General practitioner use of a C-reactive protein point-of-care test to help target antibiotic prescribing in patients with acute exacerbations of chronic obstructive pulmonary disease (the PACE study): study protocol for a randomised controlled trial

Point-of-care tests (POCTs) for acute infections are being promoted to reduce inappropriate prescribing, reduce antimicrobial resistance and to improve patient-reported outcomes [1‐4]. While POCTs are frequently subjected to evaluations of analytic performance, they have often been introduced into routine care before determining their clinical effectiveness via rigorous clinical trials and without understanding cost-effectiveness using relevant health and service delivery outcomes.

Better targeting of antibiotics in acute exacerbations of chronic obstructive pulmonary disease (AECOPD) represents a major opportunity for antimicrobial stewardship and improved patient care. Over 80% of all antibiotics are prescribed in the community [5] with high prescribing of broad-spectrum antibiotics a particular concern. AECOPD accounts for over two million antibiotic prescriptions each year in the UK [6]. Cohort studies of patients recruited in secondary care (which may not be representative of primary care but is the best relevant data available) suggest that most COPD patients in the UK will suffer between 2.5 and 3 AECOPD per year. Over 70% of patients presenting with AECOPD in primary care are prescribed an antibiotic, accounting for 4.6% of all antibacterial prescriptions every year [7] . COPD patients are an important and increasingly large group who are at risk of significant mortality, morbidity and hospitalisation and, as such, are more likely to be prescribed broad-spectrum antibiotics [8]. However, many AECOPD are triggered by non-bacterial causes, such as viral infections and environmental factors including common pollutants or weather. It has been estimated that approximately 70% of AECOPD are triggered by an infection and 30% are caused by environmental factors. Of the 70% that are triggered by an infection, potential pathogenic bacteria are only isolated in 20–58% of clinical samples, while pathogenic respiratory viruses can be detected in approximately 50% [9‐11].

Current antibiotic prescribing recommendations for general practitioners (GPs) are generally based on symptoms alone (Anthonisen criteria; [12, 13]). However, these symptoms are subjective and have insufficient diagnostic accuracy to predict which patients can safely be managed without antibiotics. Both our placebo-controlled trial of antibiotics for AECOPD in primary care [14] and systematic reviews [15] suggest that many patients with AECOPD in primary care do not benefit from antibiotic treatment. Overuse of antibiotics drives antimicrobial resistance [16] and is facilitated by the unnecessary consumption of antibiotics for COPD. Antimicrobial treatment in patients with COPD decreases the infecting load but does not usually entirely eradicate organisms in the airways, increasing the risk of resistant bacteria in COPD patients [17]. Infections with antibiotic-resistant Streptococcus. pneumoniae in patients with COPD are associated with antibiotic exposure [18, 19]. A meta-analysis of seven studies of respiratory tract bacteria that included 2605 participants showed that the pooled odds ratio (OR) for resistance was 2.4 (1.4 to 3.9) within 2 months of antibiotic treatment, and 2.4 (1.3 to 4.5) within 12 months [5]. The unnecessary use of antibiotics for AECOPD not only contributes to the increasingly pressing public health threat of antibiotic resistance, it also poses a risk for the individual, and may increase the risk of subsequent antibiotic-resistant exacerbations and hasten disease progression. Indiscriminate use of antibiotics in patients with COPD is particularly high risk because the respiratory tracts of those affected are frequently colonised with potential pathogens [20]. Unnecessary antibiotics also increase the risk of patient side effects, waste money, and undermine self-care [21].

A Cochrane systematic review of the use of antibiotics in the management of exacerbations of COPD included 16 trials (n = 2068 participants), and reported that there was insufficient evidence of effectiveness to guide antibiotic prescribing decisions in primary care [15].

C-reactive protein (CRP) is an acute-phase protein found in the blood. The serum level of CRP increases rapidly during infections, particularly in severe bacterial infections. A prospective evaluation of 36 biomarkers found that CRP was the most selective biomarker to confirm AECOPD, and in combination with Anthonisen criteria produced an area under the curve (AUC) of 0.88 (95% confidence interval (CI) 0.82–0.93), indicating that it had good diagnostic accuracy [22]. High serum CRP is correlated with sputum purulence and raised serum leucocyte counts and serum CRP is higher in the presence of bacterial infection [23, 24]. CRP rises in patients with AECOPD and is correlated with Anthonisen score and the degree of airflow limitation in hospitalised patients [25, 26]. As CRP levels are especially raised in the presence of bacterial infection the treatment effect of antibiotics increase with higher values of CRP [27]. A CRP value above 50 mg/L (mean CRP of 97 mg/L, 95% CI 49–145) in hospitalised patients with AECOPD is associated with pneumonia and such patients are likely to benefit from antibiotics [26]. CRP measurement independently distinguished between pneumonia and other causes of exacerbations in another study of patients hospitalised with asthma and COPD (cut-off value of 48 mg/L with sensitivity of 91% and specificity of 93%) [28]. In a randomised controlled trial we conducted in patients with AECOPD in primary care, we found no difference in clinical cure between antibiotics and placebo in those with a CRP < 40 (risk ratio (RR) for clinical failure = 0.72 (95% CI 0.28–1.82) p = 0.484) [23].

