The Diagnostics Advisory Committee reviewed the evidence available on the clinical and cost effectiveness of using procalcitonin testing with standard clinical practice to guide antibiotic therapy in the following populations: adults and children with confirmed or highly‑suspected sepsis in intensive care settings; and adults and children with suspected bacterial infection presenting to the emergency department.
The Committee considered the role of procalcitonin and noted that it is a more specific biomarker for bacterial infection than other biomarkers that are currently widely used for diagnosing and monitoring bacterial infection and sepsis. Other biomarkers, such as C‑reactive protein and white blood cell count are inflammatory markers that have low specificity for the diagnosis of bacterial infection.
The Committee considered the equivalence of the 5 assays for procalcitonin testing that were included in the scope. It noted that most of the studies measured procalcitonin levels using the BRAHMS PCT Sensitive Kryptor assay and 2 studies measured procalcitonin levels using the VIDAS BRAHMS PCT assay. No relevant studies were identified that used the ADVIA Centaur BRAHMS PCT assay, the Elecsys BRAHMS PCT assay or the LIAISON BRAHMS PCT assay. The Committee heard from the manufacturers that they consider all assays to be comparable because they all use the same capture and detection antibodies in an immunoassay format. It also heard from the External Assessment Group that all assays have been standardised using the BRAHMS PCT LIA assay. The Committee noted that the Kryptor analyser is not widely available in NHS laboratories and therefore there may be a cost impact for laboratories in buying the Kryptor analyser. The Committee concluded that the procalcitonin assays could be considered technically comparable and noted that this meant NHS laboratories would have flexibility in the choice of analyser.
The Committee considered the quality of the studies included in the systematic review of clinical effectiveness. It noted that the External Assessment Group generally considered the studies to be at unclear risk of bias because insufficient information was reported in the publications. The Committee also noted that the studies were heterogeneous, and most reported low event numbers. In addition, several studies reported median results that could not be incorporated into the summary estimates and the economic model. The Committee concluded that these factors contributed to the uncertainty in clinical effectiveness.
The Committee considered the implications of adding procalcitonin testing to standard clinical practice in intensive care unit settings and in the emergency department. It noted that clinical outcome results, such as mortality, infection relapse or recurrence and re‑admission rates, were not statistically significantly different between patients in the intervention groups (procalcitonin testing plus standard clinical practice) and those in the control groups (standard clinical practice alone). The Committee concluded that using procalcitonin testing with standard clinical practice is unlikely to result in worse clinical outcomes compared with standard clinical practice alone.
The Committee discussed whether antibiotic stewardship practices differ between hospitals. The Committee noted the Public Health England report on the English surveillance programme for antimicrobial utilisation and resistance (2014) showed that combined GP and hospital antibiotic prescribing had increased by 6% between 2010 and 2013, despite the publication of the 'Start smart – then focus' toolkit for antimicrobial stewardship (2011). The Committee also considered whether the lack of suitable laboratory tests to inform antibiotic prescribing decisions was a possible reason for the lack of improvement in antibiotic stewardship. The Committee concluded that improved antimicrobial stewardship and consequent reduction in antibiotic use was not currently being achieved in the NHS, and that this could reflect that guidance on antimicrobial stewardship was not being fully adhered to in all NHS hospitals.
The Committee considered the impact on resource use of adding procalcitonin testing to standard clinical practice in intensive care unit settings. The Committee noted that most studies found reductions in resource use when procalcitonin was added to standard clinical care. However, the Committee noted that studies included in the review were done in Switzerland, France, Belgium, Brazil and China; no UK studies were included. It further noted that many of the studies were done and published several years ago, and that clinical practice for antibiotic stewardship has changed considerably in the last few years. The Committee therefore considered whether the results of these studies were generalisable to the NHS. The Committee heard from the External Assessment Group that many of the studies did not clearly report how treatment decisions were made in the control arms. The Committee heard from clinical experts that when details of the control arm were reported, they did not reflect current standard clinical practice in the UK. The Committee also heard from clinical experts that in NHS intensive care units, antibiotic stewardship is considered daily during ward rounds that include microbiologists and intensivists (doctors who specialise in the care and treatment of patients in intensive care). The Committee therefore concluded that in the intensive care unit setting, the reductions in resource use reported in the included studies when procalcitonin was added to standard clinical practice were unlikely to be realised in the NHS.
