10 (eight in Australia; two in New Zealand) paediatric and five adult public teaching hospitals providing metropolitan and regional clinics. T1D was diagnosed according to ADA criteria and the date of initiation of insulin as the date of onset. Inclusion criteria were participants with T1D, aged 2 to 25 years at the beginning of ADDN follow-up. Exclusion criteria were other forms of diabetes. As guidelines differ as to their recommendations for those aged 10 years or younger or those aged 11 years or older, we analysed assessments in two groups: participants who were either younger than 11 years or those aged 11 years or older at their last visit. In recent Type 1 diabetes exchange clinic network (T1DX) and Prospective Diabetes Follow-up (DPV) cohorts, dyslipidaemia was defined if at least one lipid value was raised, total cholesterol was more than 200 mg/dL, high-density lipoprotein (HDL) was less than 35 mg/dL, low-density lipoprotein (LDL) was more than 130 mg/dL, or triglycerides was more than 150 mg/dL. Thresholds for the definition of raised outcomes were according to recently published ISPAD4 and ADA5 thresholds. Thresholds for raised systolic or diastolic BP were more than the 95th centile for age, gender, and height on at least two occasions. Thresholds for abnormal lipid profile were LDL cholesterol of more than 3.4 mmol/L, and/or total cholesterol of more than 5.2 mmol/L, and/or HDL cholesterol of less than 0.9 mmol/L on at least one occasion. Thresholds for raised ACR were more than 2.5 mg/mmol in males and more than 3.5 mg/mmol in females on at least two occasions. Height was measured using a Harpenden stadiometer, weight using a floor scale, and BP was measured sitting at rest using a sphygmomanometer with the appropriate cuff size. Insulin delivery system (continuous subcutaneous insulin infusion, multiple daily injection, or twice-daily insulin injections) was recorded at each visit, as described.13 Body mass index (BMI) standard deviation (SD) scores were calculated using the Centre for Disease Control and Prevention 2000 reference scale.15 Urine was collected as a spot early morning sample. Lipids were measured in the non-fasting state using commercial enzymatic assays on Roche Hitachi Cobas C systems. Urinary albumin was measured by immunoassay predominantly immunoturbidimetric, and urinary creatinine by an enzyme colorimetric method (Roche Cobas C501; Hitachi). HbA1c was measured using point of care or laboratory testing methods, commonly Vantage analyser (Siemens Diagnostics, Camberley, UK) or Variant analyser (Bio-Rad Laboratories, Hercules, CA). All laboratory methods participated in the ongoing Royal Australasian College of Pathologists Quality Assurance Programs. Ethics approval was obtained through the Human Research Ethics Committee for each of the participating centres. Informed written consent was obtained from parents and adults aged older than 17.9 years.13 Deidentified data from young adults transitioning to adult care were collected as a waiver. Pearson's chi-squared test was used to compare proportions between the two groups. For continuous variables, we used the independent student t-test to compare means between groups and are presented as mean (SD). In the event of departure from normality, we used the non-parametric equivalent of the Wilcoxon rank-sum test and are presented as median (interquartile range [IQR]). Data analysis was undertaken in Stata V16 (Stata Corp, College Station, TX) and the level of significance was set at 5%. The ADDN registry included 11 562 individuals with T1D aged 2 to 25 years at the beginning of their follow-up from diagnosis. All 11 562 were included in the study. Their characteristics are presented in Table 1. Frequency of assessment and treatment of cardiovascular risk factors are presented in Table 1 and Figure 1. BP measurements: of 2144 individuals, 889 (41.5%) had no recorded BP measurement during follow-up, 281 (13.1%) had one measurement, and the remainder had two or more measurements. Lipids measurements: of 2144 individuals, 1755 (81.9%) had no recorded lipids measurement, 281 (13.1%) had one measurement, and the remainder had two or more measurements. ACR measurements: of 2144 individuals, 2035 (94.9%) had no recorded ACR measurement, 90 (4.2%) had one measurement, and the remainder had two or more measurements. Frequency of assessment increased with longer duration of T1D and older age. BP measurements: of 9418 individuals, 2954 (31.4%) had no recorded BP measurement during follow-up, 853 (9.1%) had one measurement of BP, and the remainder had two or more measurements. Lipids measurements: of 9418 individuals, 4690 (49.8%) had no recorded lipids measurement, 1813 (19.3%) had one measurement, and the remainder had two