To view this email online, @link@.
NICE - National Institute for Health and Care Excellence
Eyes on Evidence
expert commentary on important new evidence
Visit Evidence Search Contact Us Previous Issues Share: Facebook Twitter Forward
Issue 55 – November 2013

A meta-analysis of randomised controlled trials in breast cancer reports that a third are biased in their reporting of treatment efficacy and two-thirds do not report toxicity appropriately.

Two systematic reviews show little evidence to support the choice of intervention to prevent and treat pressure ulcers, whereas a third suggests that bundles of interventions may be effective.

A randomised controlled trial finds that intermittent pneumatic compression reduces the risk of deep vein thrombosis in immobile inpatients who have been admitted to hospital after a stroke.

A systematic review and meta-analysis of comparative epidemiology studies finds that passive smoking more than doubles the risk of invasive meningococcal disease in children and young people.

A pragmatic randomised trial indicates that laparoscopic fundoplication for gastro-oesophageal reflux disease provides better quality of life and symptom relief than long-term drug treatment.

We would like your examples of how health and social care staff are helping to improve quality and productivity.

NICE has recently published Evidence Updates on:
  • Urinary tract infection in children
Bias in reporting of randomised controlled trials in breast cancer
Overview: Randomised controlled trials can be biased by several possible factors, which can reduce the credibility of trial results. Selection bias is when the groups of patients selected to take part in studies have different characteristics that could affect the treatment outcomes. Performance bias may result from differences in the care provided to each group. Detection bias may occur if outcomes are determined differently between groups. Attrition bias results from differences in how many people in each group withdraw from the study. Reporting bias is selective reporting of results.

Current advice: The Consolidated Standards of Reporting Trials (CONSORT) statement is an evidence-based set of recommendations for reporting of randomised controlled trials. It aims to help authors to report results completely and transparently. Many medical journals endorse the CONSORT statement and should, therefore, publish only randomised controlled trials that report their results in accordance with the standards.

New evidence: Vera-Badillo et al. (2013) conducted an analysis of 164 randomised controlled trials in breast cancer to assess bias and ‘spin’ in the reporting of efficacy and toxicity end points. Only phase III trials with 200 participants or more and a primary end point measuring time to the end point event were included. For recent trials, the reported primary end point was checked against that listed in the trials registry

Bias was defined as inappropriate reporting of the primary end point and toxicity. The authors placed emphasis on the reporting in study abstracts, because they noted that busy clinicians often read only the abstracts of publications. Studies that specified multiple primary end points in which at least one end point was significant were not assessed for bias. Spin was defined as the reporting of trials’ secondary end points as evidence that a treatment is beneficial when the primary end point did not show a statistically significant result. 

About half of included trials (49.4%) were of adjuvant treatments, and a similar proportion (50.6%) were in women with metastatic breast cancer. Most trials (83.5%) used disease-free or progression-free survival as the primary end point; the remainder (16.5%) used overall survival. Only 30 trials (18%) were registered on, and of these, 7 (23.3%) reported a different primary end point than was registered. More than half of trials (56.1%) had a negative result, with no significant difference between groups in the primary end point.

Around a third of trials (32.9%) were reported in a positive manner on the basis of non-primary end points. Studies with a negative primary outcome were more likely to omit the primary end point from the abstract than were studies with a positive primary end point (OR=5.15, 95% CI 1.86 to 14.26, p=0.001). Two-thirds of studies (67.1%) were rated as having bias in reporting of adverse events. Industry funding was not associated with bias in the reporting of primary end points or toxicity.

Commentary: “Decision-making for women with breast cancer and their clinicians is difficult. Many practical and personal aspects need factoring in. Coping with treatment toxicities, their after effects, and everything that will impinge on living life well thereafter will depend on patients and doctors being confident about the quality, reliability and completeness of the information available to share in the consultation. 

“This information may be gleaned from many sources, all of which will be dependent on data having been fully, promptly and accurately reported in trial reports and systematic reviews. The facts and findings communicated through the whole chain of reporting depend entirely on honest, objective, impartial and complete reporting by the initial investigators, as free as possible of spin and biases. Without trust in their proper accounting we are lost. Proper motivation is required of everyone involved. 