The Committee considered the potential reasons for the reduction in resource use in intensive care unit settings reported in included studies. The Committee discussed whether the reductions were directly due to the use of procalcitonin, or whether the reductions may be due to the introduction of a clinical protocol that included use of a biomarker. It heard from clinical experts that in intensive care unit settings in the UK, protocols are used to guide care. The Committee noted that it was unclear if protocols were used to guide care in the control arms in the studies included in the systematic review, whereas the intervention arms included a procalcitonin algorithm with defined thresholds. The Committee concluded that in intensive care unit settings they could not be certain whether the reductions in resource use in the included studies were caused by use of procalcitonin or by use of a clinical protocol. Therefore, because clinical protocols are standard clinical practice in the NHS, the Committee also concluded that it is uncertain whether the potential reduction in resource use in intensive care settings would be realised from introducing procalcitonin testing into protocol‑guided care.
The Committee considered the results of a survey that had been done by a company. The principal investigators of all the included studies in the assessment were contacted and asked to provide clarification on the exact methods used within the control arm referred to in section 6.8, and how this compares with the current UK approach to antibiotic stewardship. The Committee noted that although the survey provided some reassurance, the questionnaire contained a generic description of the UK standard of care and a leading question was used to obtain details of the control arms in the studies, so the survey findings were considered at risk of bias. The Committee also reviewed a list of studies done in the UK that was provided by the companies. The Committee heard from the External Assessment Group that these studies were not comparative studies and therefore cannot be used to assess the benefits of adding procalcitonin to standard clinical practice. The Committee concluded that although it was useful that more evidence had been submitted at consultation, there was still uncertainty about the generalisability of the studies and the additional benefit that procalcitonin testing would provide in the NHS.
The Committee noted that no studies were found that were done in children with confirmed or highly‑suspected sepsis in intensive care unit settings. It considered whether the data identified in an adult population could apply to children. It heard from clinical experts that a large proportion of children in paediatric intensive care units are younger than 1 year and are likely to have different care and resource needs. The Committee therefore concluded that data on adults in an intensive care unit should not be extrapolated to children.
The Committee considered trials in progress. It noted that the 'Stop Antibiotics on guidance of Procalcitonin Study (SAPS) had recently been completed and results were expected to be available after publication of this guidance. The Committee noted that this study is much larger than any of the studies included in the systematic review of clinical effectiveness (estimated 1800 patients). The Committee also considered an Australian study, published after the completion of the External Assessment Group's systematic review of clinical effectiveness (Shehabi et al. 2014), which found that using procalcitonin testing with standard clinical care in adults in an intensive care unit did not significantly reduce the duration of antibiotics. The Committee noted that the control arm of this study included antimicrobial stewardship implemented by twice‑weekly infectious diseases rounds, which is similar to standard clinical care in the UK. However, the Committee also heard from the External Assessment Group that this study used a lower threshold of procalcitonin test values to determine discontinuation of antibiotics compared with the studies included in the systematic review (0.1 nanograms per millilitre and 0.25 nanograms per millilitre), which would likely result in longer durations on antibiotics.
The Committee considered the impact on resource use of adding procalcitonin testing to standard clinical practice in the emergency department. As with the intensive care unit setting (section 6.7), the Committee noted that most studies found reductions in resource use when procalcitonin testing was added to standard clinical practice. It noted further that studies were done in Switzerland, China and Italy, and that more than half of included studies were published before 2012. The Committee heard from clinical experts that in the emergency department there is much more variation in clinical assessment and in the decision to give antibiotics, compared with the intensive care unit. The Committee heard further from clinical experts that for patients with suspected bacterial infection at low risk of sepsis, clinical assessment can be variable, and therefore the clinical benefits of adding procalcitonin to standard clinical practice are uncertain. The Committee also heard from clinical experts that severely ill patients at high risk of sepsis would have care according to a clinical protocol and in line with the Surviving Sepsis Campaign's guidelines (2012). The Committee considered the results of the survey and UK‑based studies provided by the companies and concluded that uncertainty remains around the generalisability of the studies to NHS practice (see section 6.9). The Committee therefore concluded that it was uncertain whether the reductions in resource use shown in studies done in other countries would be applicable to the NHS.
