Clinical Practice Research Datalink Database Research Articles

Weight Management with Orlistat in Type 2 Diabetes - An Electronic Healthcare Records Study.

Orlistat is recommended as an adjunct to diet and exercise for weight loss in type 2 diabetes mellitus (T2DM). The aims were to explore (i) associations between patient characteristics and orlistat prescribing and to determine (ii) associations of orlistat with weight loss in T2DM/prediabetes. This cohort study used anonymised records of T2DM/prediabetes (2016-2017) patients ≥18 years, from the UK Clinical Practice Research Datalink (CPRD) database. Multivariable logistic regression models determined associations with starting orlistat and stopping it early (<12 weeks of prescriptions) and orlistat's associations with weight loss in those not prescribed 2nd line anti-diabetic medications. Out of 100,552 patients with incident T2DM/prediabetes, 655 (0.8%) T2DM and 128 (0.7%) prediabetes patients were prescribed orlistat. Younger people, females, those in more deprived regions, current smokers, co-prescribed metformin and recorded with hypertension were significantly more likely to be prescribed orlistat while higher baseline HbA1c levels were associated with early stopping. Those who continued orlistat for >12 weeks were more likely to lose ≥5% weight (adjusted OR: 1.69; 95% CI: 1.07, 2.67) than those not on orlistat, but those who stopped orlistat early were less likely (adjusted OR: 0.56; 95% CI: 0.29, 1.09). Orlistat was significantly associated with weight loss in T2DM/prediabetes when taken for at least 12 weeks. However, orlistat is infrequently prescribed and often taken for <12 weeks. Orlistat may be a useful adjunct to lifestyle modifications in T2DM/prediabetes however barriers to continuing orlistat means it may not be effective for everyone in managing weight loss.

Burden of Disease and Treatment Patterns Amongst Patients With Vitiligo: Findings From a National, Longitudinal Retrospective Study in the United Kingdom.

This retrospective study, using UK Clinical Practice Research Datalink and Hospital Episode Statistics databases, analysed 17,239 incident patients with vitiligo. Mean incidence of vitiligo was 0.16 (2010-2021) per 1000 person-years (range: 0.10 [2010-COVID] to 0.19 [2013/2018]); prevalence increased from 0.21% (2010) to 0.38% (2021). The most common comorbidities recorded after vitiligo diagnosis were diabetes (19.4%), eczema (8.9%), thyroid disease (7.5%), and rheumatoid arthritis (6.9%). Mental health diagnoses recorded at any time were most commonly depression and/or anxiety (24.6%), depression (18.5%), anxiety (16.0%), and sleep disturbance (12.7%); recorded after vitiligo diagnosis in 6.4%, 4.4%, 5.5%, and 3.9%, respectively. Mental health comorbidities were more common among White patients (eg, depression and/or anxiety, 29.0%) than Black (18.8%), Asian (16.1%), and other ethnicities (21.4%). In adolescents, depression and/or anxiety was most commonly diagnosed after vitiligo diagnosis (7.4% vs before, 1.8%). Healthcare resources were used most frequently in the first year after vitiligo diagnosis (incident cohort), typically dermatology-related outpatient appointments (101.9/100 person-years) and general practitioner consultations (97.9/100 person-years). In the year after diagnosis, 60.8% of incident patients did not receive vitiligo-related treatments (ie, topical corticosteroids, topical calcineurin inhibitors, oral corticosteroids, phototherapy), increasing to 82.0% the next year; median (95% CI) time from diagnosis to first treatment was 34.0 (31.6-36.4) months. Antidepressants and/or anxiolytics were recorded for 16.7% of incident patients in the year after diagnosis. In 2019, 85.0% of prevalent patients did not receive vitiligo-related treatments; 16.6% had a record of antidepressant and/or anxiolytic treatments. Most patients were not on vitiligo-related treatments within a year of diagnosis, with time to first treatment >2 years, suggesting that vitiligo may be dismissed as unimportant and not treated early, in part due to limited effectiveness of available treatments. New effective treatments, early initiation, and psychological intervention and support are needed to reduce vitiligo burden on patients.

