Christopher H. Jackson

Diagnostic accuracy of coronary angiography and risk factors for post–heart-transplant cardiac allograft vasculopathy

Cardiac allograft vasculopathy (CAV) is a common cause of death after heart transplantation. Coronary angiography is used to monitor the progress of recipients. Diagnostic accuracy of angiography and risk factors for CAV have not been clearly established. Between August 1979 and January 2002, 566 1-year survivors of heart transplantation underwent 2,168 angiograms and were classified as having no CAV (0% stenosis), mild-moderate CAV (up to 70% stenosis), or severe CAV (>70% stenosis). We used serial measurements of stenosis to estimate the diagnostic accuracy of angiography and to assess the following risk factors for CAV onset, progression, and survival: recipient and donor age and sex, preoperative ischemic heart disease (IHD), acute rejection rates, cytomegalovirus (CMV) infection, and serologic status. CAV was diagnosed by angiography in 248 of 556 (45%) 1-year survivors, with a mean onset time of 8.6 years. Patients spent a mean of 3.4 years with mild-moderate disease and 3.4 years with severe disease before death. Angiography specificity was 97.8%, and sensitivity was 79.3%. The following variables were found to significantly increase the risk of CAV onset: recipient age relative rate (95% confidence interval) 1.16 (1.01–1.34), donor age by 1.27 (1.13–1.43), male recipient by 2.00 (1.11–2.57), pretransplant IHD by 1.75 (1.30–2.36), cumulative rejection by 1.13 (1.05–1.21), and CMV infection by 1.42 (1.06–1.92). Acute rejection increased risk of death by 1.48 (1.19–1.85). Angiography is highly specific and moderately sensitive for diagnosis of CAV. Risk of CAV onset is related to donor age and recipient history of pretransplant IHD and is further increased by immune-related insults of acute rejection and CMV infection.

Clinical effectiveness and cost-effectiveness of endobronchial and endoscopic ultrasound relative to surgical staging in potentially resectable lung cancer: results from the ASTER randomised controlled trial

OBJECTIVE To assess the clinical effectiveness and cost-effectiveness of endosonography (followed by surgical staging if endosonography was negative), compared with standard surgical staging alone, in patients with non-small cell lung cancer (NSCLC) who are otherwise candidates for surgery with curative intent. DESIGN A prospective, international, open-label, randomised controlled study, with a trial-based economic analysis. SETTING Four centres: Ghent University Hospital, Belgium; Leuven University Hospitals, Belgium; Leiden University Medical Centre, the Netherlands; and Papworth Hospital, UK. PARTICIPANTS INCLUSION CRITERIA known/suspected NSCLC, with suspected mediastinal lymph node involvement; otherwise eligible for surgery with curative intent; clinically fit for endosonography and surgery; and no evidence of metastatic disease. EXCLUSION CRITERIA previous lung cancer treatment; concurrent malignancy; uncorrected coagulopathy; and not suitable for surgical staging. INTERVENTIONS Study patients were randomised to either surgical staging alone (n = 118) or endosonography followed by surgical staging if endosonography was negative (n = 123). Endosonography diagnostic strategy used endoscopic ultrasound-guided fine-needle aspiration combined with endobronchial ultrasound-guided transbronchial needle aspiration, followed by surgical staging if these tests were negative. Patients with no evidence of mediastinal metastases or tumour invasion were referred for surgery with curative intent. If evidence of malignancy was found, patients were referred for chemoradiotherapy. MAIN OUTCOME MEASURES The main clinical outcomes were sensitivity (positive diagnostic test/nodal involvement during any diagnostic test or thoracotomy) and negative predictive value (NPV) of each diagnostic strategy for the detection of N2/N3 metastases, unnecessary thoracotomy and complication rates. The primary economic outcome was cost-utility of the endosonography strategy compared with surgical staging alone, up to 6 months after randomisation, from a UK NHS perspective. RESULTS Clinical and resource-use data were available for all 241 patients, and complete utilities were available for 144. Sensitivity for detecting N2/N3 metastases was 79% [41/52; 95% confidence interval (CI) 66% to 88%] for the surgical arm compared with 94% (62/66; 95% CI 85% to 98%) for the endosonography strategy (p = 0.02). Corresponding NPVs were 86% (66/77; 95% CI 76% to 92%) and 93% (57/61; 95% CI 84% to 97%; p = 0.26). There were 21/118 (18%) unnecessary thoracotomies in the surgical arm compared with 9/123 (7%) in the endosonography arm (p = 0.02). Complications occurred in 7/118 (6%) in the surgical arm and 6/123 (5%) in the endosonography arm (p = 0.78): one pneumothorax related to endosonography and 12 complications related to surgical staging. Patients in the endosonography arm had greater EQ-5D (European Quality of Life-5 Dimensions) utility at the end of staging (0.117; 95% CI 0.042 to 0.192; p = 0.003). There were no other significant differences in utility. The main difference in resource use was the number of thoracotomies: 66% patients in the surgical arm compared with 53% in the endosonography arm. Resource use was similar between the groups in all other items. The 6-month cost of the endosonography strategy was £9713 (95% CI £7209 to £13,307) per patient versus £10,459 (£7732 to £13,890) for the surgical arm, mean difference £746 (95% CI -£756 to £2494). The mean difference in quality-adjusted life-year was 0.015 (95% CI -0.023 to 0.052) in favour of endosonography, so this strategy was cheaper and more effective. CONCLUSIONS Endosonography (followed by surgical staging if negative) had higher sensitivity and NPVs, resulted in fewer unnecessary thoracotomies and better quality of life during staging, and was slightly more effective and less expensive than surgical staging alone. Future work could investigate the need for confirmatory mediastinoscopy following negative endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), the diagnostic accuracy of EUS-FNA or EBUS-TBNA separately and the delivery of both EUS-FNA or EBUS-TBNA by suitably trained chest physicians. TRIAL REGISTRATION Current Controlled Trials ISRCTN 97311620. FUNDING This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 16, No. 18. See the HTA programme website for further project information.

