A handful of facilities making 200 million single-dose units per year could fast-track the immunization of the developing world.
Countries classified by the United Nations as part of the developing world are most in need of vaccines but least able to afford them. This article explains the general requirements for manufacturing new vaccines in the developing world instead of importing them.
For all common devastating childhood diseases — tetanus, diphtheria, pertussis, polio, and measles — worldwide universal coverage has been and remains the goal.1 If vaccines for other diseases, such as mumps, rubella, pneumococcal pneumonia, meningitis, and hepatitis B, were easily available in the required quantities, they too would be targets for universal coverage.
Vaccines are currently in development for three diseases that are major threats to world public health: tuberculosis, malaria, and AIDS. Tuberculosis is re-emerging as a major public health threat with almost 2 billion people infected, 17 million with active disease, and approximately 2 million deaths each year worldwide. Malaria infects approximately 300 million persons, causes over a million deaths (mostly among children under five years of age), and debilitates and destroys the economic effectiveness of many adults throughout the world each year. HIV currently infects 40 million people and approximately 15,000 new infections occur every day. It is important to plan for the production, cost, and distribution of vaccines for the developing world so immunization can begin as soon as possible after vaccines are available.
WHO, multiple non-governmental organizations, and charitable organizations (most notably the Bill and Melinda Gates Foundation) have prioritized the development of vaccines for these diseases.2 The discovery phase of vaccine development receives considerable attention, but to achieve ultimate success, these vaccines must be manufactured and delivered to billions of people.3 This article explores the requirements for manufacturing vaccines for tuberculosis, malaria, and AIDS in the developing world where there are extreme economic constraints and other challenges to biopharmaceutical production.
Figure 1. Proportion of Indian Population Immunized by Year.
Before we discuss the problems of producing enough vaccine, we must estimate how many people need to be immunized. The following questions are our starting point.
These questions can be answered with a reasonable degree of certainty even before the vaccines have been discovered and the manufacturing processes developed.3 We will examine the process, manufacturing, and distribution requirements for a new vaccine.
Table 1. Comparison of Single-Dose and Multiple-Dose Products.
The population and growth of India provides a model of the eventual production needs for vaccines for tuberculosis, malaria, and AIDS. India has a population of approximately one billion and an annual growth rate of 1.7%. The population will grow to approximately 1.4 billion in 20 years, which is the expected lifetime of a manufacturing plant. A new vaccine-manufacturing plant should be designed and built at the correct size.
We count the vaccine market on the basis of ten-year cohorts. Up to age 20 there are ten cohorts: 0-10, 1-11, 2-12, 3-13, 4-14, 5-15, 6-16, 7-17, 8-18, 9-19, and 10-20. Approximately 50 to 100 million people will be within one of these cohorts. The lowest age of vaccination will depend on the disease. Early reports indicate that shortly after birth is best for tuberculosis and malaria vaccination, but sometime near puberty is best for HIV, but until the vaccine is approved, we can't predict which cohort will be targeted. Obviously, the exact number of people requiring vaccination will also depend on the age structure of the population.
As the new immunization program gets underway, the entire targeted cohort will be unvaccinated. Afterwards, only those entering the age group will require vaccination. This catch-up dose problem is common to all new vaccine introductions and causes a dramatic change in the dose requirements in the first few years. The dose requirement could be as much as ten times higher the first year than in later years.
Figure 2. Conceptual Layout for a Vaccine Factory.
These uncertainties need to be assessed so that capital is not wasted by building too large a factory that will sit idle a few years later and to avoid building it too small to make enough vaccine to effectively stop the disease. If the objective is purely financial, the balance point will be determined by a reasonable rate and duration of return on the capital investment, but if the objective is public health, the balance point will be shifted to optimize the protection of the population.
Immunizing the target population over five years and providing adequate vaccine supplies over 20 years has emerged as accepted policy over previous immunization campaigns. If we model the growth of the whole population, the disease, and the immunized population for this (or another) rate of immunization, we can design a plant with reasonable capacity in both its early and later years.
For India, an initial immunization rate of 50 million persons per year fulfills the basic requirements for disease control without over-expenditure. An example of this modeling exercise is given in Figure 1, which shows the expected growth of the population, the growth of the target population (a ten-year cohort), the rate of immunization, and the growth of the immunized population.
Figure 3. Timetable for Design, Construction, and Start-Up.
