The 72-Hour Window: How Post-Exposure Prevention Is Supposed to Work
With the contrast among syphilis, chlamydia and gonorrhoea in view, the next problem is the clock. Doxy-PEP is not defined merely by a 200 mg tablet; its subject is the event-driven sequence that connects a sexual exposure to a timely dose and then to a possible preventive effect. Following that sequence lets us see which part of the intervention is shared across pathogens—and which later differences cannot be explained by timing alone.
The dose is defined by the clock
Post-exposure prophylaxis begins with an event. A person has condomless sex, a possible route of bacterial exposure, and then takes doxycycline in an attempt to prevent infection from becoming established. The guidance specifies one 200 mg oral dose as soon as possible afterward: ideally within 24 hours, and no later than 72 hours. The dose should not exceed 200 mg in any 24-hour period.
Those rules describe the intervention that was actually studied. They are not interchangeable pieces of a general instruction to “take doxycycline around sex.” The difference between taking a dose soon after an exposure and taking it at an unspecified later time is the difference between a protocol with an evidence base and a behavior whose effect cannot simply be assumed. The 24-hour ideal and 72-hour limit give the intervention a temporal boundary.
WHO explains that this interval reflects both the protocols used in the randomized trials and doxycycline’s pharmacological profile, including differences in how the drug is distributed across mucosal tissues. That statement is deliberately useful without pretending to offer a complete pharmacokinetic map. The guideline does not provide a ranked list of tissues, concentrations or pathogen-specific cutoffs in this material. It tells us that location and timing matter, not that we can calculate a guaranteed result for every exposure from the clock alone.
The sequence also has a termination condition. Once 200 mg has been taken in a 24-hour period, another 200 mg dose is not supposed to be added in the hope of closing every possible gap. More antibiotic is not automatically more prevention. It may increase exposure to adverse effects without extending the evidence-supported benefit. The rule therefore protects two boundaries at once: it keeps dosing close to the studied regimen and limits unnecessary repeat exposure.
This is why event-driven PEP should not be quietly converted into daily PrEP. Daily use changes the amount and pattern of antibiotic exposure, while the current guidance does not establish that it produces the same balance of efficacy, safety and antimicrobial-resistance consequences. WHO explicitly says there is no current evidence to support daily doxycycline use for this purpose and does not recommend doxycycline as STI PrEP. A plausible extension is still an extension; it is not validated merely because the medicine is the same.
The boundary between PEP and PrEP is especially important for people with frequent exposures. Repeated events can make an event-driven pattern look almost continuous, but “almost continuous” is not itself a studied regimen. The evidence still has to answer how often a dose can be used, what cumulative exposure means for safety and resistance, and whether the resulting drug levels and behavior preserve the same preventive effect. Those questions are part of the reason WHO identifies dosing strategy as a research gap rather than filling it with an assumption.
The 24-hour ideal also should not be read as a precise switch that turns protection on at one moment and off at the next. A dose six hours after exposure and a dose 60 hours after exposure both fall inside the broad studied window, but the guideline does not provide a numerical estimate for the difference between those times. What it does establish is a direction for practice: take the dose as soon as possible, respect the 72-hour outer boundary, and do not infer certainty from merely being somewhere inside it.
The timing rule is also not a promise that every dose taken within 72 hours will prevent infection. It is the window in which the intervention has been evaluated and is expected to have the opportunity to act. A dose taken late, a missed dose, or an exposure pattern unlike those in the trials may change the result. The evidence can support an instruction about when to take the dose, but it cannot turn the instruction into certainty about an individual outcome.
The sequence shows the intervention’s supported timing and its verification step; it does not imply that follow-up testing causes the biological outcome.
From exposure to protection at a mucosal site
A useful causal model has several links. Sexual exposure creates the possibility of infection. The dose is the intervention that follows. Absorption and distribution create an opportunity for doxycycline to be present at relevant mucosal sites. The bacterial target then encounters that drug exposure. The outcome—an infection prevented or detected—appears later in clinical follow-up.
Each link matters because the links are not the same thing. The exposure is the trigger for taking PEP, not proof that infection has occurred. The dose is the intervention, not the outcome. Mucosal distribution is a mediating condition described by WHO’s pharmacological rationale, not a measurement that the sources quantify for every tissue. The infection result is what a trial can compare between a doxy-PEP group and a group without doxy-PEP; it is not a direct readout of every step that happened inside the body.
That distinction gives the reader a disciplined way to interpret an apparent failure. If infection is later detected, the result tells us that the overall prevention chain did not avert that infection. It does not, by itself, tell us whether the dose was taken late, whether exposure occurred at a site with a different distribution pattern, whether the organism was susceptible, or whether another feature of the exposure mattered. A trial can estimate the difference in incidence between groups, but it generally cannot assign every individual outcome to one broken link in the chain.
