Article Text

Download PDFPDF

DTB Select: 12 | December 2016


Azithromycin of no benefit for asthma attacks ● Risk of heart failure with NSAIDs ● Bisphosphonates and steroid-induced osteoporosis ● Triple therapy for moderate-to-severe COPD: marginal benefits ● Much data on adverse events remains unpublished ● Rising placebo response rates in antidepressant trials a ‘myth’ ● Prenatal fetal RhD testing recommended

View Full Text

Statistics from

Azithromycin of no benefit for asthma attacks

A randomised controlled trial that assessed the efficacy of azithromycin, a macrolide antibiotic, in the management of acute exacerbations of asthma has found that adding it to usual care resulted in no statistically or clinically significant benefit.1 As part of the rationale for the study, the authors cite an earlier trial of a different macrolide (telithromycin), which reported a small difference in symptoms scores (0.3 points on a 7-point scale) compared with placebo.2 However, adverse reactions limited telithromycin's potential for use.

The multicentre trial of azithromycin included 18–55-year-olds with any smoking history, 56–65-year-olds with a smoking history of less than 20 pack-years and those aged over 65 years with a smoking history of less than 5 pack-years. Participants had a history of asthma for more than 6 months and were recruited within 48 hours of presentation to medical care with an acute deterioration in asthma control requiring a course of oral and/or systemic corticosteroids. Notably, of 4,582 patients screened at 31 centres in the UK, only 199 of a planned 380 were randomised. The major reason for non-recruitment was use of antibiotics within 28 days of enrolment (2,044 [44.6%] screened patients).

Participants were randomised to either azithromycin 500mg daily or matched placebo for 3 days. The primary outcome was diary card symptom score (7-point scale) assessed 10 days after randomisation. Secondary outcomes included changes in quality of life and lung function between exacerbation and day 10, and time to a 50% reduction in symptom score. Mean (standard deviation) asthma symptom scores were 4.14 (1.38) at exacerbation and 2.09 (1.71) at 10 days with azithromycin and 4.18 (1.48) and 2.20 (1.51) with placebo, respectively. There was no significant difference in symptom scores between azithromycin and placebo at day 10 (difference, –0.17, 95% CI –0.67 to 0.34), nor on any day between exacerbation and day 10. In addition, there were no significant differences in secondary outcomes between groups. There were more gastrointestinal adverse effects with azithromycin than placebo (35 vs. 24 events).

(The study was funded by the Medical Research Council in partnership with the National Institute for Health Research.)

Comment: National guidance advises that routine prescription of antibiotics is not indicated for patients with acute asthma as most infections that precipitate an asthma attack are likely to be viral.3 The study's authors noted that azithromycin had been shown to have antiviral properties in an in vitro study.1 However, given concerns over antimicrobial stewardship and the very limited benefit seen in the telithromycin study, justification for a study such as this seems limited. Of particular interest was the finding that the study could not recruit sufficient participants (and was therefore underpowered for the primary endpoint), largely because of prior antibiotic use. This suggests that antibiotics are being overused in the management of people with acute asthma attacks. It is worth noting that long-term use of a macrolide antibiotic has also been investigated for the management of chronic asthma but such use is not recommended as it is no better than placebo for most outcomes.3,4


Risk of heart failure with NSAIDs

The authors of a large case-control study have suggested that the risk of heart failure varies between individual NSAIDs and according to the dose used.1 Concerns over adverse cardiovascular effects of NSAIDs are well documented, with much of the evidence focused on thrombotic effects of selective and non-selective drugs.2 Studies have also shown that the risk of hospitalisation due to heart failure is increased by NSAIDs compared with placebo.3 Current advice suggests avoiding NSAIDs in people with severe heart failure.4 For people with mild to moderate heart failure, ibuprofen (up to 1,200mg/day) or naproxen (up to 1,000mg/day) are recommended as first-line options.

The nested case-control study was based on five electronic health databases from the UK, the Netherlands, Italy and Germany.1 The authors identified nearly 10 million adults starting NSAID treatment during 2000–2010. Participants were excluded if they had been admitted to hospital with a primary diagnosis of heart failure in the year before the date of cohort entry.

The endpoint of interest was the first hospital admission for heart failure during follow-up and a total of 92,163 admissions were identified. Cases were matched with 8,246,403 controls according to age, sex and year of cohort entry. The mean age among cases and controls was 77 and 76 years, respectively. A total of 27 different NSAIDs were used during the study period (23 traditional NSAIDs and four selective cyclo-oxygenase-2 enzyme [COX2] inhibitors). Heart failure as a comorbidity during the 12 months before cohort entry was present in 9% of cases and 3% of controls.