“Researchers must remember always that the purpose of their joint endeavours is to faithfully reduce treatment uncertainties, which requires that they report all potential benefits and harms in full and balanced measure. The Equator Network provides meticulously devised checklists that can be used to meet every practical and ethical obligation of recording and reporting research information. Unbiased research reporting is an essential factor that helps clinicians and patients to make good treatment choices together.” – Hazel Thornton, Honorary Visiting Fellow, Department of Health Sciences, University of Leicester

Study sponsorship: Not reported.

Prevention and treatment of pressure ulcers
Overview: Between 1.1% and 1.3% of patients treated in hospital and community settings in England develop pressure ulcers during their care (NHS Safety Thermometer Report). Pressure ulcers are estimated to cost the health service £1.4 to £2.1 billion a year, equivalent to 4% of annual NHS expenditure (Bennett et al. 2004). Some groups of people are particularly vulnerable to pressure ulcers, such as seriously ill patients, people with spinal cord injuries, older people who are immobile, patients who are obese, and patients with conditions that affect blood flow, such as diabetes. 

Pressure ulcer is one of four conditions targeted by the NHS 'Harm free' care programme. This programme aims to help NHS organisations build quality improvement initiatives to eliminate pressure ulcers, falls, urinary tract infections in patients with a catheter, and new venous thromboembolism. 

See the NICE Evidence Services topic page on pressure ulcer for a general overview of this condition.

Current advice: NICE guidance on pressure ulcer prevention (currently being updated) recommends that all people at risk of pressure ulcers should, as a minimum provision, be placed on a high specification foam mattress with pressure-relieving properties.

NICE guidance on management of pressure ulcers (also being updated) notes that there is no conclusive evidence that any one pressure-relieving support technology is superior to another. 

Recommendations based on professional consensus include:
  • People with a grade 1–2 pressure ulcer should, as a minimum provision, be placed on a high specification foam mattress or cushion with pressure-reducing properties, combined with very close observation of skin changes and a documented positioning and repositioning regimen.
  • People with grade 3–4 pressure ulcers should, as a minimum provision, be placed on an alternating pressure mattress (replacement or overlay) or sophisticated continuous low pressure system.
NICE also recommends the use of modern dressings, such as hydrocolloids and foams, in preference to basic dressings, such as gauze, paraffin gauze and simple dressing pads. 

The NICE Pathway on pressure ulcer management brings together all related NICE guidance and associated products on the condition in a set of interactive topic-based diagrams.

New evidence: Three systematic reviews have looked at the risk assessment, prevention and treatment of pressure ulcers in adults in acute and long-term care. 

Chou et al. (2013) reviewed 67 randomised controlled trials and cohort studies looking at the effects of risk assessment tools and preventive strategies on the incidence and severity of pressure ulcers. Meta-analysis of the data in the studies identified was not possible because of the methodological limitations and clinical heterogeneity of the studies. Nevertheless, the authors concluded that risk assessment tools, such as the Waterlow scale, and most types of support surfaces, except for advanced static mattresses, had no effect on prevention of pressure ulcers. Likewise, nutritional support, repositioning, and dressings and pads had no clear beneficial effects, although a couple of trials reported some reduction in incident pressure ulcers with cleansers and fatty acid creams.

Smith et al. (2013) reviewed 174 studies of treatment strategies for adults with pressure ulcers, including randomised controlled trials, comparative observational studies and non-comparative series studies. Meta-analysis was done for the outcome of complete wound healing. Support surfaces, nutrition support, local wound applications and adjunctive therapies had no effect on complete wound healing compared with standard care, placebo or sham interventions. Some moderate-strength evidence suggested that air-fluidised beds, protein-containing nutritional supplements, electrical stimulation and application of radiant heat helped to reduce wound size. 

Sullivan and Schoelles' 2013 review of 26 studies focused on the factors associated with successful implementation of multicomponent preventive strategies ('skin bundles') for pressure ulcers. Common features of effective programmes were: education and training of healthcare staff; revision of protocols for assessing and documenting ulcers; audit and feedback; redesigning documentation and reporting processes; and use of risk prediction scores.

