B R I E F R E P O R T
Failure of the Milwaukee Protocol in
a Child With Rabies
Angela Aramburo,1Rodney E. Willoughby,5Andrew W. Bollen,6Carol
A. Glaser,7Charlotte J. Hsieh,2Suzanne L. Davis,3Kenneth W. Martin,4and
1Division of Critical Care Medicine, Children's Hospital & Research Center Oakland,
California,2Division of Infectious Diseases, Children's Hospital & Research Center
Oakland, California;3Division of Neurology, Children's Hospital & Research Center
Oakland, California;4Department of Diagnostic Imaging, Children's Hospital &
Research Center Oakland, California;5Division of Infectious Diseases, Children's
Hospital of Wisconsin, Milwaukee, Wisconsin;6Department of Pathology,
Neuropathology Division, University of California San Francisco; and7Encephalitis
and Special Investigation Section, Division of Communicable Disease Control,
California Department of Public Health, Richmond, California
Rabies has the highest case-fatality rate of all infectious dis-
eases, with 50 000 cases occurring annually worldwide. In
2004 an unvaccinated adolescent survived after novel therapy.
We report the management of a child with rabies. Although
the implementation of this same therapeutic protocol was
successful, the child died after 1 month of hospitalization.
Rabies encephalitis was considered universally fatal in humans
until 2004, when an unvaccinated adolescent survived with
novel therapy now dubbed the Milwaukee Protocol (MP) .
This protocol includes therapeutic coma, antiviral therapy, ce-
We report the treatment of a child with rabies, who received the
most timely and complete application of the original MP to
date, and compare this case with other MP attempts, discussing
implications for advancement in the field.
In November 2006, an 11 year-old male from the Philippines
presented to a community emergency department (ED) with
symptoms suggestive of furious rabies. Two years earlier, the
patient had been bitten by a dog in the Philippines and did not
receiverabies vaccine or other post-exposure prophylaxis (PEP);
clinical presentation has been reported elsewhere . Briefly,
sore throat, fever, and fatigue were followed by progressive
shortness of breath, dysphagia, and insomnia. In the ED, he
developed irregular mouth movements, visual hallucinations,
agitation, aerophobia, and hypersalivation.
Upon transfer to our children’s hospital ED, mental status
alternated between extreme agitation and obtundation. Marked
heart rate and blood pressure variability were compatible with
severe dysautonomia. He was intubated for airway protection.
Following thiopental for sedation, he became severely brady-
cardic, requiring brief cardiopulmonary resuscitation (CPR).
Neuromuscular blockade was administered because of pharyn-
geal and diaphragmatic spasms.
On admission to the intensive care unit (ICU), simultaneous
severe variable hypertension and heart rate suggested neurally
mediated catecholamine storm. Echocardiogram revealed severe
secondary left ventricular dysfunction. With a fosphenytoin load
responded to CPR. Inotropic support was required for 4 days.
Coma was induced with ketamine and midazolam infusions, as
recommended in the MP (version 1.1), for presumed rabies.
On hospital day 1 (HD1), direct fluorescent antibody (DFA)
detected rabies virus antigen in corneal impressions; serum and
cerebrospinal fluid (CSF) serologies were negative. From a saliva
sample on HD3, molecular testing performed at the Centers for
Disease Control and Prevention (CDC) detected rabies virus
RNA, corresponding genetically to Philippine dog rabies. Anti-
rabies immunoglobulin G (IgG) was first detected via indirect
immunofluorescence in serum and CSF on HD11 and HD13,
respectively. Initial nuchal biopsy was positive by DFA on HD3
and again on HD19.
On HD1, upon diagnosis of rabies, and after discussion with
the California Department of Public Health, CDC, and authors
of the MP, intravenous ribavirin and enteral amantidine, tet-
rahydrobiopterin (BH4), coenzyme Q10, and ascorbic acid were
initiated. All subsequent titrations and modifications of the MP
were made in direct consultation with the MP primary in-
With therapeutic coma, dysautonomia steadily improved.
Given concerns for development of cerebral electrical silence,
vasospasm, and edema, intense neurologic monitoring was ini-
tiated. This included continuous electroencephalogram (EEG),
continuous cerebral regional oxygen saturation measurement
Received 31 December 2010; accepted 6 June 2011.
