Most RCTs have compared 10-day regimens of CDI treatment agents, and it appears that a 10–day period is sufficient to resolve symptoms in most patients. If patients have improved, but do not have symptom resolution by 10 days, an extension of treatment duration to 14 days should be considered.⁷⁴

There is currently no evidence to suggest that a shorter duration of therapy could lead to higher recurrence rates and thus further research is required.⁸⁰ It is highly recommended to adhere to generally accepted dosage regimens of currently used agents.

The use of high doses of vancomycin (500 mg orally four times daily) was previously recommended by some guidelines for the management of severe complicated CDI.⁷⁴ However, there is insufficient evidence to support the use of doses > 125 mg four times daily in the absence of ileus.⁸¹

Vancomycin at high doses has previously been recommended for fulminant CDI, despite the lack of high-quality evidence.

In the presence of ileus, vancomycin can also be given rectally although it remains unclear whether adequate concentrations of the drug are reached beyond the left colon.⁶

Despite the lack of data, it seems reasonable to administer oral and/or rectal vancomycin at higher doses for patients with fulminant CDI (500 mg 6 hourly orally and 500 mg in 100 mL of normal saline by retention enema). The use of high-dose vancomycin is safe, but it is appropriate to monitor serum trough concentrations to rule out drug accumulation.⁸²

Intravenous metronidazole (500 mg 8 hourly) should be used in addition to vancomycin.⁸³ This is particularly important in the presence of ileus as intravenous metronidazole may achieve therapeutic concentrations in an inflamed colon.

In patients not responding to vancomycin and metronidazole, intravenous tigecycline (loading dose of 100 mg followed by 50 mg twice daily) has been used as adjunctive or alternative therapy, but no RCTs have been performed to date.⁸⁴ Recently published retrospective studies have shown conflicting results regarding the efficacy of tigecycline for fulminant CDI.^85,86,87,88 Because of the poor outcomes associated with fulminant CDI, there may be a role for adjunctive tigecycline.⁸⁴ Consideration of surgical intervention and faecal microbiota transplantation (FMT) in such settings is also indicated.

Surgical intervention can be life-saving in patients with fulminant disease; timely consultation with a gastroenterologist and surgeon is critical in such clinical circumstances.

A guiding principle of infectious disease management is that specimens for diagnosis should be collected before initiation of therapy whenever possible.⁴¹ Empiric therapy for CDI may result in false-negative diagnostic test results.

Empiric therapy for CDI should be considered in all patients with mild-to-moderate disease severity if there is an anticipated delay of > 48 h or if a patient presents with severe or fulminant CDI.⁶ For all other patients, antibiotic therapy should be started after diagnosis to limit the overuse of antibiotics and the associated toxicities, which include overgrowth of multidrug-resistant pathogens.⁴¹

12 and 13. What are the optimal treatment regimens for recurrent Clostridioides difficile infection?

The management of rCDI is the same for relapse (with same strain from prior CDI episode) and re-infection (with different strain).

Recurrence rates are significantly lower following initial treatment with fidaxomicin compared to vancomycin.^78,89

A first or second recurrence may be treated with fidaxomicin or a tapered and pulsed regimen of vancomycin. In a randomised, stratified sub-study of patients with a first recurrence, a subsequent second recurrence at 28 days was less common following therapy with fidaxomicin compared to a standard 10-day course of vancomycin (19.7% vs. 35.5%; p = 0.045).⁹⁰ There is limited published evidence on the efficacy of fidaxomicin for multiply rCDI. A small multicentre retrospective review showed that the efficacy of fidaxomicin for the prevention of rCDI is reduced in multiply rCDI compared to that in initial and first recurrence episodes.⁹¹ This is expected as patients with multiply rCDI are likely to have severe gut dysbiosis.