Prevalence of comorbidities with the potential to increase the risk of nonadherence to topical ocular hypotensive medication in patients with open-angle glaucoma

Objective To evaluate the prevalence of comorbidities that may limit or prevent adherence to topical ocular hypotensive therapy in patients with open-angle glaucoma (OAG). Methods The UK Clinical Practice Research Datalink (CPRD) database of primary and secondary care and prescription records was analyzed to identify patients with a first (index) diagnosis of OAG during 2016–2020. The primary care records of these patients were screened for diagnostic terms linked to prespecified (qualifying) comorbidities considered to have the potential to impact patients’ ability to instill eye drops. The prevalence of each of 10 categories of qualifying comorbidity recorded within the period from 5 years before to 2 years after the index OAG diagnosis was analyzed. Results A total of 100,968 patients with OAG were included in the analysis. Among the patients in the OAG cohort, 13,962 (13.8%) were aged 40–54 years, 32,145 (31.8%) were aged 55–69 years, 42,042 (41.6%) were aged 70–84 years, and 12,819 (12.7%) were aged 85+ years. Within the OAG population, 82.7%, 14.6%, and 2.7% of patients had no category, one category, and two or more categories of qualifying comorbidity, respectively. Qualifying comorbidities were most common in older patients. The most prevalent qualifying comorbidities were categorized as degenerative, traumatic, or pathological central nervous system disorder disrupting cognitive function (5.2%), movement disorder (4.4%), and low vision (4.1%). The prevalence of arthropathies and injuries affecting upper limbs (including arthritis in the hands) was 2.4%. Conclusions The presence of comorbidities should be considered when determining whether eye drops are suitable treatment for glaucoma. Neurodegenerative disease affecting cognition and memory, motor disease, and low vision are common comorbidities that may impact adherence to eye drops, and affected patients may benefit from non-drop treatment modalities.

Using an expanded algorithm to estimate prevalence of amyotrophic lateral sclerosis in U.S. and UK.

There is an increasing need to better understand the burden of amyotrophic lateral sclerosis (ALS) using real-world data (RWD). However, identifying ALS cases using RWD presents several challenges due to the rarity of ALS and the differences in database coding systems. MarketScan claims, and the UK Clinical Practice Research Datalink (CPRD) databases were searched for diagnosis codes of ALS or MND, the only drugs approved for treating ALS (riluzole and edaravone) and clinical visits with 12-month enrolment prior to 1 January 2011. The main algorithm required ≥ 1 ALS diagnosis code together with prescriptions or clinical visits. We expanded the existing algorithm to identify unspecific (possible) ALS group that had codes for motor neuron disease (MND) and the ALS drugs. The study period was from 1 January 2011 until 31 December 2020. We identified 16,246 patients with ≥ 1 ALS code in Marketscan (denominator n = 85,279,619), yet only 184 were found in the UK CPRD (denominator n = 21,318,589). Using the main algorithm 9,433 ALS patients were included in MarketScan, with a prevalence ranged between 4.5 per 100,000 in 2019 and 6.2 in 2015. In MarketScan, 3,658 (4.3 per 100,000) had ≥ 1 MND code and the ALS drug codes (possible cases). In CPRD, 47.9% of 2,785 patients with ≥ 1 MND code had a prescription for riluzole (6.3 per 100,000), regarded as possible ALS cases. The expanded algorithm enabled the identification of a large population with ALS, or possible ALS, and the estimation of ALS prevalence in MarketScan and CPRD.

Clostridioides difficile infections, recurrences, and clinical outcomes in real-world settings from 2015 to 2019: The RECUR England study

ObjectiveTo estimate the epidemiological and clinical burden of Clostridioides difficile infections (CDIs) and recurrences (rCDIs) in England. MethodsThis retrospective study included adult patients diagnosed with CDI (community or hospital settings) over 2015-2019 from Clinical Practice Research Datalink and Hospital Episode Statistics databases. Incidences of CDI and rCDI were determined annually. Time to subsequent rCDI was estimated by Kaplan-Meier method. Rates of complications were assessed within 12 months from index episode. Association of risk factors with complications was evaluated using a Cox regression model. ResultsA total of 52,443 CDI episodes were recorded among 36,913 patients. Of these, 75% were aged ≥65 years, 59% were women; 73% were treated in community settings. CDI incidence remained stable (111 episodes per 100,000 patients in 2019). Around 21% of patients had ≥1 rCDI. Sepsis (12%) was the most common complication, followed by colectomy and ulcerative colitis. Age, gender, comorbidities, rCDI, preindex medical procedures, hospitalizations and consultations, and CDI treatment in hospital, were found to increase the risk of complication. ConclusionsCDI remains a concern in England. The study highlights the importance of managing primary and rCDI episodes via effective and improved therapies to prevent fatal complications.