Acute and chronic onset of bronchiolitis obliterans syndrome (BOS): are they different entities?

BACKGROUND Bronchiolitis obliterans syndrome (BOS), defined as an irreversible, staged decline in forced expiratory volume in 1 second (FEV(1)), is an established marker of obliterative bronchiolitis. Potential causes of BOS include sub-clinical chronic rejection and/or exaggerated healing response following acute injury. BOS may thus result from two or more distinct processes, both acute and chronic. METHODS A total of 5,916 measurements of FEV(1) from 204 lung transplant recipients surviving at least 6 months after transplantation were analyzed. Follow-up ranged from 6 months to 13 years. By adjusting for the acute effects of rejection, pulmonary infection and measurement variation on FEV(1) trace, patients either had a linear decline characterized by a single acute drop in FEV(1) of >15% at BOS onset, or a chronic linear decline in FEV(1). The fraction having acute onset was estimated. Acute events occurring within the first 6 months were assessed as risk factors for acute onset BOS. RESULTS Of the 204 patients, 8% died before BOS onset and 18% were BOS-free at analysis. For 18% of patients, BOS onset followed a chronic linear decline in FEV(1) of 3.7% per year, with a median time of BOS onset >99 months. For 56% of patients, BOS onset followed an acute drop in FEV(1) of median 33.8% (95% CI 19.1% to 39.7%), with median onset time of 52 months. During the first 6 months, acute rejection was significantly and independently associated with acute onset of BOS (relative risk = 1.15 per episode, 95% CI [1.03 to 1.29], p = 0.01), whereas pulmonary infection and cytomegalovirus (CMV) infection were not. Acute BOS onset followed a documented acute event in the previous 6 months in 38 of 114 (33%) of cases. CONCLUSIONS BOS likely reflects more than one process. Compared with those who had a slow linear decline in lung function, acute BOS onset was associated with acute rejection in the first 6 months, was often triggered by an acute event and had poor prognosis, with obliterative bronchiolitis (OB) the main cause of death.

Accounting for uncertainty in health economic decision models by using model averaging

Health economic decision models are subject to considerable uncertainty, much of which arises from choices between several plausible model structures, e.g. choices of covariates in a regression model. Such structural uncertainty is rarely accounted for formally in decision models but can be addressed by model averaging. We discuss the most common methods of averaging models and the principles underlying them. We apply them to a comparison of two surgical techniques for repairing abdominal aortic aneurysms. In model averaging, competing models are usually either weighted by using an asymptotically consistent model assessment criterion, such as the Bayesian information criterion, or a measure of predictive ability, such as Akaikes information criterion. We argue that the predictive approach is more suitable when modelling the complex underlying processes of interest in health economics, such as individual disease progression and response to treatment.