Various vancomycin tapered and pulsed regimens have been described in the literature. One such regimen recommended by IDSA advises the following: after a 10–14-day course of vancomycin 125 mg four times a day, give vancomycin 125 mg twice daily for a week, 125 mg once daily for a week and then 125 mg every 2–3 days for 2–8 weeks.⁶ There are only a few published studies evaluating the efficacy of tapered and pulsed vancomycin treatment regimens for rCDI.^70,92,93,94 Tan and Johnson suggested that close patient follow-up and individualisation of tapered and pulsed vancomycin regimens can be associated with favourable outcomes.⁹⁴ Optimal treatment for rCDI (OpTION trial) (NCT02667418) is a randomised clinical trial currently underway. The study has 3 arms, comparing: (1) standard fidaxomicin therapy (10-day course) to (2) standard vancomycin therapy (10-day course) that is followed by tapered and pulsed vancomycin (31 days of vancomycin overall) and (3) standard vancomycin therapy (10-day course) alone, for sustained response at 59 days.⁹⁵

The rationale for the FMT recommendation is discussed in section D (The role of FMT in the treatment of CDI in adults).

Bezlotoxumab is a fully human monoclonal antibody that binds to and neutralises C. difficile toxin B.⁹⁶ This product is not yet licensed in South Africa. Two double-blind, randomised, placebo-controlled, phase-3 studies (A study of MK-3145, MK-6072 and MK-3415A in participants receiving antibiotic therapy for C. difficle infection [MODIFY I] and A study of MK-6702 and MK-3415A in participants receiving antibiotic therapy for C. difficile infection [MODIFY II]) investigated the efficacy of bezlotoxumab in the prevention of rCDI in adults with primary or rCDI who were receiving standard of care antibiotics. The difference in the rate of sustained cure (initial clinical cure of baseline CDI episode and the absence of recurrent infection for 12 weeks) between bezlotoxumab and placebo was statistically significant in MODIFY II and pooled MODIFY I and II results but not in MODIFY I.⁹⁷

The rate of rCDI was lower with bezlotoxumab compared to placebo in patients who had at least one of five predefined risk factors (age ≥ 65 years, previous CDI in last 6 months, immunocompromised, severe CDI and infection with a hypervirulent strain) for rCDI or for adverse outcomes related to CDI. Patients with three or more of these risk factors derived the optimum benefit.⁹⁸ For patients infected with a hypervirulent strain, the reduction in rCDI was not statistically significant. There was no benefit in patients who did not have any of the mentioned risk factors. Fidaxomicin was used in only 4% of patients (the majority were treated with metronidazole or vancomycin) and thus the effects of bezlotoxumab use with fidaxomicin warrant further investigation. Serious adverse events related to congestive cardiac failure (CCF) were reported more frequently in the bezlotoxumab group.⁹⁸ Whilst deemed to be an artefact by the European Medicines Agency, the reasons for the difference in CCF numbers between the bezlotoxumab and placebo groups are under investigation.

The efficacy of bezlotoxumab in special patient populations regarded as being at increased risk of rCDI requires further investigation. Some of the currently available data of post hoc analyses, with the studies not intended to show statistical significance between the groups, are as follows:

Inflammatory bowel disease: small number of IBD patients included in the study. The trend is towards reduced rCDI in the bezlotoxumab group (an absolute reduction of 27.2%; 95% CI −57.9 to 9.6).⁹⁹

Concomitant antibiotics: bezlotoxumab may be associated with a reduction in rCDI in patients receiving concomitant antibiotics for other infections during initial CDI treatment or in the 90 days following treatment for initial CDI episode.¹⁰⁰

Haematologic malignancies and solid organ tumours: lower rate of rCDI in haematologic malignancies and solid organ transplant but small patient numbers. It requires further investigation.¹⁰¹

Renal dysfunction: reduced rCDI rate in bezlotoxumab group (absolute reduction −17.1%; 95% CI −23.4 to −10.6).¹⁰²

Although CDIs have been considered nuisance infections amongst people exposed to healthcare, the epidemiology of this pathogen is changing globally. Over the last decade, the incidence of CDI has increased in European, Australian and North American countries.^14,124,140 The severity of disease and CDI-associated mortality has also increased, presumably because of the emergence and dissemination of the hypervirulent C. difficile strain ribotype 027 or NAP1.^124,141,142 Recently, C. difficile was shown to be an equally significant pathogen amongst hospitalised patients in Asian countries, although the prevalence of the hypervirulent strain was considerably lower compared to Western countries.¹⁴ In addition, community-acquired C. difficile infections amongst individuals with no healthcare exposure have now increased, accounting for almost half of all infections in some settings.¹⁸