Resource utilisation and outcomes of people with heart failure in England: a descriptive analysis of linked primary and secondary care data – the PULSE study

BackgroundHeart failure (HF) is associated with high levels of resource use and mortality, but prior UK studies have not compared outcomes by HF subtype (HF with reduced ejection fraction (HFrEF)...

Obesity, metabolic health and clinical outcomes after incident cardiovascular disease: A nationwide population-based cohort study.

The association between metabolic syndrome and increased cardiovascular disease (CVD) risk is well-established. However, in patients with incident CVD, the relationship between obesity, metabolic health, and subsequent CVD and mortality outcomes are less well-established. This study investigated the association between body mass index (BMI), metabolic health and the risk of subsequent cardiovascular mortality and morbidity outcomes in patients with incident CVD events. This cohort study identified 130685 patients from the nationwide Clinical Practice Research Datalink (CPRD GOLD) and Hospital Episode Statistics (HES) databases in the United Kingdom. Patients were ≥18years with incident CVD [coronary heart disease (CHD), stroke, or peripheral vascular disease (PVD)] between 1 January 1998 and 31 December 2017. BMI (in kg/m2 ) was categorized as underweight (<18.5), normal (18.5-24.9), overweight (25.0-29.9) and obese (≥30). Within each BMI category, patients were grouped by increasing count of 1, 2 or 3 metabolic risk factors [RF] (dyslipidaemia, diabetes mellitus and hypertension) and were regarded as metabolically unhealthy while absence of these factors was considered metabolically healthy (MH). Multivariable Cox regression was used to assess the risk (hazard ratio with 95% confidence interval) of subsequent outcomes (non-fatal CHD, stroke, PVD, incident heart failure, CVD-mortality and all-cause mortality) in BMI subgroups with incremental count of metabolic RFs. During a median follow-up of 13.0years, a higher BMI was associated with reduced risk for stroke, PVD, CVD-mortality and all-cause mortality within each metabolic risk category, while increasing metabolic RFs within each BMI subgroup accounted for increasing risks. When compared with patients with normal BMI and no RF, CVD-mortality risk in overweight patients with no RF was 0.76 (0.70-0.84), and in obese patients with no RF was 0.85 (0.76-0.96). The respective risk for all-cause mortality in patients with overweight and no RF was 0.69 (0.65-0.72), and in obese patients with no RF was 0.75 (0.70-0.79). Subsequent outcomes of stroke and PVD showed similar trends. In contrast, the risk of subsequent non-fatal CHD events and incident HF increased with higher BMI and with incremental metabolic risk factors within each BMI category. Underweight was constantly associated with increased risk for all outcomes regardless of the presence of metabolic RFs except for non-fatal CHD events. In patients with incident CVD, overweight and obesity were related to a more favourable prognosis for subsequent stroke, PVD and mortality (CVD-related and all-cause) irrespective of the presence of other metabolic risk factors.

Uncontrolled blood pressure and therapeutic inertia in treated hypertensive patients: A retrospective cohort study using a UK general practice database.

Uncontrolled blood pressure (BP) and therapeutic inertia pose significant challenges in effectively managing hypertension. This study objective was to quantify levels of uncontrolled BP and therapeutic inertia among patients treated for hypertension in primary care. This retrospective cohort study used data recorded by general practitioners from the UK Clinical Practice Research Datalink database. Adults with primary hypertension who received a recorded prescription for any antihypertensive drug between January 2015 and June 2017 (index date) were included, with a follow-up of 18 months. Primary outcomes included the percentage of patients with uncontrolled BP (defined as systolic BP ≥140mmHg or diastolic BP ≥90mmHg) and of apparent therapeutic inertia (defined as two consecutive uncontrolled BP records without treatment change) during follow-up. Finally, of 581260 patients receiving antihypertensive drug(s), 37.2% (n=216014) had uncontrolled BP at the index date and 30.3% (n=175955) had no record of BP at this date. During follow-up, 59.2% had ≥1 record of uncontrolled BP, in 22% all records showed uncontrolled BP, and 12.8% had no record of BP. Among those with uncontrolled BP at the index date, 72.9% had ≥1 record of uncontrolled BP during follow-up, and in 28.3% all records showed uncontrolled BP. Therapeutic inertia was observed in 33.1% of patients overall, and in 55.7% of those with uncontrolled BP at the index date. In conclusion, BP recording was infrequent, possibly reflecting both a low frequency of measurement and potential under-recording. Uncontrolled BP and therapeutic inertia appear to be widespread in UK general practice.

