Advantages and disadvantages of shorter antimicrobial courses

Abstract 

The rationale for shorter course antimicrobials is supported by pharmacokinetic-pharmacodynamic principles, which emphasise the importance of high initial drug concentrations for effective bacterial eradication. By limiting prolonged exposure to antibiotics, shorter courses help reduce the selection pressure on bacteria, mitigating the development of antimicrobial resistance. Studies evaluating resistance outcomes demonstrate that shorter durations are associated with decreased resistance development compared to longer courses. For many studies, clinical cure is non-inferior in recruited patients.

For uncomplicated bacterial infection in patients with a good initial clinical response, short-course antimicrobials should be the standard of care. For more complex patients, caution is required when interpreting study findings owing the inclusion/exclusion of patients studied. Tailoring antimicrobial therapy to each patient depending on initial treatment response is preferable to ensure optimal treatment outcomes. Further research and prospective studies are necessary to expand the evidence base and guide clinical practice in this area.

Key words: Shorter course antimicrobials, antimicrobial resistance, AMR, pharmacokinetic-pharmacodynamic principles, patient tailored antimicrobial prescribing

Antimicrobial resistance (AMR) is a global concern threatening patient safety directly, through infection and has a secondary impact on patients undergoing chemotherapy, surgery, transplantation and other areas of care reliant upon effective anti-infective therapy^​1​. AMR has a large impact on the most vulnerable patients in hospitals, notably those who are immunocompromised or who have other chronic health conditions. These patients are at increased risk of complications, hospitalisation, and death^​2​. It has been estimated that by 2050, if solutions to AMR are not found, up to 10 million deaths each year and a cumulative £73tn global cost will be directly attributed to this issue^​1​.

Shortening the course length of antimicrobial therapy for common infections has emerged as a potential strategy to combat AMR. By reducing the overall exposure to antimicrobials at patient level, this approach is aimed to decrease the selective pressure that drives AMR development. However, limited supporting evidence of treatment efficacy and safety in some high-risk patient groups may hinder the widespread adoption of shorter treatment regimens. This review will examine the evidence supporting shorter antimicrobial courses as a tool for mitigating AMR, exploring both potential benefits and challenges.

How does antimicrobial resistance develop?

Under the selective pressure of antimicrobial exposure, susceptible bacteria are killed or inhibited, while bacteria that are naturally (or intrinsically) resistant multiply^​3​. The most widely recognised mechanisms contributing to the development of resistance involve random mutations occurring within a population, ranging from single-base substitutions to extensive genome rearrangements. When a mutant bacterium gains a competitive advantage, often demonstrated by its ability to survive exposure to a lethal antibiotic, these resilient mutant cells persist and endure. As a consequence, they multiply and establish themselves as the prevailing population^​4​.

The primary source of AMR stems from the transmission of naturally occurring resistance mechanisms. The majority of antimicrobial drugs used clinically are derived from natural sources, such as environmental fungi and saprophytic bacteria, with only a few being wholly synthetic, like sulphonamides and fluoroquinolones^​3​. These organisms have evolved in a competitive world, constantly battling other microbes for survival. To gain an advantage, they produce substances that inhibit or kill their rivals. Others organisms have acquired protective (or resistance) mechanisms against these threats, including protecting against cell entry, the production of degrading enzymes, or altering target sites for these natural antibiotics^​3​. Horizontal gene transfer enables the rapid dissemination of these resistance traits among bacteria, providing a selective advantage. When these resistance mechanisms impose minimal fitness costs, they become endemic within pathogens, exacerbating AMR in healthcare settings^​3​.

Since the introduction of antibiotics in the 1940s, we have embarked upon a global experiment in evolutionary selection pressure, administering vast quantities of antibiotics to combat infections in patients and foster growth in animals utilised for food production^​5​. This exposure has led to the selection and amplification of naturally occurring antimicrobial molecules derived from microorganisms, through the dynamics of Darwinian competition^​6​. Consequently, continued exposure to antimicrobials has resulted in the emergence of resistant bacteria both in the environment and among individual patients^​7​.

Cumulative exposure, measured by the total number of days of antimicrobial treatment, predicts this outcome, with prolonged systemic treatments associated with an increased relative risk of future AMR and antimicrobial-associated infections, such as Clostridioides difficile infection^​8–10​.  The risk of selective pressure on both antimicrobial-resistant strains and C. difficile infections escalates with each additional day of antimicrobial treatment^​11,12​. Moreover, national-level antimicrobial prescribing practices exhibit a correlation with national-level AMR trends, as countries with high rates of antimicrobial prescribing witness elevated levels of sentinel AMR^​13,14​. This is compounded by transmission of resistant pathogens from patient to patient in some settings, often as a result of breakdown in infection control practice, most commonly inadequate hand hygiene. International travel and health tourism have accelerated the transmission of resistant organisms worldwide, compounding this issue further^​15​.

