Structural equation modelling the relationship between anti-fungal prophylaxis and Pseudomonas bacteremia in ICU patients

Abstract

Purpose: Animal models implicate candida colonization facilitating invasive bacterial infections. The clinical relevance of this microbial interaction remains undefined and difficult to study directly. Observations from studies of anti-septic, antibiotic, anti-fungal, and non-decontamination-based interventions to prevent ICU acquired infection collectively serve as a natural experiment.

Methods: Three candidate generalized structural equation models (GSEM), with Candida and Pseudomonas colonization as latent variables, were confronted with blood culture and respiratory tract isolate data derived from 464 groups from 279 studies including studies of combined antibiotic and antifungal exposures within selective digestive decontamination (SDD) interventions.

Results: Introducing an interaction term between Candida colonization and Pseudomonas colonization substantially improved GSEM model fit. Model derived coefficients for singular exposure to anti-septic agents (- 1.23; - 2.1 to - 0.32), amphotericin (- 1.78; - 2.79 to - 0.78) and topical antibiotic prophylaxis (TAP; + 1.02; + 0.11 to + 1.93) versus Candida colonization were similar in magnitude but contrary in direction. By contrast, the model-derived coefficients for singular exposure to TAP, as with anti-septic agents, versus Pseudomonas colonization were weaker or non-significant. Singular exposure to amphotericin would be predicted to more than halve candidemia and Pseudomonas bacteremia incidences versus literature benchmarks for absolute differences of approximately one percentage point or less.

Conclusion: GSEM modelling of published data supports the postulated interaction between Candida and Pseudomonas colonization towards promoting bacteremia among ICU patients. This would be difficult to detect without GSEM modelling. The model indicates that anti-fungal agents have greater impact in preventing Pseudomonas bacteremia than TAP, which has no impact.

Keywords: Anti-fungal; Candidemia; Generalized structural equation modelling; Pseudomonas bacteremia; Topical antibiotics.

Fig. 1

Theoretical model of clinical factors bearing on the interaction between Pseudomonas and candida colonization towards causing blood stream and other infections. ‘contextual’ refers to the contextual effect within each ICU setting. The blue boxes label the elements required to address the central research question here depicted by the vertical arrow labelled ‘?’. This research question would not be easily addressed within a single center study

Fig. 2

Search method, screening criteria and resulting classification of eligible studies and subsequent decant of component groups. The six steps are as follows: (1) An electronic search for systematic reviews or meta-analysis (SR/MA) containing potentially eligible studies using search terms; “ventilator associated pneumonia”, “mechanical ventilation”, “intensive care unit”, each combined with either “meta-analysis” or “systematic review” up to November 2021; (2) The systematic reviews were then searched for studies of patient populations requiring prolonged (> 24 h) ICU admission (3) The studies were triaged from the systematic reviews into one of five categories; studies in which there was no intervention (observational studies), studies of various non-decontamination methods such as methods delivered either via the gastric route, the airway route or via the oral care route, studies of anti-septic methods, studies of antibiotic-based interventions, and studies of single drug antifungal (SAF) prophylaxis. (4) All studies were reviewed for potentially eligible studies and screened against inclusion and exclusion criteria. Any duplicate or ineligible studies were removed and (5) Studies identified outside of systematic reviews were included; (6) The component groups were decanted from each study being control (rectangles), intervention (ovals) and observation (diamond) groups. The total numbers do not tally as some systematic reviews provided studies in more than one category and some studies provided groups in more than one category and some studies have unequal numbers of control and interventions groups

Fig. 3

a, b Scatter plots, on a logit scale, of the incidence proportions of Pseudomonas bacteremia (a) and candidemia (b) for groups from 289 studies as listed in Additional file 1: Tables S1 to S5. The mean proportion (and 95% CI) derived by random effect meta-analysis for each category of component (observational [Ob], control [_C] and intervention [_I]) group derived from observational [Ob], non-decontamination (non-D), antibiotic-based and single anti-fungal (SAF) studies, is displayed. In each plot, the benchmark proportion (solid vertical line) is the mean proportion derived from the observational groups. Those component groups that did (solid symbols) versus did not (open symbols) select patients with CRF’s are indicated. NCC non-concurrent control, CC concurrent control. Note that antibiotic groups received multiple exposures in association with compound regimens (e.g. SDD and SOD, which combine TAP, an antifungal together with or without PPAP)

Fig. 4

GSEM of the interaction model in relation to Pseudomonas and Candida infection data. Candida col and Pseudomonas col (ovals) are latent variables representing Candida and Pseudomonas colonization, respectively. The variables in rectangles are binary predictor variables representing the group level exposure to the following; patient selection for candidemia risk factors (CRF); trauma ICU setting (trauma50), mean or median length of ICU stay ≥ 7 days (los7), exposure to a topical anti-septic (a_S), exposure to TAP (tap), concurrency of a control group with an antibiotic-based intervention group (CC), exposure to a non-decontamination based prevention method (non-D), use of mechanical ventilation for more than 90% (mvp90) or exposure to PPAP (ppap). Note that the model factorizes exposures from compound regimens (e.g. SDD and SOD, which combine TAP, an antifungal together with or without PPAP) into singleton TAP, PPAP and anti-fungal exposures. The circles contain error terms. The three part boxes represent the binomial data for Candida and Pseudomonas VAP (v_can_n, v_ps_n) and candidemia (b_can_n) or bacteremia (b_ps_n) counts with the number of patients as the denominator which is logit transformed using the logit link function in the generalized model

Fig. 5

a, b Model predictions derived from model A (Fig. 4) for the incidence proportions of Pseudomonas bacteremia (a), and candidemia (b) for a putative group of patients in a non-trauma ICU with mean LOS > 7 days without selection for CRF. The projections are for control (top) or intervention (bottom panel) groups receiving prophylaxis with various singleton or combination interventions. In each plot, the benchmark proportion (solid vertical line) is the mean prediction derived for an equivalent NCC group without exposures. NCC non-concurrent control, CC concurrent control, non-D non-decontamination, a_s anti-septic, TAP topical antibiotic prophylaxis, amb amphotericin, ny nystatin, ppap protocolized parenteral antibiotic prophylaxis