Duration of Type 2 Diabetes and Incidence of Cancer: An Observational Study in England.

To investigate the association between duration of type 2 diabetes and cancer incidence. In the Clinical Practice Research Datalink database, we identified 130,764 individuals with type 2 diabetes aged ≥35 years at diagnosis who were linked to hospital and mortality records. We used sex-stratified Royston-Parmar models with two timescales to estimate incidence rates of all cancers, the four commonest cancers in the U.K. (colorectal, lung, prostate, breast), and the obesity-related cancers (e.g., liver, ovary) between 1 January 1998 and 14 January 2019, by age and diabetes duration. During 1,089,923 person-years, 18,977 incident cancers occurred. At the same age, rates of all cancers in men and women did not vary across durations ranging from diagnosis to 20 years; conversely, for any duration, there was a strong, positive association between age and cancer rates. In men, the rate ratio (95% CI) comparing 20 with 5 years of duration was 1.18 (0.82-1.69) at 60 years of age and 0.90 (0.75-1.08) at 80 years; corresponding ratios in women were 1.07 (0.71-1.63) and 0.84 (0.66-1.05). This pattern was observed also for the four commonest cancers. For obesity-related cancers, although rates were generally higher in individuals with a higher BMI, there was no association with duration at any level of BMI. In this study, we did not find evidence of an association between duration of type 2 diabetes and risk of cancer, with the higher risk observed for longer durations related to ageing.

P17 Baseline characteristics, disease progression rates and risk factors among female patients with uncomplicated urinary tract infection in England

Abstract Background and Objectives Uncomplicated urinary tract infection (uUTI) is a common, usually community-acquired bacterial infection. Little is known about risk factors for disease progression (defined here as hospitalization for Escherichia coli sepsis, bacteraemia or acute pyelonephritis) among patients with uUTI. We report the baseline demographic and clinical characteristics among female uUTI patients with/without disease progression in England. Methods This retrospective cohort study utilized anonymized patient data from 1 January 2018 to 31 December 2019 from the Clinical Practice Research Datalink (CPRD) database linked to English Hospital Episode Statistics data. Eligible patients were female, aged ≥12 years, had received an index diagnosis for a community-acquired uUTI, had ≥12 months’ CPRD data history, and had received ≥1 oral antibiotic within ±5 days of diagnosis. Patients were excluded if they had complicated UTI or complicating comorbidities, were hospitalized/visited an accident and emergency department 28 days prior to index or had received IV antibiotics as initial therapy. uUTI episodes were defined as the 28 days post-index (Figure 1). Descriptive patient (demographic and clinical characteristics), region and practice-level characteristics in the 12 month baseline period are presented for patients with/without disease progression. Figure 1.Study design. HES, Hospital Episode Statistics. Results Among 116793 female patients with a uUTI, 207 were hospitalized for E. coli sepsis, bacteraemia or acute pyelonephritis within 28 days of index (0.2%). The mean age (SD) for patients with/without progression was 67.3 (20.2) and 52.9 (21.1), respectively; 81% and 34% of patients with/without progression were ≥50 years old. Imbalances were observed for comorbid burden, baseline medications, initial presentation (i.e. home visit requirement) and prior antimicrobial treatment (Table 1). Few regional or practice-level differences were apparent (Table 1). Table 1.Patient (demographic and clinical characteristics), region and practice-level variable distribution between uUTI patients without and with hospitalization for uUTI disease progressionVariable description/Statistic or categoryPatient subgroupsno disease progression (N=116586)disease progression (N=207)N%N%Patient-level Age at diagnosis, years, mean (SD)52.9 (21.1)–67.3 (20.2)– Age at diagnosis ≥12 to &amp;lt;167110.600 ≥16 to &amp;lt;301986117157.2 ≥30 to &amp;lt;401668214136.3 ≥40 to &amp;lt;501518413115.3 ≥50 to &amp;lt;6024512214220 ≥65396363412661 Age at diagnosis (pooled) &amp;lt;5052438453919 ≥50396363416881 Charlson comorbidity index, mean (IQR)0 (0, 0)–0 (0, 1)– Charlson comorbidity category 0959448213465 118007155828 223002125.8 ≥33350.3&amp;lt;5 Race/ethnicity, N (%) Asian or Asian British51614.462.9 Black or Black British32872.862.9 Chinese or other group19951.7&amp;lt;5- Ethnic group not recorded22781204321 White41340357335 Number of hospital admissions in prior year, mean (IQR)0 (0, 0)–0 (0, 0)– Number of A&amp;E attendances in prior year, mean (IQR)0 (0, 1)–0 (0, 1)– Smoking status Non-smoker725466211957 Ex-smoker27084236230 Smoker16956152613 Menopausal status, yes34072296029 Obesity, yes13858123316 Mild/moderate renal impairment, yes22621.9104.8 Controlled diabetes (HgA1C &amp;lt;6.5/7%), yes24512.1146.8 History of recurrent uUTI, yes49374.2115.3 Prior antimicrobial treatment (any), yes702746015072 Cumulative antimicrobial prescriptions, median (IQR)1 (0, 2)–&amp;lt;5–Region-level Quintiles of index of multiple deprivation Q1 (least deprived)26869234220 Q223597203919 Q322273193818 Q422085194622 Q5 (most deprived)21762194220 Rural-urban area classification Rural15416133014 Urban1011708717786Practice-level Practice size Small (&amp;lt;8000 patients)30783265627 Medium (8000–11000 patients)25377224421 Large (&amp;gt;11000 patients)604265210752Percentage values do not add up to 100% for ethnicity as these data were not available for all patients and mixed ethnicity is not shown due to limited interpretability.A&amp;E, accident and emergency; HgA1C, glycated haemoglobin; Q, quintile. Conclusions Hospitalization following uUTI is rare in this study, however, at a population level, this still affects a large number of female patients. Coding limitations meant potentially linked admissions (e.g. falls and confusion) could not be identified. Hospitalizations appear to be driven by patient factors rather than regional differences or practice patterns. The role of these putative risk factors for disease progression will be explored further via multivariable logistic regression analysis.

