Susanna Dodd

The impact of outcome reporting bias in randomised controlled trials on a cohort of systematic reviews

Objective To examine the prevalence of outcome reporting bias—the selection for publication of a subset of the original recorded outcome variables on the basis of the results—and its impact on Cochrane reviews. Design A nine point classification system for missing outcome data in randomised trials was developed and applied to the trials assessed in a large, unselected cohort of Cochrane systematic reviews. Researchers who conducted the trials were contacted and the reason sought for the non-reporting of data. A sensitivity analysis was undertaken to assess the impact of outcome reporting bias on reviews that included a single meta-analysis of the review primary outcome. Results More than half (157/283 (55%)) the reviews did not include full data for the review primary outcome of interest from all eligible trials. The median amount of review outcome data missing for any reason was 10%, whereas 50% or more of the potential data were missing in 70 (25%) reviews. It was clear from the publications for 155 (6%) of the 2486 assessable trials that the researchers had measured and analysed the review primary outcome but did not report or only partially reported the results. For reports that did not mention the review primary outcome, our classification regarding the presence of outcome reporting bias was shown to have a sensitivity of 88% (95% CI 65% to 100%) and specificity of 80% (95% CI 69% to 90%) on the basis of responses from 62 trialists. A third of Cochrane reviews (96/283 (34%)) contained at least one trial with high suspicion of outcome reporting bias for the review primary outcome. In a sensitivity analysis undertaken for 81 reviews with a single meta-analysis of the primary outcome of interest, the treatment effect estimate was reduced by 20% or more in 19 (23%). Of the 42 meta-analyses with a statistically significant result only, eight (19%) became non-significant after adjustment for outcome reporting bias and 11 (26%) would have overestimated the treatment effect by 20% or more. Conclusions Outcome reporting bias is an under-recognised problem that affects the conclusions in a substantial proportion of Cochrane reviews. Individuals conducting systematic reviews need to address explicitly the issue of missing outcome data for their review to be considered a reliable source of evidence. Extra care is required during data extraction, reviewers should identify when a trial reports that an outcome was measured but no results were reported or events observed, and contact with trialists should be encouraged.

Cost of illness of inflammatory bowel disease in the UK: a single centre retrospective study

Background and aims: The potentially high costs of care associated with inflammatory bowel disease (IBD) are recognised but we have little knowledge of the scale, profile, or determinants of these costs in the UK. This study aimed to describe costs of illness for a group of IBD patients and determine factors associated with increased healthcare costs. Setting: A university hospital serving a target population of approximately 330 000. Patients and methods: A six month cohort of IBD patients receiving any form of secondary care was identified, comprising 307 cases of ulcerative (or indeterminate) colitis and 172 cases of Crohn’s disease. Demographic and clinical data were abstracted from clinical records and individual resource use was itemised for all attributable costs (including extraintestinal manifestations). Item costs were derived from national and local sources. Cost data were expressed as mean six month costs per patient (with 95% confidence interval (CI)) obtained using non-parametric bootstrapping. Determinants of cost were analysed using generalised linear regression modelling. A postal survey of patients was undertaken to examine indirect costs, out of pocket expenses, and primary care visits. Results: Inpatient services (medical and/or surgical) were required by 67 patients (14%) but accounted for 49% of total secondary care costs. Drug costs accounted for less than a quarter of total costs. Individual patient costs ranged from £73 to £33 254 per six months. Mean (95% CI) six month costs per patient were £1256 (£988, £1721) for colitis and £1652 (£1221, £2239) for Crohn’s disease. Hospitalisation, disease severity grade, and disease extent correlated positively with cost of illness but costs were independent of age or sex. Compared with quiescent cases of IBD, disease relapse was associated with a 2–3-fold increase in costs for non-hospitalised cases and a 20-fold increase in costs for hospitalised cases. Survey data suggested average six month costs were <£30 per patient for primary care visits (both diseases) and median loss of earnings were £239 for colitis and £299 for Crohn’s disease. Conclusions: This study represents the first detailed characterisation of the scale and determinants of costs of illness for IBD in a British hospital. Hospitalisation affected a minority of sufferers but accounted for half of the total direct costs falling on the healthcare system.

Minimal access retroperitoneal pancreatic necrosectomy: improvement in morbidity and mortality with a less invasive approach.