or more measurements. ACR measurements: of 9418 individuals, 5761 (61.2%) had no recorded ACR measurement, 1378 (14.6%) had one measurement, and the remainder had two or more measurements. Only 2/2144 participants aged younger than 11 years were treated for any cardiovascular risk factor; in both cases for an abnormal lipid profile with statins. In participants aged 11 years or older at their last visit there was no difference in gender, HbA1c, or remoteness (metropolitan, regional, or remote) as determined by postcode at diagnosis between those that were and were not treated for an abnormal cardiovascular risk factor. Duration of T1D was longer in those treated for any risk factor (P = .02). BMI z-score was higher in those treated for an abnormal lipid profile (P < .001). BP measurements were more frequent in those treated for any risk factor and in those treated for raised ACR (Table 1). In participants aged 11 years or older at their last visit who had any raised cardiovascular risk factor during follow-up, only 66/2441 (2.7%) received pharmaceutical treatment. Treatment of raised BP: 978/6464 (15.1%) participants who were aged 11 years or older at their last visit and who had BP assessments during follow-up met the criteria for raised BP (Table 1). Of these, 34/978 (3.5%) were prescribed one or more antihypertensive medications during follow-up, namely, angiotensin-converting enzyme (ACE) inhibitors (n = 27), angiotensin receptor blockers (n = 6), calcium channel blockers (n = 1), beta blockers (n = 4), alpha blockers (n = 3), and diuretics (n = 1). Treatment of an abnormal lipid profile: 1588/4728 (33.6%) participants who were aged 11 years or older at their last visit and who had lipid assessments during follow-up met the criteria for an abnormal lipid profile (Table 1). Of these, 14/1588 (0.9%) were prescribed statins. Treatment of raised ACR: 248/3657 (6.8%) participants who were aged 11 years or older at their last visit and who had ACR assessments during follow-up met the criteria for raised ACR. Of these, 18/248 (7.3%) participants were prescribed ACE inhibitors (n = 12) or angiotensin receptor blockers (n = 6). We report low rates of assessment and particularly low rates of pharmaceutical treatment for abnormal cardiovascular risk factors, namely, raised BP, abnormal lipid profiles, and raised ACR, in this longitudinal Australasian cohort. Assessment rates were lower in children aged younger than 11 years, as recommended. Approximately half had BP measurements recorded, and the majority had neither lipids nor ACR measured; neither of which are recommended to be regularly assessed before 10 to 11 years of age, other than baseline lipids at diagnosis5 or in those with a family history of dyslipidaemia.4 While the majority of adolescents who were 11 years or older at their last visit had BP measurements recorded, the frequency of any assessment for lipids and ACR remained at approximately 40%-50%. Overall, just under 3.0% of participants aged 11 years or older at their last visit, who had a raised cardiovascular risk factor during follow-up, had received pharmaceutical treatment. Rates of treatment were particularly low for the treatment of an abnormal lipid profile. As expected, duration of T1D was longer in those who received treatment for any risk factor, and BMI was higher in those treated for an abnormal lipid profile. Otherwise there were no clear differences in gender, HbA1c, or socioeconomic demographic of remoteness between participants who did and did not have assessments or receive treatment. These findings are noteworthy: the centres were in teaching hospital settings, and clearly the endocrinologists were circumspect about prescribing life-long statins, in particular in young people. Rates of pharmaceutical treatment in participants who were 11 years or older at their last visit were comparable with those of a recent cross-sectional audit of the T1DX and DPV registries across the United States and Europe,8 in which 5% or fewer adolescents and young adults aged 12 to 26 years with raised BP or an abnormal lipid profile received antihypertensives or statins. Further, these rates of treatment of hypertension or dyslipidaemia were not higher than those reported in a large cross-sectional DPV cohort over 15 years ago.16 The frequency of detection of microalbuminuria in young people with T1D in T1DX and SEARCH has been reported at 4.4% and 9.2%, respectively, comparable with our findings, but with higher treatment rates.17 Failure to detect raised ACR has implications for the development of both microvascular and macrovascular complications.18 The strengths of the audit are the prospectively collected longitudinal data for a comparatively long average duration of 8 years in those