P19 Burden of illness and costs among female patients with uncomplicated urinary tract infection in England

Abstract Background and Objectives Uncomplicated urinary tract infection (uUTI) is usually community-acquired. Data regarding the wider burden of illness from uUTI are scarce. We evaluated the burden of uUTI in female patients in England. Methods This retrospective comparative cohort study utilized linked anonymized patient data from 1 January 2017 to 29 February 2020 from the Clinical Practice Research Datalink (CPRD) and English Hospital Episode Statistics databases. All eligible patients were female, aged ≥12 years and had ≥12 months’ CPRD data available pre- and post-index (i.e. ≥24 months total). Patients with an index UTI diagnosis who had received ≥1 oral antibiotic within ±5 days of index formed the uUTI cohort; uUTI episodes were defined as 28 days from index diagnosis. Patients who had visited a primary care setting but did not have a UTI (uUTI or complicated UTI [cUTI]) diagnosis during the study period formed the control cohort and were matched to the index date of uUTI patients. Patients who were hospitalized or visited an accident and emergency department 28 days prior to index, had cUTI and/or had received IV antibiotics as initial therapy were excluded. Burden of illness (index/12 month follow-up all-cause and UTI-related healthcare resource use [HCRU] and costs) was evaluated in a 1:1 age- and Charlson Comorbidity Index (CCI) score-matched cohort. Results Overall, 120519 female patients with uUTI and 233670 female control patients (without uUTI) were identified and matched 1:1. Before matching, more of the uUTI cohort patients versus controls were &amp;lt;50 years old, had a higher mean CCI score, and were more often menopausal; region- and practice-level differences were modest (Table 1). HCRU and costs were significantly higher among uUTI patients versus matched controls, notably for primary care, specialist, and accident and emergency visits, and hospital admissions (P&amp;lt;0.001, Table 2). After adjusting for patient characteristics, Poisson or negative binomial models showed significant HCRU differences in primary and secondary care for uUTI patients versus matched controls (Table 3). Conclusions Patients with uUTI in England had significantly higher burden of illness, consuming more resources, during index uUTI episodes and follow-up, versus matched controls.