Despite the emergence of C. difficile as a major public health threat, its burden in most developing countries remains unknown.^143,144 In South Africa, rates and the burden of C. difficile are unknown, with limited single-centre studies showing a prevalence of 9.2% – 22.0% amongst patients with diarrhoeal illnesses, and an incidence of 8.7 cases per 10 000 admissions reported for one facility.^8,127,145 Amongst these studies, two that reported CDI origin showed that healthcare-associated CDI occurred more frequently than community-associated infections, suggesting that CDI is more of a concern amongst hospitalised patients in South Africa.^8,127 Because of the lack of reports, it is unclear whether South African facilities are monitoring CDIs and whether this is an important pathogen in this setting.

The use of standardised surveillance methods and definitions allows for effective monitoring of the occurrence and spread of health threats, detection of outbreaks and monitoring of interventions. Because of the emergence of C. difficile in the last decade, the National Health Safety Network (NHSN) of the CDC and the European Centre for Disease Control and Prevention (ECDC) have published CDI surveillance guidelines, which have been widely adopted to monitor this pathogen.^59,146,147 A case is defined as a patient with diarrhoea or toxic megacolon and a positive laboratory test result for C. difficile toxin A and/or B, or detection of a toxin-producing C. difficile organism by culture or other means, or pseudomembranous colitis diagnosed by endoscopic examination or surgery, or pseudomembranous colitis diagnosed by histopathology. In settings where laboratory testing is only done on loose stools, the laboratory criterion can be used for routine surveillance. For defining the origin of CDI, time and location at symptom onset or specimen collection as well as previous history of healthcare exposure are used. Healthcare facility-onset (HO) cases are defined as symptom onset or specimen collection > 3 days after admission. Community-onset (CO) cases are defined as symptom onset or specimen collection ≤ 3 days after admission, and community-onset healthcare facility-associated (CO-HCFA) cases are defined as symptom onset or specimen collection ≤ 3 days from admission, and previous admission or discharge from a healthcare facility in the last 28 days. Rates of CDI can be expressed using different denominators. Typically, the number of cases (numerator) occurring over the surveillance period is divided by patient days (sum of the number of days each patient had an overnight stay), admissions (sum of all patients with an overnight stay), discharges (sum of all patients discharged) or bed-days (number of hospital beds multiplied by the surveillance period). For HO-CDI, incidence density (i.e. cases per 10 000 patient days) is the most beneficial as it provides risk information of C. difficile transmission per day whilst admitted and allows for comparisons between facilities or wards, which may admit patients requiring different treatments and thus length of stay. However, alternative denominators may be used depending on availability or feasibility. For CO-CDI, healthcare exposure is not a risk factor and alternative denominators may be used. As rates are determined using only incident cases, rCDI episodes must be excluded from the numerator. Recurrent cases are defined as patients with subsequent episodes > 14 days and ≤ 56 days following the initial symptoms onset or specimen collection.

In resource-limited settings, such as low-income countries, surveillance is challenging because of problems with understaffing, lack of skilled and knowledgeable personnel, lack of laboratory diagnostic capacity and lack of national guidelines.^148,149 It is therefore important that facilities in resource-limited settings individualise C. difficile surveillance, selecting components suitable to available resources. As exposure to healthcare is the most recognisable risk factor for CDI, surveillance in healthcare facilities should primarily monitor hospital-onset infections. This will allow tracking and detection of increases in the number of CDI cases.¹⁵⁰ The European Clostridium difficile Infection Surveillance Network (ECDIS-Net) conducted a study to assess different methods of surveillance and showed that with ‘Minimal’ surveillance, which imposed the least amount of workload through collection of only aggregated numerator and denominator data, incidence rates could be determined and tracked.¹⁴⁶ However, the ‘Light’ and ‘Enhanced’ surveillance methods where additional epidemiologic case information was collected, including CDI origin, microbiological and outcome data, offered more detailed information that would allow for targeted control measures, although with more workload. Facilities may also elect to perform only laboratory-based surveillance as this approach allows for diagnostic data to be used as a proxy for clinical infection surveillance and is less labour intensive than using clinical data to define cases.¹⁴⁷