Immunizing a ten-year age cohort at an annual rate of 5% of the total population can achieve nearly full immunization of the cohort within five years, creating a wave of immunized persons within the population. For a population of one billion, a vaccine manufacturing plant meeting these criteria would have to provide 200 million doses per year (50 million full courses of immunization at four doses each). Depending on the population to be served and other logistical criteria, only three to five plants can cover the entire developing world.
A substantial manufacturing facility will be required to produce 200 million doses per year. The nature and size of the manufacturing plant depends on the exact process, batch times for each step, overall yield, and the type of filling and packaging. Relatively few key features of the manufacturing process determine most of the design and size of a vaccine manufacturing plant.
Some design parameters are purely qualitative: The plant must meet all pharmaceutical regulatory requirements for its home country, and it must meet WHO, EU, or FDA requirements if the product is to be exported. Today, these requirements have largely converged, and a well-designed manufacturing plant can meet all international requirements. The main features of cGMP manufacturing facilities are:
Furthermore, vaccine manufacturing involves growing bacteria or viruses and occasionally pathogens, adding requirements for biological containment. New experimental vaccines intended for tuberculosis, malaria, and HIV use three predominant manufacturing technologies:
In the early days of vaccine development, the active pathogen was grown and inactivated or extracted, and each product had highly specialized and often quite difficult manufacturing steps. Today, recombinant technologies allow a uniform and more predictable manufacturing approach, regardless of vaccine target. This technological advance greatly simplifies vaccine manufacturing and allows the design and use of multi-purpose manufacturing plants to produce multiple vaccines.
In this design exercise, we start with the assumption that all three technologies will be needed to make the full range of required products. If only two are needed, it will be easier to delete one type of manufacturing technology and its associated costs than to add one that may not have been considered.
The primary manufacturing stage includes the fermentation or growth of the producing organism, cell, or virus. This results in the bulk vaccine intermediate, which contains the active antigen. Yield at this stage strongly influences the size of the manufacturing plant. Early in a vaccine development program, yields can be quite low, but they can be raised to a reasonable range through process development. Yields can be calculated for live or inactivated viral vaccines, bacterial vaccines, and protein vaccines based on either bacterial or cell culture. In some cases, the final number of doses per liter is relatively high, and a few 100-L fermentors operating 20 runs/year on a two-week production cycle can supply adequate quantities of vaccine. Other products, such as DNA-based vaccines or cell-culture-based protein vaccines, have much lower yields and require much larger production volumes, ranging to multiple 10,000-L fermentors. Early estimates of bulk product yield are adequate to determine the size and cost of the final production facility.
New vaccines pose a technical problem. A choice must be made between doing the commercial production at the development scale (and running many small batches) or undertaking a process development program to scale up to larger volumes. There are significant cost advantages to large-scale processes. The operational costs per batch increase very little as the volume increases because raw materials are generally inexpensive while the costs associated with quality control and batch release testing are high and directly related to the number of batches. For instance, if the annual production schedule is 50,000 L for a bacterial DNA vaccine produced in fermentors, it is much more economical to run 50 batches of 1,000 L than 500 batches of 100 L, despite the cost of process development studies.4
Facility requirements can be calculated for bulk production based on estimates of the volumes required for growth and the number and complexity of the steps required for antigen purification. Combined, these can be used to predict the quantities of purified water and water for injection (WFI), the floor area of the production suites, the air-handling requirements, and the annual utilities requirements. The net result is a projected facility size and cost estimate for producing the bulk vaccine.
Filling and packaging are exacting, costly, and space-intensive activities that are essential to final product manufacturing. A fundamental choice in the filling and packaging of vaccines for the developing world is whether to use multi-dose or single-dose containers. Table 1 compares the risks and benefits of these filling options. Recent vaccine products released in the developed world have been presented as single doses, providing the greatest safety to the individual. However, most of the vaccines supplied to the developing world are still packaged in multi-dose vials, providing the least cost per dose.
In areas of the world where HIV is prevalent, the danger of unintentional infection of vaccine recipients by inadvertent multiple needle use must be considered. It is clear that if cost were not an issue, single-dose containers would be preferred. One appealing option is the Uniject (TM) single-use, self-destructing, plastic syringe device supplied by Becton Dickinson. This device provides an intermediate cost option and is beginning to gain acceptance in South America, India, and Southeast Asia for vaccine delivery.
