Clinical evidence (see also full evidence review)

Evidence statements

Mortality

There was very low quality evidence from four studies (Cometta et al., 2003; EORTC, 1989; Erjavec et al., 2000 and Pizzo et al., 1982) about when to change empiric antibiotics in patients with unresponsive fever (Table 7.1). No study compared changing empiric therapy at two different time points. Patients (N=461) with persistent fever were randomised to either remain on the empiric antibiotic or to primary treatment with the addition of another agent. No study detected a significant difference between the short term mortality of those who changed treatment and those who remained on the initial empiric treatment.

Table 7.1

GRADE profile: What is the optimal time to change the primary empiric treatment in unresponsive fever.

Critical care, quality of life and length of stay

The included studies did not report these outcomes.

Duration of fever

There was very low quality evidence about this outcome and none of the studies reported the influence of time of treatment change. Pizzo, et al., (1982) and Cometta, et al., (2003) reported shorter median time to defervesence in patients whose empiric therapy was changed (8 versus 6 days and 4.3 versus 3.5 days respectively), but there was no statistically significant difference. Erjavec, et al., (2000) reported similar rates of defervesence within 72 hours in patients who did or did not change empiric treatment.

Cost-effectiveness evidence

A literature review of published cost-effectiveness analyses did not identify any relevant papers. This topic focused on the optimal timing of a change in management strategy. The difference between strategies were considered unlikely to lead to large differences in cost, but rather be guided by differences in patient outcomes and other considerations such as service configuration that. It was agreed that these considerations would probably be difficult to accurately capture using economic modelling and therefore further health economic analysis was not undertaken.

Linking Evidence to Recommendations

The aim of this topic was to identify the optimal time to change the initial empiric treatment in unresponsive fever.

The GDG considered that the outcomes of over-treatment, death/critical care, length of stay, duration of fever and quality of life were clinically relevant to the question. No studies reported length of stay, the incidence of over-treatment or patients' quality of life. Limited evidence was available on mortality. Duration of fever was reported as an outcome but it was inconsistent and imprecise, and the GDG did not think this outcome was useful in agreeing recommendations.

The GDG noted that the evidence was classified by GRADE as being of ‘low’ or ‘very low’ quality. None of the studies dealt adequately with the methods of randomisation, allocation or blinding and, although some authors stated that appropriate statistics had been used for data analysis, the details were sometimes scant or absent and very few outcomes had more than a probability value reported.

The GDG were aware that there is a perception that empiric antibiotics should be changed after 48 hours in patients with unresponsive fever. However they noted that the evidence had not demonstrated a significant difference between patients kept on initial empiric antibiotics and those given an additional or different drug or drugs. The GDG also considered that it was important to prevent unnecessary extra treatment in this group of patients, which would reduce the risk of side effects associated with receiving additional drugs.

The GDG noted that consideration would have to be given to other causes of infection or fever but making recommendations on this was outside the scope of the guideline.

The GDG noted that no relevant, published economic evaluations had been identified and no additional economic analysis had been undertaken in this area. The opinion of the GDG was that there would probably be cost savings associated with reducing over-treatment and the corresponding reduction of adverse effects, in addition to cost savings from the early interventions in patients developing septic complications.

Therefore the GDG decided to recommend that empiric antibiotics should not be changed unless there was a clinical deterioration or a microbiological indication. However the GDG were concerned that this recommendation could result in patients not receiving proper clinical and laboratory surveillance. They therefore made an additional recommendation based on their clinical experience that the clinical status of the patient should be reviewed daily to prevent this from happening. The GDG noted from the evidence on paediatric risk assessment tools (section 4.3 & 7.3) that reassessment using a validated tool was an effective way of identifying those at low risk of septic complications. The GDG felt it was appropriate to extrapolate the data to the adult setting.

Clinical evidence (see also full evidence review)

Evidence statements

Death or critical care

Very low quality evidence from a Cochrane review (Vidal, et al., 2004, Table 7.2) suggested uncertainty about the relative effectiveness of the two treatment strategies for IV-to-oral versus IV-only the relative risk of short term mortality was 1.14 (95% CI 0.48 - 2.73). Critical care was not included as an outcome in any of the included studies, although one study (Paganini et al., 2003) did report that none of their patients required admission to the intensive care unit.

