I;,:l,rintcd from A~IERKAX JOL'IIKAL OF PUBLIC HEALTII, Vol. 27, No. 1, January, 1937

The Newer Epidemiology of
    Yellow Fever *

        FRED L. SOPER, M.D., DR.P.H.
International Health Division, The Rockefeeller Foundation, New York, N. Y.

T HE extent to which present beliefs
regarding the epidemiology of yellow
fever differ from those of just a few
years ago is not generally known out-
side of the small group of workers
actively engaged in the study and con-
trol of this disease. It must come as a
great surprise to the majority of the
members of the American Public Health
.\ssociation, meeting here in New
Orleans, one of the cities of the United
States previously most exposed to the
ravages of yellow fever, but since 1905
free of this scourge of the American
tropics, to learn that the epidemiology
of yellow fever is still a sufficiently im-
pc-rtant problem to merit a place on a
program devoted to public health prob-
lems of immediate interest to health

`Read at a Special Session on Mosquito-Borne
n&vases of the American Public Health Association
. the Sixty-fifth Annual Meeting in New Orleans,
`.a.. October 23, 1936.
j The studies and observations on which this report
is `Jased were made under the auspices of the Inter-
national Health Division of The Rockefeller Founda-
!;I+ in coSperation with the Governments of Brazil,
r  /ia, Pawgay, Peru, Ecuador, Colombia, and
muela. The report covers the work of numerous
eagues of the staff of the International Health
hvirion and of the staffs of the Co6perativ.e Yellow
Fever Services in the respective countries.

workers in the United States. Your
surprise, however great, cannot equal
the reluctance with which those engaged
in the study and control of yellow fever
in South America have relinquished their
faith in the simple man-il&des aegypti-
man epidemiology, consecrated by the
almost miraculous results obtained in
the West Indies and in North, Central,
and South America during a full quarter
of a century by a brilliant coterie of
workers drawing their inspiration from
the early work of Gorgas in Havana
and of Oswald0 Cruz in Rio de Janeiro.

The period from the first observation
of yellow fever to the demonstration in
1900 l that its virus could be transmit-
ted from man to man by the mosquito,
AZdes (Stegomyia) aegypti, may well be
called the Dark Age in the history of
the disease.  More or less constant
endemicity existed in many of the ports
of tropical America, and the disease
made not infrequent summer excursions
into both the North and South Tem-
perate Zones.  Quarantine and other
restrictive measures were `common, and
trade with the American tropics was
difhcult and costly.


2

AMERICAN JOURNAL OF PUBLIC HEALTH   Jan., 1937

The discovery of the urban, domestic
vector of yellow fever was followed by
a Golden Age of magnificent achieve-
ment and glorious dreams based on that
achievement. During this Golden Age
which covers the period 1901 to 1925,
anti-mosquito services organized in many
of the most notorious key-centers of
infection on the continent resulted in
the rapid disappearance of yellow fever
from the morbidity and mortality sta-
tistics not only of the centers where
such services were installed but also
from those of large tributary regions.
To the sanitation of Havana and
Panama was attributed the great reduc-
tion in the incidence of yellow fever
throughout the West Indies and the
littorals of the Caribbean Sea and the
Gulf of Mexico; the cleaning of Rio de
Janeiro and Santos resulted in the
disappearance of the disease in
Southern Brazil; and short campaigns
in several of the principal cities of the
Amazon River were followed by the
apparently complete blotting out of
yellow fever in the entire Amazon
Valley! Upon these observed results
there was built a simple epidemiology
of yellow fever, and from this simple
epidemiology the dream took shape
that yellow fever might disappear for-
ever from the American continent
through the temporary organization of
anti-aegypti measures in the largest
cities of the few remaining endemic
foci.

Shortly after the founding of The
Rockefeller Foundation, the Interna-
tional Health Board became interested
in the early `realization of this dream,
and General Gorgas was invited to un-
dertake the task of determining which
were the remaining key-centers of in-
fection to be sanitated, and of col-
laborating with the respective govern-
ments in the organization of the neces-
sary campaigns.  Between 1915 and
1919, in spite of the interruption caused

by the `participation of the United
States in the World War, preliminary
surveys were carried out, and what was
to have been the final drive for the
elimination of th  disease was begun.
A noteworthy res onse
   b   was made by
the governments of the continent, and
adequate authority and funds were
available to carry out the proposed
measures.

During the first decade of Founda-
tion collaboration in the control. of
yellow fever, 1915-1925; the disease
receded on all fronts and, with only
minor and apparently unimportant ex-
ceptions, obeyed implicitly all the rules
which had been' laid down for its
behavior. These rules have fortunately
been outlined for us by Dr. Henry Rose
Carter in a work written shortly before
his death in 1925? This careful and
devoted student of yellow fever, co-
worker with General Gorgas in many
important campaigns, during a lifetime
spent in yellow fever work, had an un-
equalled opportunity for close personal
contact with the disease. Carter says:

Author's Preface-We know quite well the
conditions necessary for the existence of yellow
fever, for it to become epidemic and to con-
tinue at any place; and when we find its
existence or an epidemic reported where the
conditions necessary therefbr do not exist,
we know that the disease reported as such
is not yellow fever.

