Evaluation of Epidemiological Cut-Off Values Indicates that Biocide Resistant Subpopulations Are Uncommon in Natural Isolates of Clinically-Relevant Microorganisms

If biocides kill bacteria, and if low-level biocide exposure can train bacteria to resist them, and if resisting biocides also confers resistance to antibiotics, then every hospital handwash, every disinfected surface, and every antibacterial soap could be quietly breeding superbugs. That chain of reasoning has driven real regulatory concern for over a decade. But here's the question almost nobody has answered with actual data from actual clinical populations: Is that fear grounded in what's circulating in hospitals and communities right now, or has it been running ahead of the evidence? Morrissey and colleagues set out to find out. The four biocides at the center of this study are triclosan, benzalkonium chloride, chlorhexidine, and sodium hypochlorite. You know these compounds, even if not by name. Triclosan has been used in hand soaps and toothpastes, benzalkonium chloride is the active ingredient in many hospital surface wipes, chlorhexidine is the antiseptic used in surgical scrubs, and sodium hypochlorite is bleach. These agents are applied across healthcare, agriculture, food production, and consumer products, meaning their environmental exposure is essentially continuous. That ubiquity is precisely what makes the resistance question urgent. The scientific worry has a mechanistic backbone. Levy and others showed that active efflux pumps — molecular machinery that bacteria use to pump out toxic compounds — can confer reduced susceptibility to both biocides and antibiotics simultaneously. Sanchez and colleagues demonstrated that triclosan exposure can select Stenotrophomonas mutants that overproduce a multidrug efflux pump. Mobile genetic elements can carry resistance genes for both classes of agents at once. So the fear isn't paranoid; it has real molecular plausibility. What it lacked was population-level evidence. There was also a more basic problem: There's no agreed clinical definition of what "biocide resistant" even means. With antibiotics, decades of data on clinical outcomes, pharmacokinetics, and minimum inhibitory concentration distributions have produced breakpoints — thresholds that tell you whether a pathogen will respond to treatment. No equivalent system exists for biocides. So before you can ask how common biocide resistance is, you need a ruler. Morrissey and colleagues built one. The tool they used is called an epidemiological cut-off value, or ECOFF. An ECOFF marks the upper statistical boundary of the wild-type susceptibility distribution for a given species and agent. To calculate it, you first measure two things for each isolate: the minimum inhibitory concentration — the lowest biocide concentration that stops the microbe from growing — and the minimum bactericidal concentration, which is the concentration required to actually kill it. If you plot those values across thousands of isolates and the distribution forms a single bell curve, your ECOFF sits at the concentration that captures ninety-nine point nine percent of the population. If the distribution has two humps — two distinct clusters of isolates — you place the ECOFF between them. Anything above it is flagged as potentially resistant. It's the same logic as an antibiotic breakpoint, applied to biocides for the first time at this scale. The dataset they assembled to do this was large and deliberately diverse: three thousand three hundred twenty-seven natural clinical isolates across eight microbial species. The two headline species were Staphylococcus aureus, with one thousand six hundred thirty-five isolates collected worldwide between two thousand two and two thousand three from both hospital and community infections, and Salmonella, with nine hundred one isolates from European veterinary sources collected between nineteen ninety-nine and two thousand three. The rest of the panel included Escherichia coli, Candida albicans, Klebsiella pneumoniae, Enterobacter species, Enterococcus faecium, and Enterococcus faecalis. These were not lab-adapted strains optimized for experimental convenience. They were isolates pulled from real infections and surveillance collections — which is exactly the point. The ECOFFs this study produces are calibrated to what's actually circulating. The central result is clean: for the great majority of species-biocide combinations, the minimum inhibitory concentration and minimum bactericidal concentration distributions formed a single, unimodal bell curve. One population, clustered around a consistent susceptibility level, with no resistant tail lurking above. Across the board, minimum bactericidal concentrations were higher than minimum inhibitory concentrations. You need more biocide to kill bacteria than to merely stop them growing, which is expected — but the shape of the distributions was mostly normal. No hidden subpopulations. No evidence of widespread resistance in the wild. There were exceptions, and they matter. Bimodal distributions — two distinct humps, signaling a resistant subpopulation separate from the wild type — appeared in specific combinations. Enterobacter and chlorhexidine showed bimodality in both the minimum inhibitory concentration and minimum bactericidal concentration. Triclosan produced bimodal distributions in Enterobacter at the minimum bactericidal concentration level, in E. coli at both minimum inhibitory and minimum bactericidal concentration, and in S. aureus at both. To anchor those with numbers: the ECOFF for S. aureus triclosan was zero point five milligrams per liter for minimum inhibitory concentration and two milligrams per liter for minimum bactericidal concentration; for E. coli, two milligrams per liter and sixteen milligrams per liter, respectively; for Enterobacter chlorhexidine, sixteen milligrams per liter for both. Two humps in the data mean two biologically distinct populations — some isolates far more tolerant than the wild-type majority. Triclosan keeps appearing in those exceptions, and there's a clear reason why. Unlike chlorhexidine or bleach, which attack multiple cellular targets simultaneously, making resistance extremely difficult to evolve, triclosan hits a single enzyme: the FabI protein, a component of the bacterial fatty acid synthesis pathway. A single mutation in the fabI gene, a change in the fabI promoter, or even an extra copy of the gene can raise triclosan minimum inhibitory concentrations enough to push an isolate into the resistant subpopulation. Morrissey and colleagues confirmed exactly this in S. aureus: every isolate with a high triclosan minimum bactericidal concentration carried one of those fabI alterations. None of the phenotypically susceptible strains carried them. The mechanism is specific, heritable, and consistent with the pattern across species — E. coli isolates with reduced triclosan susceptibility also showed changes in fabI sequences. Now, the co-selection question. If a strain develops triclosan resistance through a fabI mutation, does that also make it harder to kill with antibiotics? This is the fear that started the whole regulatory conversation. What Morrissey and colleagues found was that in S. aureus, the fabI duplications linked to triclosan resistance did not correlate with increased resistance to antibiotics currently in clinical use. More broadly, because biocide-resistant subpopulations were rare across the dataset, the paper concludes that co-selection of antibiotic resistance through biocide exposure should also be uncommon in natural populations. The data, as they stand, do not support the worst-case scenario. But "uncommon" is not the same as "absent," and the authors are careful about this. They raise the concept of minimum inhibitory concentration creep — the gradual upward drift in the baseline susceptibility of an average isolate over time. A cross-sectional study captures a snapshot, not a trajectory. If biocide use is slowly nudging populations toward higher minimum inhibitory concentrations, that shift won't be obvious until you have longitudinal data to compare against. The ECOFFs proposed in this study provide exactly the baseline you'd need for that kind of surveillance. They also flag that even if resistant mutants carry fitness costs that eventually see them outcompeted, there's a window — between emergence, spread, and clinical infection — where they could still do harm. The practical upshot is this: the widespread fear that biocides are silently selecting for antibiotic-resistant superbugs across clinical microbial populations is not supported by these data. In three thousand three hundred twenty-seven clinical isolates tested across four biocides and eight species, resistant subpopulations were the exception, not the rule. The ECOFFs Morrissey and colleagues propose give the field something it has never had — a standardized, species-specific threshold for calling a microbe biocide resistant, built from real clinical isolates rather than laboratory strains. That tool now exists. What comes next is using it — routinely, over time — to watch whether the picture changes, especially for triclosan, where the mechanistic story suggests the potential for resistance is real, targeted, and worth monitoring. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.