Table 7.2

GRADE profile: When is the optimal time to switch from intravenous to oral antibiotic therapy for patients with neutropenic sepsis.

Overtreatment, length of stay and quality of life

These outcomes were not reported in any of the included studies.

Duration of fever / treatment failure

Duration of fever was not reported in the systematic review (Vidal, et al., 2004). Three of the included trials reported this outcome but none of these reported a statistically significant difference in the duration of fever between treatment groups.

Vidal, et al., (2004) reported treatment failure as a composite outcome comprising one or more of the following: death; persistence, recurrence or worsening of clinical signs or symptoms of presenting infection; any addition to or modification of the assigned intervention. Low quality evidence suggested no significant difference in the rate of treatment failure in the IV-to-oral group compared to the IV only group, RR 1.07 (95% C.I. 0.9 to 1.27).

Cost-effectiveness evidence

A literature review of published cost-effectiveness analyses did not identify any relevant papers. This topic focused on the optimal timing of a change in management strategy. The difference between strategies were considered unlikely to lead to large differences in cost, but rather be guided by differences in patient outcomes and other considerations such as service configuration. It was agreed that these considerations would probably be difficult to accurately capture using economic modelling and therefore further health economic analysis was not undertaken.

Linking Evidence to Recommendations

The aim of this topic was to identify when is the optimal time to switch (step down) from intravenous to oral antibiotic therapy.

The GDG considered the outcomes of over treatment, critical care, length of stay and quality of life to be clinically relevant to the question. No evidence was available for any of the outcomes required. Limited evidence was found relating to duration of fever. The available evidence largely reported an outcome of treatment failure, which was a composite outcome comprising one or more of death; persistence, recurrence or worsening of clinical signs or symptoms of presenting infection; any addition to or modification of the assigned intervention. The GDG agreed that this was an important and relevant outcome and used this as the basis for their recommendation.

The overall quality of the evidence classified by GRADE was ‘low’ for addressing mortality and treatment failure, and ‘very low’ in relation to adverse outcomes.

The GDG noted that mortality for patients switching to oral antibiotics was low and equivalent to that of patients receiving intravenous antibiotics. In addition, the clinical experience of the GDG was that switching to oral antibiotics would probably be beneficial to patients because they would spend less time in hospital and have reduced exposure to broad spectrum IV antibiotics – with a corresponding reduction in side effects and risk of developing antimicrobial resistance.

The GDG also noted that the evidence only included patients who had been classified as low risk at the time of the decision to switch to oral antibiotics. The clinical experience of the GDG was that switching to oral antibiotics was not appropriate for patients at high risk of complications. The GDG recognised that in studies which undertook an early switch patients were more likely to have treatment failure than those with a later time of switch. Based on their clinical experience the GDG agreed that most adverse events would be clinically apparent within the first 48 hours of admission, so there would be less risk associated with switching after this time.

The recommendations allow for stepping down to oral antibiotics with or without discharge of patients. This is because while most patients who could be discharged early are able to tolerate oral antibiotics, some may have a specific contraindication which requires IV antibiotics. The social circumstances of some patients may mean they are not able to be discharged but are still able to step down to less resource intensive regimens for example oral antibiotics.

The GDG noted that a literature review of published cost effectiveness analyses identified one relevant paper, however this paper was excluded due to serious selection bias. No additional economic analysis had been undertaken. The GDG agreed based on their opinion that a continued intravenous strategy would probably be more costly than switching to oral antibiotics.

Therefore the GDG decided to recommend that patients who have reassessed as being low risk of severe sepsis using a validated risk scoring system should switch to oral antibiotics after 48 hours. Since the studies appraised did not show striking differences in outcomes according to age, the GDG decided not to make a separate recommendation for children. The GDG were aware that local microbiological resistance patterns vary and choice of antibiotics may be influenced by prior quinolone prophylaxis. Consequently they were unable to recommend a specific antibiotic strategy but for those patients without prior prophylaxis oral antibiotic regimes with a quinolone and/or co-amoxiclav have been most frequently used.

The GDG noted that there was potential to achieve very large gains in improved patient experience by switching to oral antibiotics after an even shorter time period than recommended (for example after 8-16 hours). However there is currently no strong evidence in this area. The GDG therefore decided to recommend further research.