The Committee noted that most of the studies included in the systematic review of clinical effectiveness in an emergency department setting were done in people with suspected bacterial infection in the lungs. The Committee considered whether the data identified would be applicable for suspected bacterial infections in other sites. The Committee heard from clinical experts that for bacterial infections such as urinary tract infections, other tests are available that can be used to make a diagnosis. It also heard that lower respiratory tract infections are more difficult to diagnose compared with bacterial infections in other sites. The Committee concluded that the results from the systematic review could not be applied to people presenting to the emergency department with suspected bacterial infection outside the lungs.
The Committee noted that the studies done in emergency department settings included many patients who were sent home without antibiotics. This reduced the number of patients from whom resource use data could be gathered and therefore increased the uncertainty in the results, leading to fewer statistically significant reductions in resource use compared with studies done in intensive care units. The Committee was also concerned that there was limited longer term follow‑up of people who were sent home without antibiotics; therefore the risk of these people having the incorrect treatment and then re‑attending medical services is unknown.
The Committee considered the heterogeneity of the populations in the emergency department studies and noted that they included a wide range of conditions (lower respiratory tract infections including community‑acquired pneumonia, chronic obstructive pulmonary disease exacerbations, severe asthma exacerbations, urinary tract infection). The Committee noted that some studies reported conflicting results when analysing different subgroups of patients. For example, the study by Esposito et al. (2011) found that using procalcitonin testing with standard clinical practice was associated with a greater reduction in antibiotic use for children with mild community‑acquired pneumonia than for children with severe community‑acquired pneumonia. The Committee also heard from an expert on the Committee that people presenting to the emergency department with chronic obstructive pulmonary disease exacerbations are highly likely to be admitted and given antibiotics, but that most of these patients do not have proven bacterial infection. The Committee discussed whether the benefits of procalcitonin testing in severely ill patients at high risk of sepsis might be different from the benefits in less severely ill patients. The Committee concluded that the heterogeneity in the populations resulted in a high‑level uncertainty about the benefits of procalcitonin testing that would be realised in clinical practice. It considered that focused studies in explicitly‑defined patient groups were needed to address this uncertainty.
The Committee noted that there were limited data on children presenting to the emergency department with suspected bacterial infection. It heard from clinical experts that over the last decade, the number of children admitted to hospital with suspected bacterial infection after presenting to an emergency department has increased by 28% because of increased uncertainty and lack of confidence in clinical assessment. The Committee noted the study by Baer et al. (2013) that reported a trend towards increased antibiotic use when procalcitonin test results were added to standard clinical practice. The Committee also noted that there was no agreed threshold for procalcitonin that should be used to encourage or discourage antibiotic use in children with suspected bacterial infection. The Committee considered whether the results of a laboratory test could cause clinical judgement to be overruled; potentially leading to people with test results just above the threshold having antibiotics, whereas clinical judgement alone would have led to antibiotics being withheld or stopped. The Committee was concerned that without robust evidence, the introduction of procalcitonin testing into standard clinical practice may not reduce the use of antibiotic therapy, but could potentially increase antibiotic exposure and hospital admissions. The Committee concluded that more robust evidence is needed on the use of procalcitonin testing with standard clinical practice in children presenting to the emergency department with suspected bacterial infection.
The Committee heard from experts on the Committee that hospital laboratories perform a large volume of C‑reactive protein tests, many of which may be requested unnecessarily. The Committee was concerned that without the correct clinical risk assessment protocols in place, procalcitonin testing could be done in all people presenting to the emergency department with a respiratory tract infection. It noted that some test results will be false positives and may influence clinical judgement resulting in people having antibiotics unnecessarily. The Committee further noted that procalcitonin testing is unlikely to replace C‑reactive protein testing for clinical decision‑making. It also noted that the costs of unnecessary tests were not included in the cost‑effectiveness analysis. The Committee concluded that if procalcitonin testing were to be recommended for routine use in the emergency departments, it is possible that many unnecessary tests would be performed and clear guidance on their use would be needed to minimise this risk.
The Committee discussed the use of length of stay as a key outcome measure in studies on procalcitonin testing, both in intensive care unit settings and after admission through the emergency department. It noted that length of stay may be influenced by factors other than procalcitonin testing, such as differences in local protocols and preferences of treating clinicians. The Committee concluded that measuring length of stay may not reliably capture the potential benefits of procalcitonin testing. It encouraged further research using clinically meaningful outcomes.
