Statistical models for the deterioration of kidney function in a primary care population: A retrospective database analysis

Background: Evidence for kidney function monitoring intervals in primary care is weak, and based mainly on expert opinion. In the absence of trials of monitoring strategies, an approach combining a model for the natural history of kidney function over time combined with a cost-effectiveness analysis offers the most feasible approach for comparing the effects of monitoring under a variety of policies. This study aimed to create a model for kidney disease progression using routinely collected measures of kidney function.

Methods: This is an open cohort study of patients aged ≥18 years, registered at 643 UK general practices contributing to the Clinical Practice Research Datalink between 1 April 2005 and 31 March 2014. At study entry, no patients were kidney transplant donors or recipients, pregnant or on dialysis. Hidden Markov models for estimated glomerular filtration rate (eGFR) stage progression were fitted to four patient cohorts defined by baseline albuminuria stage; adjusted for sex, history of heart failure, cancer, hypertension and diabetes, annually updated for age.

Results: Of 1,973,068 patients, 1,921,949 had no recorded urine albumin at baseline, 37,947 had normoalbuminuria (<3mg/mmol), 10,248 had microalbuminuria (3–30mg/mmol), and 2,924 had macroalbuminuria (>30mg/mmol). Estimated annual transition probabilities were 0.75–1.3%, 1.5–2.5%, 3.4–5.4% and 3.1–11.9% for each cohort, respectively. Misclassification of eGFR stage was estimated to occur in 12.1% (95%CI: 11.9–12.2%) to 14.7% (95%CI: 14.1–15.3%) of tests. Male gender, cancer, heart failure and age were independently associated with declining renal function, whereas the impact of raised blood pressure and glucose on renal function was entirely predicted by albuminuria.

Conclusions: True kidney function deteriorates slowly over time, declining more sharply with elevated urine albumin, increasing age, heart failure, cancer and male gender. Consecutive eGFR measurements should be interpreted with caution as observed improvement or deterioration may be due to misclassification.

Associations between statins and adverse events in primary prevention of cardiovascular disease: systematic review with pairwise, network, and dose-response meta-analyses

OBJECTIVE

To assess the associations between statins and adverse events in primary prevention of cardiovascular disease and to examine how the associations vary by type and dosage of statins.

DESIGN

Systematic review and meta-analysis.

DATA SOURCES

Studies were identified from previous systematic reviews and searched in Medline, Embase, and the Cochrane Central Register of Controlled Trials, up to August 2020.

REVIEW METHODS

Randomised controlled trials in adults without a history of cardiovascular disease that compared statins with non-statin controls or compared different types or dosages of statins were included.

MAIN OUTCOME MEASURES

Primary outcomes were common adverse events: self-reported muscle symptoms, clinically confirmed muscle disorders, liver dysfunction, renal insufficiency, diabetes, and eye conditions. Secondary outcomes included myocardial infarction, stroke, and death from cardiovascular disease as measures of efficacy.

DATA SYNTHESIS

A pairwise meta-analysis was conducted to calculate odds ratios and 95% confidence intervals for each outcome between statins and non-statin controls, and the absolute risk difference in the number of events per 10 000 patients treated for a year was estimated. A network meta-analysis was performed to compare the adverse effects of different types of statins. An E_max model based meta-analysis was used to examine the dose-response relationships of the adverse effects of each statin.

RESULTS

62 trials were included, with 120 456 participants followed up for an average of 3.9 years. Statins were associated with an increased risk of self-reported muscle symptoms (21 trials, odds ratio 1.06 (95% confidence interval 1.01 to 1.13); absolute risk difference 15 (95% confidence interval 1 to 29)), liver dysfunction (21 trials, odds ratio 1.33 (1.12 to 1.58); absolute risk difference 8 (3 to 14)), renal insufficiency (eight trials, odds ratio 1.14 (1.01 to 1.28); absolute risk difference 12 (1 to 24)), and eye conditions (six trials, odds ratio 1.23 (1.04 to 1.47); absolute risk difference 14 (2 to 29)) but were not associated with clinically confirmed muscle disorders or diabetes. The increased risks did not outweigh the reduction in the risk of major cardiovascular events. Atorvastatin, lovastatin, and rosuvastatin were individually associated with some adverse events, but few significant differences were found between types of statins. An E_max dose-response relationship was identified for the effect of atorvastatin on liver dysfunction, but the dose-response relationships for the other statins and adverse effects were inconclusive.

CONCLUSIONS

For primary prevention of cardiovascular disease, the risk of adverse events attributable to statins was low and did not outweigh their efficacy in preventing cardiovascular disease, suggesting that the benefit-to-harm balance of statins is generally favourable. Evidence to support tailoring the type or dosage of statins to account for safety concerns before starting treatment was limited.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42020169955.

Impact of changes to national guidelines on hypertension-related workload: an interrupted time series analysis in English primary care

BACKGROUND

In 2011, National Institute for Health and Care Excellence (NICE) guidelines recommended the routine use of out-of-office blood pressure (BP) monitoring for the diagnosis of hypertension. These changes were predicted to reduce unnecessary treatment costs and workload associated with misdiagnosis.

AIM

To assess the impact of guideline change on rates of hypertension-related consultation in general practice.

DESIGN AND SETTING

A retrospective open cohort study in adults registered with English general practices contributing to the Clinical Practice Research Datalink between 1 April 2006 and 31 March 2017.

METHOD

The primary outcome was the rate of face-to-face, telephone, and home visit consultations related to hypertension with a GP or nurse. Age- and sex-standardised rates were analysed using interrupted time-series analysis.

RESULTS

In 3 937 191 adults (median follow-up 4.2 years) there were 12 253 836 hypertension-related consultations. The rate of hypertension-related consultation was 71.0 per 100 person-years (95% confidence interval [CI] = 67.8 to 74.2) in April 2006, which remained flat before 2011. The introduction of the NICE hypertension guideline in 2011 was associated with a change in yearly trend (change in trend -3.60 per 100 person-years, 95% CI = -5.12 to -2.09). The rate of consultation subsequently decreased to 59.2 per 100 person-years (95% CI = 56.5 to 61.8) in March 2017. These changes occurred around the time of diagnosis, and persisted when accounting for wider trends in all consultations.

CONCLUSION

Hypertension-related workload has declined in the last decade, in association with guideline changes. This is due to changes in workload at the time of diagnosis, rather than reductions in misdiagnosis.

Long-term health outcomes of people with reduced kidney function in the UK: A modelling study using population health data

BACKGROUND

People with reduced kidney function have increased cardiovascular disease (CVD) risk. We present a policy model that simulates individuals' long-term health outcomes and costs to inform strategies to reduce risks of kidney and CVDs in this population.

METHODS AND FINDINGS

We used a United Kingdom primary healthcare database, the Clinical Practice Research Datalink (CPRD), linked with secondary healthcare and mortality data, to derive an open 2005-2013 cohort of adults (≥18 years of age) with reduced kidney function (≥2 measures of estimated glomerular filtration rate [eGFR] <90 mL/min/1.73 m2 ≥90 days apart). Data on individuals' sociodemographic and clinical characteristics at entry and outcomes (first occurrences of stroke, myocardial infarction (MI), and hospitalisation for heart failure; annual kidney disease stages; and cardiovascular and nonvascular deaths) during follow-up were extracted. The cohort was used to estimate risk equations for outcomes and develop a chronic kidney disease-cardiovascular disease (CKD-CVD) health outcomes model, a Markov state transition model simulating individuals' long-term outcomes, healthcare costs, and quality of life based on their characteristics at entry. Model-simulated cumulative risks of outcomes were compared with respective observed risks using a split-sample approach. To illustrate model value, we assess the benefits of partial (i.e., at 2013 levels) and optimal (i.e., fully compliant with clinical guidelines in 2019) use of cardioprotective medications. The cohort included 1.1 million individuals with reduced kidney function (median follow-up 4.9 years, 45% men, 19% with CVD, and 74% with only mildly decreased eGFR of 60-89 mL/min/1.73 m2 at entry). Age, kidney function status, and CVD events were the key determinants of subsequent morbidity and mortality. The model-simulated cumulative disease risks corresponded well to observed risks in participant categories by eGFR level. Without the use of cardioprotective medications, for 60- to 69-year-old individuals with mildly decreased eGFR (60-89 mL/min/1.73 m2), the model projected a further 22.1 (95% confidence interval [CI] 21.8-22.3) years of life if without previous CVD and 18.6 (18.2-18.9) years if with CVD. Cardioprotective medication use at 2013 levels (29%-44% of indicated individuals without CVD; 64%-76% of those with CVD) was projected to increase their life expectancy by 0.19 (0.14-0.23) and 0.90 (0.50-1.21) years, respectively. At optimal cardioprotective medication use, the projected health gains in these individuals increased by further 0.33 (0.25-0.40) and 0.37 (0.20-0.50) years, respectively. Limitations include risk factor measurements from the UK routine primary care database and limited albuminuria measurements.

