Number Needed to Vaccinate with a Novel Tuberculosis Vaccine to Prevent Tuberculosis in High-Risk Populations, United States

This study estimates that targeted vaccination of *Mycobacterium tuberculosis*-infected high-risk individuals in the United States with an M72/AS01E-like vaccine would require between 217 and 2,486 vaccinations to prevent a single tuberculosis case, a range comparable to established adult vaccines.

The Gist

Imagine the United States is a large city where a sneaky burglar called Tuberculosis (TB) occasionally breaks into homes. For decades, the number of break-ins was going down, but recently, the trend has reversed, and break-ins are happening more often.

Most of these break-ins aren't caused by new burglars moving in from outside; they are caused by "sleeping" burglars already inside the house who suddenly wake up and start causing trouble. These sleeping burglars are the Mycobacterium tuberculosis bacteria living quietly inside people who have been infected but aren't sick yet.

The Problem with the Old Alarm System

Currently, doctors try to stop these burglars by giving people a long course of medicine (like a very strict, difficult-to-follow alarm system) to keep the burglars asleep. However, many people struggle to take the medicine every day, or the medicine makes them feel sick, so the system isn't perfect.

The New "Shield" (The Vaccine)

Scientists are testing a new kind of shield called the M72/AS01E vaccine. Think of this vaccine as a super-strong security guard that stands watch over the sleeping burglars to make sure they never wake up and cause a break-in. In recent trials, this guard was about 50% effective—meaning it successfully stopped the burglars half the time.

The big question this paper asks is: How many people do we need to give this shield to in order to stop just ONE break-in?

In the world of public health, this number is called the Number Needed to Vaccinate (NNV). Think of it like a "cost per saved house" metric.

The Two Strategies: Throwing Shields Everywhere vs. Targeted Protection

The researchers looked at two different ways to use this shield in the U.S.:

1. The "All-Comers" Strategy (Throwing Shields Everywhere)
Imagine you decide to give a security shield to every single person in a high-risk neighborhood, regardless of whether they have a sleeping burglar inside or not.

• The Result: This is very inefficient. Because most people in these groups don't have the sleeping burglar, you are wasting a lot of shields.
• The Numbers: To stop just one case of TB, you would need to vaccinate 70,350 people in the general U.S. population. That's like giving a shield to 70,000 people to save one house. For people born in the U.S., you'd need to shield over 250,000 people to save one case.

2. The "Targeted" Strategy (Finding the Burglars First)
Imagine you first check every house to see if a sleeping burglar is actually inside. You only give the shield to the houses that actually have a burglar.

• The Result: This is much more efficient. You aren't wasting shields on empty houses.
• The Numbers:
  • For people with HIV (who are at very high risk), you only need to vaccinate 217 people to prevent one TB case.
  • For people with kidney failure (ESRD), you need to vaccinate 369 people.
  • For people with diabetes, you need to vaccinate 1,991 people.
  • For people born outside the U.S. (who make up most TB cases), you need to vaccinate 2,158 people.
  • For people born in the U.S., you need to vaccinate 2,486 people.

Why This Matters: The "Shield" Comparison

The researchers compared these numbers to other well-known vaccines we already use, like the flu shot, the HPV vaccine, or the pneumonia vaccine.

• The Verdict: The "Targeted" strategy for TB is actually just as efficient as these established vaccines.
  • For example, the flu vaccine might require vaccinating 9 to 11 people to prevent one case of the flu.
  • The pneumonia vaccine might require 234 to 1,620 people.
  • The new TB vaccine (when used only on those with the infection) falls right in that same "good efficiency" range (217 to 2,486).

The Catch (Limitations)

The paper notes a few important things to keep in mind:

• The Time Limit: These numbers are based on a 3-year window. If the shield stops working after 3 years, the numbers would get worse (you'd need to vaccinate more people).
• The "Real World" Test: The 50% effectiveness comes from trials in other countries. It might work slightly differently in the U.S., especially for people with weakened immune systems (like those with HIV), where the shield might be a bit less strong.
• The Screening Cost: The "Targeted" strategy only works if we can easily and cheaply find the people who have the sleeping burglar first. If the test to find them is too expensive or hard to do, the "All-Comers" strategy might be the only option, even though it's less efficient.

The Bottom Line

This study suggests that if we can find the people who are already infected with TB bacteria and give them this new vaccine, it would be a very smart, efficient move. It would be about as effective as the flu or pneumonia shots we already trust, potentially stopping the rise of TB cases in high-risk groups without wasting resources on people who don't need it.

