Publications
scientifiques

The purpose of this analysis was to assess potential predictors of intra-cranial bleeding (ICB) and gastrointestinal bleeding (GIB) in patients with symptomatic peripheral artery disease (PAD) in UK primary care. Patients with symptomatic PAD diagnosed from 2000 to 2010 were identified from The Health Improvement Network (THIN;  = 28,484). A nested case-control analysis, adjusted for potential confounders, was performed to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for potential predictors of ICB or GIB. For GIB, follow-up was restricted to Hospital Episode Statistics-linked THIN practices. Median follow-up was 6 years. For ICB (153 cases), the OR (95% CI) was 3.85 (1.33-11.13) for previous ICB, 0.90 (0.61-1.34) for treated hypertension, 1.59 (0.65-3.87) for untreated hypertension and 1.38 (0.80-2.36) for current smoking. ORs for ICB were 0.78 (0.50-1.21), 0.40 (0.09-1.82) and 1.27 (0.47-3.47) with use of acetylsalicylic acid (ASA), clopidogrel and warfarin monotherapy, respectively, compared with non-use of such therapy. For GIB (506 cases), the OR was 1.40 (1.05-1.86) for peptic ulcer disease, 3.20 (1.81-5.64) for dual anti-platelet therapy use, 1.96 (1.46-2.64) for non-steroidal anti-inflammatory drug (NSAID) use and 1.01 (0.80-1.28) for proton pump inhibitor use. ORs for GIB were 1.78 (1.39-2.30), 2.03 (1.05-3.93) and 1.25 (0.72-2.16) with ASA, clopidogrel and warfarin monotherapy, respectively, compared with non-use. Previous ICB was a risk factor for ICB. Use of anti-platelet therapy or NSAIDs increased GIB risk. Identifying bleeding predictors could help optimize treatment strategies for patients with PAD.

BACKGROUND: Undiagnosed comorbid diseases that independently lead to weight loss before type 2 diabetes mellitus (T2DM) diagnosis could explain the observed increased mortality risk in T2DM patients with normal weight. OBJECTIVES: To evaluate the impact of weight change patterns before the diagnosis of T2DM on the association between body mass index (BMI) at diagnosis and mortality risk. METHODS: This was a longitudinal cohort study using 145,058 patients from UK primary care, with newly diagnosed T2DM from January 2000. Patients aged 18-70, without established disease history at diagnosis (defined as the presence of cardiovascular diseases, cancer, and renal diseases on or before diagnosis) were followed up to 2014. Longitudinal 6-monthly measures of bodyweight three years before (used to define groups of patients who lost bodyweight or not before diagnosis) and 2 years after diagnosis were obtained. The main outcome was all-cause mortality. RESULTS: At diagnosis, mean (SD) age was 52 (12) years, 56% were male, 52% were current or ex-smokers, mean BMI was 33 kg/m , and 66% were obese. Normal weight and overweight patients experienced a small but significant reduction in body weight 6 months before diagnosis. Among all categories of obese patients, consistently increasing body weight was observed within the same time window. Among patients who did not lose body weight pre-diagnosis (n = 117,469), compared with the grade 1 obese, normal weight patients had 35% (95% CI of HR: 1.17, 1.55) significantly higher adjusted mortality risk. However, among patients experiencing weight loss before diagnosis (n = 27,589), BMI at diagnosis was not associated with mortality risk (all p > 0.05). CONCLUSIONS: Weight loss before the diagnosis of T2DM was not associated with the observed increased mortality risk in normal weight patients with T2DM. This emphasises the importance of addressing risk factors post diagnosis for excess mortality in this group.