Current use of any NSAID (defined as use in the preceding 14 days) was associated with a 19% increase in risk of hospital admission for heart failure (adjusted odds ratio 1.19, 95% CI 1.17 to 1.22), compared with past use of any NSAID (more than 183 days in the past). The study found that the risk of admission for heart failure increased with current use of seven traditional NSAIDs (diclofenac, ibuprofen, indometacin, ketorolac, naproxen, nimesulide and piroxicam) and two COX-2 inhibitors (etoricoxib and rofecoxib). Odds ratios ranged from 1.16 (95% CI 1.07 to 1.27) for naproxen to 1.83 (1.66 to 2.02) for ketorolac. The association with heart failure was confirmed regardless of whether there was recorded evidence of a previous diagnosis of heart failure. For most NSAIDs there was an increasing dose-dependent risk of heart failure hospitalisations. Celecoxib was not associated with a raised risk of hospital admissions for heart failure at commonly used doses. However, the authors could not rule out an association at very high doses.

(The study was based on research funded by the European Community's seventh Framework Programme.)

Comment: These findings, based on real-world data from a large cohort, provide a reminder of the risk of admission for heart failure associated with commonly used NSAIDs. In general, NSAIDs should be avoided in people with heart failure or those at high risk of heart failure. If an NSAID is needed it should be used at the lowest possible dose for the shortest time and reviewed regularly.


Bisphosphonates and steroid-induced osteoporosis

An updated Cochrane review suggests that bisphosphonates are beneficial in reducing the risk of vertebral fractures and preventing and treating bone loss caused by long-term corticosteroid treatment.1 However, they appear to have little or no effect in preventing non-vertebral fractures.

The review identified 27 randomised controlled trials with 3,075 adults taking a mean steroid dose of at least 5mg/day. As a result of more rigorous inclusion criteria, only two of the trials from the original review (published in 1999) were retained. Most of the trials combined bisphosphonates with calcium and/or vitamin D, which were then also used in the comparator treatment, either alone or with placebo.

The major outcomes of interest were the incidence of vertebral and non-vertebral fractures after 12–24 months; the change in bone mineral density at the lumbar spine and femoral neck after 12 months; serious adverse events; withdrawals due to adverse events; and quality of life.

In the pooled analysis for incident vertebral fractures, bisphosphonates reduced the total number of fractures compared with control by 43% (risk ratio 0.57, 95% CI 0.35 to 0.91), with a number-needed-to-treat (NNT) for an additional beneficial outcome of 31 over 12–24 months. This finding was based on high-certainty evidence with low risk of bias.

In contrast, the pooled analysis for incident non-vertebral fractures did not find a significant beneficial effect with bisphosphonates, suggesting that bisphosphonates make little or no difference in preventing non-vertebral fractures. However, this analysis was based on low-certainty evidence.

Overall, the review found that bone density stabilised or increased in the bisphosphonate groups over the study period, while it decreased in the control groups. There were statistically significant and clinically meaningful absolute increases in bone density at the lumbar spine (mean difference [MD] 3.50%, 95% CI 2.90 to 4.10) and femoral neck (MD 2.06%, 95% CI 1.45 to 2.68) after 12 months of bisphosphonate use, compared with control. The evidence for both outcomes was moderate certainty with low risk of bias.

The review found no statistically significant differences in the incidence of serious adverse events or withdrawals due to adverse events between the bisphosphonate and control groups. This was based on low-certainty evidence. None of the trials assessed quality of life as an outcome.

Comment: National guidance recommends considering bisphosphonates to prevent vertebral fractures in men and women on prednisolone doses ≥7.5mg daily (or an equivalent dose of glucocorticoid) for ≥3 months.2 This updated review agrees with previous findings that bisphosphonates are beneficial in preventing vertebral fractures associated with steroid-induced bone loss. Although the review found no evidence of serious adverse events with bisphosphonate use, these drugs are associated with osteonecrosis of the jaw.3 Steroid therapy is also a risk factor for osteonecrosis.4,5


Triple therapy for moderate-to-severe COPD: marginal benefits

The effects of single inhaler triple therapy compared with dual therapy have been investigated in a randomised controlled trial in 1,368 people with moderate-to-severe chronic obstructive pulmonary disease (COPD).1