Objective:Comparison of minimal access retroperitoneal pancreatic necrosectomy (MARPN) versus open necrosectomy in the treatment of infected or nonresolving pancreatic necrosis. Summary of Background Data:Infected pancreatic necrosis may lead to progressive organ failure and death. Minimal access techniques have been developed in an attempt to reduce the high mortality of open necrosectomy. Methods:This was a retrospective analysis on a prospective data base comprising 189 consecutive patients undergoing MARPN or open necrosectomy (August 1997 to September 2008). Outcome measures included total and postoperative ICU and hospital stays, organ dysfunction, complications and mortality using an intention to treat analysis. Results:Overall 137 patients underwent MARPN versus open necrosectomy in 52. Median (range) age of the patients was 57.5 (18–85) years; 118 (62%) were male. A total of 131 (69%) patients were tertiary referrals, with a median time to transfer from index hospital of 19 (2–76) days. Etiology was gallstones or alcohol in 129 cases (68%); 98 of 168 (58%) patients had a positive culture at the first procedure. Of the 137 patients, 34 (31%) had postoperative organ failure in the MARPN group, and 39 of 52 (56%) in the open group (P < 0.0001); 59/137 (43%) versus 40/52 (77%), respectively, required postoperative ICU support (P < 0.0001). Of the 137 patients 75 (55%) had complications in the MARPN group and 42 of 52 (81%) in the open group (P = 0.001). There were 26 (19%) deaths in the MARPN group and 20 (38%) following open procedure (P = 0.009). Age (P < 0.0001), preoperative multiorgan failure (P < 0.0001), and surgical procedure (MARPN, P = 0.016) were independent predictors of mortality. Conclusion:This study has shown significant benefits for a minimal access approach including fewer complications and deaths compared with open necrosectomy.

Alemtuzumab in Combination With Methylprednisolone Is a Highly Effective Induction Regimen for Patients With Chronic Lymphocytic Leukemia and Deletion of TP53: Final Results of the National Cancer Research Institute CLL206 Trial

PURPOSE In chronic lymphocytic leukemia (CLL), TP53 deletion/mutation is strongly associated with an adverse outcome and resistance to chemotherapy-based treatment. In contrast, TP53 defects are not associated with resistance to the anti-CD52 monoclonal antibody alemtuzumab or methylprednisolone. In an attempt to improve the treatment of TP53-defective CLL, a multicenter phase II study was developed to evaluate alemtuzumab and methylprednisolone in combination. PATIENTS AND METHODS Thirty-nine patients with TP53-deleted CLL (17 untreated and 22 previously treated) received up to 16 weeks of treatment with alemtuzumab 30 mg three times a week and methylprednisolone 1.0 g/m(2) for five consecutive days every 4 weeks. Antimicrobial prophylaxis consisted of cotrimoxazole, itraconazole, and aciclovir (or valganciclovir for asymptomatic cytomegalovirus viremia). The primary end point was response as assigned by an end-point review committee. Secondary end points were safety, progression-free survival (PFS) and overall survival (OS). RESULTS The overall response rate, complete response rate (including with incomplete marrow recovery), median PFS, and median OS were 85%, 36%, 11.8 months, and 23.5 months, respectively, in the entire cohort and 88%, 65%, 18.3 months, and 38.9 months, respectively, in previously untreated patients. Grade 3 to 4 hematologic and glucocorticoid-associated toxicity occurred in 67% and 23% of patients, respectively. Grade 3 to 4 infection occurred in 51% of the overall cohort and in 29% of patients less than 60 years of age. Treatment-related mortality was 5%. CONCLUSION Alemtuzumab plus methypredisolone is the most effective induction regimen hitherto reported in TP53-deleted CLL. The risk of infection is age related and, in younger patients, seems only marginally higher than that associated with rituximab, fludarabine, and cyclophosphamide.