BACKGROUND: Prior studies found patients treated with sodium-glucose co-transporter-2 inhibitors (SGLT-2i) had lower rates of death and heart failure (HF). Whether the benefits of SGLT-2i vary based upon the presence of cardiovascular disease (CVD) is unknown. OBJECTIVES: This study sought to determine the association between initiation of SGLT-2i therapy and HF or death in patients with and without CVD. METHODS: The CVD-REAL (Comparative Effectiveness of Cardiovascular Outcomes in New Users of SGLT-2 Inhibitors) study was a multinational, observational study in which adults with type 2 diabetes were identified. Patients prescribed an SGLT-2i or other glucose-lowering drugs (GLDs) were matched based on a propensity score for initiation of an SGLT-2i. Hazard ratios (HRs) for the risk of death, HF, and HF or death in patients with and without established CVD were estimated for each country and pooled. RESULTS: After propensity score matching, 153,078 patients were included in each group. At baseline, 13% had established CVD. Compared with therapy using other GLDs, initiation of an SGLT-2i was associated with lower risk of death in patients with and without CVD (HR: 0.56; 95% confidence interval [CI]: 0.44 to 0.70; and HR: 0.56; 95% CI: 0.50 to 0.63, respectively). There were also associations between SGLT-2i and lower risk of HF (HR: 0.72; 95% CI: 0.63 to 0.82; and HR: 0.61; 95% CI: 0.48 to 0.78, respectively) and the composite of HF or death (HR: 0.63; 95% CI: 0.57 to 0.70; and HR: 0.56; 95% CI: 0.50 to 0.62, respectively) observed in patients with and without established CVD. CONCLUSIONS: In this large, multinational, observational study, initiation of SGLT-2i was associated with lower risk of death and HF regardless of pre-existing CVD. Ongoing clinical trials will provide further evidence regarding the benefit of SGLT-2i in patients without established CVD. (Comparative Effectiveness of Cardiovascular Outcomes in New Users of SGLT-2 Inhibitors [CVD-REAL]; NCT02993614).

OBJECTIVE: To evaluate the relation between penicillin allergy and development of meticillin resistant (MRSA) and . DESIGN: Population based matched cohort study. SETTING: United Kingdom general practice (1995-2015). PARTICIPANTS: 301 399 adults without previous MRSA or enrolled in the Health Improvement Network database: 64 141 had a penicillin allergy and 237 258 comparators matched on age, sex, and study entry time. MAIN OUTCOME MEASURES: The primary outcome was risk of incident MRSA and . Secondary outcomes were use of β lactam antibiotics and β lactam alternative antibiotics. RESULTS: Among 64 141 adults with penicillin allergy and 237 258 matched comparators, 1365 developed MRSA (442 participants with penicillin allergy and 923 comparators) and 1688 developed (442 participants with penicillin allergy and 1246 comparators) during a mean 6.0 years of follow-up. Among patients with penicillin allergy the adjusted hazard ratio for MRSA was 1.69 (95% confidence interval 1.51 to 1.90) and for was 1.26 (1.12 to 1.40). The adjusted incidence rate ratios for antibiotic use among patients with penicillin allergy were 4.15 (95% confidence interval 4.12 to 4.17) for macrolides, 3.89 (3.66 to 4.12) for clindamycin, and 2.10 (2.08 to 2.13) for fluoroquinolones. Increased use of β lactam alternative antibiotics accounted for 55% of the increased risk of MRSA and 35% of the increased risk of . CONCLUSIONS: Documented penicillin allergy was associated with an increased risk of MRSA and that was mediated by the increased use of β lactam alternative antibiotics. Systematically addressing penicillin allergies may be an important public health strategy to reduce the incidence of MRSA and among patients with a penicillin allergy label.

BACKGROUND: The epidemiology of type 1 diabetes mellitus (T1DM) suggests diagnostic delays may contribute to children developing diabetic ketoacidosis at diagnosis. We sought to quantify opportunities for earlier diagnosis of T1DM in primary care. METHODS: A matched case-control study of children (0-16 years) presenting to UK primary care, examining routinely collected primary care consultation types and National Institute for Health and Care Excellence (NICE) warning signs in the 13 weeks before diagnosis. RESULTS: Our primary analysis included 1920 new T1DM cases and 7680 controls. In the week prior to diagnosis more cases than controls had medical record entries (663, 34.5% vs 1014, 13.6%, odds ratio 3.46, 95% CI 3.07-3.89; p<0.0001) and the incidence rate of face-to-face consultations was higher in cases (mean 0.32 vs 0.11, incidence rate ratio 2.90, 2.61-3.21; p<0.0001). The preceding week entries were found in 330 cases and 943 controls (17.2% vs 12.3%, OR 1.49, 1.3-1.7, p<0.0001), but face-to-face consultations were no different (IRR 1.08 (0.9-1.29, p=0.42)). INTERPRETATION: There may be opportunities to reduce time to diagnosis for up to one third of cases, by up to two weeks. Diagnostic opportunities might be maximised by measures that improve access to primary care, and public awareness of T1DM.