The same model helps interpret success. Fewer infections in the doxy-PEP group support a preventive effect under the studied conditions. They do not prove that every participant had identical adherence, identical tissue exposure or identical bacterial susceptibility. A population-level result is an average produced by many exposure events. Keeping the causal links separate makes the result more credible, not less: it says exactly what was observed while leaving room for the conditions that made the observation possible.
This separation prevents a common mistake in interpreting the regimen. If a trial finds fewer infections, it is reasonable to say that doxy-PEP reduced incidence under the studied conditions. It is not automatically reasonable to say that the pill created the same drug exposure at every site, blocked every early stage of infection, or would work identically against every organism. The observed endpoint is real, but the pathway to it has conditions and unmeasured parts.
The phrase “relevant mucosal sites” also explains why the 72-hour window cannot be treated as a universal countdown detached from exposure. Sexual exposures can involve different anatomical sites, and WHO specifically notes differences in drug distribution across mucosal tissues. The available guidance uses that fact to justify a cautious timing rule, not to claim that one tissue is always protected and another is never protected. It is a reason to respect the studied schedule and to keep detailed tissue-level conclusions within the evidence boundary.
The pathogen enters the chain after the common input has been defined. Syphilis, chlamydia and gonorrhoea are all bacterial STI targets in the trials, but they are not the same biological target. If doxycycline reaches a relevant site on time, the resulting protection can still differ because the organisms differ in susceptibility and resistance. The clock establishes an opportunity; it does not erase the target’s characteristics. That is the key reason timing can be common to all three infections while their measured reductions remain unequal.
There is a second practical implication. A person may take the dose correctly and still acquire an infection. That outcome does not by itself prove that the regimen was taken too late, that the drug failed to reach a site, or that the organism was resistant. Follow-up testing can establish that infection occurred, but a single clinical outcome usually cannot identify which link in the chain broke. The same caution applies at the population level: a lower or higher effect in one study can reflect several interacting conditions, not one visible cause.
For this reason, the dose should be understood as a constrained opportunity rather than an eraser of risk. It acts after exposure, within a specified window, and as part of a system that still requires testing and prevention services. The system is designed around reducing the probability of infection, not guaranteeing that no infection will occur and not replacing diagnosis or treatment.
Why the common input does not produce a common output
Holding the regimen constant is what makes a comparison possible. If syphilis, chlamydia and gonorrhoea were studied with different doses and different timing rules, a difference in outcomes could be attributed to the intervention design itself. With the event-driven regimen held broadly constant, the comparison exposes downstream differences among the targets and the contexts in which they circulate.
The contrast already visible in the evidence is substantial. Pooled randomized trials found high-certainty reductions for syphilis and chlamydia, but a smaller, moderate-certainty reduction for gonorrhoea. Timing is part of every one of those results. It is therefore a necessary condition for interpreting the gradient, but it is not a sufficient explanation for it. The next layer is the organism: doxycycline must be able to act against the particular bacterial population encountered.
This is also why a dose taken within the correct window should not be described as “protecting against bacterial STIs” in an undifferentiated way. That shorthand may be acceptable when describing the broad policy aim, but it hides the question that matters in practice: which infection is the person or program trying to prevent, and how strong is the evidence for that target in this setting? A regimen can be clinically valuable even when its effect is not identical across outcomes.
A simple study thought experiment makes the logic visible. Hold the dose, timing and follow-up schedule constant, then compare the three pathogens. If syphilis and chlamydia show large reductions while gonorrhoea shows a smaller one, the difference cannot be assigned to a different dosing clock; the downstream target and its context deserve attention. If the same pathogen produces different results in different settings, the investigator must then examine susceptibility, exposure patterns, adherence and population context rather than assuming that the medicine changed. The event-driven regimen is the common starting point, not the complete explanation.
This is also why a dose taken within the correct window should not be used as a personal test of whether infection has occurred. The dose is prevention after a possible exposure, not a diagnostic result. It does not establish that an exposure was infected, and it does not make follow-up testing unnecessary. Those distinctions keep the causal sequence honest: prevention changes probability, while testing supplies information about what happened afterward.
The same logic applies to the relationship between PEP and PrEP. Both use a medicine to reduce infection risk, but they place the dose at different points in the exposure timeline. PEP is event-driven and post-exposure; PrEP is taken before exposure. Changing that timing changes the intervention’s pharmacological and behavioral pattern. It may eventually be possible to develop evidence for other strategies, especially for people with frequent exposures, but current guidance does not let readers fill that gap by analogy.
The clock, then, is the first gate between exposure and prevention. A dose must occur within the evidence-supported interval, and repeat dosing must stay within the stated limit. But passing that gate does not settle the result. The drug must also create a useful opportunity at the relevant site, and the bacterial target must remain susceptible enough for that opportunity to matter. The next stage of the analysis asks why gonorrhoea is particularly sensitive to the last condition.