Discussion
Vaccination against CT can significantly reduce infection rates in the USA, even if the vaccine has a moderate efficacy of only 50% and does not provide sterilising immunity. The vaccine’s benefits become apparent within a few years of its launch, and its impact increases with time. With the relatively small number of vaccinations needed to prevent one infection, the vaccine may have the potential to be cost-effective (if not cost-saving) under plausible scenarios, even at moderate efficacy and significantly costly vaccine, in line with a health economics analysis.66 By 2035, 18 vaccinations can prevent one infection, and this number could drop to just 12 by 2050. Vaccination against CT should be considered a priority public health measure to control its spread and reduce the disease and economic burdens associated with it.
To optimise the cost-effectiveness of the vaccine, it can be targeted towards specific population groups. Vaccinating the 15–19 age group is the most effective, with only eight vaccinations needed by 2035 to prevent one infection. Adolescents between the ages of 10 and 14 are also a good target group, with only 11 vaccinations needed to prevent one infection. Populations at high risk of infection, such as female sex workers and men who have sex with men, can benefit greatly from vaccination, with only two vaccinations needed to prevent one infection.
It is expected that the CT vaccine will be developed to prevent acquisition of the infection with a specific value for 
. However, breakthrough infections among vaccinated individuals are likely to have a modified natural history. For instance, vaccinated individuals who acquire SARS-CoV-2 experience lower viral load and shorter duration of infection.67–69 It is plausible that the CT vaccine will have additional biological effects that are related to the mechanisms of action behind its 
, such as reducing the infectiousness ( 
) and duration ( 
) of the infection, as supported in laboratory studies on animal models.26 45 62 The study findings indicate that these ‘breakthrough’ effects could be as impactful as the conventional vaccine effect against infection acquisition, confirming earlier findings.45 A CT vaccine developed to protect against acquisition of the infection is likely to have a larger population impact than expected based solely on the measured value of 
. These findings emphasise the importance of measuring, in vaccine trials, the effects of the vaccine on bacterial load and duration of infection in addition to its effect on acquisition of the infection.
The impact of the three efficacies, 
, 
and 
, was generally comparable, but with some differences. 
had a smaller impact on incidence of infection (figures 4 and 5), but a larger impact on prevalence (figure 3), due to shorter duration of infection. This shorter duration may decrease the likelihood of serious disease sequelae, such as PID,45 making 
particularly important for public health. 
had an overall smaller impact than 
, but targeting the vaccine to those under 20 years old and populations at highest risk may increase its impact since they have higher rates of secondary transmissions (online supplemental figure 3).
The study has limitations due to uncertainties and assumptions in modelling CT infection. The validity and generalisability of input data are important for our model estimations, but the natural history and transmissibility of this infection are still inadequately understood.16 17 34 70 71 This challenge primarily arises due to the ethical complexities associated with conducting studies to directly measure these aspects, given the treatable nature of CT infection. Consequently, researchers often resort to indirect methods to estimate these parameters, resulting in varying results among different studies. Nonetheless, vaccine impact assessment relies on metrics that gauge relative changes, making it less susceptible to the constraints imposed by an incomplete understanding of the infection’s natural history and transmissibility. This is evident in the different additional analyses conducted with varying assumptions for the model parameters, where the resultant vaccine impact estimate displayed generally minimal variations.
The model calibration relied on nine publicly available NHANES rounds, and recent rounds were not included, although this is unlikely to affect the predictions given the largely stable prevalence of CT in the USA.47 72–74 We did not consider possible risk behaviour changes that may occur after vaccination due to the absence of concrete evidence supporting this possibility.75 The definition of a ‘sexual risk group’ is somewhat ambiguous,40 76 77 making the results for vaccine effectiveness among these groups approximate.
The model did not explicitly account for the effects of CT testing and treatment programmes. This approach is rooted in the debate surrounding whether CT testing and treatment programmes have had a considerable impact on CT prevalence.2 However, in practice, we indirectly account for the effects of testing and treatment programmes. This is accomplished through our model’s calibration to observed prevalence rates, which inherently reflects the dynamics of CT transmission in the presence of testing and treatment.
Our estimated number of incident CT infections in women is lower than the most recent CDC model estimate72 but comparable to the earlier CDC round estimate.73 This difference may have arisen due to an underestimation of the proportion of women who receive diagnosis and treatment,8 leading to effectively a longer duration of infection in the population than in reality. The number of women diagnosed and treated for CT annually in the USA8 is substantially higher than what is implicit in our model assumptions and its baseline results. This difference with CDC estimates may have also occurred because of variations in the assumed natural history parameters for this infection, which remain inadequately characterised.16 17 34 70 71
However, while estimating incidence can be challenging, the same does not hold true for prevalence. Prevalence is reliably captured through the high-quality NHANES,47 whereas incidence is generated indirectly through modelling estimations with additional assumptions. Our model was fitted to these NHANES prevalence estimates. As a result, our estimates related to vaccine impact on prevalence should be robust, while those pertaining to absolute incidence among women could be conservative. Notably, even when we increased the probability of asymptomatic women receiving treatment by eightfold, thereby indirectly considerably increasing estimated incidence, this adjustment had a limited effect on the estimated vaccine impact in terms of the relative reduction in incidence following vaccination.