All three reviews highlighted the weakness of the evidence base, which comprises mostly small studies with considerable methodological limitations, and called for more high-quality studies on preventive and treatment interventions for pressure ulcers.

Commentary: "The conclusions of the high quality reviews by Chou et al. (2013) and Smith et al. (2013) will be disappointing for health professionals. Although it is already widely appreciated that advanced static support surfaces are better at preventing pressure ulcers than standard mattresses, it may surprise some that the Chou et al. (2013) review found no evidence in support of pressure ulcer risk scoring over clinical judgement. The positive conclusions of the Smith et al. (2013) treatment review are largely based on clinically irrelevant comparisons (for example, comparing air fluidised beds with standard mattresses) or surrogate outcomes, such as change in ulcer area, rather than time to complete healing. 

"By contrast, the Sullivan and Schoelles (2013) review concluded that bundling prevention interventions together in multicomponent programmes reduces pressure ulceration – a surprising finding given the absence of effects detected for the single interventions studied by Chou et al. (2013) There are several possible explanations for the contrasting findings. One is that bundling several interventions together and raising awareness of pressure ulcer risk, together with education, audit and feedback, is much more effective than single interventions. Another explanation is that the Sullivan and Schoelles review took a more relaxed view of the risk of bias in the underlying studies, so the conclusions may be misleading. 

"That said, identification of patients at risk, use of appropriate and effective support surfaces, and, possibly most importantly, regular inspection of skin condition with appropriate modification of prevention strategies, are essential components of best practice that should be employed anywhere that patients are at risk of pressure ulcers." – Professor Dame Nicky Cullum, Professor of Nursing, University of Manchester

Study sponsorship: US Agency for Healthcare Research and Quality.

Intermittent pneumatic compression for deep vein thrombosis in patients with stroke
Overview: Patients who have been admitted to hospital with a medical illness and have significantly reduced mobility are at increased risk of developing deep vein thrombosis (DVT). The risk is particularly high in immobile patients admitted to hospital with a stroke: incidence of DVT is 50% in these patients compared with 13% in general medical patients.

People with stroke may be at high risk of bleeding (Whiteley et al. 2013), so anticoagulants, such as low molecular weight heparin, are not always suitable for these patients. Mechanical prophylaxis, such as intermittent pneumatic compression (IPC), is effective at preventing DVT in patients who have undergone surgery (Roderick et al. 2005), but the benefits of this approach in patients with stroke are not clear.

See the NICE Evidence Services topic page on venous thromboembolism for a general overview of this condition.

Current advice: NICE guidance on reducing the risk of venous thromboembolism in patients admitted to hospital recommends that adults admitted for stroke should be offered a foot impulse or IPC device until start of pharmacological prophylaxis with low molecular weight heparin (unfractionated heparin in patients with renal failure). However, pharmacological prophylaxis can be given to patients with stroke only if their risk of bleeding is assessed to be low and they meet one or more of the following criteria: major restriction of mobility; previous history of venous thromboembolism; dehydration; and comorbidities (such as malignant disease). 

The NICE Pathway on venous thromboembolism brings together all related NICE guidance and associated products on the condition in a set of interactive topic-based diagrams.

New evidence: In this third study from the CLOTS Trials Collaboration (2013), patients who were immobile and had been admitted to hospital within 3 days of acute stroke were randomly assigned to receive or not receive IPC. Compression was administered continuously day and night to both legs for a minimum of 30 days from randomisation or until the patient regained mobility, was discharged or died. To detect DVT, compression duplex ultrasound was performed at 7–10 days by a technician blinded to treatment allocation. If practical, compression duplex ultrasound was performed again at 25–30 days. Patients were followed up at 6 months to determine survival and late symptomatic venous thromboembolism. 

Patients (n=2876) at 94 centres in the UK were randomised to IPC or no compression (n=1438 in each group, median age 76 years). IPC was administered for a mean of 12.5 days, with only around a third (31%) of patients receiving compression for 30 days. 