Since 2017, C. difficile became a NMC. All laboratories that identify the C. difficile are required to report patients from whom the organism is isolated on a monthly basis to the National Institute for Communicable Diseases.¹⁵¹ As the main purpose of the NMC system is to conduct surveillance on a national level, it remains important for healthcare facilities to monitor their own CDI rates for timely monitoring and implementation of control measures.

Two tiers of precautions are recommended for all patients known or presumed to be colonised or infected with C. difficile.

Standard precautions are used for the protection of all persons exposed to a hazardous biological agent; they are applied to all patients and in all situations, regardless of diagnosis or presumed infection or colonisation status. Because all patients can serve as reservoirs of infectious agents, adherence to standard precautions during the care of all patients is essential in interrupting the transmission of microorganisms. Detailed information on standard precautions is available in the Occupational Health and Safety Act, 1993: Regulations for Hazardous Biological Agents.

The potential transmission role of patients asymptomatically colonised with C. difficile has recently been highlighted and illustrates the need for using standard precautions universally.⁴³

Standard precautions apply to: (1) blood; (2) all body fluids, secretions and excretions except sweat, regardless of whether they contain visible blood or not; (3) non-intact skin; (4) mucous membranes; and (5) contaminated items, whether or not gloves are worn.

Contact precautions are applied in addition to standard precautions for patients known or presumed to be infected or colonised with epidemiologically important micro-organisms that can be transmitted by direct contact with the patient or indirect contact with environmental surfaces or patient care items in the patient’s environment.

Clostridioides difficile transmission occurs most likely from direct contact through person-to-person spread, either from symptomatic patients or from asymptomatic carriers. Another reservoir that facilitates the spread is exposure to a contaminated environment. McDonald et al. suggested that the hands of healthcare workers (HCWs) are probably the main means through which the organism is spread in healthcare settings.⁶ It is thus prudent to identify patients and environments where a high bioburden would contribute to sustained transmission through horizontal transfer. Placing patients on contact precautions theoretically alerts HCWs to a heightened transmission risk and reinforces interventions, such as hand hygiene, to mitigate this risk. However, the evidence supporting this is of low quality and may reflect not only the contribution of other variables to transmission but also the rigour with which contact precautions are implemented and adhered to.^152,153 A single-centre study has demonstrated that contact precautions are not necessary for all patients and that there is a low risk of transmission (1.3%) of CDI in contact patients exposed for > 24 h through sharing a common room.¹⁵²

Despite the potential role of asymptomatic carriers in the transmission of CDI, neither the IDSA guidelines of 2018 nor the Association for Professionals in Infection Control and Epidemiology (APIC) guidelines of 2013 recommend placing asymptomatic carriers on contact precautions because of insufficient evidence on the efficacy of this measure.

Current guidelines suggest that patients with presumptive CDI should be placed on pre-emptive contact precautions pending test results and that early identification is key; ensure appropriate prompt testing of patients with an acute diarrhoeal illness, not otherwise explained (refer to the CDI diagnosis section).^6,154

Patients with CDI should ideally be accommodated in a single (private) room with a dedicated or en-suite toilet.⁶ In a resource-constrained setting, this is not always feasible and a decision as to which patients require most isolation may need to be made. In consideration of which patients require isolation, once again one needs to consider which patient poses the greatest transmission risk. In the context of transmission risk, there are potential host and microbe factors to consider. Patients with incontinence or poor hygiene would increase the risk of transmission through extensive environmental contamination. There is also evidence to suggest that patients harbouring a hypervirulent strain are more likely to transmit their strain and cause CDI.¹⁵² In the absence of isolation facilities, it is recommended to have dedicated toilets for CDI patients.