CONCLUSIONS

The CKD-CVD policy model is a novel resource for projecting long-term health outcomes and assessing treatment strategies in people with reduced kidney function. The model indicates clear survival benefits with cardioprotective treatments in this population and scope for further benefits if use of these treatments is optimised.

Impact of Changes to National Hypertension Guidelines on Hypertension Management and Outcomes in the United Kingdom

In recent years, national and international guidelines have recommended the use of out-of-office blood pressure monitoring for diagnosing hypertension. Despite evidence of cost-effectiveness, critics expressed concerns this would increase cardiovascular morbidity. We assessed the impact of these changes on the incidence of hypertension, out-of-office monitoring and cardiovascular morbidity using routine clinical data from English general practices, linked to inpatient hospital, mortality, and socio-economic status data. We studied 3 937 191 adults with median follow-up of 4.2 years (49% men, mean age=39.7 years) between April 1, 2006 and March 31, 2017. Interrupted time series analysis was used to examine the impact of changes to English hypertension guidelines in 2011 on incidence of hypertension (primary outcome). Secondary outcomes included rate of out-of-office monitoring and cardiovascular events. Across the study period, incidence of hypertension fell from 2.1 to 1.4 per 100 person-years. The change in guidance in 2011 was not associated with an immediate change in incidence (change in rate=0.01 [95% CI, -0.18-0.20]) but did result in a leveling out of the downward trend (change in yearly trend =0.09 [95% CI, 0.04-0.15]). Ambulatory monitoring increased significantly in 2011/2012 (change in rate =0.52 [95% CI, 0.43-0.60]). The rate of cardiovascular events remained unchanged (change in rate =-0.02 [95% CI, -0.05-0.02]). In summary, changes to hypertension guidelines in 2011 were associated with a stabilisation in incidence and no increase in cardiovascular events. Guidelines should continue to recommend out-of-office monitoring for diagnosis of hypertension.

Validation of clinical prediction models: what does the "calibration slope" really measure?

BACKGROUND AND OBJECTIVES

Definitions of calibration, an aspect of model validation, have evolved over time. We examine use and interpretation of the statistic currently referred to as the calibration slope.

METHODS

The history of the term "calibration slope", and usage in papers published in 2016 and 2017, were reviewed. The behaviour of the slope in illustrative hypothetical examples and in two examples in the clinical literature was demonstrated.

RESULTS

The paper in which the statistic was proposed described it as a measure of "spread" and did not use the term "calibration". In illustrative examples, slope of 1 can be associated with good or bad calibration, and this holds true across different definitions of calibration. In data extracted from a previous study, the slope was correlated with discrimination, not overall calibration. Many authors of recent papers interpret the slope as a measure of calibration; a minority interpret it as a measure of discrimination or do not explicitly categorise it as either. Seventeen of thirty-three papers used the slope as the sole measure of calibration.

CONCLUSION

Misunderstanding about this statistic has led to many papers in which it is the sole measure of calibration, which should be discouraged.

Variability in body temperature in healthy adults and in patients receiving chemotherapy: prospective observational cohort study

Between-individual variability of body temperature has been little investigated, but is of clinical importance: for example, in detection of neutropenic sepsis during chemotherapy. We studied within-person and between-person variability in temperature in healthy adults and those receiving chemotherapy using a prospective observational design involving 29 healthy participants and 23 patients undergoing chemotherapy. Primary outcome was oral temperature. We calculated each patient's mean temperature, standard deviation within each patient (within-person variability), and between patients (between-person variability). Secondary analysis explored temperature changes in the three days before admission for neutropenic sepsis. 1,755 temperature readings were returned by healthy participants and 1,765 by chemotherapy patients. Mean participant temperature was 36.16 C (95% CI 36.07-36.26) in healthy participants and 36.32 C (95% CI 36.18-36.46) in chemotherapy patients. Healthy participant within-person variability was 0.40 C (95% CI 0.36-0.44) and between-person variability was 0.26 C (95% CI 0.16-0.35). Chemotherapy patient within-person variability was 0.39 C (95% CI 0.34-0.44) and between-person variability was 0.34 C (95% CI 0.26-0.48). Thus, use of a population mean rather than personalised baselines is probably sufficient for most clinical purposes as between-person variability is not large compared to within-person variability. Standardised guidance and provision of thermometers to patients might help to improve recording and guide management.

Statistical models for the deterioration of kidney function in a primary care population: A retrospective database analysis

Background: Evidence for kidney function monitoring intervals in primary care is weak, and based mainly on expert opinion. In the absence of trials of monitoring strategies, an approach combining a model for the natural history of kidney function over time combined with a cost-effectiveness analysis offers the most feasible approach for comparing the effects of monitoring under a variety of policies. This study aimed to create a model for kidney disease progression using routinely collected measures of kidney function. Methods: This is an open cohort study of patients aged ≥18 years, registered at 643 UK general practices contributing to the Clinical Practice Research Datalink between 1 April 2005 and 31 March 2014. At study entry, no patients were kidney transplant donors or recipients, pregnant or on dialysis. Hidden Markov models for estimated glomerular filtration rate (eGFR) stage progression were fitted to four patient cohorts defined by baseline albuminuria stage; adjusted for sex, history of heart failure, cancer, hypertension and diabetes, annually updated for age. Results: Of 1,973,068 patients, 1,921,949 had no recorded urine albumin at baseline, 37,947 had normoalbuminuria (<3mg/mmol), 10,248 had microalbuminuria (3–30mg/mmol), and 2,924 had macroalbuminuria (>30mg/mmol). Estimated annual transition probabilities were 0.75–1.3%, 1.5–2.5%, 3.4–5.4% and 3.1–11.9% for each cohort, respectively. Misclassification of eGFR stage was estimated to occur in 12.1% (95%CI: 11.9–12.2%) to 14.7% (95%CI: 14.1–15.3%) of tests. Male gender, cancer, heart failure and age were independently associated with declining renal function, whereas the impact of raised blood pressure and glucose on renal function was entirely predicted by albuminuria. Conclusions: True kidney function deteriorates slowly over time, declining more sharply with elevated urine albumin, increasing age, heart failure, cancer and male gender. Consecutive eGFR measurements should be interpreted with caution as observed improvement or deterioration may be due to misclassification.

The Impact of Using Emotionally Arousing Stimuli on Muscle Response Testing Accuracy

INTRODUCTION

Muscle response testing (MRT) is an assessment method used by 1 million practitioners worldwide, yet its usefulness remains uncertain. The aim of this study, one in a series assessing the accuracy of MRT, was to determine whether emotionally arousing stimuli influence its accuracy compared to neutral stimuli.

METHODS

To assess diagnostic test accuracy 20 MRT practitioners were paired with 20 test patients (TPs). Forty MRTs were performed as TPs made true and false statements about emotionally arousing and neutral pictures. Blocks of MRT alternated with blocks of intuitive guessing (IG).

RESULTS

MRT accuracy using emotionally arousing stimuli was different than when using neutral stimuli. However, MRT accuracy was found to be significantly better than IG and chance. Similar to previous studies in this series, this study failed to detect any characteristic that consistently influenced MRT accuracy.

CONCLUSION

Using emotionally arousing stimuli had no effect on MRT accuracy compared to using neutral stimuli. This study would have been strengthened by adding personally relevant lies instead of impersonal stimuli. A limitation of this study is its lack of generalizability to other applications of MRT. This study shows that a simple yet robust methodology for assessing MRT as a diagnostic tool can be implemented effectively.

Quality of diabetes care in breast, colorectal, and prostate cancer

Purpose

Overlooking other medical conditions during cancer treatment and follow-up could result in excess morbidity and mortality, thereby undermining gains associated with early detection and improved treatment of cancer. We compared the quality of care for diabetes patients subsequently diagnosed with breast, colorectal, or prostate cancer to matched, diabetic non-cancer controls.