At 30 days' follow-up, patients allocated to IPC were less likely to have symptomatic or asymptomatic DVT in the proximal veins than were patients not in the compression group (adjusted OR=0.65, 95% CI 0.51 to 0.84, p=0.001). Patients in the IPC group were also less likely to have any DVT at 6 months' follow-up (adjusted OR=0.72, 95% CI 0.60 to 0.87, p=0.001). Fewer patients in the IPC group than in the no compression group had died at 30 days (11% vs 13%, p=0.057) or at 6 months (22% vs 25%, p=0.059). Skin breaks on the legs occurred more frequently in the IPC group than in the no compression group (3% vs 1%, p=0.002). 

The authors concluded that IPC reduces the risk of DVT and is a safe prophylactic technique in patients with stroke. They acknowledged that all patients and investigators were aware of treatment allocation and adherence was low, which could bias their findings.

Commentary: "Patients with acute stroke are at risk of complications of immobility, such as DVT, and it is the prevention and treatment of such complications that potentially accounts for the effectiveness of stroke units. However, no study has thus far clearly demonstrated the independent benefits of a single specific intervention. 

"The previous CLOTS studies have benefited clinical practice by allowing removal of the routine use of graduated compression stockings for DVT, which were shown to be ineffective at preventing thrombosis (CLOTS Trials Collaboration 2009 and 2010). This measure has delivered cost savings, not only from the price of stockings but also in terms of freeing up the time of healthcare staff (for example, from taking stockings on and off patients). Pharmacological treatments are not effective in practice because the patients who would be at risk of DVT, and, therefore, recommended to be on anticoagulant drugs, cannot take anticoagulants because they are also at high risk of having bleeding complications.

"The benefits of IPC demonstrated in the third CLOTS trial bring to the fore an alternative approach to DVT prevention. However, implementation of the findings will be a challenge because the equipment is expensive and not readily available in the NHS. In addition, the intervention is not favoured by patients or staff, and the concomitant lack of adherence was clear in this study because the suggested treatment period of 30 days was not achieved in the majority of patients." – Caroline Watkins, Professor of Stroke & Older People's Care and Director of Research, School of Health, University of Central Lancashire

Study sponsorship: National Institute of Health Research Health Technology Assessment programme, Chief Scientist Office of the Scottish Government and Covidien (MA, USA).

Passive smoking and meningococcal disease in children and young people
Overview: Bacterial meningitis is the leading infectious cause of death in early childhood, making its control a priority for clinical management and public health. It usually presents as septicaemia, meningitis or both. 

In the UK, 2–6 cases of bacterial meningitis occur a year per 100,000 people, with the case-fatality rate about 10%. The incidence is highest in infants around 6 months, lower in those aged 4 years, and peaks again in teenagers aged 17–18 years (Jones and Mallard 1993). The incidence is also highest in the winter months. Apart from age, risk factors include passive smoking, preceding influenza A infection and overcrowding.

See the NICE Evidence Services topic page on meningitis for a general overview of this condition.

Current advice: NICE has guidance on the management of bacterial meningitis and meningococcal septicaemia in children and young people under 16 years. In addition to treating the disease, clinicians have a legal requirement to notify a proper officer of the local authority urgently on suspicion of meningitis or meningococcal septicaemia. The Health Protection Agency (now part of Public Health England) has issued guidance for public health management of meningococcal disease in the UK

The NICE pathway on bacterial meningitis and meningococcal septicaemia brings together all related NICE guidance and associated products on the condition in a set of interactive topic-based diagrams.

New evidence: Murray et al. (2012) did a systematic review and meta-analysis of comparative epidemiological studies to investigate the association between passive smoking and invasive meningococcal disease in children. Fifteen case-control studies (n=14,862) and 2 cohort studies (n=283,579) were included assessing passive smoking in children and young people aged under 18 years. Passive smoking was measured by self-report or biochemical markers, and the outcome of meningococcal disease was diagnosed clinically, confirmed by laboratory tests or both.

Pooled analysis of all 17 studies showed that exposure to second-hand smoke by any smoker in the household more than doubled the risk of invasive meningococcal disease (odds ratio [OR]=2.18, 95% confidence interval [CI] 1.63 to 2.92, p<0.00001). Maternal smoking during pregnancy and after birth were similarly associated with a doubling of the risk of invasive meningococcal disease, but studies of paternal smoking were too different to allow pooling, and results of individual studies were conflicting. After accounting for a small publication bias, the authors calculated that 350 cases of invasive meningococcal disease per year in the UK could arise from exposure to smoking in the home.

