COMMENTARY
Acceptable Levels of Endotoxin in Vaccine Formulations During
Preclinical Research
LUIS A. BRITO, MANMOHAN SINGH
Novartis Vaccines and Diagnostics, 350 Massachusetts Ave., Cambridge, Massachusetts 02139
Received 26 March 2010; revised 7 May 2010; accepted 10 May 2010
Published online 23 June 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22267
ABSTRACT: This brief commentary reviews endotoxin levels of commercial vaccines and puts
them into context for the preclinical researcher working in vaccines. Vaccines are not required to
adhere to endotoxin levels as outlined in the United States Pharmacopoeia. Vaccine manufac-
turers have to show that the vaccine is safe and efficacious in clinical trials. Endotoxin limits are
typically lot release specifications for most vaccines, but these values are not available to most
researchers designing preclinical experiments. The limits outlined are calculated from endo-
toxin levels found in a variety of vaccine types such as gene vectors, recombinant subunits,
polysaccharide, live attenuated, inactivated and toxoid vaccines. It is clear that certain families
of vaccines such as toxoids contain much higher levels of endotoxin, where others such as
purified recombinant subunits and gene vectors may contain very low levels. ß 2010 Wiley-Liss,
Inc. and the American Pharmacists Association J Pharm Sci 100:34 37, 2011
Keywords: endotoxin; vaccines; formulation; excipients; injectables; preformulation
INTRODUCTION vaccine, a live attenuated viral vaccine, or a
polysaccharide conjugate.
Researchers involved in preclinical vaccine research
typically have very little guidance on acceptable
endotoxin levels for preclinical evaluation. Previously
ENDOTOXIN TESTING
we defined endotoxin levels for preclinical evaluation
by using formulas released by the USP and calculat-
There are a variety of tests available to measure
ing maximal endotoxin levels based on animal weight
endotoxin levels, the gel clot technique, chromogenic
using the USP recomended endotoxin limit of 5 USP-
and end point turbidimetric or kinetic turbidimetric
EU/kg.1 While this is technically correct for most
techniques each with its own advantages or dis-
drugs, CFR610.13 states that,   the test for pyrogenic
advantages. The gel clot technique allows for testing
substances is not required for the following products:
of a very wide range of samples including samples
Products containing formed blood elements; Cryo-
that are opaque such as emulsions, nanoparticle and
precipitate; Plasma; Source Plasma; Normal Horse
microparticle suspensions and alum; these samples
Serum; bacterial, viral, and rickettsial vaccines and
are unable to be tested with standard turbidimetric or
antigens; toxoids; toxins; allergenic extracts; venoms;
chromogenic techniques. The limitation of this assay
diagnostic substances and trivalent organic arseni-
is that it is labor intensive and only a small number of
cals.  2 This update to our previous publication will
samples can be analyzed at once, without automation.
focus on reviewing the literature for endotoxin levels
Chromogenic and turbidimetric assays can be run in a
found in commercial vaccines to allow the preclinical
96-well format. This allows for a higher throughput of
researcher involved in vaccine development to set
samples but are limited to samples that are clear
appropriate endotoxin levels for the type of vaccine
solutions. There are now units made specifically to
they are developing such as a recombinant subunit
measure endotoxin levels such as Endosafe1 by
Charles River Labs. This unit facilitates rapid testing
of samples with minimal user interaction, though it is
Correspondence to: Manmohan Singh (Telephone: 617-871-
based on a turbidimetric system, so it does have
8100; Fax: [phone]; E-mail: [e-mail])
limitations on the types of samples that can be tested
Journal of Pharmaceutical Sciences, Vol. 100, 34 37 (2011)
ß 2010 Wiley-Liss, Inc. and the American Pharmacists Association on it.
34 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 1, JANUARY 2011
VACCINE FORMULATIONS DURING PRECLINICAL RESEARCH 35
ENDOTOXIN IN THE CLINIC ENDOTOXIN LEVELS IN COMMERCIAL
VACCINES
Endotoxin has been characterized as an agonist of
TLR 4. It is well known that even a small amount Vaccine manufacturers do not release their internal
(4 ng/kg) of endotoxin can cause a change in lot release specifications on endotoxin levels of their
physiology in humans.3 Reports of adverse events finished product. There are reports in the literature of
with endotoxin exposure have been well documen- independent laboratories analyzing endotoxin levels
ted.4 These findings justify the recommendations on for a variety of vaccines. Table 1 outlines the findings
the amount of endotoxin in finished products passed from the literature for a number of different
by the United States Pharmacopeia (USP), United commercial vaccines. These values should not be
States Food and Drug Administration (USFDA), and taken as limits set by companies, but only what has
the European Medicines Agency (EMEA), in addition been detected in a series of vaccines analyzed
to other regulatory agencies throughout the world. throughout the years.