aged older than 11 years at their last visit, in comparison with other audits in this age group. There are also several limitations. It is difficult to separate failure to enter data in the ADDN database from a lack of assessment, or assessment by an external pathology provider. An audit of participant case notes across the 15 sites to resolve this was beyond the scope of this project. While BP is measured and entered at the time of the clinic visit, pathology results are uploaded regularly prior to 6-monthly ADDN data reviews. In addition, we could not audit the uptake of lifestyle intervention as recommended as a first-line measure for an abnormal lipid profile before beginning medication, nor the use of ambulatory BP monitoring in the case of raised clinic BP, as is also recommended.5, 6 Finally, while representative of the ethnicity, urban, and regional demographic, as well as healthcare systems in Australia and New Zealand, ADDN currently represents approximately 60% of all young people in Australia and New Zealand with T1D. In conclusion, our findings highlight the need to prioritize discussion around the uptake of screening and pharmaceutical intervention guidelines for the prevention of premature cardiovascular disease in young people with T1D, in addition to interventions to improve metabolic control. Further investigation will explore solutions to increase the frequency of assessments and to guide appropriate treatment for cardiovascular risk,19 both from the physicians' and the patients' perspectives. This research was conducted as part of the Australasian Diabetes Data Network. We are grateful to Juvenile Diabetes Research Foundation Australia, and to the children and young people with diabetes and their families who provided the data. This research was supported by JDRF Australia, the recipient of the Australian Research Council Special Research Initiative in Type 1 Juvenile Diabetes. There is no conflict of interest or disclosure for any author. JJC and AE are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors have given final approval of the version to be published. Claire A. Robertson: conception and design, analysis and interpretation of data, drafting the manuscript; Arul Earnest: conception and design, analysis and interpretation of data, drafting the manuscript; Melissa Chee: analysis and interpretation of data, drafting the manuscript; Maria E Craig: acquisition of data, revising manuscript critically for important intellectual content; Peter Colman: acquisition of data, revising manuscript critically for important intellectual content; Helen L Barrett: acquisition of data, revising manuscript critically for important intellectual content; Philip Bergman: acquisition of data, revising manuscript critically for important intellectual content; Fergus Cameron: acquisition of data, revising manuscript critically for important intellectual content; Elizabeth E Davis: acquisition of data, revising manuscript critically for important intellectual content; Kim C Donaghue: acquisition of data, revising manuscript critically for important intellectual content; P Gerry Fegan: acquisition of data, revising manuscript critically for important intellectual content; P Shane Hamblin: acquisition of data, revising manuscript critically for important intellectual content; D Jane Holmes–Walker : acquisition of data, revising manuscript critically for important intellectual content; Craig Jefferies: acquisition of data, revising manuscript critically for important intellectual content; Stephanie Johnson: acquisition of data, revising manuscript critically for important intellectual content; Meng Tuck Mok: acquisition of data, revising manuscript critically for important intellectual content; Bruce R King: acquisition of data, revising manuscript critically for important intellectual content; Richard Sinnott: repository of data, analysis; Glenn Ward: acquisition of data, revising manuscript critically for important intellectual content; Benjamin J Wheeler: acquisition of data, revising manuscript critically for important intellectual content; Anthony Zimmermann: acquisition of data, revising manuscript critically for important intellectual content; Timothy W Jones: conception and design, analysis and interpretation of data, drafting the manuscript, acquisition of data; Jenny J Couper: conception and design, analysis and interpretation of data, drafting the manuscript, acquisition of data The peer review history for this article is available at https://publons.com/publon/10.1111/dom.14584. The data that support the findings of this study are available from Australasian diabetes data network (ADDN) . Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the authors with the permission of ADDN.