Commentary: "This study shows clear evidence that passive smoking doubles the risk of invasive meningococcal disease, with an estimated 350 cases a year in children under the age of 16 years in the UK attributable to it. This is a significant public health issue, and the information from this paper should be used to formulate public health messages. 

"This study has distinct findings about the risk of meningococcal infection associated with smoking in pregnancy, which add to the already clear literature that smoking in pregnancy makes the infant more susceptible to serious respiratory infections. There is no doubt that people working to prevent meningococcal disease, together with those working with mothers and young families, should be thinking of new ways to target information to smokers to reduce the likelihood of meningococcal infection in young children. 

"Smoking is linked to lower socioeconomic status, as shown in a review by Hiscock and colleagues (2012). Smoking is more common in more disadvantaged groups, and this study has conclusively shown that second hand smoke caused by anyone in the household is linked to a doubling of the risks of developing invasive meningococcal disease. Failure to reduce smoking contributes further to health inequalities and the adverse health experienced by disadvantaged groups." – Philip Monk, Consultant in Public Health, Public Health England

Study sponsorship: The University of Nottingham, Cancer Research UK and the UK Clinical Research Collaboration.

Minimally invasive surgery for gastro-oesophageal reflux disease
Overview: Gastro-oesophageal reflux disease is a common cause of dyspepsia in which stomach acid leaks out of the stomach via the lower oesophageal sphincter and into the oesophagus. In some patients, gastro-oesophageal reflux disease can irritate the lining of the oesophagus and cause mucosal breaks, known as oesophagitis. Endoscopy is used to diagnose oesophagitis caused by reflux, but 40–60% of patients with symptoms have no signs of reflux on endoscopy (endoscopy-negative reflux disease).

In England, 73,135 patients admitted to hospital in 2011–12 were diagnosed with gastro-oesophageal reflux disease, 34,235 of whom had oesophagitis. Gastro-oesophageal reflux disease is more common in older people, women and people who are obese. 

Current advice: NICE guidance on managing dyspepsia in adults in primary care (currently being updated) states that surgery is not recommended for the routine management of persistent gastro-oesophageal reflux disease. Individual patients whose quality of life remains significantly impaired may, however, value this form of treatment.

Instead, people with gastro-oesophageal reflux disease should undergo diagnostic endoscopy and should initially be offered a full dose proton pump inhibitor for 1 or 2 months. If symptoms recur following initial treatment, proton pump inhibitors should be continued at the lowest dose possible to control symptoms, with a limited number of repeat prescriptions. An H2 receptor antagonist or prokinetic therapy should be offered if response to a proton pump inhibitor is inadequate. 

New evidence: Grant et al. (2013) conducted a pragmatic randomised trial to assess quality of life and symptom relief after laparoscopic fundoplication or medical management in people with gastro-oesophageal reflux disease. Participants who had been on drug treatment for more than 12 months and who were eligible for both surgery and continued medical management were randomly allocated to either treatment strategy. Those patients who had a strong preference for either surgery or medical management were invited to join one of two non-randomised preference arms. Surgery was full or partial laparoscopic fundoplication at the discretion of the surgeon, whereas medical management entailed treatment review and adjustment by a local gastroenterologist. 

Between March 2001 and June 2004, 357 patients at 21 UK centres were randomised to surgery (n=178) or medical management (n=179). A further 453 patients were recruited to the non-randomised preference groups for surgery (n=261) and medical management (n=192). Around two-thirds of patients (69.9% overall) completed follow-up postal questionnaires at 5 years.