We have developed a conceptual design for a vaccine manufacturing facility capable of producing 50 million immunizations per year and which encompasses all three production technologies. This design is large and flexible to accommodate changes in product specifications and process implementation for multiple products. If the manufacturing facility is too large for the actual demand, it is easy and inexpensive to use part of it or to reduce the rate of use. However, if the plant is undersized, another plant must be built — with the accompanying time delay and at considerable cost. The flexible design will allow other vaccines to be produced once the catch-up dose requirement for the first vaccine is met. Whether a malaria, tuberculosis, or HIV vaccine is the first into production, a wisely designed facility can produce the next vaccine with little modification, as current designs for these vaccines use similar production technology. Regardless of product flexibility, a single plant will not meet the total developing world demand, and multiple plants will be required.
Figure 2 is a sketch of a conceptual vaccine manufacturing facility accommodating filling and packaging, warehousing for controlled raw materials and finished product (at room temperature, 2-8°C, and frozen), separate wings for each of three production technologies as well as all associated labs and utilities. Current thinking on these diseases suggests the initial dose will involve one technology, and booster doses will involve another. Technologies not required can easily be deleted from the design prior to construction, but from a practical engineering perspective, it is easier and more economical to design for all possibilities.
Multiple scenarios based on combinations of technologies and a range of possible yields can be generated along the conceptual lines described here. Each can be factored into areas of different classification and corresponding costs per unit area. In exploring a variety of combinations based on current costs in northeastern United States, we estimate an average cost of approximately $800/ft2 (inclusive of all structural, utility, and equipment costs). This estimate includes administrative areas, with fairly low unit costs, and Class 100 (Class A) cleanroom areas, with relatively high unit costs. Therefore, the final cost will depend on the facility's exact design.
A full-capacity facility in India designed for 200 million single-dose units per year using multiple production technologies will require approximately 200,000-300,000 ft2 and will cost $200-300 million to the point of start-up. In order to estimate costs for other locations, it is necessary to consider local costs, such as land and construction fees, as well as the costs of imported equipment and supplies that will be constant across all locations.
Other factors will impact project and operational success. The facility should be similar in size and design to known pharmaceutical plants in order to minimize construction and operational difficulties. Its design also should be matched to the availability of power, water, waste discharge, and potential employees. Care in design can improve operational efficiency by providing time- and energy-efficient process and product flows and maximizes the ease of equipment maintenance. Energy efficiency in cooling and process steam and WFI production (possibly using cogeneration capabilities) can minimize energy costs and ultimately, product costs.
To minimize design and construction costs and maximize the success of plant start-up and operation, multiple plants around the world should be nearly identical. Cross-licensed, identical manufacturing facilities could maintain uninterrupted supply of multiple vaccines to many countries and markets around the world. Figure 3 illustrates a conceptual timetable for the design, construction, validation, and start-up of multiple identical vaccine plants around the world. This concept has been used successfully in the semiconductor industry.
Advance planning, as well as the maintenance of "live" plans that can be adapted to varying circumstances, will contribute immensely to decision-making and mobilization, ensuring the fastest possible availability of new vaccines for these debilitating diseases. By scheduling a one-year interval between plants, we are able to move key personnel at the best time to the next project. This optimizes experience gained during engineering, construction, and plant start-up and validation. Execution of this type of systematic manufacturing approach will greatly improve both the availability and standardization of vaccines worldwide.
1. World Health Organization. State of the World's Vaccines and Immunization, Revised Edition. Geneva, Switzerland: World Health Organization; 2003.
2. Bill & Melinda Gates Foundation. Responding to the Needs of Others — 2003 Annual Report. Seattle (WA): Bill & Melinda Gates Foundation; 2003.
3. Wechsler J. Manufacturing capability key to global health advances. Pharmaceutical Technology 2003; 27(2):24.
4. Gerson DF, et al. Transfer of processes from development to manufacturing. Drug Information Journal 1998; 32(1):19-26.
Corresponding author Bhawani Mukherjee is senior vice president of biopharmaceutical operations at Paulus, Sokolowski and Sartor, LLC, 67A Mountain Boulevard Extension, Warren, NJ 07059, 732.584.0210, fax 732.764.6667, email@example.com, www.psands.com. Donald F. Gerson, Ph.D. is president of Axenic, Inc., 12 Charnwood Dr., Montebello, NY 10901, 845.368.8157, fax 845.368.8157, firstname.lastname@example.org