Expansion of the Kidney Donor Pool by Using Cardiac Death Donors with Prolonged Time to Cardiorespiratory Arrest

Donation after Cardiac Death (DCD) is an increasingly important source of kidney transplants, but because of concerns of ischemic injury during the agonal phase, many centers abandon donation if cardiorespiratory arrest has not occurred within 1 h of controlled withdrawal of life‐supporting treatment (WLST). We report the impact on donor numbers and transplant function using instead a minimum ‘cut‐off’ time of 4 h. The agonal phase of 173 potential DCD donors was characterized according to the presence or absence of: acidemia; lactic acidosis; prolonged (>30 min) hypotension, hypoxia or oliguria, and the impact of these characteristics on 3‐ and 12‐month transplant outcome evaluated by multivariable regression analysis. Of the 117 referrals who became donors, 27 (23.1%) arrested more than 1 h after WLST. Longer agonal‐phase times were associated with greater donor instability, but surprisingly neither agonal‐phase instability nor its duration influenced transplant outcome. In contrast, 3‐ and 12‐month eGFR in the 190 transplanted kidneys was influenced independently by donor age, and 3‐month eGFR by cold ischemic time. DCD kidney numbers are increased by 30%, without compromising transplant outcome, by lengthening the minimum waiting time after WLST from 1 to 4 h.

A Framework for Addressing Structural Uncertainty in Decision Models

Decision analytic models used for health technology assessment are subject to uncertainties. These uncertainties can be quantified probabilistically, by placing distributions on model parameters and simulating from these to generate estimates of cost-effectiveness. However, many uncertain model choices, often termed structural assumptions, are usually only explored informally by presenting estimates of cost-effectiveness under alternative scenarios. The authors show how 2 recent research proposals represent parts of a framework to formally account for all common structural uncertainties. First, the model is expanded to include parameters that encompass all possible structural choices. Uncertainty can then arise because these parameters are estimated imprecisely from data, for example, a treatment effect of doubtful significance. Uncertainty can also arise if there are no relevant data. If there are relevant data, uncertainty can be addressed by averaging expected costs and effects generated from probabilistic analysis of the models with and without the parameter. The weights used for averaging are related to the predictive ability of each model, assessed against the data. If there are no data, additional parameters can often be informed by eliciting expert beliefs as probability distributions. These ideas are illustrated in decision models for antiplatelet therapies for vascular disease and new biologic drugs for the treatment of active psoriatic arthritis.

Risk factors for the development and progression of dyslipidemia after heart transplantation.

Background. Hyperlipidemia is an important complication after organ transplantation and contributes to the development of posttransplant accelerated coronary artery diseases. Methods. We have retrospectively evaluated the relative contribution of various risk factors associated with the development and progression of hyperlipidemia in 194 heart transplant recipients by the use of mixed effects multiple linear regression analysis. The demographic characteristics evaluated were primary diagnosis of ischemic heart disease (IHD), gender, and age. Postoperative characteristics included number of treated rejections, dosage of cyclosporine (CYA), tacrolimus (TAC), prednisolone and azathioprine, and concentration of serum creatinine and glucose. The effects of administration of antihypertensive agents, diuretics, and lipid lowering agents were also studied. Results. The total cholesterol concentration increased significantly in the first 3 months posttransplant but gradually decreased thereafter. Total cholesterol and the ratio of low density lipoprotein (LDL) cholesterol to high density lipoprotein (HDL) cholesterol (LDL-C/HDL-C) increased to a greater extent in patients with IHD although female transplant recipients had a greater increase in the total cholesterol concentration. Each episode of rejection increased serum cholesterol by 0.306 mmol/liter (0.258, 0.355) [mean (95% C.I.)] and serum triglyceride by 0.164 mmol/liter (0.12, 0.209) although switching to TAC improved total cholesterol and LDL-C/HDL-C. Administration of frusemide, increased the total cholesterol and LDL-C/HDL-C whereas administration of bumetanide or metolazone increased the concentration of serum triglyceride. Serum glucose was associated with hypertriglyceridemia whereas serum creatinine was associated with increases in the total cholesterol, LDL-C/HDL-C and triglyceride. Conclusions. We have identified demographic and postoperative covariables that predispose heart transplant recipients to hyperlipidemia. Some of these risk factors, such as the effect of diuretics, have not been identified before in this group of patients and may be amenable to modification or closer control. TAC rather than CYA may be the immunosuppressive of choice for patients who are at greater risk of developing hyperlipidemia.