Cohorting of patients is also an option in a limited resource setting although the evidence for this is of low quality. When cohorting patients, avoid cohorting patients infected or colonised with other multidrug-resistant (MDR) organisms, for example, methicillin-resistant Staphylococcus aureus (MRSA) or VRE.⁶

There are varied recommendations on the duration of precautions in preventing CDI. The CDC and the South African Regulations for Hazardous Biological Agents recommend that contact precautions should be maintained for the duration of the illness, whilst others recommend continuing contact precautions for at least 48 h after diarrhoea has resolved.^6,155

The recommendation from the American Society of Gastroenterology is that contact precautions should be maintained until bowel movements have returned to normal.¹⁵⁶

In a prospective study of 52 patients, C. difficile was suppressed to undetectable levels in stool samples for most patients by the time diarrhoea resolved, with a mean of 4.2 days.¹⁵⁷ However, at the time of resolution of diarrhoea, skin contamination remained high at 60% and environmental contamination at 37%. Also, C. difficile was again detectable in 56% of stool specimens 1–4 weeks after treatment.

There are no studies demonstrating a reduction in CDI incidence by extension of contact precautions. However, as an additional control measure, contact precautions should be extended until discharge if CDI rates remain high despite standard control measures being implemented.

Isolation is a prevention measure used by most healthcare facilities; however, single or private rooms are not always available for infectious CDI patients.

Private rooms facilitate better infection control practices: in a cohort study of healthcare-associated CDI acquisition, higher rates of CDI were demonstrated in patients sharing a double room than in single rooms and there was a significantly higher risk of acquisition after exposure to a roommate with a positive culture result.¹⁵⁸ Admission to a C. difficile cohort ward was shown to be an independent predictor of recurrence of CDI, despite adjustment for potential risk factors, such as age, comorbidities and continued antibiotic use.¹⁵⁹ It is well recognised that the risk of acquisition of C. difficile and other MDR organisms is increased when the previous occupant of a room was colonised with the specific organism in question.¹⁶⁰ This highlights the potential role of environmental contamination in the transmission and underscores the need for appropriate isolation and environmental decontamination.

Thus, in resource-constrained settings where limited private rooms are available, prioritise patients with stool incontinence for placement in private rooms. If private rooms are not available, cohorting is indicated following the principles outlined above.

Hand hygiene is considered to be the cornerstone for the prevention of transmission of most contact-driven infections, including CDI. In routine or endemic settings, perform hand hygiene before and after contact with a patient with CDI and after removal of gloves, with soap and water. Handwashing must occur after direct contact with faeces or potentially faecally contaminated areas of the body (e.g. the perineal region).⁶

Transmission of C. difficile commonly occurs via the hands of HCWs. When gloves are not worn, 14% – 59% of HCWs’ hands are contaminated after contact with a patient.⁶ Furthermore, studies have demonstrated low rates of handwashing by HCWs, especially when washbasins are not available.⁶ Bacterial spores can be removed from hands by the physical action of washing and rinsing. Hand hygiene compliance has been vastly improved by the introduction of AHRs, which are easy to use at the point of care and which kill most vegetative bacteria and many viruses. However, C. difficile spores are highly resistant to inactivation by alcohol and ethanol treatment of stool samples is a recommended step to facilitate the culture of C. difficile from stool.¹⁶¹ Despite this, studies have not shown an association between the use of AHR and an increased incidence of CDI. It is therefore not conclusive that AHR is inferior to handwashing, despite the lack of in vitro activity of ethanol against C. difficile spores.¹⁵⁴

Thus, it is important to confirm and monitor compliance with glove use (including the safe removal of gloves) and with hand hygiene. It is recommended to wash hands with soap and water before and after providing care for CDI patients, rather than using AHR alone although this should not supplant AHR where compliance is high. Although gloving will significantly reduce the degree of contamination of HCWs’ hands, there is still an absolute requirement for optimal hand hygiene after removal of gloves.

Handwashing by patients should be actively encouraged, in particular, after using the toilet and before eating.

