There is a range of preclinical study evidence that suggests that interactions between Candida with other bacteria in the patient microbiome has the potential to promote invasive bacterial infections. The basis for the interaction may be molecular 7 or mechanical.8
However, there are several obstacles to defining the clinical relevance of any interactions between Candida and bacteria in the patient microbiome. Candida colonization has several predictors 1, 2 some of which, such as prolonged antibiotic exposure, have broad effects on the microbiome. Candida and bacterial colonization are problematic to define. VAP is an imprecise and somewhat subjective endpoint. Various specific Gram-positive bacteremias are each uncommon, as is candidemia. Data arising from single center clinical studies, and especially so infection data, will exhibit dependency and may not be generalizable. Finally, for S aureus bacteremia, being a relatively rare end point with a benchmark incidence of 1.8%, even multi-center studies may be underpowered to show any interaction with Candida colonization.45 Moreover, other Gram-positive bacteria such as Enterococci are even less common than S aureus bacteremia. Finally, CNS bacteremia is variably defined in the studies with or without using CDC defining criteria.
Presumably as a result of these obstacles, attempts to define the clinical relevance of any interaction between Candida with other bacteria are scant and relate mostly to interactions with Gram negative bacteria. Some have reported conflicting results.
There is some evidence that the risk of VAP in association with Pseudomonas aeruginosa is more common in patients colonized by C. albicans46 and that antifungal treatments can reduce this likelihood.47 One study found that Candida colonization of the respiratory tract is associated with Acinetobacter VAP but not Pseudomonas VAP.48
In contrast, other attempts to define the clinical relevance of any interaction between Candida with other bacteria through either retrospective studies of the association with anti-fungal use or through studies of either pre-emptive or intensified prophylactic anti-fungal treatment 49-51 have failed to resolve the question. Several have questioned the specificity of the association and whether any association is simply a reflection of confounding by illness severity.1,52
The approach here is to circumvent these obstacles by using as a natural experiment data from > 400 patient groups from > 250 studies of infection prevention interventions among ICU patients. The various groups of these studies have been exposed to infection prevention interventions which, in conjunction with other exposures, modify the patient microbiome. Of note, any one group here could experience multiple concurrent exposures such as concomitant CRF, TAP, PPAP, anti-fungal and ICU-LOS>10 days. This is reflected in the wide range in incidences of infections across the >250 groups here.
SEM is emerging as a method to model the relationships among multiple simultaneously observed variables in order to provide a quantitative test of any theoretical model proposed within the literature.10, 11, 53 An ability to test the validity and inferred relationship of conceptual variables that cannot be directly quantified is achieved by using latent variables within the model. GSEM allows generalized linear response functions in addition to the linear response functions allowed by SEM.
Limitations.
There are six key limitations to this analysis, the first being that this analysis is a group level modelling of four latent variables; colonization with each of Candida, Staphylococcus aureus, CNS and Enterococci. These latent variables and the coefficients derived in the GSEM models are indicative and intended for internal reference only. They have no counterpart at the level of any one patient or study and cannot be directly measured.
The GSEM analysis takes a structural rather than statistical approach to the question of any interactions between Candida and Gram-positive bacteria. The structural approach means that a limited number of conceptually key group level factors were entered as simple binary variables into intentionally simplistic GSEM models. There was no ability nor purpose to adjust for the underlying patient level risk. The true relationships between exposures and outcomes will likely be center specific, complex, graded and with multiple expoure interactions. A statistical approach would use more conventional analytic methods such as meta-analysis which, being based on an assumption of exchangability between control and intervention groups that randomized assignement of exposures provides, allows more precise effect size estimates for specific individual interventions under study. However, a randomized assignement of individual exposure to Candida colonization is neither an ethical nor a practical intervention outside of a natural experiement resulting from group level exposures to various ICU infection prevention interventions.
The second limitation is that there was considerable heterogeneity in the interventions, populations, and study designs among the studies here as the inclusion criteria for the various studies have been intentionally broadly specified. This breadth is both a strength, in that the breadth of the group wide exposures is the basis for the natural experiment here, and a limitation, in that the associations for a group wide exposure may not equate to associations at the level of an individual patient exposure.
Thirdly, several assumptions have been made for studies that failed to report key exposure and outcome variables in the analysis. For example, missing data for ICU-LOS and percent receiving MV has been broadly imputed. The extracted data is drawn mostly from studies located in systematic reviews. The data is provided in sufficient detail in the ESM to enable replication of the analysis.
Fourth, there are a large number of studies not included here because the required infection count data was not reported. However, the differences between control and intervention group mean infection incidences noted here in the scatter plots (Fig 1 - 4, Fig S2-S4) are similar to the summary effect sizes for each of the three broad categories of TAP, anti-septic and non-decontamination methods, against both overall VAP and against overall bacteremia which in turn are similar to prior published effect estimates sizes seen in systematic reviews of these interventions from which most of the studies examined here were derived.19-41
Fifth, the various regimens of TAP, anti-septic and anti-fungal intervention that have been used within the various studies have been considered as similar within each category. This is a deliberate simplification as some, for example the anti-fungal regimens, targeted different body sites. Also, the duration of application of the regimens varied among the studies. On the other hand, a strength of this analysis is that it attempts to unpack the separate associations between the infection incidences and exposure to the variable exposure to the various SDD components (TAP, PPAP, anti-fungal).
Finally, the studies have been predominately undertaken within ICU’s in first world countries. It is uncertain how representative of the microbiome elsewhere in the world. There is some evidence that the bacteria that cause VAP vary in different parts of the world.54-56
A strength of this analysis is that the sensitivity of the model to the inclusion of groups with ICU-LOS < 5 days, both with and without MV, is tested. These patient groups are of interest given the targeting of these interventions to broader categories of ICU patient groups including those with an overall shorter length of stay and with or without MV.