BACKGROUND: Glycaemic targets for older people have been revised in recent years because of concern that more stringent targets are associated with increased mortality. We aimed to investigate the association between glycaemic control (mean HbA ) and variability (variability of HbA over time) and mortality in older people with diabetes. METHODS: We did a 5-year retrospective cohort study using The Health Improvement Network database, which includes data from 587 UK primary care practices. We included patients of either sex who were aged 70 years and older with type 1 or type 2 diabetes. The primary outcome was time to all-cause mortality. Our primary exposure variables were mean HbA and variability of HbA over time. The observation included a 4-year run-in period (from 2003) as a baseline, with a 5-year follow-up (from 2007 to 2012). We assessed mean HbA in three models: a baseline mean HbA for 2003-06 (model 1), the mean across the whole follow-up period (model 2), and a time-varying yearly updated mean (model 3). A variability score (from 0 [low] to 100 [high]) was calculated on the basis of number of changes in HbA of 0·5% (5·5 mmol/mol) or more from 2003 to 2012 or to the point of mortality, based on changes in the annual mean as per each model with a minimum of six readings. FINDINGS: The cohort consisted of 54 803 people, of whom 17 680 (8614 [30·7%] of 28 017 women and 9066 [33·8%] of 26 786 men) died during the observation period. The overall mortality rate was 77 per 1000 person-years (73 per 1000 person-years for women and 80 per 1000 person-years for men). The data showed a J-shaped distribution for mortality risk in both sexes, with significant increases with HbA values greater than 8% (64 mmol/mol) and less than 6% (42 mmol/mol), although excess mortality risk was non-significant in model 1 for men at HbA values of 8% (64 mmol/mol) to less than 8·5% (<69 mmol/mol) and in models 1 and 3 for both sexes assessed individually at HbA values less than 6% (42 mmol/mol). Mortality increased substantially with increasing HbA variability in all models (overall and for both sexes). For the model 2 HbA measure, the adjusted hazard ratios comparing patients with a glycaemic variability score of more than 80 to 100 with those with a score of 0 to 20 were 2·47 (95% CI 2·08-2·93) for women and 2·21 (1·87-2·61) for men. Fitting the mean HbA models with the glycaemic variability score altered the risk distribution; this observation was most marked in the model 2 analysis, in which a significant increased risk was only apparent with HbA values greater than 9·5% (80 mmol/mol) in women and 9% (75 mmol/mol) in men. INTERPRETATION: Both low and high levels of glycaemic control were associated with an increased mortality risk, and the level of variability also seems to be an important factor, suggesting that a stable glycaemic level in the middle range is associated with lower risk. Glycaemic variability, as assessed by variability over time in HbA , might be an important factor in understanding mortality risk in older people with diabetes. FUNDING: King's College London and Diabetes Frail.

BACKGROUND: Approximately 80 000 children and young people in the UK suffer from depression, but many are untreated because of poor identification of early warning signs and risk factors. AIMS: This study aimed to derive and to investigate discrimination characteristics of a prediction model for a first recorded diagnosis of depression in young people aged 15-24 years. METHOD: This study used a matched case-control method using electronic primary care records. Stepwise conditional logistic regression modelling investigated 42 potential predictors including symptoms, co-morbidities, social factors and drug and alcohol misuse. RESULTS: Of the socio-economic and symptomatic predictors identified, the strongest associations were with depression symptoms and other psychological conditions. School problems and social services involvement were prominent predictors in men aged 15-18 years, work stress in women aged 19-24 years. CONCLUSION: Our model is a first step in the development of a predictive model identifying early warning signs of depression in young people in primary care.