Methods

Longitudinal cohort study using primary care records from the Clinical Practice Research Datalink, United Kingdom. Patients with pre-existing diabetes were followed for up to 5 years after cancer diagnosis, or after an assigned index date (non-cancer controls). Quality of diabetes care was estimated based on Quality and Outcomes Framework indicators. Mixed effects logistic regression analyses were used to compare the unadjusted and adjusted odds of meeting quality measures between cancer patients and controls, overall and stratified by type of cancer.

Results

3382 cancer patients and 11,135 controls contributed 44,507 person-years of follow-up. In adjusted analyses, cancer patients were less likely to meet five of 14 quality measures, including: total cholesterol ≤ 5 mmol/L (odds ratio [OR] = 0.82; 95% confidence interval [CI], 0.75–0.90); glycosylated hemoglobin ≤ 59 mmol/mol (adjusted OR = 0.77; 95% CI, 0.70–0.85); and albumin creatinine ratio testing (adjusted OR = 0.83; 95% CI, 0.75–0.91). However, cancer patients were as likely as their matched controls to meet quality measures for other diabetes services, including retinal screening, foot examination, and dietary review.

Conclusions

Although in the short-term, cancer patients were less likely to achieve target thresholds for cholesterol and HbA1c, they continued to receive high-quality diabetes primary care throughout 5 years post diagnosis.

Implications for Cancer Survivors

These findings are important for cancer survivors with pre-existing diabetes because they indicate that high-quality diabetes care is maintained throughout the continuum of cancer diagnosis, treatment, and follow-up.

Electronic supplementary material

The online version of this article (10.1007/s11764-018-0717-5) contains supplementary material, which is available to authorized users.

Is "chronic kidney disease" a disease?

Several philosophers of medicine have attempted to answer the question "what is disease?" In current clinical practice, an umbrella term "chronic kidney disease" (CKD) encompasses a wide range of kidney health states from commonly prevalent subclinical, asymptomatic disease to rare end-stage renal disease requiring transplant or dialysis to support life. Differences in severity are currently expressed using a "stage" system, whereby stage 1 is the least severe, and stage 5 the most. Early stage CKD in older patients is normal, of little concern, and does not require treatment. However, studies have shown that many patients find being informed of their CKD distressing, even in its early stages. Using existing analyses of disease in the philosophy literature, we argue that the most prevalent diagnoses of CKD are not, in fact, diseases. We conclude that, in many diagnosed cases of CKD, diagnosing a patient with a "disease" is not only redundant, but unhelpful.

Association of guideline and policy changes with incidence of lifestyle advice and treatment for uncomplicated mild hypertension in primary care: a longitudinal cohort study in the Clinical Practice Research Datalink

OBJECTIVES

Evidence to support initiation of pharmacological treatment in patients with uncomplicated (low risk) mild hypertension is inconclusive. As such, clinical guidelines are contradictory and healthcare policy has changed regularly. The aim of this study was to determine the incidence of lifestyle advice and drug therapy in this population and whether secular trends were associated with policy changes.

DESIGN

Longitudinal cohort study.

SETTING

Primary care practices contributing to the Clinical Practice Research Datalink in England.

PARTICIPANTS

Data were extracted from the linked electronic health records of patients aged 18-74 years, with stage 1 hypertension (blood pressure between 140/90 and 159/99 mm Hg), no cardiovascular disease (CVD) risk factors and no treatment, from 1998 to 2015. Patients exited if follow-up records became unavailable, they progressed to stage 2 hypertension, developed a CVD risk factor or received lifestyle advice/treatment.

PRIMARY OUTCOME MEASURES

The association between policy changes and incidence of lifestyle advice or treatment, examined using an interrupted time-series analysis.

RESULTS

A total of 108 843 patients were defined as having uncomplicated mild hypertension (mean age 51.9±12.9 years, 60.0% female). Patientsspent a median 2.6 years (IQR 0.9-5.5) in the study, after which 12.2% (95% CI 12.0% to 12.4%) were given lifestyle advice, 29.9% (95% CI 29.7% to 30.2%) were prescribed medication and 19.4% (95% CI 19.2% to 19.6%) were given both. The introduction of the quality outcomes framework (QOF) and subsequent changes to QOF indicators were followed by significant increases in the incidence of lifestyle advice. Treatment prescriptions decreased slightly over time, but were not associated with policy changes.

CONCLUSIONS

Despite secular trends that accord with UK guidance, many patients are still prescribed treatment for mild hypertension. Adequately powered studies are needed to determine if this is appropriate.

Standard and competing risk analysis of the effect of albuminuria on cardiovascular and cancer mortality in patients with type 2 diabetes mellitus

BACKGROUND

Competing risks occur when populations may experience outcomes that either preclude or alter the probability of experiencing the main study outcome(s). Many standard survival analysis methods do not account for competing risks. We used mortality risk in people with diabetes with and without albuminuria as a case study to investigate the impact of competing risks on measures of absolute and relative risk.

METHODS

A population with type 2 diabetes was identified in Clinical Practice Research Datalink as part of a historical cohort study. Patients were followed for up to 9 years. To quantify differences in absolute risk estimates of cardiovascular and cancer, mortality standard (Kaplan-Meier) estimates were compared to competing-risks-adjusted (cumulative incidence competing risk) estimates. To quantify differences in measures of association, regression coefficients for the effect of albuminuria on the relative hazard of each outcome were compared between standard cause-specific hazard (CSH) models (Cox proportional hazards regression) and two competing risk models: the unstratified Lunn-McNeil model, which estimates CSH, and the Fine-Gray model, which estimates subdistribution hazard (SDH).

RESULTS

In patients with normoalbuminuria, standard and competing-risks-adjusted estimates for cardiovascular mortality were 11.1% (95% confidence interval (CI) 10.8-11.5%) and 10.2% (95% CI 9.9-10.5%), respectively. For cancer mortality, these figures were 8.0% (95% CI 7.7-8.3%) and 7.2% (95% CI 6.9-7.5%). In patients with albuminuria, standard and competing-risks-adjusted estimates for cardiovascular mortality were 21.8% (95% CI 20.9-22.7%) and 18.5% (95% CI 17.8-19.3%), respectively. For cancer mortality, these figures were 10.7% (95% CI 10.0-11.5%) and 8.6% (8.1-9.2%). For the effect of albuminuria on cardiovascular mortality, regression coefficient values from multivariable standard CSH, competing risks CSH, and competing risks SDH models were 0.557 (95% CI 0.491-0.623), 0.561 (95% CI 0.494-0.628), and 0.456 (95% CI 0.389-0.523), respectively. For the effect of albuminuria on cancer mortality, these values were 0.237 (95% CI 0.148-0.326), 0.244 (95% CI 0.154-0.333), and 0.102 (95% CI 0.012-0.192), respectively.

CONCLUSIONS

Studies of absolute risk should use methods that adjust for competing risks to avoid over-stating risk, such as the CICR estimator. Studies of relative risk should consider carefully which measure of association is most appropriate for the research question.

Systematic Review and Metaanalysis Comparing the Bias and Accuracy of the Modification of Diet in Renal Disease and Chronic Kidney Disease Epidemiology Collaboration Equations in Community-Based Populations

BACKGROUND

The majority of patients with chronic kidney disease are diagnosed and monitored in primary care. Glomerular filtration rate (GFR) is a key marker of renal function, but direct measurement is invasive; in routine practice, equations are used for estimated GFR (eGFR) from serum creatinine. We systematically assessed bias and accuracy of commonly used eGFR equations in populations relevant to primary care.

CONTENT

MEDLINE, EMBASE, and the Cochrane Library were searched for studies comparing measured GFR (mGFR) with eGFR in adult populations comparable to primary care and reporting both the Modification of Diet in Renal Disease (MDRD) and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations based on standardized creatinine measurements. We pooled data on mean bias (difference between eGFR and mGFR) and on mean accuracy (proportion of eGFR within 30% of mGFR) using a random-effects inverse-variance weighted metaanalysis. We included 48 studies of 26875 patients that reported data on bias and/or accuracy. Metaanalysis of within-study comparisons in which both formulae were tested on the same patient cohorts using isotope dilution-mass spectrometry-traceable creatinine showed a lower mean bias in eGFR using CKD-EPI of 2.2 mL/min/1.73 m² (95% CI, 1.1-3.2; 30 studies; I² = 74.4%) and a higher mean accuracy of CKD-EPI of 2.7% (1.6-3.8; 47 studies; I² = 55.5%). Metaregression showed that in both equations bias and accuracy favored the CKD-EPI equation at higher mGFR values.