Clostridioides difficile produces spores that are resistant to drying, heat, detergents and some disinfectants and can therefore survive for months to years in the hospital environment. Affected patients shed spores into their immediate environment, which can serve as the source of transmission to other patients. Clostridioides difficile spores have been cultured from commodes, toilets, floors, bed rails, call buttons, sinks and over-bed tables.⁶ It is well documented that environmental contamination occurs as a result of active infection, particularly when patients have large amounts of liquid stool or stool incontinence. Environmental contamination has been found to be the highest in rooms of patients with CDI, lower in rooms of asymptomatic carriers and the lowest in rooms of culture-negative patients.⁶ Heavy contamination occurs on floors, commodes, toilets, bedpans and bed frames.¹⁵⁵

Single-use disposable equipment should be used to prevent CDI transmission wherever available. Any re-usable equipment should be dedicated to the isolation area until the patient has been discharged and/or it has been thoroughly cleaned and effectively disinfected. It is recommended that the cleaning and disinfection methodology and responsibility for decontamination are clearly addressed in standard operating procedures. Clostridioides difficile transmission has been associated with contaminated commodes, blood pressure cuffs and oral and rectal electronic thermometers.⁶

Cleaning with detergents may be insufficient for environments contaminated with C. difficile and sub-inhibitory concentrations of disinfectants may enhance sporulation.¹⁵⁵ Daily sporicidal disinfection has been associated with reductions of CDI in outbreak settings (in conjunction with other measures). Mayfield et al. demonstrated that a hypochlorite-based solution at 5000 parts per million (ppm) available chlorine (0.5% solution) reduced the incidence of CDI in a bone marrow transplant unit, where the baseline incidence of CDI was relatively high. When the original quaternary ammonium compound was re-introduced, CDI increased to almost the baseline level.¹⁶² A study by Orenstein et al. found that daily use of bleach wipes containing 0.55% active chlorine decreased CDI by 85% in two units with hyperendemic rates and another study showed a reduction in contamination of HCW’s hands when high-touch surfaces were disinfected daily with a peracetic acid-based disinfectant.^163,164 However, no studies have compared daily cleaning against terminal cleaning only with a sporicidal agent, which has also been shown to decrease the incidence of CDI.¹⁶⁵

Reductions in viable C. difficile spores in the environment have been associated with improved compliance with thoroughness of cleaning by a trained, dedicated team of environmental cleaners. Barriers to effective cleaning included insufficient time, inadequate cleaning supplies, inadequate education and poor communication.

Terminal disinfection with a sporicidal agent (in conjunction with other measures to prevent CDI) has been associated with reductions in CDI in outbreak settings. This has not consistently been shown for CDI in non-outbreak, endemic settings, where the turnover of CDI patients in a room is lower and the contamination of the room is not sufficient to cause transmission, especially when there is daily cleaning, which removes C. difficile spores. Other confounding variables in studies include the use of different disinfection products (sodium hypochlorite, phenols, peroxides and UV irradiation), applied manually or by automated systems, with different regimens of cleaning: daily cleaning alone, daily and terminal cleaning, terminal cleaning alone and ‘deep cleaning’.

A modelling study evaluating environmental strategies aimed at different agents of C. difficile transmission within a hospital environment identified daily sporicidal cleaning and screening for asymptomatic carriage as the two most effective single interventions for reducing rates of hospital-onset CDI.¹⁶⁶ The strength of this study was that different transmission events were simulated to account for the complexity of hospital IPC dynamics, and to highlight the potential role of future mathematical modelling in IPC interventional studies. The study evaluated nine different IPC interventions and the results suggested that unidentified environmental sources of C. difficile should be the primary target for intervention.

Automated disinfection technologies such as ultraviolet radiation (UVR) and hydrogen peroxide vapour (HPV) have been found to reduce viable C. difficile spores in patient rooms.¹⁶⁷ According to the IDSA guidelines, most studies have at least one significant limitation precluding any definitive statement on the use of automated disinfection technology as a core component of a CDI prevention programme.⁶ Moreover, further data subsequent to the 2018 IDSA guidelines including the Benefits of Enhanced Terminal Room (BETR) Disinfection study did not demonstrate any reduction in C. difficile acquisition and infection by the addition of UVR compared to terminal disinfection with bleach.¹⁶⁸ In a secondary analysis of this data set, which assessed the hospital-wide, hospital-acquired incidence of CDI, addition of UVR to standard disinfection reduced the incidence of CDI. It is unclear however why this benefit was not seen in the bleach and UVR study period, although it suggests some indirect effect of UVR on the acquisition of C. difficile when targeted high-risk rooms are exposed to UVR.

