Fig 1.
Literature selection process.
Fig 2.
Linear regression of log2(MIC) values over time for Ps. aeruginosa.
Values used in linear regression were from human clinical samples collected from approximately 1940 to 2016 (open circles). The dotted line is the linear regression line derived from the human clinical isolates. Small grey symbols are log2(MIC) values from non-clinical strains or strains isolated from animals and were not included in the linear regression shown. There are 523 log2(MIC) values compiled from 21 separate reports used in linear regression.
Fig 3.
Ps. aeruginosa with single log-normal (black dashed line) and sum of two log-normal distributions (red dashed line) fit to log2(MIC)s for 523 strains for all dates.
The summary of ten permutation runs is provided in S2 Table in S2 Appendix.
Table 1.
Population parameters for CHX susceptibility of Ps. aeruginosa.
Fig 4.
Time course of K. pneumoniae susceptibility to CHX.
The results are from a single representative linear regression analysis from among 50 permutation runs. Details of linear regression parameters are given in S5 Table in S2 Appendix. Replicate permutations sampling within the range of uncertainties for isolation dates and MIC values as reported in the literature for the 21 studies compiled for this report indicated a small slope consistent with a small degree of decrease in susceptibility to CHX over the approximately 90 year time interval of isolation dates for K. pneumoniae strains.
Fig 5.
Comparison of fits by single and sum of two log-normal cumulative distribution functions to pooled K. pneumoniae data.
The improved fit of the sum of two log-normal distributions compared to a single log-normal was indicated by differences in AIC ranging from 11 to 68 and p-values for the F-test ranging from 1 x 10−5 to 1 x 10−3. See S6 Table in S2 Appendix for details.
Fig 6.
Time course of MICs for A. baumannii for strains from all isolation dates.
Individual log2(MIC) values are shown for the 1277 readings reported in the publications compiled for this report. A representative example of linear regression from among 50 permutation runs (dashed line) is shown superimposed on the individual measurements. The parameters for linear regression and details of statistical analyses are given in detail in S8 Table in S2 Appendix.
Fig 7.
Fits of a single log-normal and the sum of two log-normal distributions to MIC data from A. baumannii strains isolated on all dates.
Parameter values for the best fit are presented in the main text and details for the comparison of fits is given in S9 Table in S2 Appendix.
Fig 8.
Time course of E. coli susceptibility to CHX over time.
A large number of E. coli strains (n = 2179) have been tested for susceptibility to CHX with results shown for strains isolated from as early as the 1940s. Linear regression indicated a negative slope of relatively low magnitude as shown by the dashed line. Parameters are given in the text and in S13 Table in S2 Appendix. Subpopulations of E. coli strains with different susceptibilities to CHX were not detected (S6 Fig, S14 Table in S2 Appendix). That is, no evidence was detected of a CHX-resistant subpopulation among more than 2000 strains evaluated.
Fig 9.
Susceptibility of S. aureus to CHX over time.
Two sets of results are shown: Strains specified as MRSA (dark blue open circles) and strains specified as MSSA (dark red open circles). Corresponding representative linear regression lines are shown as dashed lines in the same colors as the symbols. Details of linear regression analyses are given in S20 and S21 Tables in S2 Appendix.