Standardized antibacterial honey (Medihoney™) with standard therapy in wound care: randomized clinical trial

AIM This paper is a report of a study to compare a medical grade honey with conventional treatments on the healing rates of wounds healing by secondary intention. BACKGROUND There is an increasing body of evidence to support the use of honey to treat wounds, but there is a lack of robust randomized trials on which clinicians can base their clinical judgement. METHOD A sample of 105 patients were involved in a single centre, open-label randomized controlled trial in which patients received either a conventional wound dressing or honey. Data were collected between September 2004 and May 2007. RESULTS The median time to healing in the honey group was 100 days compared with 140 days in the control group. The healing rate at 12 weeks was equal to 46.2% in the honey group compared with 34.0% in the conventional group, and the difference in the healing rates (95% confidence interval, CI) at 12 weeks between the two groups was 12.2% (-13.6%, 37.9%). The unadjusted hazard ratio (95% CI) from a Cox regression was equal to 1.30 (0.77, 2.19), P = 0.321. When the treatment effect was adjusted for confounding factors (sex, wound type, age and wound area at start of treatment), the hazard ratio increased to 1.51 but was again not statistically significant. CONCLUSION Wound area at start of treatment and sex are both highly statistically significant predictors of time to healing. These results support the proposition that there are clinical benefits from using honey in wound care, but further research is needed.

Trends in deaths from respiratory illness in children in England and Wales from 1968 to 2000

Background: Childhood mortality has decreased markedly over the last three decades. A study was undertaken to determine trends in deaths from respiratory illness in children in England and Wales. Methods: Mortality data collected by the Office for National Statistics were analysed. The data included all deaths registered from all causes in children aged between 28 days and 16 years in England and Wales from 1 January 1968 to 31 December 2000. The main outcome measures were overall and age-specific mortality rates due to all respiratory disorders and specific rates for pneumonia, asthma, cystic fibrosis (CF), and bronchiolitis. Results: In children aged 1–16 years the overall mortality rate (per 100 000 children) declined from 49.9 in 1968 to 16.3 in 2000, and rates due to respiratory illness fell from 8.6 to 1.3. The proportion of all deaths caused by respiratory illness in children aged 28 days to 16 years fell from 30.8% in 1968 to 9.9% in 2000. In post-neonatal infants (aged 28–364 days), the “all cause” mortality rate fell from 592.8 in 1968 to 176 in 2000 and the rates due to respiratory illness fell from 280 to 22.8. In 2000, pneumonia, asthma and CF together accounted for 73% of all respiratory deaths in 1–16 year olds. In this age group, mortality rates per 100 000 for pneumonia fell from 4.22 to 0.57, for asthma from 0.83 to 0.25, and for CF from 0.66 to 0.12 between 1968 and 2000. Over the same period mortality rates for pneumonia in post-neonatal infants fell from 165 to 6.78 per 100 000 and for CF from 4.88 to 0.33. Bronchiolitis mortality rates per 100 000 in post-neonatal infants fell from 21.47 in 1979 to 1.82 in 2000. Conclusions: Mortality rates due to all respiratory illnesses in children have fallen markedly in the last three decades. This decline has been more rapid than the overall decline in childhood mortality and respiratory diseases are now responsible for a smaller proportion of deaths in children. These data could provide a foundation for assessing the impact on mortality of future health initiatives such as the introduction of a universal pneumococcal vaccination programme in England and Wales.

Confounding effect of obstructive jaundice in the interpretation of proteomic plasma profiling data for pancreatic cancer.

It is well established that variation in sampling, processing and storage protocols can alter the levels of potential biomarkers in serum and plasma. Here, using pancreatic cancer as an example, we demonstrate that consideration of clinical parameters related to the patients illness is equally important when seeking cancer-specific biomarkers. Bile duct-obstruction is a feature of pancreatic disease that can cause jaundice. Comparing patients with pancreatic cancer, chronic pancreatitis or biliary duct obstruction, we observed that the plasma levels of apolipoprotein A1, transthyretin, and apolipoprotein E, when examined in isolation, were each associated with pancreatic cancer. However, when the effect of bile duct obstruction was considered, only transthyretin levels were independently associated with cancer likelihood. Our results demonstrate the importance of accounting for disease-related confounding factors when analyzing data for the detection of cancer biomarkers.

Nonadherence to treatment protocol in published randomised controlled trials: a review