An outbreak is defined as an increase in the number of cases above a certain threshold of what is expected. The identification, subsequent investigation and then implementation of interventional measures require surveillance data. A risk assessment and analysis of potential contributory factors should be undertaken using surveillance data. Given the influence of diagnostics on defining a case of CDI, it is important to establish clinically whether cases truly represent CDI versus colonisation as the outbreak may be spurious, an artefact of increased testing and detection especially where highly sensitive NAAT are used as a sole means to diagnose CDI. In a suspected outbreak situation, it is recommended to ascertain and stratify all cases in terms of true hospital-onset, healthcare facility-associated community onset and community onset. Furthermore, additional surveillance rates stratification by location is recommended. Communication and relaying of information is crucial to all relevant stakeholders.

The identification of an outbreak should prompt an evaluation of all relevant IPC measures aimed at reducing rates of CDI. It is imperative that all IPC measures are reinforced and any breaches are rectified as soon as possible. It is often possible to halt an outbreak through prompt identification of lapses in IPC measures, and subsequent reinforcement of these measures. Reported studies addressing CDI outbreaks typically employ a ‘bundle’ approach to contain the outbreak. Because of the retrospective observational design of these studies, it is difficult to ascertain if any single intervention carries more weight and should be preferentially implemented over others. Furthermore, there are no studies from Africa or South Africa that take into account the unique challenges and varied hospital/healthcare facility settings, which are prevalent throughout the continent. Based on this, all the provided recommendations have a low to very low quality of evidence base, yet the strength of the recommendation is supported by these interventions having either established benefit in controlling CDI in endemic settings or a good clinical practice basis.¹⁵⁴

The rationale for including CDI rates in ASP’s metrics relates to the following:

A recent systematic review and meta-analysis found that ASPs effectively reduced the incidence of both infection and colonisation with multidrug-resistant (MDR) Gram-negative bacteria, MRSA and CDI.¹⁶⁹

In-hospital CDI rates may also be used in goal setting as national targets for reducing antibiotic-resistant infections by a certain time – in the case of the USA, a 50% reduction by 2020.¹⁷⁰

Antibiotic misuse and overuse facilitate the development of multidrug-resistant organisms (MDROs), as well as CDI infections – an antibiotic-associated adverse drug event – making antimicrobial stewardship an important synergistic HAI prevention and control strategy.^169,171

Benchmarking rates of CDI across multiple hospitals, identification of the causes of inter-hospital variability in rates and the best strategies to reduce rates seem to be important areas of outcomes research in stewardship.¹⁷²

Intuitively, the strong association between antibiotic use and CDI makes it appealing to use CDI rates as a measure of the effectiveness of antimicrobial stewardship interventions. However, as outlined in the aforementioned sections of this guidance document, there are a number of potential pitfalls that must be considered prior to utilising CDI rates as a measure of ASP. These relate not specifically to the ASP intervention(s) itself but rather to the surveillance and diagnosis aspects of CDI. These two aspects of CDI are highly variable and not currently standardised, although the purpose of this document is primarily to provide guidance on these issues, which may facilitate standardisation.

In the context of surveillance, it is crucial to understand the burden of CDI within a facility, both in terms of incident cases (hospital-acquired) and admission prevalence (community-acquired or healthcare facility-associated community onset). Without distinguishing these cases at a facility level, the impact of any ASP interventions will be unclear. Thus, the use of CDI rates as a tool to monitor effectiveness of an ASP requires a robust surveillance system that is able to accurately categorize CDI in terms of facility/ward incidence and prevalence as outlined in the current CDC MDRO protocol.¹⁴⁷ This degree of surveillance is beyond that which is recommended in this guidance document although the recommendations made here are designated the minimum requirement and facilities which have capacity to perform enhanced surveillance should aim to do so.