Correspondence: Arup Roy-Burman, MD, Division of Critical Care Medicine, Children's
Hospital & Research Center Oakland, 747 52nd St, Oakland, CA 94609 (arup.roy-
Clinical Infectious Diseases
? The Author 2011. Published by Oxford University Press on behalf of the Infectious
Diseases Society of America. All rights reserved. For Permissions, please e-mail:
d CID 2011:53 (15 September)
d BRIEF REPORT
by guest on October 28, 2015
by near-infrared spectroscopy (NIRS), and daily transcranial
Cerebral vasospasm was suggested by TCD on HD6 and
HD10–HD12, as reported elsewhere ; vasospasm-targeted
therapies included escalating doses of BH4, followed by milri-
burst suppression by HD13; anticonvulsant prophylaxis in-
cluded fosphenytoin, midazolam, and phenobarbital. On HD15,
pupils became fixed and dilated. Prolonged electrographic
seizures developed on HD17. On HD21 topiramate was ad-
ministered for multifocal spikes; subsequently, the EEG became
isoelectric. Medications potentially responsible for this finding
were weaned. Computed tomography on HD24 revealed severe
cerebral edema; prior neuroimaging, including CT on HD1 and
HD13, along with magnetic resonance imaging on HD6 and
HD12, was normal. Right frontal lobe biopsy on HD25 showed
perivascular cuffing with lymphocytes extending into the sur-
rounding brain and associated microglia, consistent with en-
cephalitis. Support was withdrawn on HD27.
The hospital course included multiple secondary complica-
tions previously described in rabies: transient hyponatremia,
hemidiaphragmatic paralysis, pancreatitis, hypothyroidism, and
heart block requiring transvenous pacing. Additional complica-
tions included ventilator-associated pneumonia and peripheral
thrombophlebitis. Hospitalization charges exceeded $700 000.
encephalitis and widespread loss of neurons in the cortex, with
many residual neurons undergoing necrosis. The cerebellum
demonstrated extensive loss of Purkinje cells and internal granule
cell neurons. Negri bodies were identified (Figure 1). Scattered
microinfarcts and small perivascular hemorrhages were also seen,
possibly representing a component of vasospasm and associated
thromboembolism. Overall, pathology showed evidence of severe
lymphocytic encephalitis with superimposed secondary hypoxic
ischemic encephalopathy. CDC confirmed the presence of rabies
particles in brain tissue by DFA and molecular studies.
Rabies encephalitis is almost universally fatal, and therapy has
been classically palliative. Prior to 2004, there were only 5 docu-
mented human survivors, all of whom had received PEP, albeit
incomplete or late . In 2004, after a novel protocol was applied
(MP), the first survival without PEP was reported . Since then,
the MP has been viewed as a potentially effective therapy.
As reported to date in the MP Rabies Registry, variations of
the MP have been applied to 25 additional patients, with 3
minimally documented ‘‘survivors’’ . A girl from Colombia
 and a man from Peru died (HD76 and HD70, respectively)
after presumedviralclearanceanddischarge fromtheICU,from
nonrabiesattributable medicalcomplications(R. E. Willoughby,
personal communication, 2011). The third survivor, a Brazilian
boy with bat rabies  who received partial PEP before onset of
symptoms, is living at home. Of all MP cases reported to date,
ourmanagementmostcloselymirrorsthat oftheindex MPcase,
given the early diagnosis and initiation of therapy, avoidance of
immunizations, and direct comanagement with the MP PI.
The MP has evolved and is currently in version 3.1 . Its
original assumptions, however, remain unchanged. First, wild-
type rabies infection confers little viral or immune-mediated
cytopathic effectandistherefore theoreticallyreversible.Second,
a natural immune response is sufficient to clear the virus. Third,
BH4 deficiency and generalized cerebral vasospasm may be in
the causal pathway of rabies mortality. Four general principles
guide therapy: (1) prolonged therapeutic coma to prevent early
life-threatening dysautonomia; (2) antiviral therapy; (3) pro-
phylaxis, monitoring, and treatment of cerebral vasospasm; and
(4) avoidance of immune prophylaxis. Of these principles, only
the first has been used consistently with success.
Unfortunately, antiviral therapy with combined ribavirin,
ketamine, and amantadine has not shown antiviral effect in MP
patients . Additionally, ribavirin may delay the antibody re-
sponse and is no longer recommended . Conversely, given its
favorable side effect profile, amantadine is still recommended.
Similarly, management of cerebral vasospasm has shown little
benefit . Although suggested by TCD, confirmatory imaging
was not performed, and autopsy did not show evidenceof major
vasospasm. The MP has used multiple vasodilator agents, but
none appear to have prevented brain injury .
The rabies antibody response typically appears by 2 weeks.
The MP postulates that a vigorous immune response may exac-
erbate the disease process , and it avoids active immunization
Photomicrograph of the cerebellum (hematoxylin-eosin stain; original
Negri body (arrow) is identified within a Purkinje cell.
d CID 2011:53 (15 September)
by guest on October 28, 2015
once symptoms develop. Of note, the Brazilian survivor 
received 4 of 5 recommended PEP vaccinations before symp-
toms. Furthermore, the index survivor, in whom rabies virus
was never detected, had antibodies at 1 week . The MP also
avoids passive immunization with rabies immune globulin
(RIG) because of theoretical interference with the native im-
mune response and unclear CNS penetration. However, RIG
may have a protective role in the heart and other organs, where
rabies virus has been demonstrated .