P18 Treatment patterns and extent of inappropriate antimicrobial treatment among female patients with uncomplicated urinary tract infection in England

P18 Treatment patterns and extent of inappropriate antimicrobial treatment among female patients with uncomplicated urinary tract infection in England

P69 Baseline characteristics of patients with physician-diagnosed hidradenitis suppurativa: results from a retrospective observational study using data from the Clinical Practice Research Datalink database

Abstract Hidradenitis suppurativa (HS) is a difficult-to-treat, painful, chronic inflammatory skin disorder characterized by nodules, abscesses and skin tunnels. Delay in diagnosis increases disease severity and overall disease burden. In England, HS is associated with a high healthcare resource use, with an estimated prevalence of 1.2% in the general UK population (Ingram JR, Jenkins-Jones S, Knipe DW et al. Population-based Clinical Practice Research Datalink study using algorithm modelling to identify the true burden of hidradenitis suppurativa. Br J Dermatol 2018; 178:917–24). Improved understanding of HS patient profiles in England will help optimize patient care and treatment outcomes. This study aimed to characterize the patient profiles, journey, treatment pathways and treatment outcomes of patients with HS in England using data from the Clinical Practice Research Datalink database. Here, we present the baseline characteristics of patients with HS from this database. Overall, 39 860 patients with physician-diagnosed HS were identified from 1 January 2009 to 31 March 2021. The mean (SD) age at diagnosis was 34.49 (13.07) years, and the majority were female (73.8%). Twenty-seven per cent of patients were White, 7.5% were Asian, 5.0% were Black and 27.9% were of mixed race, while 2.1% were from other races and the race was unknown for 30.7%. Overall, 44.6% were current smokers and 12.7% were ex-smokers. Depression (31.7%), anxiety (21.5%) and acne (20.6%) were the most commonly reported comorbidities related to HS. Chronic pulmonary disease was reported in 26.2% of patients; this may, in part, be attributed to relatively high rates of smoking in the HS population. The most frequently prescribed medications prior to HS diagnosis were pain medication, especially ‘non-opioid’ analgesics (21.8%) and opioids (19.8%); oral antibiotics, particularly tetracyclines (15.7%); combined contraceptive pills (7.2%) and metformin (4.8%). Only a few patients (&amp;lt; 0.05%) received a biologic prior to their HS diagnosis for other chronic inflammatory conditions. The various comorbidities, including psychiatric, metabolic, cardiovascular, gynaecological and pulmonary and associated treatments, observed in this real-world study in England suggest that optimal management should involve a multidisciplinary approach to improve outcomes in patients with HS.

Advancing heart failure research using machine learning

Advancing heart failure research using machine learning

Comparison of characteristics of patients with lung cancer in U.K. primary care databases: Clinical Practice Research Datalink Aurum and GOLD.

In recent years, the number of general practices contributing to the Clinical Practice Research Datalink (CPRD) database GOLD is decreasing. Therefore, for research questions addressing for instance novel treatments requiring up-to-date data, sample size will become an important consideration in study feasibility. In recent years, CPRD Aurum, containing information of practices that use EMIS software, has become an additional data source that is being used for CPRD studies. In order to establish whether Aurum is suited to act as data source for future studies in the field of lung cancer research, we aimed to compare characteristics between patients with lung cancer in Aurum and GOLD. A retrospective study was performed comparing characteristics and overall survival (OS) of patients with lung cancer in Aurum and GOLD. To further evaluate similarity, hypothetical eligibility of these patients in Aurum and GOLD was compared for 11 randomized clinical trials (RCTs). Baseline characteristics registered in Aurum and GOLD were largely similar, with some clinically irrelevant differences for previous malignancies, deviant laboratory values and drug use. Median OS was 9.8 and 9.0 months for patients in Aurum and GOLD, respectively. Potential RCT eligibility varied between 49.4% and 79.5% and 49.1% and 78.1% for patients in Aurum and GOLD, respectively. Mortality rates and the comparison of the obtained HRs per hypothetical eligibility cohort per RCT were similar in Aurum and GOLD. This study showed that data of patients with lung cancer in Aurum and GOLD are largely comparable, suggesting that Aurum is suitable for future epidemiological lung cancer research.

Preconception Management of Hyperthyroidism and Thyroid Status in Subsequent Pregnancy: A Population-Based Cohort Study.