Guidance from the National Institute for Health and Care Excellence (NICE) and the Global Initiative for Chronic Obstructive Lung Disease (GOLD) recommends a long-acting muscarinic antagonist (LAMA) or an inhaled corticosteroid plus a long-acting beta2 agonist (LABA) for patients with moderate-to-severe disease.2,3 Triple therapy with LAMA and LABA+ICS is an option for people who remain breathless or have exacerbations despite taking LABA+ICS. However, there are relatively few published trials of triple therapy and none are thought to have used a single inhaler.1

The trial (carried out across 159 sites in 14 countries) compared the efficacy of a single inhaler combination of beclometasone dipropionate, formoterol fumarate and glycopyrronium bromide (BDP/FF/GB) with a single inhaler combination of beclometasone dipropionate and formoterol fumarate (BDP/FF) over 26 weeks. Trial participants had post-bronchodilator forced expiratory volume in 1 second (FEV1) of <50%, at least one moderate-to-severe COPD exacerbation in the previous 12 months, COPD Assessment Test total score of 10 or more, and a Baseline Dyspnea Index focal score of 10 or less. The study's three co-primary endpoints were pre-dose FEV1, 2-hour post-dose FEV1 and Transition Dyspnea Index (TDI) focal score, all measured at week 26 in the intention-to-treat population. Assessment of exacerbations was a secondary endpoint. An exacerbation was defined as a worsening of symptoms that required systemic corticosteroids, antibiotics or hospital admission (or a combination of them).

At week 26, triple therapy improved pre-dose FEV1 by 0.08L (95% CI 0.05 to 0.11) and 2-hour post-dose FEV1 by 0.12L (95% CI 0.09 to 0.15) compared with standard treatment. There was no significant difference between triple therapy and standard treatment in mean TDI focal scores at week 26. The percentage of people with moderate-to-severe exacerbations was lower in those using triple therapy (31% vs. 35%). The adjusted annual rate of moderate-to-severe exacerbation was 0.41 for triple therapy and 0.53 for dual therapy (rate ratio 0.77, 95% CI 0.65 to 0.92). Rates of adverse events were similar between groups, with one serious treatment-related adverse event (atrial fibrillation) in the triple therapy group.

(Study funded by Chiesi Farmaceutici SpA.)

Comment: Overall, compared with LABA/ICS therapy, LAMA/LABA/ICS therapy produced a small improvement in measures of lung function but not in symptoms of breathlessness. The lower rate of exacerbations with triple therapy compared with dual therapy represents a difference of approximately one exacerbation per patient every 8 years. Most of the benefit came from an effect on moderate exacerbations. The importance to patients of this potential reduction in exacerbations is questionable.


Much data on adverse events remains unpublished

The AllTrials campaign and similar initiatives have drawn attention to the problem of missing and unavailable trial data, and how this can distort our understanding of the benefits and harms of treatments.1 A systematic review has attempted to quantify under-reporting of adverse events in trials, suggesting that a substantial amount of information on the harms of interventions remains unpublished.2

The researchers identified 28 studies (31 publications) that compared adverse event data in published journal articles and other sources (e.g. websites, trial registries, conference proceedings and industry-held data). The review focused on quantifying the amount of unpublished data compared with published data and assessed the impact of including unpublished adverse event data on the results of systematic reviews. This included instances where the published and unpublished sources were matched (i.e. different versions of the same study) or unmatched (i.e. separate sets of studies on the same topic area). All types of healthcare intervention were eligible for inclusion but the majority involved drugs.

Nine studies compared the proportion of trials reporting adverse events by publication status. For matched comparisons, the median percentage of published documents with adverse event information was 46% compared with 95% in the corresponding unpublished documents. The results were similar for unmatched comparisons, with 43% of published studies containing adverse event information compared with 83% of unpublished studies.

Eleven studies compared the numbers of adverse events in matched published and unpublished documents. The percentage of adverse events that would have been missed had each analysis relied only on the published versions varied between 43% and 100%, with a median of 64%. These same 11 studies also included 24 comparisons of named adverse events such as death, suicide, or respiratory adverse events. In 18 of the 24 comparisons, the number of named adverse events was higher in unpublished than published documents. Additionally, two other studies found that substantially more types of adverse event were reported in matched unpublished than published documents.