Structural and parameter uncertainty in Bayesian cost-effectiveness models

Health economic decision models are subject to various forms of uncertainty, including uncertainty about the parameters of the model and about the model structure. These uncertainties can be handled within a Bayesian framework, which also allows evidence from previous studies to be combined with the data. As an example, we consider a Markov model for assessing the cost-effectiveness of implantable cardioverter defibrillators. Using Markov chain Monte Carlo posterior simulation, uncertainty about the parameters of the model is formally incorporated in the estimates of expected cost and effectiveness. We extend these methods to include uncertainty about the choice between plausible model structures. This is accounted for by averaging the posterior distributions from the competing models using weights that are derived from the pseudo-marginal-likelihood and the deviance information criterion, which are measures of expected predictive utility. We also show how these cost-effectiveness calculations can be performed efficiently in the widely used software WinBUGS.

Cost-effectiveness of initial stress cardiovascular MR, stress SPECT or stress echocardiography as a gate-keeper test, compared with upfront invasive coronary angiography in the investigation and management of patients with stable chest pain: Mid-term outcomes from the CECaT randomised controlled trial

Objectives To compare outcomes and cost-effectiveness of various initial imaging strategies in the management of stable chest pain in a long-term prospective randomised trial. Setting Regional cardiothoracic referral centre in the east of England. Participants 898 patients (69% man) entered the study with 869 alive at 2 years of follow-up. Patients were included if they presented for assessment of stable chest pain with a positive exercise test and no prior history of ischaemic heart disease. Exclusion criteria were recent infarction, unstable symptoms or any contraindication to stress MRI. Primary outcome measures The primary outcomes of this follow-up study were survival up to a minimum of 2 years post-treatment, quality-adjusted survival and cost-utility of each strategy. Results 898 patients were randomised. Compared with angiography, mortality was marginally higher in the groups randomised to cardiac MR (HR 2.6, 95% CI 1.1 to 6.2), but similar in the single photon emission CT-methoxyisobutylisonitrile (SPECT-MIBI; HR 1.0, 95% CI 0.4 to 2.9) and ECHO groups (HR 1.6, 95% CI 0.6 to 4.0). Although SPECT-MIBI was marginally superior to other non-invasive tests there were no other significant differences between the groups in mortality, quality-adjusted survival or costs. Conclusions Non-invasive cardiac imaging can be used safely as the initial diagnostic test to diagnose coronary artery disease without adverse effects on patient outcomes or increased costs, relative to angiography. These results should be interpreted in the context of recent advances in imaging technology. Trial registration ISRCTN 47108462, UKCRN 3696.

Survival Models in Health Economic Evaluations: Balancing Fit and Parsimony to Improve Prediction

Health economic decision models compare costs and health effects of different interventions over the long term and usually incorporate survival data. Since survival is often extrapolated beyond the range of the data, inaccurate model specification can result in very different policy decisions. However, in this area, flexible survival models are rarely considered, and model uncertainty is rarely accounted for. In this article, various survival distributions are applied in a decision model for oral cancer screening. Flexible parametric models are compared with Bayesian semiparametric models, in which the baseline hazard can be made arbitrarily complex while still enabling survival to be extrapolated. A fully Bayesian framework is used for all models so that uncertainties can be easily incorporated in estimates of long-term costs and effects. The fit and predictive ability of both parametric and semiparametric models are compared using the deviance information criterion in order to account for model uncertainty in the cost-effectiveness analysis. Under the Bayesian semiparametric models, some smoothing of the hazard function is required to obtain adequate predictive ability and avoid sensitivity to the choice of prior. We determine that one flexible parametric survival model fits substantially better than the others considered in the oral cancer example.