Clinical evidence (see also full evidence review)

Evidence statements

Two randomised trials compared early discharge with continued inpatient care in adults (Innes, et al., 2003) or children (Santolaya, et al., 2004) treated for neutropenic sepsis. There was very sparse evidence about the relative effectiveness of early discharge and continued inpatient care in terms of short term mortality and hospital readmission. This evidence is summarised in Table 7.3.

Table 7.3

GRADE profile: Is early discharge more effective than continued inpatient care in patients receiving empiric treatment for neutropenic sepsis.

Early discharge rates

In four observational studies the percentage of adult patients meeting the criteria for early hospital discharge ranged from 38% to 90% (Cherif. et al., 2006; Girmenia, et al., 2007; Klastersky, et al., 2006 and Tomiak, et al., 1994). In order to be discharged early, low risk patients were required to meet additional criteria including ability to tolerate oral antibiotics, no history of poor compliance and ability to read a thermometer. The percentage of patients who were actually discharged early ranged from 13% to 69% (Cherif, et al., 2006; Girmenia, et al., 2007; Klastersky. et al., 2006 and Tomiak. et al., 1994).

In eleven observational studies the percentage of paediatric patients meeting the criteria for early hospital discharge ranged from 27% to 63% (Lau, et al., 1994; Dommett, et al., 2009; Lehrnbecher, et al., 2002; Bash, et al., 1994; Tordecilla, et al., 1994; Aquino, et al., 1997; Mullen, et al., 1990; Griffin, et al., 1992; Wakcker, et al., 1997; Hodgson-Veiden, et al., 2005 and Santos-Muchado, et al., 1999). Most of these studies were retrospective and patients were not prospectively assigned to high/low risk groups. These studies reported the outcomes of those who were actually discharged early, which ranged from 19% to 68%.

Hospital readmission

In the Innes, et al., (2003) randomised trial, 5% of patients discharged early required hospital readmission.

In four observational studies (Cherif. et al., 2006; Girmenia, et al., 2007; Klastersky, et al., 2006 and Tomiak, et al., 1994) the rate of hospital re-admission for adult patients discharged early ranged from 0% - 13%. Re-admission rates ranged from 0% - 9% in eleven observational studies of paediatric patients (Lau, et al., 1994; Dommett, et al., 2009; Lehrnbecher, et al., 2002; Bash, et al., 1994; Tordecilla, et al., 1994; Aquino, et al., 1997; Mullen, et al., 1990; Griffin, et al., 1992; Wakcker, et al., 1997; Hodgson-Veiden, et al., 2005 and Santos-Muchado, et al., 1999).

Short term mortality

Patients selected for early discharge were at low risk of adverse events thus mortality data were sparse: in the Innes, et al., (2003) trial there were no deaths during follow-up. The reported short term (within 30 days of follow up) mortality rate was 0% for patients discharged early from hospital in all but one study of adult patients (Tomiak, et al., 1994). This study reported one death (a mortality rate of 3%). This was the only study of adult patients that did not use the MASCC criteria to stratify patients according to risk.

The reported short term mortality rate was 0% for patients discharged early from hospital in all studies of paediatric patients.

Quality of life and overtreatment

These outcomes were not reported by any of the identified studies of adult or paediatric patients.

Cost-effectiveness evidence

A literature review of published cost-effectiveness analyses did not identify any relevant papers. As there was no definition of what constitutes a specific inpatient management strategy for this question, costing and evaluating health outcomes using economic modelling was not feasible. The GDG anticipated that the different management strategies were unlikely to result in large differences in patient outcomes and those strategies that minimise or reduce the duration of inpatient care would generally be less costly. Given that there was little uncertainty surrounding the economics of this question, further health economic analysis was not undertaken.

Linking Evidence to Recommendations

The aim of this topic was to define the optimal duration of inpatient care for adults and children with neutropenic sepsis to avoid any adverse experiences or outcome. For this topic the GDG considered the outcomes of overtreatment, death/critical care, quality of life, re-admission rate and adverse events (hospital acquired infection) to be the most relevant.

No evidence was found for overtreatment, quality of life or adverse events. Evidence was reported on the re-admission rate and death/critical care for those patients that were discharged early. The overall quality of the evidence as classified by GRADE across all outcomes was “low” to “very low”.

