Mechanisms of hypervirulent Clostridium difficile ribotype 027 displacement of endemic strains: an epidemiological model

Abstract

Following rapid, global clonal dominance of hypervirulent ribotypes, Clostridium difficile now constitutes the primary infectious cause of nosocomial diarrhoea. Evidence indicates at least three possible mechanisms of hypervirulence that facilitates the successful invasion of these atypical strains: 1) increased infectiousness relative to endemic strains; 2) increased symptomatic disease rate relative to endemic strains; and 3) an ability to outcompete endemic strains in the host's gut. Stochastic simulations of an infection transmission model demonstrate clear differences between the invasion potentials of C. difficile strains utilising the alternative hypervirulence mechanisms, and provide new evidence that favours certain mechanisms (1 and 2) more than others (3). Additionally, simulations illustrate that direct competition between strains (inside the host's gut) is not a prerequisite for the sudden switching that has been observed in prevailing ribotypes; previously dominant C. difficile strains can be excluded by hypervirulent ribotypes through indirect (exploitative) competition.

Figure 1. The Clostridium difficile epidemiological state transitions simulated by the stochastic model.

The left column details colonisation with ‘typical’ endemic strains and the right column details colonisation with hypervirulent strains. From reviewing the literature, three possible areas were identified that differentiated transitions in hypervirulent strains from endemic strains: 1) Hypervirulent strains are more infectious (modelled by increasing the transmission coefficient); 2) Hypervirulent strains give rise to a higher symptomatic rate; and 3) Hypervirulent strains can outcompete endemic strains in the host’s gut.

Figure 2. Proportion of hypervirulent Clostridium difficile introductions that successfully invade (along with fitted logistic regression curves; black left y-axis) and the new level to which they equilibrate (per 10,000 individuals, red right y-axis).

Output from 1000 simulated introductions are displayed for each parameter level under all mechanistic scenarios. Left plot: hypervirulent strains are more infectious (modelled by increasing the transmission coefficient); Middle plot: hypervirulent strains give rise to a higher symptomatic rate; Right plot: hypervirulent strains can outcompete endemic strains in the host’s gut.

Figure 3. Frequency distribution of time (in months) taken for a newly introduced hypervirulent strain of Clostridium difficile to establish new equilibrium under the different mechanisms of hypervirulence.

Output from 1000 simulated introductions are displayed for each parameter level under all mechanistic scenarios (note the reversed parameter axes for β and for α which was done to display results more clearly).

Figure 4. Successfully invading hypervirulent strains of Clostridium difficile (left column, red) competitively exclude pre-existing endemic strains (right column, blue).

The fold increase in infectiousness relative to endemic strains is shown in the top left of the hypervirulent plots. A one-month moving average of the successful hypervirulent strain invasions from 1000 simulated introductions are displayed (with shaded 95% confidence intervals).