INTRODUCTION: Signal detection remains a cornerstone activity of pharmacovigilance. Routine quantitative signal detection primarily focuses on screening of spontaneous reports. In striving to enhance quantitative signal detection capability further, other data streams are being considered for their potential contribution as sources of emerging signals, one of which is longitudinal observational databases, including electronic medical record (EMR) and transactional insurance claims databases. Quantitative signal detection on such databases is a nascent field-with published methods being primarily based either on individual metrics, which may not effectively represent the complexity of the longitudinal records and their necessary variation for analysis for drug-outcome pairs, or on visualization discovery approaches leveraging multiple aspects of the records, which are not particularly tractable to high-throughput hypothesis-free signal detection. One extensively tested example of the latter is chronographs. METHODS: We apply a disturbance detection algorithm to chronographs using UK EMR The Health Improvement Network (THIN) data. The algorithm utilizes autoregressive integrated moving average (ARIMA)-based time series methodology designed to find disturbances that occur outside the normal pattern trends of the ARIMA model for the chronograph. Chronographs currently highlight drug-event pairs as potentially worthy of further clinical assessment, via filter-based increases in disproportionality scores from before to after the index drug exposure, tested across a range of case and control windows. RESULTS: We replicate the findings on six exemplar chronographs from a previous publication, and show how disturbances can be effectively detected across this set of pairs. Further, 692 disturbances were detected in analysis of all 384 individual READ code outcomes ever recorded 50 or more times for patients prescribed sibutramine. The disturbances are algorithmically further broken into subsets of clinical interest. CONCLUSION: Overall, the disturbance algorithm approach shows promising capacity for detecting outliers, and shows tractability of the algorithmic approach for large-scale screening. The method offers an array of pattern types for detection and clinical review.

BACKGROUND: Weight loss is a non-specific cancer symptom for which there are no clinical guidelines about investigation in primary care. AIM: To summarise the available evidence on weight loss as a clinical feature of cancer in patients presenting to primary care. DESIGN AND SETTING: A diagnostic test accuracy review and meta-analysis. METHOD: Studies reporting 2 × 2 diagnostic accuracy data for weight loss (index test) in adults presenting to primary care and a subsequent diagnosis of cancer (reference standard) were included. QUADAS-2 was used to assess study quality. Sensitivity, specificity, positive likelihood ratios, and positive predictive values were calculated, and a bivariate meta-analysis performed. RESULTS: A total of 25 studies were included, with 23 (92%) using primary care records. Of these, 20 (80%) defined weight loss as a physician's coding of the symptom; the remainder collected data directly. One defined unexplained weight loss using objective measurements. Positive associations between weight loss and cancer were found for 10 cancer sites: prostate, colorectal, lung, gastro-oesophageal, pancreatic, non-Hodgkin's lymphoma, ovarian, myeloma, renal tract, and biliary tree. Sensitivity ranged from 2% to 47%, and specificity from 92% to 99%, across cancer sites. The positive predictive value for cancer in male and female patients with weight loss for all age groups ≥60 years exceeded the 3% risk threshold that current UK guidance proposes for further investigation. CONCLUSION: A primary care clinician's decision to code for weight loss is highly predictive of cancer. For such patients, urgent referral pathways are justified to investigate for cancer across multiple sites.

INTRODUCTION: The Self-Controlled Case Series (SCCS) method has been widely used for hypothesis testing, but there is limited evidence of its performance for safety signal detection. OBJECTIVE: The objective of this study was to evaluate SCCS for signal detection on recently approved products. METHODS: A retrospective study covered the period after three recently marketed drugs were launched through to 31 December 2010 using The Health Improvement Network, a UK primary care database, and Optum, a US claims database. The SCCS method was applied to examine five heterogenous outcomes with desvenlafaxine and escitalopram and six outcomes with adalimumab for Signals of Disproportional Recording (SDRs); a positive finding was determined to be when the lower bound of 95% Confidence Interval of the incidence rate ratio (IRR) estimate was >  1. Multiple design choices were tested and the trend in IRR estimates over calendar time for one drug event pair was examined. RESULTS: All six outcomes with adalimumab, three of five outcomes with desvenlafaxine, and four of five outcomes with escitalopram had SDRs. SCCS highlighted all acute events in the primary analysis but was less successful with slower-onset outcomes. Performance varied by risk period definition. Changes in IRR estimates over quarterly intervals for adalimumab with herpes zoster showed marked higher SDR within 9 months of drug launch. CONCLUSION: SCCS shows promise for signal detection: it may highlight known associations for recent marketed products and has potential for early signal identification. SCCS performance varied by design choice and the nature of both exposure and event pair. Future work is needed to determine how effective the approach is in prospective testing and determining the performance characteristics of the approach.