Technical Summary

Technical Summary: Number Needed to Vaccinate with a Novel Tuberculosis Vaccine in High-Risk U.S. Populations

Problem Statement
U.S. tuberculosis (TB) incidence has risen annually since 2021, reversing a decades-long decline. The disease burden is disproportionately concentrated among non-U.S.-born individuals (approximately 75% of cases) and those with specific medical comorbidities, including diabetes, HIV, and end-stage renal disease (ESRD), which collectively account for roughly 40% of cases. Approximately 85% of U.S. TB cases result from the reactivation of latent Mycobacterium tuberculosis (Mtb) infection rather than recent transmission. While preventive drug therapy (TPT) can mitigate progression to active disease, its uptake remains suboptimal due to adherence issues and medication side effects. Consequently, there is a need to evaluate whether novel vaccines, specifically the M72/AS01E candidate currently in Phase III trials, could serve as an effective alternative for reducing TB incidence in these high-risk U.S. populations.

Methodology
The study estimated the Number Needed to Vaccinate (NNV) to prevent one case of TB in U.S. high-risk groups under two distinct vaccination scenarios:

1. All-comers: Vaccination of all individuals within a risk group regardless of Mtb infection status.
2. Targeted: Vaccination restricted only to individuals with confirmed Mtb infection.

Data were derived from the National Tuberculosis Surveillance System (NTSS), utilizing TB case counts stratified by risk group and the proportion attributed to reactivation (determined via CDC's plausible source case method). The authors adapted a back-calculation methodology to estimate the size of the Mtb-infected subpopulation within each risk group by dividing reactivation-attributed cases by published group-specific reactivation rates.

Key assumptions included:

• Vaccine Efficacy (VE): A baseline of 50% efficacy, based on M72/AS01E Phase IIb trial data, with a 3-year time horizon matching the trial follow-up.
• Sensitivity Analysis: Calculations were repeated assuming VE of 40%, 60%, 70%, and 80%.
• Statistical Approach: 95% confidence intervals (CIs) were constructed using parametric bootstrapping (1,000,000 iterations) in R version 4.5.2.

Key Results
The analysis yielded significantly different NNVs depending on the vaccination strategy and the specific risk group:

• Targeted Vaccination (Mtb-infected only):

  • Lowest NNV: Persons with HIV and Mtb infection required vaccinating 217 individuals to prevent one TB case (95% CI: 95–755).
  • Other Medical Conditions: NNVs were 369 for persons with ESRD (95% CI: 133–1,370) and 1,991 for persons with diabetes (95% CI: 960–6,800).
  • Nativity: For non-U.S.-born persons with Mtb infection, the NNV was 2,158 (95% CI: 1,200–7,100), comparable to the 2,486 (95% CI: 660–10,100) estimated for U.S.-born persons.
  • Total U.S. Population (Targeted): The overall NNV was 2,240 (95% CI: 1,100–7,600).

• All-Comers Vaccination (No screening):

  • NNVs were substantially higher across all groups. For the total U.S. population, 70,350 individuals would need vaccination to prevent one case (95% CI: 41,500–229,000).
  • Among nativity groups, the NNV for non-U.S.-born persons (13,185) was approximately 19-fold lower than for U.S.-born persons (252,021).
  • Among medical conditions, the lowest all-comers NNV was for HIV (5,429), followed by ESRD (6,361) and diabetes (28,365).

• Sensitivity and Comparisons:

  • NNVs scaled inversely with vaccine efficacy; under a conservative 40% VE assumption, targeted NNVs ranged from 269 (HIV) to 3,088 (U.S.-born).
  • The targeted NNVs fell within the range of established U.S. adult vaccines, including pneumococcal (PCV-13: 234–1,620), HPV (324), and influenza (9–11).

Significance and Claims
The paper posits that targeted vaccination of Mtb-infected high-risk persons with an M72/AS01E-like vaccine is a potentially efficient strategy for TB control in the United States. The authors claim that the resulting NNVs (ranging from 217 to 2,486) are comparable to those of widely accepted adult vaccines, suggesting that such a program could be considered efficient relative to existing public health interventions.

The study highlights that targeted vaccination yields NNVs 6- to 101-fold lower than all-comers approaches, indicating that incorporating Mtb infection screening (e.g., interferon-gamma release assay testing) into vaccination programs could substantially improve efficiency. However, the authors maintain a modest tone regarding the broader implications, noting that:

• The analysis captures only direct vaccine effects, as indirect transmission effects are unlikely to be significant in the low-transmission U.S. setting.
• Efficacy in immunocompromised groups (HIV, ESRD) may be lower than the 50% baseline, making the sensitivity analyses particularly relevant.
• The 3-year horizon assumes protection does not wane; if protection declines, long-term NNVs would be higher.
• No universal NNV threshold defines program utility, but the comparison places TB vaccination within the efficiency range of current programs.

The authors conclude that future work must evaluate the cost-effectiveness of these strategies relative to TPT and update estimates as Phase III efficacy data becomes available.