SUMMARY

Both equations underestimated mGFR, but CKD-EPI gave more accurate estimates of GFR.

How can point-of-care HbA1c testing be integrated into UK primary care consultations? - A feasibility study

AIMS

Point-of-care (POC) HbA1c testing gives a rapid result, allowing testing and treatment decisions to take place in a single appointment. Trials of POC testing have not been shown to improve HbA1c, possibly because of how testing was implemented. This study aimed to identify key components of POC HbA1c testing and determine strategies to optimise implementation in UK primary care.

METHODS

This cohort feasibility study recruited thirty patients with type 2 diabetes and HbA1c>7.5% (58mmol/mol) into three primary care clinics. Patients' clinical care included two POC HbA1c tests over six months. Data were collected on appointment duration, clinical decisions, technical performance and patient behaviour.

RESULTS

Fifty-three POC HbA1c consultations took place during the study; clinical decisions were made in 30 consultations. Five POC consultations with a family doctor lasted on average 11min and 48 consultations with nurses took on average 24min. Five POC study visits did not take place in one clinic. POC results were uploaded to hospital records from two clinics. In total, sixty-three POC tests were performed, and there were 11 cartridge failures. No changes in HbA1c or patient behaviour were observed.

CONCLUSIONS

HbA1c measurement with POC devices can be effectively implemented in primary care. This work has identified when these technologies might work best, as well as potential challenges. The findings can be used to inform the design of a pragmatic trial to implement POC HbA1c testing.

Short- and medium-term effects of light to moderate alcohol intake on glycaemic control in diabetes mellitus: a systematic review and meta-analysis of randomized trials

BACKGROUND

People with diabetes are told that drinking alcohol may increase their risk of hypoglycaemia.

AIMS

To report the effects of alcohol consumption on glycaemic control in people with diabetes mellitus.

METHODS

Medline, EMBASE and the Cochrane Library databases were searched in 2015 to identify randomized trials that compared alcohol consumption with no alcohol use, reporting glycaemic control in people with diabetes. Data on blood glucose, HbA_1c and numbers of hypoglycaemic episodes were pooled using random effects meta-analysis.

RESULTS

Pooled data from nine short-term studies showed no difference in blood glucose concentrations between those who drank alcohol in doses of 16-80 g (median 20 g, 2.5 units) compared with those who did not drink alcohol at 0.5, 2, 4 and 24 h after alcohol consumption. Pooled data from five medium-term studies showed that there was no difference in blood glucose or HbA_1c concentrations at the end of the study between those who drank 11-18 g alcohol/day (median 13 g/day, 1.5 units/day) for 4-104 weeks and those who did not. We found no evidence of a difference in number of hypoglycaemic episodes or in withdrawal rates between randomized groups.

CONCLUSIONS

Studies to date have not provided evidence that drinking light to moderate amounts of alcohol, with or without a meal, affects any measure of glycaemic control in people with Type 2 diabetes. These results suggest that current advice that people with diabetes do not need to refrain from drinking moderate quantities of alcohol does not need to be changed; risks to those with Type 1 diabetes remain uncertain.

Performance of point-of-care HbA1c test devices: implications for use in clinical practice - a systematic review and meta-analysis

BACKGROUND

Point-of-care (POC) devices could be used to measure hemoglobin A1c (HbA1c) in the doctors' office, allowing immediate feedback of results to patients. Reports have raised concerns about the analytical performance of some of these devices. We carried out a systematic review and meta-analysis using a novel approach to compare the accuracy and precision of POC HbA1c devices.

METHODS

Medline, Embase and Web of Science databases were searched in June 2015 for published reports comparing POC HbA1c devices with laboratory methods. Two reviewers screened articles and extracted data on bias, precision and diagnostic accuracy. Mean bias and variability between the POC and laboratory test were combined in a meta-analysis. Study quality was assessed using the QUADAS2 tool.

RESULTS

Two researchers independently reviewed 1739 records for eligibility. Sixty-one studies were included in the meta-analysis of mean bias. Devices evaluated were A1cgear, A1cNow, Afinion, B-analyst, Clover, Cobas b101, DCA 2000/Vantage, HemoCue, Innovastar, Nycocard, Quo-Lab, Quo-Test and SDA1cCare. Nine devices had a negative mean bias which was significant for three devices. There was substantial variability in bias within devices. There was no difference in bias between clinical or laboratory operators in two devices.

CONCLUSIONS

This is the first meta-analysis to directly compare performance of POC HbA1c devices. Use of a device with a mean negative bias compared to a laboratory method may lead to higher levels of glycemia and a lower risk of hypoglycaemia. The implications of this on clinical decision-making and patient outcomes now need to be tested in a randomized trial.

Estimating the accuracy of muscle response testing: two randomised-order blinded studies

Background

Manual muscle testing (MMT) is a non-invasive assessment tool used by a variety of health care providers to evaluate neuromusculoskeletal integrity, and muscular strength in particular. In one form of MMT called muscle response testing (MRT), muscles are said to be tested, not to evaluate muscular strength, but neural control. One established, but insufficiently validated, application of MRT is to assess a patient’s response to semantic stimuli (e.g. spoken lies) during a therapy session. Our primary aim was to estimate the accuracy of MRT to distinguish false from true spoken statements, in randomised and blinded experiments. A secondary aim was to compare MRT accuracy to the accuracy when practitioners used only their intuition to differentiate false from true spoken statements.

Methods

Two prospective studies of diagnostic test accuracy using MRT to detect lies are presented. A true positive MRT test was one that resulted in a subjective weakening of the muscle following a lie, and a true negative was one that did not result in a subjective weakening of the muscle following a truth. Experiment 2 replicated Experiment 1 using a simplified methodology. In Experiment 1, 48 practitioners were paired with 48 MRT-naïve test patients, forming unique practitioner-test patient pairs. Practitioners were enrolled with any amount of MRT experience. In Experiment 2, 20 unique pairs were enrolled, with test patients being a mix of MRT-naïve and not-MRT-naïve. The primary index test was MRT. A secondary index test was also enacted in which the practitioners made intuitive guesses (“intuition”), without using MRT. The actual verity of the spoken statement was compared to the outcome of both index tests (MRT and Intuition) and their mean overall fractions correct were calculated and reported as mean accuracies.

Results

In Experiment 1, MRT accuracy, 0.659 (95% CI 0.623 - 0.695), was found to be significantly different (p < 0.01) from intuition accuracy, 0.474 (95% CI 0.449 - 0.500), and also from the likelihood of chance (0.500; p < 0.01). Experiment 2 replicated the findings of Experiment 1. Testing for various factors that may have influenced MRT accuracy failed to detect any correlations.

Conclusions

MRT has repeatedly demonstrated significant accuracy for distinguishing lies from truths, compared to both intuition and chance. The primary limitation of this study is its lack of generalisability to other applications of MRT and to MMT.

Study registration

The Australian New Zealand Clinical Trials Registry (ANZCTR; www.anzctr.org.au; ID # ACTRN12609000455268, and US-based ClinicalTrials.gov (ID # NCT01066312).

Electronic supplementary material

The online version of this article (doi:10.1186/s12906-016-1416-2) contains supplementary material, which is available to authorized users.

Blood pressure variability and cardiovascular disease: systematic review and meta-analysis

OBJECTIVE

 To systematically review studies quantifying the associations of long term (clinic), mid-term (home), and short term (ambulatory) variability in blood pressure, independent of mean blood pressure, with cardiovascular disease events and mortality.

DATA SOURCES

 Medline, Embase, Cinahl, and Web of Science, searched to 15 February 2016 for full text articles in English.

ELIGIBILITY CRITERIA FOR STUDY SELECTION

 Prospective cohort studies or clinical trials in adults, except those in patients receiving haemodialysis, where the condition may directly impact blood pressure variability. Standardised hazard ratios were extracted and, if there was little risk of confounding, combined using random effects meta-analysis in main analyses. Outcomes included all cause and cardiovascular disease mortality and cardiovascular disease events. Measures of variability included standard deviation, coefficient of variation, variation independent of mean, and average real variability, but not night dipping or day-night variation.