OBJECTIVES: To determine rates of stroke or transient ischaemic attack (TIA) and all cause mortality in patients with a diagnosis of "resolved" atrial fibrillation compared to patients with unresolved atrial fibrillation and without atrial fibrillation. DESIGN: Two retrospective cohort studies. SETTING: General practices contributing to The Health Improvement Network, 1 January 2000 to 15 May 2016. PARTICIPANTS: Adults aged 18 years or more with no previous stroke or TIA: 11 159 with resolved atrial fibrillation, 15 059 controls with atrial fibrillation, and 22 266 controls without atrial fibrillation. MAIN OUTCOME MEASURES: Primary outcome was incidence of stroke or TIA. Secondary outcome was all cause mortality. RESULTS: Adjusted incidence rate ratios for stroke or TIA in patients with resolved atrial fibrillation were 0.76 (95% confidence interval 0.67 to 0.85, P<0.001) versus controls with atrial fibrillation and 1.63 (1.46 to 1.83, P<0.001) versus controls without atrial fibrillation. Adjusted incidence rate ratios for mortality in patients with resolved atrial fibrillation were 0.60 (0.56 to 0.65, P<0.001) versus controls with atrial fibrillation and 1.13 (1.06 to 1.21, P<0.001) versus controls without atrial fibrillation. When patients with resolved atrial fibrillation and documented recurrent atrial fibrillation were excluded the adjusted incidence rate ratio for stroke or TIA was 1.45 (1.26 to 1.67, P<0.001) versus controls without atrial fibrillation. CONCLUSION: Patients with resolved atrial fibrillation remain at higher risk of stroke or TIA than patients without atrial fibrillation. The risk is increased even in those in whom recurrent atrial fibrillation is not documented. Guidelines should be updated to advocate continued use of anticoagulants in patients with resolved atrial fibrillation.

OBJECTIVES: Imaging studies in patients with cutaneous psoriasis have demonstrated asymptomatic bone and tendon changes, commonly of the foot and ankle. We sought to determine if patients with cutaneous psoriasis have an increased risk of clinically significant foot and ankle tendinopathy or enthesopathy compared with the general population. METHODS: Patients with cutaneous psoriasis and a general population cohort were identified in The Health Improvement Network, a general practice medical records database from the UK. All patients with psoriatic arthritis were excluded. Cox proportional-hazards models (α=0.05) estimated the HR for development of foot and ankle tendinopathy or enthesopathy among patients with psoriasis, with adjustment for numerous covariates. RESULTS: In total, 78 630 patients with cutaneous psoriasis and 5 983 338 persons from the general population were identified. In an unadjusted model, patients with cutaneous psoriasis had a 25% increased risk of developing foot and ankle tendinopathy or enthesopathy compared with the general population (HR 1.25, 95% CI 1.20 to 1.30, p<0.0001). The HR remained unchanged and statistically significant after adjusting for covariates, and in sensitivity analyses. CONCLUSIONS: These data suggest that patients with psoriasis can have foot and ankle tendinopathy or enthesopathy without having psoriatic arthritis, presenting a diagnostic challenge to physicians. Further research is needed to elucidate mechanisms contributing to this increased risk.

AIMS: To estimate the risk of developing long-term major cardiovascular and renal complications in relation to levels of body mass index (BMI) in a population of White European (WE), African-Caribbean (AC), and South Asian (SA) patients with type 2 diabetes mellitus (T2DM). MATERIALS AND METHODS: Patients with new diagnosis of T2DM, aged ≥ 18 years from January 2000 (n = 69,436) and their age-sex-ethnicity matched non-diabetic controls (n = 272,190) were identified from UK primary care database. Incidence rates ratios (IRRs) for non-fatal major cardiovascular events (MACE) and chronic kidney disease (CKD) in patients with T2DM compared to controls were estimated using multivariate Mantel-Cox model. RESULTS: Among normal weight patients with T2DM, WEs had significantly higher prevalence of cardiovascular multi-morbidity (95% CI 9.5, 11.3), compared to SAs (95% CI 4.8, 9.5). AC and SA overweight and obese patients had similar prevalence, while obese WEs had significantly higher prevalence. During a median 7 years of follow-up, risk of MACE was significantly higher for overweight (95% CI of IRR 1.50, 2.46) and obese (95% CI of IRR 1.49, 2.43) SAs compared to their WE counterparts. However, similar risk levels were observed for normal weight WEs and SAs, respectively. Risk of CKD was higher and uniform for BMI ≥ 25 kg/m amongst WEs and ACs, whereas only overweight patients had significantly higher risk of CKD amongst SA [IRR 2.08 (95% CI 1.49, 2.93)]. CONCLUSION: Risk of MACE/CKD varies over levels of BMI within each ethnic group, with overweight SAs having a disproportionate risk of CKD.