Reducing antimicrobial selective pressure through shorter courses

In response to this known association with prolonged and excessive treatment durations, there has been a renewed interest is optimising treatment durations. The traditional treatment durations for common infections such as pneumonia, urinary tract infection and bacteraemias are derived from expert opinion, rather than empirical evidence-based recommendations. These ‘optimal’ treatment durations are passed down from generation to generation in local treatment guidelines and as part of training programmes^​10​. Most durations are derived not from clinical efficacy studies but rather the days in a week or number of fingers on one’s hand (e.g. durations of 5, 7, 14 or 28 days)^​10​. Clinicians believed that longer treatment courses were necessary to ensure complete eradication of pathogens and prevent disease relapse^​16​. This approach was rooted in the limited understanding of pharmacokinetics, pharmacodynamics and the complex nature of bacterial infections at the time. Physicians aimed to eliminate the last remaining bacteria, fearing that a shorter duration would leave behind a reservoir of potentially resistant organisms^​3​. More information on how to evaluate the clinical appropriateness of an antimicrobial can be found here.

The advent of controlled clinical trials and the understanding of concentration-dependent killing and post-antibiotic effects challenged the traditional beliefs advocating longer treatment courses (10). Early studies questioned the necessity of prolonged durations, paving the way for evidence-based medicine and the gradual accumulation of scientific evidence supporting shorter treatment courses^​17,18​.

The mantra of ‘shorter is better’ publications have caught the imagination of the infectious diseases and antimicrobial stewardship community. Many vocal advocates are looking to translate this mantra into clinical practice^​10,19​.

Evidence supporting shorter course antimicrobials

The evidence base supporting shorter course antimicrobials as non-inferior options to reduce AMR has grown substantially in recent years. Several clinical trials, systematic reviews and meta-analyses have demonstrated that shorter treatment durations can achieve comparable clinical outcomes to longer courses while minimising the risk of AMR development^{​10,19–22​}. Here, we will review this evidence focusing on common conditions. 

COPD exacerbation and acute uncomplicated bronchitis

Most infections are self-limiting and treated in the community setting. Rates of confirmed bacterial infection are low with most caused by viral pathogens^​23,24​. The true benefits of antimicrobial therapy for these conditions may be debated; however, more than 25 randomised control studies (n =>10,000 patients) have demonstrated non-inferior clinical outcomes using a five-day course length compared with seven days or longer, thus where antimicrobials are used the shortest effective course length is advised^{​21,25,26​}. This practice has been widely adopted within the UK with National Institute for Health and Care Excellence (NICE) guidance advising this short-course option^​27​. Improved diagnostic stewardship to exclude bacterial infections and thus avoid any antibacterial usage is desirable, but it is often impractical in busy clinical settings^​28​.

Urinary tract infection

In women with uncomplicated bacterial cystitis, clinicians should prescribe short-course antibiotics with either nitrofurantoin for five days, trimethoprim for three days, ciprofloxacin for three days, a beta-lactam for three days or fosfomycin as a single dose^​29​. Published data supporting short course therapy are weighted to young healthy females with cystitis; less in known for older and more co-morbid patients^​17​. Complicated urinary tract infections, including patients with structural abnormalities, urinary catheters, pregnancy and male patients, are expected to benefit from longer treatment courses (seven days)^​30​.

National guidance for uncomplicated UTIs is supportive of these shorter course durations and UK practice has adopted these changes^​31​. Supporting evidence for three-day nitrofurantoin in uncomplicated UTI is limited with studies showing inferior outcomes against five day or longer^​32​. However, in clinical practice, nitrofurantoin is widely prescribed for three days across the UK in line with NICE guidance. 

For UTIs with signs of systemic symptoms (e.g. fever, haemodynamically unstable), including pyelonephritis with or without a concurrent bacteraemia, the emerging literature supports further opportunities for short-course options^{​10,22,30,33–35​}. Several studies have demonstrated that 5 to 7 days of antimicrobials are non-inferior to longer courses (typically 14 days)^​33,36​. Quinolones are the most commonly utilised therapies in these studies; beta-lactam and co-trimoxazole have also been included^​31​and show non-inferior outcomes when used for durations of 7 days or less^{​37–40​}.

These studies have strict inclusion and exclusion criteria that limit their applicability to clinical practice. Patients with comorbidities — such as malignancies, structural abnormalities of the urinary tract, immunosuppression or pregnancy — are generally excluded from these prospective studies^{​33,36,40​}. Additionally, the trial designs may introduce study bias favouring the intervention arm. Most study designs randomise patients after 48 to 96 hours post treatment, including only those patients with evidence of early clinical response. The more complex patients and those with delayed clinical response are often excluded. The applicability of these studies and short-course therapies for the wider population has been debated^​41​. How we translate these new research findings and apply to our patients in the clinical setting remains challenging. Shorter courses should be used wherever possible, but more complex cases must receive tailored treatment durations in response to their clinical response. 