RESULTS

 41 papers representing 19 observational cohort studies and 17 clinical trial cohorts, comprising 46 separate analyses were identified. Long term variability in blood pressure was studied in 24 papers, mid-term in four, and short-term in 15 (two studied both long term and short term variability). Results from 23 analyses were excluded from main analyses owing to high risks of confounding. Increased long term variability in systolic blood pressure was associated with risk of all cause mortality (hazard ratio 1.15, 95% confidence interval 1.09 to 1.22), cardiovascular disease mortality (1.18, 1.09 to 1.28), cardiovascular disease events (1.18, 1.07 to 1.30), coronary heart disease (1.10, 1.04 to 1.16), and stroke (1.15, 1.04 to 1.27). Increased mid-term and short term variability in daytime systolic blood pressure were also associated with all cause mortality (1.15, 1.06 to 1.26 and 1.10, 1.04 to 1.16, respectively).

CONCLUSIONS

 Long term variability in blood pressure is associated with cardiovascular and mortality outcomes, over and above the effect of mean blood pressure. Associations are similar in magnitude to those of cholesterol measures with cardiovascular disease. Limited data for mid-term and short term variability showed similar associations. Future work should focus on the clinical implications of assessment of variability in blood pressure and avoid the common confounding pitfalls observed to date.

SYSTEMATIC REVIEW REGISTRATION

 PROSPERO CRD42014015695.

The use of statistical methodology to determine the accuracy of grading within a diabetic retinopathy screening programme

AIMS

We aimed to use longitudinal data from an established screening programme with good quality assurance and quality control procedures and a stable well-trained workforce to determine the accuracy of grading in diabetic retinopathy screening.

METHODS

We used a continuous time-hidden Markov model with five states to estimate the probability of true progression or regression of retinopathy and the conditional probability of an observed grade given the true grade (misclassification). The true stage of retinopathy was modelled as a function of the duration of diabetes and HbA1c .

RESULTS

The modelling dataset consisted of 65 839 grades from 14 187 people. The median number [interquartile range (IQR)] of examinations was 5 (3, 6) and the median (IQR) interval between examinations was 1.04 (0.99, 1.17) years. In total, 14 227 grades (21.6%) were estimated as being misclassified, 10 592 (16.1%) represented over-grading and 3635 (5.5%) represented under-grading. There were 1935 (2.9%) misclassified referrals, 1305 were false-positive results (2.2%) and 630 were false-negative results (1.0%). Misclassification of background diabetic retinopathy as no detectable retinopathy was common (3.4% of all grades) but rarely preceded referable maculopathy or retinopathy.

CONCLUSION

Misclassification between lower grades of retinopathy is not uncommon but is unlikely to lead to significant delays in referring people for sight-threatening retinopathy.

Methods for meta-analysis of pharmacodynamic dose-response data with application to multi-arm studies of alogliptin

Standard methods for meta-analysis of dose-response data in epidemiology assume a model with a single scalar parameter, such as log-linear relationships between exposure and outcome; such models are implicitly unbounded. In contrast, in pharmacology, multi-parameter models, such as the widely used E_max model, are used to describe relationships that are bounded above and below. We propose methods for estimating the parameters of a dose-response model by meta-analysis of summary data from the results of randomized controlled trials of a drug, in which each trial uses multiple doses of the drug of interest (possibly including dose 0 or placebo). We assume that, for each randomized arm of each trial, the mean and standard error of a continuous response measure and the corresponding allocated dose are available. We consider weighted least squares fitting of the model to the mean and dose pairs from all arms of all studies, and a two-stage procedure in which scalar inverse-variance meta-analysis is performed at each dose, and the dose-response model is fitted to the results by weighted least squares. We then compare these with two further methods inspired by network meta-analysis that fit the model to the contrasts between doses. We illustrate the methods by estimating the parameters of the E_max model to a collection of multi-arm, multiple-dose, randomized controlled trials of alogliptin, a drug for the management of diabetes mellitus, and further examine the properties of the four methods with sensitivity analyses and a simulation study. We find that all four methods produce broadly comparable point estimates for the parameters of most interest, but a single-stage method based on contrasts between doses produces the most appropriate confidence intervals. Although simpler methods may have pragmatic advantages, such as the use of standard software for scalar meta-analysis, more sophisticated methods are nevertheless preferable for their advantages in estimation.

Quantifying the effects of diuretics and β-adrenoceptor blockers on glycaemic control in diabetes mellitus - a systematic review and meta-analysis

AIMS

Although there are reports that β-adrenoceptor antagonists (beta-blockers) and diuretics can affect glycaemic control in people with diabetes mellitus, there is no clear information on how blood glucose concentrations may change and by how much. We report results from a systematic review to quantify the effects of these antihypertensive drugs on glycaemic control in adults with established diabetes.

METHODS

We systematically reviewed the literature to identify randomized controlled trials in which glycaemic control was studied in adults with diabetes taking either beta-blockers or diuretics. We combined data on HbA1c and fasting blood glucose using fixed effects meta-analysis.

RESULTS

From 3864 papers retrieved, we found 10 studies of beta-blockers and 12 studies of diuretics to include in the meta-analysis. One study included both comparisons, totalling 21 included reports. Beta-blockers increased fasting blood glucose concentrations by 0.64 mmol l(-1) (95% CI 0.24, 1.03) and diuretics by 0.77 mmol l(-1) (95% CI 0.14, 1.39) compared with placebo. Effect sizes were largest in trials of non-selective beta-blockers (1.33, 95% CI 0.72, 1.95) and thiazide diuretics (1.69, 95% CI 0.60, 2.69). Beta-blockers increased HbA1c concentrations by 0.75% (95% CI 0.30, 1.20) and diuretics by 0.24% (95% CI -0.17, 0.65) compared with placebo. There was no significant difference in the number of hypoglycaemic events between beta-blockers and placebo in three trials.

CONCLUSIONS

Randomized trials suggest that thiazide diuretics and non-selective beta-blockers increase fasting blood glucose and HbA1c concentrations in patients with diabetes by moderate amounts. These data will inform prescribing and monitoring of beta-blockers and diuretics in patients with diabetes.

Chronic renal disease is not chronic kidney disease: implications for use of the QRISK and Joint British Societies risk scores

BACKGROUND

Chronic kidney disease (CKD) is a major risk factor for cardiovascular disease (CVD) and European guidelines advocate assessment of CVD risk. QRISK and JBS3 risk calculators do not use the consensus definition of CKD stages 3-5 but instead use a definition referring to renal pathologies and CKD stages 4 and 5. Consequently, there is potential for doctors to misclassify their patients when using these risk calculators.

OBJECTIVES

To quantify the number of people who may be affected by such misclassifications.

METHODS

Database analysis using the Clinical Practice Research Datalink (CPRD).We identified 2512053 adults aged 25-84 without prior history of CVD on 1st January 2014. We identified those with 'chronic renal disease' and/or CKD by searching medical event history data.

RESULTS

The study population was 48.7% male with mean age of 50.2 years. A total of 80718 had diagnostic READ codes for CKD stages 3, 4 or 5. Of these, 6585 individuals (8.2%) were classified as having 'chronic renal disease' according to the updated QRISK 2014, up from 3365 according to QRISK 2013. Whilst the updated QRISK definition of 'chronic renal disease' in total identified 62% more people than previously and had improved sensitivity for CKD stages 3 to 5, sensitivity remained poor (8.16%; 95% CI: 7.97-8.35%).

CONCLUSION

Misuse of risk scores by general practitioners could result in clinically important differences in risk estimates. Users of risk scores should recognize the potential for error and developers should aim to label risk factors more clearly.

Heterogeneity of prognostic studies of 24-hour blood pressure variability: systematic review and meta-analysis

In addition to mean blood pressure, blood pressure variability is hypothesized to have important prognostic value in evaluating cardiovascular risk. We aimed to assess the prognostic value of blood pressure variability within 24 hours. Using MEDLINE, EMBASE and Cochrane Library to April 2013, we conducted a systematic review of prospective studies of adults, with at least one year follow-up and any day, night or 24-hour blood pressure variability measure as a predictor of one or more of the following outcomes: all-cause mortality, cardiovascular mortality, all cardiovascular events, stroke and coronary heart disease. We examined how blood pressure variability is defined and how its prognostic use is reported. We analysed relative risks adjusted for covariates including the appropriate mean blood pressure and considered the potential for meta-analysis. Our analysis of methods included 24 studies and analysis of predictions included 16 studies. There were 36 different measures of blood pressure variability and 13 definitions of night- and day-time periods. Median follow-up was 5.5 years (interquartile range 4.2–7.0). Comparing measures of dispersion, coefficient of variation was less well researched than standard deviation. Night dipping based on percentage change was the most researched measure and the only measure for which data could be meaningfully pooled. Night dipping or lower night-time blood pressure was associated with lower risk of cardiovascular events. The interpretation and use in clinical practice of 24-hour blood pressure variability, as an important prognostic indicator of cardiovascular events, is hampered by insufficient evidence and divergent methodologies. We recommend greater standardisation of methods.