Endotoxin values are listed as entotoxin units (EU) These levels are noted as EU/mL. As mentioned
since there is heterogeneity in the molecular weight of previously the USP sets endotoxin limits by EU/kg.
endotoxin (it is not a single component). Given the average volume of 0.5 mL per human
There are few cases of endotoxin contamination in vaccine dose the values can be back calculated to
parenteral products that have been administered to compare them to the USP limits. Since vaccines are
patients.5 One such example occurred when genta- not bound by the endotoxin limits stated in the USP
micin that was contaminated with endotoxin and a we are recommending the following endotoxin level
nonendotoxin contaminate known to potentate an recommendation for various vaccine types summar-
endotoxin response. There were 210 adverse events ized in Table 2 based on our findings from the
filed. The lots were tested by the FDA and 10% of literature. A recommendation of <10 EU/mL for
the lots tested showed elevated endotoxin levels genetic vectors was chosen due to the extensive
(above 5 EU/kg).6,7 This event highlights that syner- purification needed for these types of vaccines. A level
gistic contaminates can potentate the effect of of <20 EU/mL for recombinant subunits and poly-
endotoxin.5 saccharide vaccines was recommended due to the
A major reason for changing from a whole cell high purity and highly controlled method of manu-
pertussis vaccine to an acceular vaccine was to facture for these vaccine types. A recommendation of
decrease the reactogenicity of the vaccine.8,9 The <200 EU/mL for live attenuated vaccines is based on
high levels of endotoxin may have been one of the the complex nature of these vaccines containing
causes of the reactogenicity, but this is not able to be multiple antigens. A similar rational goes into the
confirmed due to the over 3000 different components recommendation of <500 EU/mL for inactivated
in the whole cell vaccine.10 This example highlights vaccines. Toxoid vaccines have the highest recom-
how difficult it is for certain vaccines to distinguish a mended values of <200,000 EU/mL due to the
response due to endotoxin or a response due to complexity of these vaccines and the fact that
another component or excipient in the vaccine. they are all derived from bacteria with minimal
Table 1. Reported endotoxin levels for various vaccines
Vaccine Amount of Endotoxin References
9,11
DTaP Less than 100 EU per dose, 0.288 1390.8 EU/mL
9,12
DTwP 7600 181,640 EU/mL
12,13
Influenza 10 100 ng/mL, 0.304 380 EU/mL
14,15
HIB <1.2 <12 EU/dose, 10.9 220.4 EU/dose
16
Prevnar <0.21 EU/dose
12
Small Pox Undetected
12
Rubeola Undetected
12
Rubella 10 EU/mL
12
Polio Undetected 100 EU/mL
12
Rabies 100 EU/mL
12
Mumps 100 EU/mL
12
Typhus 100,000 EU/mL
12
Cholera 1,000,000 EU/mL
17,18
Viral/nonviral vector <10 EU/mL
1 ng of endotoxinź10 EU. This is a general comment that allows readers to convert the levels in the table to ng
of endotoxin based on reference 1. There is no specific link within the table.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 1, JANUARY 2011
36 BRITO AND SINGH
Table 2. Recommended endotoxin levels for different able steps have been performed to try and remove any
vaccine types
residual endotoxin. However as the purification
process gets further optimized, the residual endotoxin
Maximum Recommended
levels may drop off and the subsequent immunogeni-
Vaccine Type Endotoxin Levels (EU/mL)
city data with the same dose of antigen may show a
Gene vector <10
reduction. The levels outlined above should only act
Recombinant subunit <20
as a starting point for researchers. Ideally endotoxin
Polysacharide <20
levels should be as low as possible to allow for an
Live attenuated <200
appropriate evaluation of the antigens being tested.
Inactivated <500
Toxoida <200,000
With the evaluation of adjuvants, it is even more
important to dissect the response of the adjuvant by
a
Based on values from DTwP, and Cholera vaccines, through purification
and process changes endotoxin levels should fall below 2000 EU/dose (based having a low residual endotoxin coming from the
on DTaP).
purified antigen. The endotoxin levels should also be
closely monitored due to the potential of a synergistic
effect.
purification. Though advances in purification and
process have been continuously improving the
endotoxin levels in these types of vaccines. We are
REFERENCES
proposing these values as guidelines for evaluating
antigens in all animal studies during early develop-
1. Malyala P, Singh M. 2008. Endotoxin limits in formulations for
ment. As vaccine development progresses to the
preclinical research. J Pharm Sci 97:2041 2044.
clinic, safety and efficacy will be the ultimate readout. 2. FDA. 2009. General Biological Products Standards Purity.
Code of Federal Regulations Title 21(610.13).