Our modelling approach assumed a prophylactic vaccine with benefits directed toward individuals who have not previously been infected with CT. It is reasonable to assume that once a person has been infected with CT, the vaccine may not offer additional benefits. This assumption is supported by both human and laboratory animal data on CT immunity.19–21 However, in the actual implementation of the vaccine, it may not be feasible to restrict vaccination to those who have not been exposed to the infection. Determining whether individuals with prior exposure can benefit from vaccination and whether the benefits outweigh potential risks may require empirical data. These factors should be taken into consideration in future research, especially when data become available for a specific CT vaccine and its indications and target populations. It is possible that CT vaccination will follow a model similar to the human papillomavirus vaccination programme,75 targeting adolescents before their initial exposure to the infection.
The model operated under the assumption that individuals losing their vaccine protection would be revaccinated to maintain continuous immunity. Although this assumption might appear idealised, it is grounded in the concept that sustaining long-term vaccine protection for such a bacterial infection could necessitate an initial vaccination series supplemented by periodic booster shots administered every few years. These boosters are typically administered according to a schedule, without relying on individualised laboratory testing to assess the waning of immunity. Notably, all vaccinations, including both the primary series and booster shots, were counted in the presented modelling scenarios. The results presented in this study, such as the number of vaccinations needed to avert one infection, also incorporate the total number of implemented vaccinations.
The durability of vaccine immunity and how this immunity will wane remain unknown. If 20-year vaccine protection can be achieved solely through the primary vaccination series, the model-estimated number of vaccinations, implicitly incorporating the count of booster vaccinations, exceeds the actual requirement. This overestimation stems from the model’s assumption that the duration of protection follows an exponentially distributed pattern. Consequently, a significant portion of individuals would lose their vaccine protection within a few years of vaccination and would subsequently be revaccinated, while others would retain immunity for more than 20 years. The model’s estimated absolute impact on prevalence and incidence would also be overestimated due to these revaccinations. It should be noted, however, that the inclusion or exclusion of these revaccinations had a generally small effect on vaccine impact up to 2050 (online supplemental figure 13).
This study was exclusively focused on evaluating the epidemiological impact of CT vaccination. It did not include a health economics analysis that takes into account various cost components, such as diagnosis, clinic visits, treatment for individuals with asymptomatic and symptomatic infections, management of women with PID, addressing women with ectopic pregnancies, and the expenses associated with individuals seeking assisted reproductive technologies due to infertility. A natural extension of this research involves gathering comprehensive cost information and conducting detailed health economics analyses, optimally when cost data related to the vaccine becomes available. This approach would provide a comprehensive perspective on the economic implications of CT vaccination, significantly augmenting our understanding of its epidemiological impact.
Ideally, it would have been best to use CIs or ranges for the parameters in the uncertainty analysis based on actual empirical values.78 However, such confidence intervals or ranges are not available due to the inadequate understanding of the natural history and transmissibility of this infection. Therefore, we employed a commonly applied approach in the epidemiological modelling literature44 59 79–82 of applying a uniform (±30%) uncertainty around the point estimates of the parameters.
This study has strengths. Our model was complex enough to account for the complexity of CT transmission and different vaccine characteristics, yet also tailored to the available data. The results are robust to a wide range of model assumptions and are not overly sensitive to imprecision in knowledge of the infection’s natural history parameters. Our model generated conservative estimates. For instance, if the duration of protection against reinfection following natural infection is shorter than our assumption, the vaccine’s impact is higher (online supplemental figure 9).
In conclusion, a moderately efficacious CT vaccine can significantly reduce infection rates and control the disease burden of this infection. The benefits of the vaccine become apparent within a few years of its launch and increase with time. With the relatively small number of vaccinations needed to prevent one infection, the vaccine may have the potential to be cost-effective, even at moderate efficacy levels. Targeting specific population groups can further maximise the vaccine cost-effectiveness, with adolescents and populations at high risk of infection benefiting greatly. The potential ‘breakthrough’ effects of the vaccine, namely reducing infectiousness and duration of infection, could further enhance its impact. Vaccine development and vaccination against CT should be considered a public health priority. To thoroughly understand a vaccine’s impact, it is critical for vaccine trials to measure not only its effect on the acquisition of the infection but also its potential effects on bacterial load and duration of infection.