Optimal thyroid status in pregnancy is essential in reducing the risk of adverse outcomes. The management of hyperthyroidism in women of reproductive age poses unique challenges and it is unclear how preconception treatment strategies impact on thyroid status in subsequent pregnancy. We aimed to determine trends in the management of hyperthyroidism before and during pregnancy and to assess the impact of different preconception treatment strategies on maternal thyroid status. We utilized the Clinical Practice Research Datalink database to evaluate all females aged 15-45 years with a clinical diagnosis of hyperthyroidism and a subsequent pregnancy (January 2000 to December 2017). We compared thyroid status in pregnancy according to preconception treatment, namely, (1) antithyroid drugs up to or beyond pregnancy onset, (2) definitive treatment with thyroidectomy or radioiodine before pregnancy, and (3) no treatment at pregnancy onset. Our study cohort comprised 4712 pregnancies. Thyrotropin (TSH) was measured in only 53.1% of pregnancies, of which 28.1% showed suboptimal thyroid status (TSH >4.0 mU/L or TSH <0.1 mU/L plus FT4 >reference range). Pregnancies with prior definitive treatment were more likely to have suboptimal thyroid status compared with pregnancies starting during antithyroid drug treatment (odds ratio 4.72, 95% CI 3.50-6.36). A steady decline in the use of definitive treatment before pregnancy was observed from 2000 to 2017. One-third (32.6%) of first trimester carbimazole-exposed pregnancies were switched to propylthiouracil while 6.0% of propylthiouracil-exposed pregnancies switched to carbimazole. The management of women with hyperthyroidism who become pregnant is suboptimal, particularly in those with preconception definitive treatment, and needs urgent improvement. Better thyroid monitoring and prenatal counseling are needed to optimize thyroid status, reduce teratogenic drug exposure, and ultimately reduce the risk of adverse pregnancy outcomes.

Prevalence and treatment patterns of adult atopic dermatitis in the UK Clinical Practice Research Datalink.

The prevalence of active atopic dermatitis (AD) in adults in the UK according to disease severity shows variability. This study evaluated disease prevalence and treatment patterns among the adult UK population with AD. Data were obtained from the Clinical Practice Research Datalink (CPRD) database. Adults with active AD were identified by an AD-related prescription or general practitioner visit within the same calendar year. Prevalence was defined as the number of patients with active AD on 1 January of each year as a percentage of the number of adults in the CPRD population on that date. Moderate-to-severe disease was classified as either referral to a specialist or prescription(s) for topical calcineurin inhibitors, phototherapy, or systemic treatment. Patient characteristics and treatment and referral patterns were analysed for patients with active AD in 2019. The overall prevalence of AD was stable at 2.4% per year during the period 2015-2019. In 2019, mean patient age (± standard deviation) was 52.6±21.0years, 58.2% of patients were female and mean disease duration was 9.4±5.9years. The most prescribed treatment was topical corticosteroids, in 78.5% of patients. 36.7% of patients with moderate-to-severe AD were prescribed systemic agents and 59.8% (vs. 32.3% of patients with mild AD) were referred to any secondary care or specialist treatment. The prevalence of active AD in the adult UK population was stable over the 5-year period (2015-2019) and was comparable to estimates from similar studies based on UK primary healthcare records.

Life expectancy following a cardiovascular event in individuals with and without type 2 diabetes: A UK multi-ethnic population-based observational study

Background and aimsWe aimed to evaluate the life expectancy following the first cardiovascular disease (CVD) event by type 2 diabetes (T2D) status and ethnicity. Methods and resultsWe used the Clinical Practice Research Datalink database in England (UK), linked to the Hospital Episode Statistics information, to identify individuals with and without T2D who survived a first CVD event between 1st Jan 2007 and 31st Dec 2017; subsequent death events were extracted from the Office for National Statistics database. Ethnicity was categorised as White, South Asian (SA), Black, or other. Flexible parametric survival models were used to estimate survival and predict life expectancy. 59,939 individuals with first CVD event were included: 7596 (12.7%) with T2D (60.9% men; mean age at event: 69.7 years [63.2 years in SA, 65.9 in Black, 70.2 in White]) and 52,343 without T2D (56.7% men; 65.9 years [54.7 in Black, 58.2 in SA, 66.3 in White]). Accounting for potential confounders (sex, deprivation, lipid-lowering medication, current smoking, and pre-existing hypertension), comparing individuals with vs without T2D the mortality rate was 53% higher in White (hazard ratio [HR]: 1.53 [95% CI: 1.44, 1.62]), corresponding to a potential loss of 3.87 (3.30, 4.44) life years at the age of 50 years in individuals with T2D. No evidence of a difference in life expectancy was observed in individuals of SA (HR: 0.82 [0.52, 1.29]; −1.36 [-4.58, 1.86] life years), Black (HR: 1.26 [0.59, 2.70]; 1.21 [-2.99, 5.41] life years); and other (HR: 1.64 [0.80, 3.39]; 3.89 [-2.28, 9.99] life years) ethnic group. ConclusionFollowing a CVD event, T2D is associated with a different prognosis and life years lost among ethnic groups.