Higher endotoxin levels than outlined in Table 2 may
3. Wilson M, Blum R, Dandona P, Mousa S. 2001. Effects in
be acceptable in certain instances, though research-
humans of intravenously administered endotoxin on soluble
ers should ensure they are adequately addressing
cell-adhesion molecule and inflammatory markers: A model of
endotoxin levels.
human diseases. Clin Exp Pharmacol Physiol 28:376 380.
The FDA has also released guidelines on how 4. Burrell R. 1994. Human responses to bacterial endotoxin. Circ
Shock 43:137 153.
combination vaccines should be developed and
5. Cooper JF, Williams KL. 2009. Endotoxins pyrogens, LAL
tested.19 The guidance states that   Each bulk
testing and depyrogenation. New York: Informa Healthcare.
component of a combined vaccine should meet purity
6. Fanning MM, Wassel R, Piazza-Hepp T. 2000. Pyrogenic reac-
characteristics appropriate for that component. For
tions to gentamicin therapy. N Engl J Med 343:1658 1659.
example, the Haemophilus b conjugate component 7. CDC. 1998. Endotoxin-like reactions associated with intrave-
nous gentamicin California. Morb Mortal Wkly Rep 47:877
should meet the acceptance criteria for pyrogenicity
880.
testing even if combined with a vaccine such as DTP
8. Braun MM, Mootrey GT, Salive ME, Chen RT, Ellenberg SS.
which is known to contain endotoxin. In this case,
2000. Infant immunization with acellular pertussis vaccines in
final product will not be pyrogen tested.  19 This
the United States: Assessment of the first two years data from
further highlights that endotoxin levels are specific to the Vaccine Adverse Event Reporting System (VAERS). Pedia-
trics 106:E51.
each licensed vaccine and a general statement as
9. Geier DA, Geier MR. 2002. Clinical implications of endotoxin
what is found in the USP guidance for endotoxin do
concentrations in vaccines. Ann Pharmacother 36:776 780.
not apply. It should be noted that a pyrogen is a
10. Moylett EH, Hanson IC. 2004. Mechanistic actions of the risks
substance endogenous or exogenous that induces a
and adverse events associated with vaccine administration.
febrile response, endotoxin is only one type of J Allergy Clin Immunol 114:1010 1020. quiz 1021.
11. Eur. P. 2008. Bacterial endotoxins. 60 5.0:2.6.14.
pyrogen.20 Pyrogenicity testing looks at the febrile
12. Geier MR, Stanbro H, Merril CR. 1978. Endotoxins in com-
response typically in rabbits, and does not look at the
mercial vaccines. Appl Environ Microbiol 36:445 449.
specific cause of the increased temperature.5
13. Geier MR, Geier DA, Zahalsky AC. 2003. Influenza vaccination
and Guillain Barre syndrome small star, filled. Clin Immunol
107:116 121.
14. Bolgiano B, Mawas F, Burkin K, Crane DT, Saydam M, Rigsby
CONCLUSION
P, Corbel MJ. 2007. A retrospective study on the quality of
Haemophilus influenzae type b vaccines used in the UK
During preclinical development of the next genera-
between 1996 and 2004. Hum Vaccin 3:176 182.
tion vaccines endotoxin levels should be closely
15. Ochiai M, Kataoka M, Toyoizumi H, Yamamoto A, Kamachi K,
monitored despite the heterogeneity in licensed Arakawa Y, Kurata T, Horiuchi Y. 2004. Endotoxin content in
Haemophilus influenzae type b vaccine. Jpn J Infect Dis 57:
vaccines. Routine measurement of endotoxin levels
58 59.
is necessary to properly understand preclinical
16. Sen G, Khan AQ, Chen Q, Snapper CM. 2005. In vivo humoral
animal studies results. Importantly high endotoxin
immune responses to isolated pneumococcal polysaccharides
levels in a preclinical setting should not preclude its
are dependent on the presence of associated TLR ligands.
use for immunogenicity studies assuming all reason- J Immunol 175:3084 3091.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 1, JANUARY 2011 DOI 10.1002/jps
VACCINE FORMULATIONS DURING PRECLINICAL RESEARCH 37
17. Wright JF. 2008. Manufacturing and characterizing AAV- 19. FDA. 1997. Guidance for Industry for the Evaluation of Com-
based vectors for use in clinical studies. Gene Ther 15:840 bination Vaccines for Preventable Diseases: Production, Test-
848. ing and Clinical Studies. In Research CfBEa, editor. U.S.
18. Przybylowski M, Bartido S, Borquez-Ojeda O, Sadelain M, Department of Health and Human Services.
Riviere I. 2007. Production of clinical-grade plasmid DNA for 20. Blatteis Clark M, Sehic E, Li S. 2000. Pyrogen sensing and
human Phase I clinical trials and large animal clinical studies. signaling: Old views and new concepts. Clin Infect Dis 31:
Vaccine 25:5013 5024. S168 S177.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 1, JANUARY 2011