The science of clinical practice: disease diagnosis or patient prognosis? Evidence about "what is likely to happen" should shape clinical practice

BACKGROUND

Diagnosis is the traditional basis for decision-making in clinical practice. Evidence is often lacking about future benefits and harms of these decisions for patients diagnosed with and without disease. We propose that a model of clinical practice focused on patient prognosis and predicting the likelihood of future outcomes may be more useful.

DISCUSSION

Disease diagnosis can provide crucial information for clinical decisions that influence outcome in serious acute illness. However, the central role of diagnosis in clinical practice is challenged by evidence that it does not always benefit patients and that factors other than disease are important in determining patient outcome. The concept of disease as a dichotomous 'yes' or 'no' is challenged by the frequent use of diagnostic indicators with continuous distributions, such as blood sugar, which are better understood as contributing information about the probability of a patient's future outcome. Moreover, many illnesses, such as chronic fatigue, cannot usefully be labelled from a disease-diagnosis perspective. In such cases, a prognostic model provides an alternative framework for clinical practice that extends beyond disease and diagnosis and incorporates a wide range of information to predict future patient outcomes and to guide decisions to improve them. Such information embraces non-disease factors and genetic and other biomarkers which influence outcome.

SUMMARY

Patient prognosis can provide the framework for modern clinical practice to integrate information from the expanding biological, social, and clinical database for more effective and efficient care.

Diagnostic accuracy study of three alcohol breathalysers marketed for sale to the public

OBJECTIVES

To assess the diagnostic accuracy of three personal breathalyser devices available for sale to the public marketed to test safety to drive after drinking alcohol.

DESIGN

Prospective comparative diagnostic accuracy study comparing two single-use breathalysers and one digital multiuse breathalyser (index tests) to a police breathalyser (reference test).

SETTING

Establishments licensed to serve alcohol in a UK city.

PARTICIPANTS

Of 222 participants recruited, 208 were included in the main analysis. Participants were eligible if they were 18 years old or over, had consumed alcohol and were not intending to drive within the following 6 h.

OUTCOME MEASURES

Sensitivity and specificity of the breathalysers for the detection of being at or over the UK legal driving limit (35 µg/100 mL breath alcohol concentration).

RESULTS

18% of participants (38/208) were at or over the UK driving limit according to the police breathalyser. The digital multiuse breathalyser had a sensitivity of 89.5% (95% CI 75.9% to 95.8%) and a specificity of 64.1% (95% CI 56.6% to 71.0%). The single-use breathalysers had a sensitivity of 94.7% (95% CI 75.4% to 99.1%) and 26.3% (95% CI 11.8% to 48.8%), and a specificity of 50.6% (95% CI 40.4% to 60.7%) and 97.5% (95% CI 91.4% to 99.3%), respectively. Self-reported alcohol consumption threshold of 5 UK units or fewer had a higher sensitivity than all personal breathalysers.

CONCLUSIONS

One alcohol breathalyser had sensitivity of 26%, corresponding to false reassurance for approximately three people in four who are over the limit by the reference standard, at least on the evening of drinking alcohol. The other devices tested had 90% sensitivity or higher. All estimates were subject to uncertainty. There is no clearly defined minimum sensitivity for this safety-critical application. We conclude that current regulatory frameworks do not ensure high sensitivity for these devices marketed to consumers for a decision with potentially catastrophic consequences.

Changes in HbA1c level over a 12-week follow-up in patients with type 2 diabetes following a medication change

Background

Current guidance about the interval needed before retesting HbA1c when monitoring for glycaemic control is based on expert opinion rather than well-powered studies. The aim of our work was to explore how fast HbA1c changes after a change in glucose-lowering medication. This has implications for whether routine HbA1c testing intervals before 12 weeks could inform diabetes medication adjustments.

Methods

This 12-week cohort study recruited patients from 18 general practices in the United Kingdom with non-insulin treated diabetes who were initiating or changing dose of oral glucose-lowering medication. HbA1c was measured at baseline and 2, 4, 8 and 12 weeks after recruitment. HbA1c levels at earlier time intervals were correlated with 12-week HbA1c. A ROC curve analysis was used to identify the 8-week threshold above which medication adjustment may be clinically appropriate.

Results

Ninety-three patients were recruited to the study. Seventy-nine patients with no change in medication and full 12-week follow-up had the following baseline characteristics: mean±standard deviation age of 61.3±10.8 years, 34% were female and diabetes duration of 6.0±4.3 years. Mean HbA1c at baseline, 2, 4, 8 and 12 weeks was 8.7±1.5%, (72.0±16.8 mmol/mol) 8.6±1.6% (70.7±17.0 mmol/mol), 8.4±1.5% (68.7±15.9 mmol/mol), 8.2±1.4% (66.3±15.8 mmol/mol) and 8.1±1.4% (64.8±15.7 mmol/mol) respectively. At the end of the study 61% of patients had sub-optimal glycaemic control (HbA1c>7.5% or 59 mmol/mol). The 8-week change correlated significantly with the 12-week change in HbA1c and an HbA1c above 8.2% (66 mmol/mol) at 8 weeks correctly classified all 28 patients who had not achieved glycaemic control by 12 weeks.

Conclusions/interpretation

This is the first study designed with sufficient power to examine short-term changes in HbA1c. The 12-week change in HbA1c can be predicted 8 weeks after a medication change. Many participants who had not achieved glycaemic control after 12 weeks may have benefitted from an earlier review of their HbA1c and medication.

All-cause and cardiovascular mortality in middle-aged people with type 2 diabetes compared with people without diabetes in a large U.K. primary care database

OBJECTIVE

Middle-aged people with diabetes have been reported to have significantly higher risks of cardiovascular events than people without diabetes. However, recent falls in cardiovascular disease rates and more active management of risk factors may have abolished the increased risk. We aimed to provide an up-to-date assessment of the relative risks associated with type 2 diabetes of all-cause and cardiovascular mortality in middle-aged people in the U.K.

RESEARCH DESIGN AND METHODS

Using data from the General Practice Research Database, from 2004 to 2010, we conducted a cohort study of 87,098 people, 40-65 years of age at baseline, comparing 21,798 with type 2 diabetes and 65,300 without diabetes, matched on age, sex, and general practice. We produced hazard ratios (HRs) for mortality and compared rates of blood pressure testing, cholesterol monitoring, and use of aspirin, statins, and antihypertensive drugs. RESULTS People with type 2 diabetes, compared with people without diabetes, had a twofold increased risk of all-cause mortality (HR 2.07 [95% CI 1.95-2.20], adjusted for smoking) and a threefold increased risk of cardiovascular mortality (3.25 [2.87-3.68], adjusted for smoking). Women had a higher relative risk than men, and people <55 years of age had a higher relative risk than those >55 years of age. Monitoring and medication rates were higher in those with diabetes (all P < 0.001).

CONCLUSIONS

Despite efforts to manage risk factors, administer effective treatments, and develop new therapies, middle-aged people with type 2 diabetes remain at significantly increased risk of death.

Simulating lifetime outcomes associated with complications for people with type 1 diabetes

OBJECTIVES

The aim of this study was to develop a discrete-time simulation model for people with type 1 diabetes mellitus, to estimate and compare mean life expectancy and quality-adjusted life-years (QALYs) over a lifetime between intensive and conventional blood glucose treatment groups.

METHODS

We synthesized evidence on type 1 diabetes patients using several published sources. The simulation model was based on 13 equations to estimate risks of events and mortality. Cardiovascular disease (CVD) risk was obtained from results of the DCCT (diabetes control and complications trial). Mortality post-CVD event was based on a study using linked administrative data on people with diabetes from Western Australia. Information on incidence of renal disease and the progression to CVD was obtained from studies in Finland and Italy. Lower-extremity amputation (LEA) risk was based on the type 1 diabetes Swedish inpatient registry, and the risk of blindness was obtained from results of a German-based study. Where diabetes-specific data were unavailable, information from other populations was used. We examine the degree and source of parameter uncertainty and illustrate an application of the model in estimating lifetime outcomes of using intensive and conventional treatments for blood glucose control.