At 5 years after enrolment, 334 (76%) patients in the two surgery groups and 30 (8%) patients in the two medical management arms had undergone fundoplication. People randomised to, or with a preference for, surgery had significantly better scores on the 100-point REFLUX questionnaire – a measure of quality of life, gastrointestinal symptoms and side effects of treatment – than patients in the two medical management groups (mean difference 8.3, 95% confidence interval 3.2 to 13.4, p=0.001). Patients in the surgery groups also less frequently reported individual symptoms, such as heartburn and dysphagia, and scored better on the Short Form (SF)-36 and EuroQol 5 dimension (EQ-5D) heath status measures. A lower proportion of patients randomised to surgery were taking anti-reflux drugs, such as omeprazole and lansoprazole, at 5 years than were patients randomised to medical management (56 of 127 [44.1%] vs 98 of 119 [82.4%]).

The authors concluded that laparoscopic fundoplication provides better long-term symptom relief and health-related quality of life than medical management in patients with gastro-oesophageal reflux disease. They suggested that their intention-to-treat analysis may have underestimated the benefits of fundoplication because of the low number of patients randomised to surgery who were actually operated on (63%) and the significantly lower baseline REFLUX scores in the 24 (13%) participants randomised to medical management who subsequently had surgery.

Commentary: "Reflux often responds well to acid suppression medication, but because the disorder is so common, the small proportion of unresponsive cases is a burden on primary care. Accurately diagnosing the problem (to ensure that it really is reflux) and then fixing the problem physically allows patients to have better quality of life and come off medication, reducing the burden of their care. 

"The multicentre nature of this study means that the results provide a fair snapshot of the standards and outcomes of anti-reflux surgery in UK patients with long-term reflux symptoms inadequately managed by medication. The inclusion of the preference groups allows the results to be extrapolated widely, because patients with gastro-oesophageal reflux disease do commonly come to consultations with treatment preferences. However, it is a pity that the trial failed to completely randomise participants, because full randomisation is possible (and has been achieved elsewhere) and would have produced true unbiased results.

"Anti-reflux surgery is not universally available in the UK. In addition, it is not always successful and in a small percentage of cases leaves patients with troublesome side effects. The provision of this surgery needs to increase in the UK, and any expansion needs to take place in recognised centres with good outcomes. New surgical developments to reduce side effects may help to improve outcomes, and specialist centres should be involved in the process of quality improvement." – Professor Stephen Attwood, Professor of Surgery at the Department of Health Services Research, Durham University and Consultant Upper Gastrointestinal Surgeon, Northumbria Healthcare NHS Foundation Trust

Study sponsorship: National Institute for Health Research Health Technology Assessment Programme.

QIPP case studies wanted
We would like your examples of how health and social care staff are helping to improve quality and productivity. If you have a good initiative that you are currently implementing or have implemented already, we want to hear from you. Our analysts will be happy to talk to you and provide advice on the suitability of your initiative for the collection and the type and level of information required. Case studies are evaluated on the basis of quality improvements, savings, evidence and implementation and are peer reviewed. 

Submit your QIPP initiative to the QIPP collection

Evidence Updates
NICE has recently published an Evidence Update on:
  • Urinary tract infection in children
This Evidence Update highlights and provides commentary on selected new evidence published since the NICE guidance was issued. For each topic, the evidence was considered by an Evidence Update Advisory Group (EUAG), a panel of experts, most of whom were involved in developing the original NICE guidance. 

The Evidence Update on urinary tract infection in children was published by NICE in October 2013. It includes commentary from the EUAG on 19 new articles (relevant to NICE clinical guideline 54), covering the following topics:
  • Risk factors for urinary tract infection, methods for urine collection, and laboratory tests for C-reactive protein.
  • Antibiotic treatment and prophylaxis, use of methylprednisolone for renal scarring, and the effects of cranberries.
  • Ultrasound, dimercaptosuccinic acid scintigraphy scanning, imaging algorithms detection of vesicoureteric reflux, and imaging and management of atypical urinary tract infections. 
  • Surgical management of vesicoureteric reflux.
  • Follow-up of children who have had a urinary tract infection.
Eyes on Evidence helps contextualise important new evidence, highlighting areas that could signal
a change in clinical practice. It does not constitute formal NICE guidance. The commentaries
included are the opinions of contributors and do not necessarily reflect the views of NICE.
Visit Evidence Search
Copyright © 2013 National Institute for Health and Care Excellence. All Rights Reserved.
This message was sent to by NICE using NewZapp Email Marketing. Follow this link to Unsubscribe.