In our case, ongoing viral effects with associated devastating
brain injury were observed. This questions the premise that
intensive supportive care allows the immune response to clear
the virus, while retaining reversibility of neurologic disease.
Through 2008, of the 7 reported rabies survivors to hospital
discharge [1, 4, 7], only the index MP case did not receive PEP.
Rabies virus was detected in 1 case ; all others were diagnosed
by rabies antibody. More recently, abortive rabies was described
in an adolescent female who developed neurologic symptoms
after bat exposure and had high rabies antibody titer, without
isolation of rabies virus . She survived without ICU care,
receiving both active and passive immunization only after a late
diagnosis; this and the index MP case were both infected with
bat rabies[1, 10]. Genetic variabilitiesin thehostandviruslikely
contribute to survival. Indeed, there are reports of animals
surviving rabies without therapy . Thus, application of the
MP may be more successful in specific subgroups of patients.
As one of the oldest and deadliest infectious diseases, rabies is
long overdue for development of a successful treatment. Six
years ago, when the first rabies survivor (without PEP) was
described, there was new hope for rabies victims. Unfortunately,
subsequent cases illustrate the uncertainties surrounding rabies
management and the tremendous resources expended in ag-
gressive supportive care . This case, when taken together
with other MP cases to date, suggests that an early immune
response may be better correlated with survival, the efficacy of
MP antiviral activity is unclear, and ribavirin itself may be im-
munosuppressive. Aggressive supportive care has resulted in
longer survival times and consequently a wealth of clinical and
laboratory data, helping to better understand the natural history
of rabies and develop specific questions regarding its patho-
physiology. Animal models are urgently needed to address these
questions, which may ultimately lead to successful outcomes in
We are grateful to the staff of the Viral and Rickettsial Disease Labo-
ratory at the California Department of Public Health, particularly Dave
Schnurr, Debra Wadford, Sharon Messenger, Tasha Padilla, Wanda Wong,
and Elaine Yeh, for their support with timely diagnostics. We would also
like to thank Charles Rupprecht, Chief, CDC Rabies Program, for assis-
tance in the care of this patient as well as insight into the history and future
development of rabies therapy.
Potential conflicts of interest.All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest. Conflicts that the editors consider relevant to the
content of the manuscript have been disclosed in the Acknowledgments
1. Willoughby RE Jr, Tieves KS, Hoffman GM, et al. Survival after
treatment of rabies with induction of coma. N Engl J Med 2005;
2. CDC. Human rabies—Indiana and California, 2006. MMWR Morb
Mortal Wkly Rep 2007; 56:361–5.
3. Willoughby RE, Roy-Burman A, Martin KW, et al. Generalised cranial
artery spasm in human rabies. Dev Biol (Basel) 2008; 131:367–75.
4. Wilde H, Hemachudha T, Jackson AC. Viewpoint: management of
human rabies. Trans R Soc Trop Med Hyg 2008; 102:979–82.
5. Children’s Hospital of Wisconsin. Rabies registry home page. Children’s
Hospital of Wisconsin, Available at: http://www.chw.org/rabies. Accessed
originally 14 November 2006, and most recently 15 December 2010.
6. Juncosa B. Hope for rabies victims: unorthodox coma therapy shows
promise. Scientific American. Available at: http://www.scientificamer-
therapy-shows-promise. Accessed 25 November 2008.
7. ProMED-mail. Rabies, human survival, bat—Brazil (Pernambuco).
2008. Available at: http://www.promedmail.org. 1114.3599. Accessed
25 November 2008.
8. LauJY, TamRC, Liang TJ,et al. Mechanism of actionofribavirin in the
combination treatment of chronic HCV infection. Hepatology 2002;
9. Jogai S, Radotra BD, Banerjee AK. Rabies viral antigen in extracranial
organs: a post-mortem study. Neuropathol Appl Neurobiol 2002;
10. CDC. Presumptive abortive human rabies—Texas, 2009. MMWR
Morb Mortal Wkly Rep 2010; 59:185–90.
11. Jackson AC, Reimer DL, Ludwin SK. Spontaneous recovery from the
encephalomyelitis in mice caused by street rabies virus. Neuropathol
Appl Neurobiol 1989; 15:459–75.
12. McDermid RC, Saxinger L, Lee B, et al. Human rabies encephalitis
following bat exposure: failure of therapeutic coma. CMAJ 2008;
d CID 2011:53 (15 September)
d BRIEF REPORT
by guest on October 28, 2015