This review aimed to ascertain the extent to which nonadherence to treatment protocol is reported and addressed in a cohort of published analyses of randomised controlled trials (RCTs). One hundred publications of RCTs, randomly selected from those published in BMJ, New England Journal of Medicine, the Journal of the American Medical Association and The Lancet during 2008, were reviewed to determine the extent and nature of reported nonadherence to treatment protocol, and whether statistical methods were used to examine the effect of such nonadherence on both benefit and harms analyses. We also assessed the quality of trial reporting of treatment protocol nonadherence and the quality of reporting of the statistical analysis methods used to investigate such nonadherence. Nonadherence to treatment protocol was reported in 98 of the 100 trials, but reporting on such nonadherence was often vague or incomplete. Forty-two publications did not state how many participants started their randomised treatment. Reporting of treatment initiation and completeness was judged to be inadequate in 64% of trials with short-term interventions and 89% of trials with long-term interventions. More than half (51) of the 98 trials with treatment protocol nonadherence implemented some statistical method to address this issue, most commonly based on per protocol analysis (46) but often labelled as intention to treat (ITT) or modified ITT (23 analyses in 22 trials). The composition of analysis sets for their benefit outcomes were not explained in 57% of trials, and 62% of trials that presented harms analyses did not define harms analysis populations. The majority of defined harms analysis populations (18 out of 26 trials, 69%) were based on actual treatment received, while the majority of trials with undefined harms analysis populations (31 out of 43 trials, 72%) appeared to analyse harms using the ITT approach. Adherence to randomised intervention is poorly considered in the reporting and analysis of published RCTs. The majority of trials are subject to various forms of nonadherence to treatment protocol, and though trialists deal with this nonadherence using a variety of statistical methods and analysis populations, they rarely consider the potential for bias introduced. There is a need for increased awareness of more appropriate causal methods to adjust for departures from treatment protocol, as well as guidance on the appropriate analysis population to use for harms outcomes in the presence of such nonadherence.

The variable phenotype of the p.A16V mutation of cationic trypsinogen (PRSS1) in pancreatitis families

Objective To characterise the phenotypes associated with the p.A16V mutation of PRSS1. Design Clinical and epidemiological data were collected for any family in which a p.A16V mutation was identified, either referred directly to the European Registry of Hereditary Pancreatitis and Familial Pancreatic Cancer or via a collaborator. DNA samples were tested for mutations in PRSS1, SPINK1, CFTR and CTRC. Patients Participants were recruited on the basis of either family history of pancreatitis (acute or chronic) or the results of genetic testing. Families were categorised as having hereditary pancreatitis (HP), idiopathic disease or pancreatitis in a single generation. HP was defined as ≥2 cases in ≥2 generations. Main outcome measures Onset of painful episodes of pancreatitis, death from pancreatic cancer, diagnosis of diabetes mellitus and exocrine pancreatic failure. Results Ten families with p.A16V mutations were identified (22 affected individuals): six HP families, three with idiopathic disease and one with only a single generation affected. The median age of onset, ignoring non-penetrants, was 10 years (95% CI 5 to 25). There were eight confirmed cases of exocrine failure, four of whom also had diabetes mellitus. There were three pancreatic cancer cases. Two of these were confirmed as p.A16V carriers, only one of whom was affected by pancreatitis. Those with p.A16V pancreatitis were compared to affected individuals with p.R122H, p.N29I and no PRSS1 mutation. No significant differences were proven using logrank or Mann–Whitney U tests. Conclusions Penetrance of p.A16V is highly variable and family dependent, suggesting it contributes to multigenic inheritance of a predisposition to pancreatitis.

Expression of survivin, platelet-derived growth factor A (PDGF-A) and PDGF receptor α in primary central nervous system lymphoma

PurposePrimary central nervous system lymphomas (PCNSL) are rare tumours occurring in the brain. Their biology and the factors predicting survival are not well known. This study investigated expression of the antiapoptotic protein survivin and platelet-derived growth factor A (PDGF-A) and receptor (PDGFRα) in PCNSL.Experimental designA total of 44 patients with histologically confirmed PCNSL treated between 1992 and 2004 were included in this study, and tumour specimens were investigated immunohistochemically for expression of survivin, PDGF-A and PDGFRα. Protein expression and clinical variables were analyzed statistically.ResultsOf the 44 tumours 43(98%) were diffuse large B-cell non-Hodgkin’s lymphomas (NHL) and one was a T-cell NHL. Around 37 (84%) of the examined PCNSL specimens showed expression of survivin, 16 (36%) of PDGF-A and 34 (77%) of PDGFRα. Tumours expressing surviving co-expressed PDGFRα frequently and PDGF-A occasionally. Expression of the above proteins was not predictive for survival in this patient group. Except for age and therapy, no other clinical variables correlated significantly with overall survival.ConclusionsPCNSL express survivin and PDGFRα in the majority of investigated cases. PDGF-A is expressed less frequently. Immunohistochemical detection of these proteins does not correlate with overall survival and cannot be used as a prognostic factor.