Understandable enthusiasm for adopting these short course therapies for pyelonephritis has been seen nationally. Quinolones are recommended for seven days in national guidance, whilst longer course (seven to ten days) are advised for beta-lactam options owing to the lower numbers of studied patients receiving non-quinolone options^​31​.

Trimethoprim-based options are recommended for treatment for 14 days in national guidelines^​29​. Shorter courses of trimethoprim (often as co-trimoxazole) have been used in clinical practice based on the emerging supporting evidence base^​33,38​. Yet some exceptions may exist to these empiric recommendations. In patients with ongoing evidence of systemic infection beyond 72 hours and  those with uncontrolled source of infection, defined here as the presence of renal abscess, stones or retention of prosthetic material, are not discussed in these guidelines and they may warrant treatment of up to 14 days, or more, owing to high expected complication rates^​42​. Personalising treatment duration dependant on individual patient response and time to adequate source control is therefore recommended^​43​.

Community-acquired pneumonia

Many randomised controlled studies have demonstrated non-inferior outcomes with short-course antimicrobial therapy for the treatment of community-acquired pneumonia (CAP) (n = >12,000 patients)^{​25,44,45​}. The data are skewed by paediatric-focused studies where a lower mortality burden, fewer comorbidities and a greater viral co-infection rate limit wider interpretation^{​20,45,46​}. Studies focusing on hospitalised adult patients, which accounts for the majority of treated CAP cases, is less readily available but the following recent studies have helped address this^{​47–49​}.  

In 2006, el Moussaoui et al. showed non-inferior outcomes with three (n=63) versus eight days (n=56) of antimicrobials in adults admitted with non-severe CAP^​49​. In this study design, antimicrobial therapy was stopped after 72 hours in patients in the intervention arm demonstrating a good clinical response. Clinical cure rates were high for both groups (>90%) and strict exclusion criteria (e.g. recent admissions, immunosuppressed patients and previous antibacterial treatments) limit wider applicability of this study.

This study design was replicated in a similar adult CAP population by Dinh et al., again randomising to cessation of antibacterials in patients with clinical resolution of infection after three days (n=310)^​47​. Here, in hospitalised patients with a CAP diagnosis, a short-course therapy (three days) in patients with initial clinical response to empiric antimicrobial therapy was sufficient^​47​. Uranga et al. compared a five (n=162) versus ten-day course (n=150) of antimicrobials for adults with CAP and found non-inferior clinical outcomes with the shorter course therapy^​48​. Rates of clinical cure are similar between the two study groups and further support for shortening treatment durations. Therefore, where adult patient with CAP are seen to be responding to empirical antimicrobial therapy, short course therapy should be recommended^​44​.

Five-day course lengths are currently advised in national guidelines for mild-to-severe pneumonia^​27​. Based on the emerging evidence, even shorter courses or early cessation of antimicrobial therapy may be suitable for adults with early clinical response or proven non-bacterial infection, respectively. 

Short-course (three days) antibacterials for paediatrics have more robust supporting data and can be recommended for non-severe infection^​45​. The UK based CAP-IT study shows non-inferior outcomes with three days (n=814)^​46​. Generally, rates of bacterial infection in children are low with viral infection usually implicated^​50​. Where antibacterials are used, the shortest effective duration in children is advised.

Cellulitis

For skin and soft tissue bacterial infection, four prospective studies have assessed the effectiveness of short-course antibacterial durations^{​50–53​}. These studies are dominated by quinolone or oxazolidinone-based options (e.g. linezolid and tedizolid), which exhibit unique pharmacokinetic properties^​53,54​. Oxazolidinones are commonly restricted to hospital-only usage and rarely used as first-line therapy for cellulitis; quinolones are reserved for complex Gram negative infections in UK practice^​55​. Instead, beta-lactams such as flucloxacillin are recommended for empiric management of cellulitis in the UK.

The DANCE study investigated clinical cure in adults randomised to 6 versus 12 days of flucloxacillin (n=151)^​51​. While cure rates were similar, rates of re-treatment were higher in patients stopping treatment at day 6 (24% versus 6%). The study was under recruited, therefore statistical analysis on outcomes should be interpreted with caution. The national guidance recommends 5 to 7 days of beta-lactams and macrolides for cellulitis^​56​. For severe infection, delayed clinical response to treatment or in patients with immunocompromised status (e.g. diabetes, chronic kidney disease) extended course lengths (10 to 14 days) may be required depending on clinical response^​56​.