The evidence identified two RCTs that addressed the question of inpatient duration in the management of suspected bacterial infection in children and adults with low-risk febrile neutropenia. However the majority of the evidence for this topic was derived from large retrospective case series. The GDG acknowledged that much of the evidence base for this question came from specialist centres and were cautious as to how the findings should be extrapolated across all settings.

From the available evidence the GDG were unable to define an optimum duration of inpatient care for patients receiving empiric treatment for neutropenic sepsis. Instead the GDG focused their discussion on when these patients could be safely discharged from hospital.

The recommendations allow for discharge of patients and/or stepping down to oral antibiotics. This is because while most patients who could be discharged early are able to tolerate oral antibiotics, some may have a specific contraindication which requires IV antibiotics. However, these patients can be discharged if facilities exist to deliver outpatient IV antibiotics.

The GDG noted that no relevant, published economic evaluations had been identified and no additional economic analysis had been undertaken in this area. The opinion of the GDG was that there may be potential cost implications for carrying out appropriate risk assessment in secondary care. However they also expected that discharging patients early could bring cost savings particularly via a reduction in hospital stay.

Therefore the GDG recommended that patients receiving empiric treatment for neutropenic sepsis and who have been reassessed as being low risk of complications using a validated risk assessment tool (Section 4.3) and taking into account their social and clinical circumstances can be discharged from inpatient care.

Clinical evidence (see also full evidence review)

The evidence is summarised in Table 7.4.

Table 7.4

GRADE profile: What is the optimal duration of empiric antibiotic therapy in patients with neutropenic sepsis.

Cost-effectiveness evidence

A literature review of published cost effectiveness analyses did not identify any relevant papers. This topic focused on the optimal timing of a change in management strategy. The difference between strategies were considered unlikely to lead to large differences in cost, but rather be guided by differences in patient outcomes and other considerations such as service configuration. It was agreed that these considerations would probably be difficult to accurately capture using economic modelling and therefore further health economic analysis was not undertaken.

Linking Evidence to Recommendations

The aim of this topic was to identify the optimal duration of empiric antibiotic therapy in patients with neutropenic sepsis.

The GDG considered the outcomes of over-treatment, death/critical care, length of stay, duration of fever and quality of life to be important to the question. Over-treatment and quality of life were not reported in the evidence. There was limited data on mean length of stay and duration of fever. The main outcome reported by the evidence was death. However, due to very low event rates and methodologically compromised trials, the evidence on this outcome was classified by GRADE as being of ‘very low’ quality.

The GDG noted that the evidence was insufficient to determine whether stopping empiric antibiotics early was more or less effective than continuing empiric antibiotics until the patient was afebrile with a recovered neutrophil count. Nor did the evidence indicate whether or not these two strategies were equivalent. The GDG noted that consideration would have to be given to other causes of infection or fever but making recommendations on this was outside the scope of the guideline.

Based on their clinical experience, the GDG agreed that prolonged antibiotic treatment was associated with organ toxicity, increased side effects and increased risk of super-infection with fungi and/or multi-resistant organisms. Conversely, early discontinuation of treatment risked patients having relapsed/recurrent infection or significant complications such as endocarditis, osteomyelitis and abscesses. The GDG noted that relapsed infection needs to be distinguished from a new infective episode, and the studies reviewed were inadequate to assess this.

The clinical experience of the GDG was that stopping antibiotics earlier would probably be beneficial to patients because they would have reduced exposure to antibiotics, a corresponding reduction in side effects and reduced risk of developing antibiotic resistance. The patient experience of the GDG was that spending less time in hospital was preferable.

The GDG noted that no relevant, published economic evaluations had been identified and no additional economic analysis had been undertaken in this area.. The GDG considered based on their clinical experience that stopping antibiotics earlier would also probably reduce costs because patients would spend less time in hospital and there would be a reduction in spend on antibiotics and treating their associated side effects. The GDG felt that this reduction in cost would probably be greater than any additional costs associated with patients discontinuing treatment too early.

Therefore the GDG decided to recommend that empiric antibiotics should be continued in persistently febrile, but clinically stable patients, unless an alternative source of fever is established. The GDG also agreed to recommend that antibiotics could be discontinued in patients who have clinically responded, irrespective of neutrophil count.