In the context of diagnosis, detection of toxigenic strains by highly sensitive NAAT assays does not necessarily imply active CDI with disease. The asymptomatic colonised patients or patients with diarrhoeal illness attributable to other causes who test positive for a toxigenic strain are currently included in the calculation of CDI rates. Again, the inclusion of these cases may mask any ASP interventions and the impact of CDI rates. On this basis, a conditional recommendation for inclusion of CDI rates as an ASP measurement tool is made, the caveat being that surveillance of CDI for calculation of rates must take into account both the diagnostic elements and the detailed categorisation of CDI events.

Judicious use of antimicrobial agents is a key principle of AMS and thus any avoidance of unnecessary use of antimicrobials and reducing the duration of use should be strongly and actively encouraged. Clostridioides difficile infection is an unintended consequence of antimicrobial use and, thus, limiting frequency and duration of antimicrobial use will reduce CDI rates.⁶ Although virtually all antimicrobials have been linked to CDI, certain classes are more commonly linked and have a higher propensity for the development of CDI. These include extended-spectrum cephalosporins, clindamycin and fluoroquinolones. The widespread empiric use of these agents should be avoided and they should be reserved for targeted applications. The use of multiple antimicrobials has also been shown to be an important risk factor for the development of CDI and thus it is important to reduce cumulative exposure.⁶ It is important to note that ASP-targeting antimicrobial use has been shown to have a greater impact on CDI rates when co-implemented with IPC measures, especially hand hygiene.¹⁷¹ In this context, an ASP should be seen as part of the greater IPC programme and not as a stand-alone programme. The greatest benefit in reduction of CDI rates will be realised through interventions that capture the synergistic effects of antimicrobial stewardship and IPC.

Proton pump inhibitors (PPI) are commonly prescribed drugs. However, between 25% and 70% of these prescriptions are inappropriate.¹⁷³ Meta-analyses have shown an association between PPI use and the risk of developing CDI.^174,175 These analyses, however, reported substantial clinical and statistical heterogeneity. Contributing factors may be the inclusion of studies with non-PPI acid-suppressing drugs, different definitions of CDI, community- and hospital-acquired cases, varying hospital departments, and initial and recurrent CDI.¹⁷⁶ Proton pump inhibitors can also cause diarrhoea, resulting in more frequent investigation of CDI in comparison to that in patients who are not receiving PPIs.²⁴ Confounding because of greater severity of illness in CDI patients (with more ill patients being more likely to be prescribed PPIs) compared to selected controls is an additional factor that needs consideration.²⁴

A cumulative meta-analysis that included 50 studies from 2000 to 2016 found that since 2011 the association between PPI use and CDI risk has been relatively constant, with an OR of 1.20–1.26.¹⁷⁴

A lack of RCTs or other appropriate evidence to show the causality remains, preventing the formulation of clear recommendations regarding the restriction of PPIs for CDI prevention when the PPI use is otherwise indicated. However, inappropriate PPI use must be stopped.

A Cochrane review undertaken in 2017 found that in populations with a > 5% baseline risk of CDI, probiotics were associated with a 70% reduction in CDI risk (p = 0.01; moderate quality of evidence).¹⁷⁷ In lower clinical risk scenarios, probiotics showed no benefit. A CDI incidence of > 5% is unusual even in an outbreak setting.¹⁷⁸ Limitations of this review: trials from differing clinical settings (inpatients and outpatients), and varying probiotic formulations and doses were analysed together.¹⁷⁹ In addition, the trials in the review did not include immunocompromised patients. No severe adverse events (SAE) were reported in the included trials. The authors concluded that probiotics seem to be safer in non-immunocompromised and non-severely debilitated patients.

An individual patient data meta-analysis similarly reported that probiotics are useful for CDI prevention in hospitalised patients, with a ≥ 5% risk for CDI and a lack of SAE.¹⁸⁰

The administration of probiotics can result in infection.¹⁸¹ Further research is needed to establish which patient populations will benefit most from probiotic prophylaxis and the risk–benefit ratio in immunocompromised patients.¹⁷⁹