Burden of Pertussis in Adults Aged 50 Years and Older: A Retrospective Database Study in England.

Pertussis, a highly infectious respiratory disease caused by Bordetella pertussis, affects people of all ages. Older adults are particularly susceptible to its severe outcomes and complications. In this retrospective cohort study, the incidence rate of pertussis among individuals aged ≥ 50years was assessed during 2009-2018 using Clinical Practice Research Datalink and Hospital Episode Statistics databases, UnitedKingdom. Health care resource utilisation (HCRU) and direct medical costs (DMCs) were compared between patients with a pertussis diagnosis and propensity score-matched controls (matched on demographic and clinical variables). Among 5,222,860 individuals, 1638 had a pertussis diagnosis (incidence rate: 5.8 per 100,000 person-years; 95% confidence interval 5.5-6.0). Baseline (- 18 to - 6months) HCRU and DMC were similar among 1480 pertussis patients and 1480 matched controls. However, there were increases in HCRU in the pertussis vs. matched cohort around the pertussis diagnosis (from months - 6 to - 1 to 5-11). The most notable increases (pertussis vs. controls) were in the rates of general practitioner (GP)/nurse visits (4.7-fold), clinical assessments (4.1-fold), and accident and emergency visits (3.0-fold) during the month before diagnosis and GP/nurse visits during the 2months after diagnosis (2.5-fold) (all p < 0.001). DMCs were significantly higher in the pertussis cohort (p < 0.001). Total excess DMC in the pertussis cohort during months - 1 to + 11 was £318 per patient. A pertussis diagnosis among adults aged ≥ 50years resulted in significant increases in HCRU and DMC across several months around diagnosis. These results highlight the need for increased awareness of pertussis infection among adults aged ≥ 50years and suggest that pertussis booster doses among this population should be considered.

A66 PROTON PUMP INHIBITORS AND THE RISK OF INFLAMMATORY BOWEL DISEASE: A REAL WORLD POPULATION-BASED COHORT STUDY

Abstract Background Proton pump inhibitors (PPIs) and histamine-2 receptor antagonists (H2RAs) are acid suppressant drugs indicated in highly prevalent conditions such as gastroesophageal reflux disease and peptic ulcer disease. A recent observational study reported an increased risk of inflammatory bowel disease (IBD) with the use of PPIs. However, this association may have been driven by certain methodological shortcomings, such as failure to properly account for protopathic bias, a conclusion-altering bias. Purpose To determine whether the use of PPIs compared with the use of H2RAs is associated with an increased risk of IBD. Method Using the longitudinal primary care records from the United Kingdom Clinical Practice Research Datalink database, we identified 1,498,416 initiators of PPIs and 322,474 initiators of H2RAs from January 1, 1990, through December 31, 2018, with follow-up until December 31, 2019. Standardized morbidity ratio weighted Cox proportional hazards models were used to estimate marginal hazard ratios (HRs) and 95% confidence intervals (CIs). Patients were analyzed according to the timing of the IBD events after treatment initiation (early versus late). In the early-event analysis, IBD events were assessed within the first two years of treatment initiation, an analysis subject to potential protopathic bias. In the late-event analysis, all exposures were lagged by two years to account for latency and minimize protopathic bias. Result(s) In the early-event analysis, the use of PPIs was associated with an increased risk of IBD within the first two years of treatment initiation, compared with H2RAs (HR: 1.39, 95% CI: 1.14-1.69). In contrast, the use of PPIs was not associated with an increased risk of IBD in the late-event analysis (HR: 1.05, 95% CI: 0.90 to 1.22). The results remained consistent in several sensitivity analyses. Image Conclusion(s) Compared with H2RAs, PPIs were not associated with an increased risk of IBD, after accounting for protopathic bias. Please acknowledge all funding agencies by checking the applicable boxes below CIHR Disclosure of Interest R. Yanofsky: None Declared, D. Abrahami Employee of: pfizer, R. Pradhan: None Declared, H. Yin: None Declared, E. McDonald: None Declared, A. Bitton Consultant of: Member of Advisory Boards for Abbie, Pfizer, Takeda, Jansen, and Merck, Speakers bureau of: Has received speaker fees for Abbie, Jansen, Takeda, and Pfizer, L. Azoulay Consultant of: Has received consults fees from Jansen, Pfizer, and Roche, Speakers bureau of: Has received speaker fees from Jansen, Pfizer, and Roche