Limitation of available short-course antimicrobial studies

The evidence from these randomised control trials supports shorter course lengths of antimicrobials for many of the common infections; however, there are some concerns around the applicability of this work to certain patient sub-groups^​41​. Some patient populations — such as the critically ill, immunocompromised and multi-morbid patients — are not well represented in these studies owing to design and recruitment, and thus the safety of shorter course durations for these high-risk groups in clinical practice is less clear. Fixed treatment durations, whether adopting the shorter or traditional longer courses, are likely imperfect for all patients^​43​. Instead, work is required to implement shortened antimicrobial courses into clinical practice in a safe and effective manner. A patient-tailored approach is preferable, adjusting the treatment duration depending on individual patient response. Yet operationalising this into clinical practice is challenging owing to the scale of infected patients and lack of routine follow-up for many treated patients. Clinical decision support tools (CDSS) and targeted biomarkers (e.g. procalcitonin [see below]) may help address this gap in the future^{​57–59​}. The use of CDSS has been described as the response to several AMS challenges but actual data is currently lacking.

Using biomarkers such as procalcitonin and C-reactive protein to guide treatment duration

There is much interest in utilising biomarkers to exclude bacterial infection and/or guide treatment duration^​60​. Procalcitonin, the precursor of the hormone calcitonin, has been used as a biomarker to aid in diagnosis of bacterial infection or sepsis, as well as in differentiating bacterial pneumonia from viral pneumonia and chronic obstructive pulmonary disease (COPD)^​61​. Prospective studies have demonstrated that use of procalcitonin as part of structured antimicrobial reviews can reduce the duration of antimicrobials; one prospective randomised controlled trial in 15 hospitals in the Netherlands demonstrated a reduction in antimicrobial durations of 1.22 days (p<0.0001)^​59,62​. A negative serum procalcitonin (<0.25pg/mL) is associated with a low probability of invasive bacterial infection, which can help differentiate non-bacterial causes of inflammation in patients.

While currently not recommended by NICE, procalcitonin use has increased in UK hospitals in response to the COVID-19 pandemic as a mechanism for excluding bacterial co-infection in patients with SARS CoV-2^{​57,63–65​}. Access to procalcitonin in hospitals allows antimicrobial stewardship teams the opportunity of expanding use for excluding bacterial infection in patients with non-specific signs of inflammation and guiding antimicrobial durations.

C-reactive protein (CRP) is an acute-phase reactant that can be produced in response to inflammation, infection, tissue damage and malignancy. In healthy patients, a low serum CRP concentration is found (<5 mg/L); higher levels (>100 mg/L) are associated with bacterial infections and can be used as part of initial diagnosis and treatment response of infections; however, CRP is not specific to bacterial infection and levels can be raised as part of many systemic inflammatory disease states (e.g. viral infection, pancreatitis and auto-immune diseases). Despite its widespread use, limited prospective studies are available to confirm its utility in clinical practice^​60​. Where studied, the use of CRP is associated with a reduction in antimicrobial duration^​66,67​. Other experimental biomarkers (e.g. IL-6) have yet to be validated in clinical practice^​68​.

Implementing short-course antimicrobial evidence base into clinical practice.

Host factors — such as age, co-morbidities, and immune status — will impact on patient treatment response. Additionally, pathogen factors — such as resistance, bacterial burden and virulence —  also impact on treatment response^​3​. The heterogenous nature of bacterial infection, through these variables, contradicts the ‘one size fits all’ approach to antimicrobial durations. Determining the optimum antibacterial treatment and its duration at the time of initial presentation is increasingly challenging. National programmes such as the ‘Start Smart then Focus’ toolkit (for secondary care) have helped prescribers address this^​69​. After 48 to 72 hours of empiric therapy, the initial clinical response can be observed and an informed duration of treatment can be applied^​70​.

This practice is more in keeping with the described prospective studies above where decisions made on duration are based on initial treatment response. Local antimicrobial guidelines can be adapted to provide variable durations depending on the individual patient’s response at this 48 to 72 hour review. While this works well in the inpatient setting, adapting this to outpatient or community prescribing is less straightforward. Here, empiric durations must be set by the prescriber at day zero without the benefit of being able to observe the initial clinical response. Therefore, more cautious prescribing may occur with longer durations selected. This remains a focus for future research.

Conclusion

The evidence base for shorter course antimicrobial therapy has grown significantly, showing non-inferior outcomes compared to longer courses for common infections, such as pneumonia, urinary tract infection and cellulitis. This supports the introduction of short-course antibacterial therapy into clinical practice to reduce unnecessary antibacterial use. There are limitations to the current evidence base, particularly regarding the applicability of these studies to specific patient sub-groups, such as the critically ill, immunocompromised and multi-morbid patients. The optimum duration of antibacterial therapy in these complex populations remains unclear. Here, a patient-tailored approach, adjusting the treatment duration based on the patient’s response, is likely to be needed.