RESULTS

From 15 years of age, male and female patients had an estimated life expectancy of 47.2 (95 % CI 35.2-59.2) and 52.7 (95 % CI 41.7-63.6) years in the intensive treatment group. The model produced estimates of the lifetime benefits of intensive treatment for blood glucose from the DCCT of 4.0 (95 % CI 1.2-6.8) QALYs for women and 4.6 (95 % CI 2.7-6.9) QALYs for men. Absolute risk per 1,000 person-years for fatal CVD events was simulated to be 1.37 and 2.51 in intensive and conventional treatment groups, respectively.

CONCLUSIONS

The model incorporates diabetic complications risk data from a type 1 diabetes population and synthesizes other type 1-specific data to estimate long-term outcomes of CVD, end-stage renal disease, LEA and risk of blindness, along with life expectancy and QALYs. External validation was carried out using life expectancy and absolute risk for fatal CVD events. Because of the flexible and transparent nature of the model, it has many potential future applications.

Estimating the effect of sulfonylurea on HbA1c in diabetes: a systematic review and meta-analysis

AIMS/HYPOTHESIS

Sulfonylureas are widely prescribed glucose-lowering medications for diabetes, but the extent to which they improve glycaemia is poorly documented. This systematic review evaluates how sulfonylurea treatment affects glycaemic control.

METHODS

Medline, EMBASE, the Cochrane Library and clinical trials registries were searched to identify double-blinded randomised controlled trials of fixed-dose sulfonylurea monotherapy or sulfonylurea added on to other glucose-lowering treatments. The primary outcome assessed was change in HbA1c, and secondary outcomes were adverse events, insulin dose and change in body weight.

RESULTS

Thirty-one trials with a median duration of 16 weeks were included in the meta-analysis. Sulfonylurea monotherapy (nine trials) lowered HbA1c by 1.51% (17 mmol/mol) more than placebo (95% CI, 1.25, 1.78). Sulfonylureas added to oral diabetes treatment (four trials) lowered HbA1c by 1.62% (18 mmol/mol; 95% CI 1.0, 2.24) compared with the other treatment, and sulfonylurea added to insulin (17 trials) lowered HbA1c by 0.46% (6 mmol/mol; 95% CI 0.24, 0.69) and lowered insulin dose. Higher sulfonylurea doses did not reduce HbA1c more than lower doses. Sulfonylurea treatment resulted in more hypoglycaemic events (RR 2.41, 95% CI 1.41, 4.10) but did not significantly affect the number of other adverse events. Trial length, sulfonylurea type and duration of diabetes contributed to heterogeneity.

CONCLUSIONS/INTERPRETATION

Sulfonylurea monotherapy lowered HbA1c level more than previously reported, and we found no evidence that increasing sulfonylurea doses resulted in lower HbA1c. HbA1c is a surrogate endpoint, and we were unable to examine long-term endpoints in these predominately short-term trials, but sulfonylureas appear to be associated with an increased risk of hypoglycaemic events.

Pain over speed bumps in diagnosis of acute appendicitis: diagnostic accuracy study

OBJECTIVE

To assess the diagnostic accuracy of pain on travelling over speed bumps for the diagnosis of acute appendicitis.

DESIGN

Prospective questionnaire based diagnostic accuracy study.

SETTING

Secondary care surgical assessment unit at a district general hospital in the UK.

PARTICIPANTS

101 patients aged 17-76 years referred to the on-call surgical team for assessment of possible appendicitis.

MAIN OUTCOME MEASURES

Sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios for pain over speed bumps in diagnosing appendicitis, with histological diagnosis of appendicitis as the reference standard.

RESULTS

The analysis included 64 participants who had travelled over speed bumps on their journey to hospital. Of these, 34 had a confirmed histological diagnosis of appendicitis, 33 of whom reported increased pain over speed bumps. The sensitivity was 97% (95% confidence interval 85% to 100%), and the specificity was 30% (15% to 49%). The positive predictive value was 61% (47% to 74%), and the negative predictive value was 90% (56% to 100%). The likelihood ratios were 1.4 (1.1 to 1.8) for a positive test result and 0.1 (0.0 to 0.7) for a negative result. Speed bumps had a better sensitivity and negative likelihood ratio than did other clinical features assessed, including migration of pain and rebound tenderness.

CONCLUSIONS

Presence of pain while travelling over speed bumps was associated with an increased likelihood of acute appendicitis. As a diagnostic variable, it compared favourably with other features commonly used in clinical assessment. Asking about speed bumps may contribute to clinical assessment and could be useful in telephone assessment of patients.

Cancer outcomes and all-cause mortality in adults allocated to metformin: systematic review and collaborative meta-analysis of randomised clinical trials

AIMS/HYPOTHESIS

Observational studies suggest that metformin may reduce cancer risk by approximately one-third. We examined cancer outcomes and all-cause mortality in published randomised controlled trials (RCTs).

METHODS

RCTs comparing metformin with active glucose-lowering therapy or placebo/usual care, with minimum 500 participants and 1-year follow-up, were identified by systematic review. Data on cancer incidence and all-cause mortality were obtained from publications or by contacting investigators. For two trials, cancer incidence data were not available; cancer mortality was used as a surrogate. Summary RRs, 95% CIs and I (2)statistics for heterogeneity were calculated by fixed effects meta-analysis.

RESULTS

Of 4,039 abstracts identified, 94 publications described 14 eligible studies. RRs for cancer were available from 11 RCTs with 398 cancers during 51,681 person-years. RRs for all-cause mortality were available from 13 RCTs with 552 deaths during 66,447 person-years. Summary RRs for cancer outcomes in people randomised to metformin compared with any comparator were 1.02 (95% CI 0.82, 1.26) across all trials, 0.98 (95% CI 0.77, 1.23) in a subgroup analysis of active-comparator trials and 1.36 (95% CI 0.74, 2.49) in a subgroup analysis of placebo/usual care comparator trials. The summary RR for all-cause mortality was 0.94 (95% CI 0.79, 1.12) across all trials.

CONCLUSIONS/INTERPRETATION

Meta-analysis of currently available RCT data does not support the hypothesis that metformin lowers cancer risk by one-third. Eligible trials also showed no significant effect of metformin on all-cause mortality. However, limitations include heterogeneous comparator types, absent cancer data from two trials, and short follow-up, especially for mortality.

Quantifying the effect of metformin treatment and dose on glycemic control

OBJECTIVE

Metformin is the first-line oral medication recommended for glycemic control in patients with type 2 diabetes. We reviewed the literature to quantify the effect of metformin treatment on glycated hemoglobin (HbA(1c)) levels in all types of diabetes and examine the impact of differing doses on glycemic control.

RESEARCH DESIGN AND METHODS

MEDLINE, EMBASE, and the Cochrane Library were searched from 1950 to June 2010 for trials of at least 12 weeks' duration in which diabetic patients were treated with either metformin monotherapy or as an add-on therapy. Data on change in HbA(1c) were pooled in a meta-analysis. Data from dose-comparison trials were separately pooled.

RESULTS

A total of 35 trials were identified for the main analysis and 7 for the dose-comparison analysis. Metformin monotherapy lowered HbA(1c) by 1.12% (95% CI 0.92-1.32; I(2) = 80%) versus placebo, metformin added to oral therapy lowered HbA(1c) by 0.95% (0.77-1.13; I(2) = 77%) versus placebo added to oral therapy, and metformin added to insulin therapy lowered HbA(1c) by 0.60% (0.30-0.91; I(2) = 79.8%) versus insulin only. There was a significantly greater reduction in HbA(1c) using higher doses of metformin compared with lower doses of metformin with no significant increase in side effects.

CONCLUSIONS

Evidence supports the effectiveness of metformin therapy in a clinically important lowering of HbA(1c) used as monotherapy and in combination with other therapeutic agents. There is potential for using higher doses of metformin to maximize glycemic control in diabetic patients without increasing gastrointestinal effects.

Establishing an evidence base for frequency of monitoring glycated haemoglobin levels in patients with Type 2 diabetes: projections of effectiveness from a regression model

AIMS

Glycated haemoglobin (HbA1c) is monitored to guide treatment decisions in relation to glycaemic goals over time. Changes between two consecutive HbA1c tests result not only from deterioration or improvement in glycaemic control, but also from biovariability and measurement error. We model how this short-term variability impacts on HbA1c monitoring.