The review also included 20 meta-analyses that reported the odds ratios and/or risk ratios for adverse events both with and without the addition of unpublished data. Inclusion of unpublished data increased the precision of the pooled estimates, with narrower 95% confidence intervals in 15 of the 20 pooled analyses. However, this did not markedly change the direction or statistical significance of the risk in most cases, owing to the breadth of the original confidence intervals for the published data alone.

Given that serious or important adverse events occur rarely for some treatments, systematic reviews and meta-analyses that synthesise data from numerous sources, including unpublished datasets, can provide greater insight, say the authors, who encourage researchers to search beyond journal publications for harms information.

Comment: The extent of the discrepancy between published and unpublished information on harms is of concern as it may have a significant effect on the findings of systematic reviews and meta-analyses. Although pooling published and unpublished data did not affect the direction and statistical significance of the risk estimates for rare adverse effects, the authors highlight two examples of reviews that resulted in changes to their conclusions once unpublished data was included. Treatment choices need to be based on a full assessment of the balance of harms and benefits,3 which furthers the argument for greater access to unpublished data.


Rising placebo response rates in antidepressant trials a ‘myth’

The long-accepted wisdom that placebo response rates in trials of antidepressants have been steadily rising for many years is questioned in a new systematic review.1 The study assessed 252 double-blind randomised placebo-controlled trials (26,324 patients on placebo) done between 1978 and 2015. Trials included first- and second-generation antidepressants used for acute treatment of major depression in adults. The authors concluded that placebo response rates have remained constant since 1991, and that studies suggesting an increasing response rate have been based on outdated or limited datasets and used inappropriate statistical methods.

The authors concede that before 1991, response rates appeared to rise over time. However, since 1991 the average placebo response rates in antidepressant trials have remained constant in the range 35–40%. Analysis of studies published before 1991 is thought to have been confounded by methodological factors more common in trials in the earlier period and less common after it. Such factors included shorter trial duration and preponderance of single-centre studies. After controlling for these variables associated with lower placebo response rates, the association between response rate and study year was no longer significant.

The authors discuss the high placebo response rate and highlight conditions inherent in many trials that contribute to the objective response observed in patients assigned to placebo (the expectation of improvement, classic conditioning, and contact with a healthcare environment with supportive and therapeutic features). They note that that these non-pharmacological factors are usually not present to the same extent in standard clinical practice.

(Article funded by Japan Society for Promotion of Science, Great Britain Sasakawa Foundation.)

Comment: That a response rate of 35–40% can be achieved in large part through “contact with a healthcare environment with supportive and therapeutic features” should encourage clinicians to try to replicate such an environment in clinical practice, and sets the marginal benefits of antidepressants into context.


Prenatal fetal RhD testing recommended

Guidance from the National Institute for Health and Care Excellence (NICE) has recommended non-invasive prenatal testing for fetal rhesus-D (RhD) status, potentially reducing the need for an injection of anti-D immunoglobulin.1

Currently, all pregnant women whose blood tests show that they are RhD negative are given anti-D immunoglobulin to prevent RhD isoimmunisation. This blanket approach to prevention of haemolytic disease of the newborn (HDN) means that RhD negative women carrying an RhD negative baby unnecessarily receive treatment. According to NHS Maternity Statistics, 40,000 women needlessly received antenatal anti-D prophylaxis in England from April 2013 to March 2014.2

The prenatal test recommended in NICE's guidance analyses fetal DNA from blood taken from the mother during routine antenatal appointments, to establish the rhesus status of the baby. This non-invasive test offers several advantages, according to NICE, including:

  • potentially saving the NHS more than £500,000 each year by ensuring that anti-D immunoglobulin is only given to women who need it

  • preserving stocks of anti-D immunoglobulin

  • ensuring that women aren't unnecessarily exposed to the small risk of contracting blood-borne infections from anti-D immunoglobulin

  • allowing women who are RhD negative and carrying an RhD positive baby to make an informed choice about whether to have anti-D prophylaxis to prevent HDN in future pregnancies.

Comment: In an editorial in 2015, we concluded that maternal blood testing for fetal RhD status “would represent a significant advance in personalised obstetric care by allowing both routine and post-sensitisation antenatal anti-D to be directed only to those women carrying RhD positive fetuses”.3 NICE ‘s recommendation is predicated on the basis that the overall cost of testing is £24 or less. Savings are also dependent on the cost of additional care pathways and implementation costs.1


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.