In an as yet unpublished study we found that over 50% of COPD patients experiencing an exacerbation had a CRP < 8 mg/L [29]. Our recent placebo-controlled trial of antibiotics for AECOPD in primary care found marginal benefit from antibiotic treatment in patients with only one or two Anthonisen criteria. Using Anthonisen criteria to predict benefit from antibiotic treatment produced an area under the curve (AUC) of 0.708 (95% CI 0.616–0.801). Adding CRP increased this to an AUC of 0.842 (95% CI 0.76–0.924). Based on these data we anticipate that using a CRP test alongside clinical assessment will make it possible to safely reduce the antibiotic prescription rate for this condition to around 45%.

CRP POCTs are widely available and are already commonly used to help guide antibiotic prescribing decisions, including for lower respiratory tract infections (LRTIs) and AECOPD in primary care in a number of European countries (mostly Scandinavian) [30]. In two trials evaluating the use of a CRP POCT to help target antibiotic treatment for LRTIs in primary care [31, 32], antibiotics were prescribed to 68% and 48% in the usual care groups, respectively, and to 39% and 33% of patients managed by clinicians using a CRP POCT (with training). CRP POCT was cost-effective in reducing antibiotic prescribing for LRTIs when there are no tests, or low willingness to pay for the tests [31, 33]. However, the benefits of CRP POCT in conjunction with clinical examination has not yet been evaluated for AECOPD in primary care in a randomised controlled trial. Now that better and more rapid CRP POCTs are available [34], there is potential for this technology to be widely used for a variety of acute infections in primary care to better guide antibiotic prescribing and in doing so, help reduce unnecessary antibiotic consumption and thus contain antibiotic resistance. PACE seeks to establish whether a CRP POCT can safely and cost-effectively be used to better target antibiotic treatment for AECOPD in primary care to those that are most likely to benefit, so that overall antibiotic use is decreased without compromising COPD-related health status.

AECOPD lead to increased morbidity, emergency hospital attendances, hospitalisations, health care costs, and more rapid disease progression and deterioration in quality of life [46]. Most exacerbations are managed without recourse to POCTs and most patients are prescribed antibiotics, despite evidence that these prescriptions do not always benefit them and may cause harm to the individual patients and contribute to antimicrobial resistance. POCT CRP testing may help clinicians better target antibiotic prescribing, ensuring that antibiotics are prescribed for those patients who benefit and other interventions are targeted to those who will not.

POCTs have been widely promoted for improving the care of patients who have acute infections [1]. Most evaluations of diagnostic devices consider analytic performance without evaluating impact on patient outcomes or costs. However, new tests should not be introduced into routine clinical care if they do not improve outcomes that matter to patients individually or to society, including consideration of impact on recovery and quality of life, antibiotic prescribing and resistance [2‐4].

The PACE study will be the first pragmatic, randomised controlled, clinical trial to evaluate the clinical and cost-effectiveness of the addition of CRP POCT to clinical assessment for AECOPD in primary care. It will establish whether a CRP POCT can safely and cost-effectively be used to target antibiotic treatment for AECOPD so that overall antibiotic prescribing and thus consumption is decreased without compromising patients’ COPD health status. If the results of the study are positive (i.e. significant reduction in antibiotic prescribing with no evidence of significant impairment in symptomatic recovery) then implementation into practice is likely to be achieved swiftly. NICE guidelines on pneumonia already recognise a role for CRP POCT in deciding whether to prescribe antibiotics, and the use of CRP POCT is increasing in the UK. Potential barriers to uptake include the cost of testing, training, quality control and integration with laboratory systems, connectivity with electronic medical records, practitioner attitudes, and impact on clinical workflow. However, these barriers have been successfully mitigated in many European countries where CRP POCT use is now widespread [47].