The Committee considered the long‑term outcomes that may occur after sepsis and noted that no evidence was identified on these long‑term outcomes. It heard from a patient expert on the Committee that post‑sepsis syndrome affects up to 50% of survivors, and constitutes: insomnia; nightmares, vivid hallucinations and panic attacks; disabling muscle and joint pains; extreme fatigue, poor concentration and decreased cognitive functioning; and loss of self‑esteem. In some cases, post‑sepsis syndrome can lead to post‑traumatic stress disorder, which, without the correct support, can last for years. Additionally, survivors of sepsis may be more vulnerable to developing respiratory viral infections than the general population. The Committee concluded that future research should consider including long‑term outcomes to more fully capture the impact of sepsis on patients.
The Committee noted that outcomes relating to antibiotic stewardship were not included in the model. The Committee discussed the potential contribution that procalcitonin testing could make to improving antimicrobial stewardship by reducing the incidence of antibiotic‑resistant infections. The Committee noted its earlier conclusions about the uncertainty of whether the reductions in resource use reported in the studies would be realised in clinical practice in the NHS (sections 6.7 to 6.8 and section 6.12). It concluded that procalcitonin testing may contribute to improving antibiotic stewardship in the NHS if it leads to a reduction in antibiotic use. The Committee also heard from clinical experts that procalcitonin testing would only be 1 component of an antimicrobial stewardship approach and other factors may also contribute substantially to improving antimicrobial stewardship. The Committee concluded that the benefit procalcitonin would provide in addition to other measures to improve antibiotic stewardship was not clear.
The Committee considered the assumptions used in the economic model. Firstly, the Committee discussed the assumption that there were no differences in disease‑specific complications between the intervention and control groups. It heard from the External Assessment Group that studies generally reported these complications as a compound outcome that could not be separated out, and that no statistically significant difference was found between the groups. The Committee concluded that the assumption was valid. Secondly, the Committee considered the assumption that a laboratory would perform 272 procalcitonin tests per day. The Committee considered that this might be an overestimate, especially for smaller hospital laboratories. The Committee heard from the External Assessment Group that this value was used to calculate the overhead costs per test, and that in sensitivity analyses varying the total test cost by 10 fold did not affect the cost‑effectiveness results. An alternative estimate of 30 to 40 tests per day was suggested by a company based on sales data. The Committee heard from the External Assessment Group that using a value of 30 tests per laboratory per day increased the cost of procalcitonin testing by £4 in the emergency department setting and £7 in the intensive care unit setting, but this did not affect the model results.
The Committee considered the time horizon used in the model. It noted that the time horizon used in the model is 6 months (183 days), divided into an initial short‑term (28 days) phase and a subsequent phase lasting 155 days. It noted that this time horizon reflected the outcomes reported in the studies included in the systematic review of clinical effectiveness, which did not report data on long‑term outcomes. The Committee concluded that the time horizon used was appropriate given the clinical data available.
The Committee considered the cost effectiveness of adding procalcitonin testing to standard clinical practice in intensive care unit settings and in the emergency department. It noted that in all scenarios assessed, procalcitonin testing with standard clinical practice dominated standard clinical practice alone, that is, it was more effective and less costly. However, the Committee acknowledged that the change in quality‑adjusted life years (QALYs) was extremely small. The Committee noted its earlier conclusion that the use of procalcitonin testing with standard clinical practice is unlikely to result in worse clinical outcomes compared with standard clinical practice alone (section 6.5). It therefore concluded that the QALYs for each group could be considered broadly equal, that is, adding procalcitonin testing to standard care does not result in a meaningful change in QALYs.
The Committee considered whether the cost savings generated through reductions in resource use were large enough to offset the additional cost of procalcitonin testing. It noted its earlier conclusion about the uncertainty of whether the reductions in resource use reported in the studies included in the systematic review would be realised in clinical practice in the NHS (sections 6.7 to 6.8 and section 6.12). The Committee therefore concluded that it was uncertain whether savings would be large enough in a UK setting to offset the cost of procalcitonin testing.
The Committee noted that procalcitonin testing is currently used with standard clinical care in a few centres in the UK to guide decisions on antibiotic treatment in emergency departments and intensive care units. The Committee acknowledged that the evidence suggested that the addition of procalcitonin testing to standard clinical care was unlikely to result in worse clinical outcomes (section 6.5). The Committee encouraged centres currently using procalcitonin testing to guide decisions on antibiotic treatment in emergency departments and intensive care units to take part in relevant data collection and research.