METHODS

Using data from a randomized trial of non-insulin treated patients with Type 2 diabetes we fitted a random-effects model for progression and variability of HbA1c. We estimated how many tests where HbA1c ≥ 7.5% (58.5 mmol/mol) would be false-positive (underlying HbA1c < 7.5% but test ≥ 7.5% owing to variability) vs. true-positive, in people with initial HbA1c between 6.5% and 7.3% (48 mmol/mol and 56 mmol/mol).

RESULTS

Participants (n = 320) had mean (SD) age 66 (10) years, BMI 31.3 (6.0) kg/m2 and median HbA1c was 7.1% (54 mmol/mol) with interquartile range 6.6% (49 mmol/mol) to 7.7% (61 mmol/mol). Mean (95% CI) change in HbA1c was 0.1% (1 mmol/mol) with 95% confidence interval 0.05% (0.5 mmol/mol) to 0.15% (2 mmol/mol) per 6 months. The minimum interval at which a true-positive test is more likely than a false positive test is 270 days for a starting HbA1c of 6.9% (52 mmol/mol) and 360 days at a starting value of 6.5% (48 mmol/mol).

CONCLUSION

In patients with initial HbA1c close to treatment goal, retesting at 6 months would yield more true-positive than false-positive tests. For patients with lower initial HbA1c, retesting at 6 months would yield more false than true-positive tests. In all patients, retesting at 12 months yields more true than false-positive tests. In very few patients would retesting at 3 months be justified.

Evidence for a dose effect of renin-angiotensin system inhibition on progression of microalbuminuria in Type 2 diabetes: a meta-analysis

AIMS

Renin-angiotensin inhibitors in Type 2 diabetes and microalbuminuria reduce renal and cardiovascular risk, but evidence supporting use of maximal tolerated dose is unclear. We aimed to determine the extent of renin-angiotensin inhibitor dose-dependent effects from randomized trials carried out in a clinical setting.

METHODS

In a meta-analysis of randomized clinical trials, alternate doses of angiotensin receptor blockers or angiotensin converting enzyme inhibitors in patients with Type 2 diabetes and microalbuminuria were compared. MEDLINE, EMBASE and the Cochrane Register of Controlled Trials were searched from January 2006 to August 2010. Trials prior to January 2006 were identified from a prior systematic review. Identified outcomes were albumin excretion rate, progression and regression of albuminuria and adverse events.

RESULTS

Four trials including 1051 patients compared doses of angiotensin receptor blockers. No trials compared doses of angiotensin converting enzyme inhibitor. The percentage decline in albumin excretion rate from baseline was greater with higher doses (18% higher, 95% CI 8-28%), the regression to normoalbuminuria was greater (OR 1.66, 95% CI 1.22-2.27), with less progression to macroalbuminuria (OR 0.62, CI 0.38-1.02). Adverse events were fewer with lower-dose angiotensin receptor blockers (OR 1.32, 95% CI 0.90-1.92).

CONCLUSIONS

Higher-dose compared with lower-dose angiotensin receptor blockers in Type 2 diabetes with microalbuminuria are associated with significantly reduced albumin excretion rate and increased regression to normoalbuminuria. Adverse events are more frequent, but not significantly so. There is potential for trials to determine clinical cardiovascular and renal outcomes at differing doses. Our findings support current recommendations to titrate renin-angiotensin inhibitors to maximum dose whilst considering risk of adverse side effects with higher doses.

Statistical models for the control phase of clinical monitoring

The rise in the prevalence of chronic conditions means that these are now the leading causes of death and disability worldwide, accounting for almost 60% of all deaths and 43% of the global burden of disease. Management of chronic conditions requires both effective treatment and ongoing monitoring. Although costs related to monitoring are substantial, there is relatively little evidence on its effectiveness. Monitoring is inherently different to diagnosis in its use of regularly repeated tests, and increasing frequency can result in poorer rather than better statistical properties because of multiple testing in the presence of high variability. We present here a general framework for modelling the control phase of a monitoring programme, and for the estimation of quantities of potential clinical interest such as the ratio of false to true positive tests. We show how four recent clinical studies of monitoring cardiovascular disease, hypertension, diabetes and HIV infection can be thought as special cases of this framework; as well as using this framework to clarify the choice of estimation and calculation methods available. Noticeably, in each of the presented examples over-frequent monitoring appears to be a greater problem than under-frequent monitoring. We also present recalculations of results under alternative conditions, illustrating conceptual decisions about modelling the true or observed value of a clinical measure.

Reproductive history and pancreatic cancer incidence and mortality in a cohort of postmenopausal women

There is inconsistent evidence about the effect of reproductive history on women's risk of pancreatic cancer. In the Million Women Study, a prospective cohort of middle-aged women in the United Kingdom, we examined associations between reproductive history and pancreatic cancer incidence and mortality, controlling for age, socioeconomic status, geographic region, body mass index, smoking, and history of diabetes. During 7.1 million person-years of follow-up in 995,192 postmenopausal women, there were 1,182 incident pancreatic cancers. Pancreatic cancer incidence and mortality did not vary significantly with age at menarche, number of children, age at first birth, breast-feeding, type of menopause, age at menopause, or time since menopause. Any effect of reproductive history and pancreatic cancer risk in women is likely to be weak, if it exists at all.

Impact of using a non-diabetes-specific risk calculator on eligibility for statin therapy in type 2 diabetes

AIMS/HYPOTHESIS

The aim of this study was to investigate the impact of using a non-diabetes-specific cardiovascular disease (CVD) risk calculator to determine eligibility for statin therapy according to current UK National Institute for Health and Clinical Excellence (NICE) guidelines for those patients with type 2 diabetes who are at an increased risk of CVD (10 year risk >or=20%).

METHODS

The 10 year CVD risks were estimated using the UK Prospective Diabetes Study (UKPDS) Risk Engine and the Framingham equation for 4,025 patients enrolled in the Lipids in Diabetes Study who had established type 2 diabetes and LDL-cholesterol <4.1 mmol/l.

RESULTS

The mean (SD) age of the patients was 60.7 (8.6) years, blood pressure 141/83 (17/10) mmHg and the total cholesterol:HDL-cholesterol ratio was 3.9 (1.0). The median (interquartile range) diabetes duration was 6 (3-11) years and the HbA(1c) level was 8.0% (7.2-9.0%). The cohort comprised 65% men, 91% whites, 4% Afro-Caribbeans, 5% Asian Indians and 15% current smokers. More patients were classified as being at high risk by the UKPDS Risk Engine (65%) than by the Framingham CVD equation (63%) (p < 0.0001). The Framingham CVD equation classified fewer men and people aged <50 years old as high risk (p < 0.0001). There was no difference between the UKPDS Risk Engine and Framingham classification of women at high risk (p = 0.834).

CONCLUSIONS/INTERPRETATION

These results suggest that the use of Framingham-derived rather than UKPDS Risk Engine-derived CVD risk estimates would deny about one in 25 patients statin therapy when applying current NICE guidelines. Thus, under these guidelines the choice of CVD risk calculator is important when assessing CVD risk in patients with type 2 diabetes, particularly for the identification of the relatively small proportion of younger people who require statin therapy.

A method for assessing quality of control from glucose profiles

AIM

As the practice of multiple assessments of glucose concentration throughout the day increases for people with diabetes, there is a need for an assessment of glycaemic control weighted for the clinical risks of both hypoglycaemia and hyperglycaemia.

METHODS

We have developed a methodology to report the degree of risk which a glycaemic profile represents. Fifty diabetes professionals assigned risk values to a range of 40 blood glucose concentrations. Their responses were summarised and a generic function of glycaemic risk was derived. This function was applied to patient glucose profiles to generate an integrated risk score termed the Glycaemic Risk Assessment Diabetes Equation (GRADE). The GRADE score was then reported by use of the mean value and the relative percent contribution to the weighted risk score from the hypoglycaemic, euglycaemic, hyperglycaemic range, respectively, e.g. GRADE (hypoglycaemia%, euglycaemia%, hyperglycaemia%).

RESULTS

The GRADE scores of indicative glucose profiles were as follows: continuous glucose monitoring profile non-diabetic subjects GRADE = 1.1, Type 1 diabetes continuous glucose monitoring GRADE = 8.09 (20%, 8%, 72%), Type 2 diabetes home blood glucose monitoring GRADE = 9.97 (2%, 7%, 91%).

CONCLUSIONS

The GRADE score of a glucose profile summarises the degree of risk associated with a glucose profile. Values < 5 correspond to euglycaemia. The GRADE score is simple to generate from any blood glucose profile and can be used as an adjunct to HbA1c to report the degree of risk associated with glycaemic variability.