(page 3) Yellow fever is contracted by man
in nature from the bite of a mosqzdto, AZdes
(Stegomyia) aegypti (Linnaeus), itself in-
fected by having fed on a nzan sick of that
disease, and, so far as known, is in nature
only thus contracted.

(page 7) . . . and, owing, to the invariable
disappearance of yellow fever when this
species (AZdes aegypti)-and this speae-
                       * G atone
-is sufficiently controlled, it seems quite cer-
tain that in the Americas no other mosquito
closely associated with man is a vector.
(page 50) . . . It is a disease of collections
of men-civic rather than rural.
(page 71) . . . The duration of an epidemic,
or even the infection of yellow fever at any
place is limited unless stew human materid


vat. 27      EPIDEMIOLOGY 0~ YELLOW FEVER

3

be supp&zd, and this independently of climatic
conditions.

But during the decade 1926 to 1936,
which may well be termed the age of
disillusion and partial enlightenment for
yellow fever epidemiologists, yellow fever
has refused to abide by these rules and
has not continued to disappear from
the continent in spite of the most
conscientious application of methods
previously so successful. The develop-
ment of methods for the routine diag-
nosis of unsuspected fatal cases and
for the identification by the protection
test of those who have suffered infec-
tion has resulted in the revelation that
yellow fever exists widespread through-
out vast areas in the interior of
South America under conditions pre-
viously considered incompatible with its
existence.

Although the tenets of Gorgas and
Carter are undoubtedly true for certain
regions, it has been amply proved dur-
ing the past few years that they do
not hold for yellow fever in South
America.  Yellow fever does not in-
variably disappear when the A&fes
aegytpi mosquito is sufficiently con-
trolled, and the duration of infection
in many districts is not dependent upon
the introduction of non-immune human
elements into these districts.  Yellow
fever may occur under a great variety
of conditions,  and the factors de-'
termining its occurrence are in large part
still undetermined.  Yellow fever has
been observed in the complete absence
of the AZdes aegypti mosquito as scat-
tered cases in such sparsely populated
rural and jungle areas that the con-
tinuance of the disease in these areas
cannot be attributed to the occurrence
of human infections.

This report presents a short resume
of the efforts to control yellow fever
in South America since 1920 and of the
observations which have resulted in im-

portant modifications of our knowledge
of the epidemiology of this disease.

RlkUMd OF YELLOW FEVER CONTROL IN

SOUTH AMERICA, 192-1929

In 1919 it was believed by the closest
students of the disease that the only
important distributing foci of yellow
fever infection in South America were
the city of Guayaquil, Ecuador, on the
Pacific Coast, and, on the Atlantic
Coast, a small area of Northeast Brazil,
extending `as far south as the city of
Salvador, Bahia. Antimosquito meas-
ures had been organized in Guayaquil in
191 8,3 and these measures resulted in
the disappearance of the disease in 1919
from this city which since 1750 had
been notorious as a center of infection.
Campaigns in Northern Peru in 1920-
192 1 4 were equally successful, and no
known outbreaks of yellow fever have
since occurred on the Pacific Coast of
South America.

The known remaining focus of infec-
tion on the Atlantic seaboard had been
attacked in 1919 through a federal
sanitary commission organized by the
Brazilian Government for work against
yellow fever in all states lying within
the known endemic focus.   Unfor-
tunately no adequate report has ever
been published of the work done by
the CommissZo Sanitaria Federal during
1919 and 1920, and only those who
came into personal contact with that
effort can fully appreciate its scope and
sincerity. But a partial report of the
work in Pernambuco s made before the
failure of this effort was apparent, con-
tains the following significant para-
graph, which is quoted in. translation
from the Portuguese:

Yellow fever constitutes in the north of the
country (Brazil) a problem much more diffi-
cult of solution than was believed; the dis-
semination of the virus extends over a large
area, both within and outside the cities, in
sparsely populated points, and, at times,
failing to obey in a most disconcerting manner


4      AMERICAN JOURNAL OF PUBLIC HEALTH

the epidemiological rules which belong to
yellow fever, to the point of suggesting a
morbid entity not yet catalogued, were it not
for the unusual dissemination of Stegomyia
calopus (At?des (Stegomyia) aegypti) in those
regions together with the clinical observations
and autopsies made by persons well trained
in the subject of yellow fever.

During 1922 it became very apparent
that this federal campaign, which had
been largely if not totally discontinued
at the end of 1920, had failed of its
objective and that yellow fever was
still an important problem in Northeast
Brazil.   In 1923 The Rockefeller
Foundation accepted an invitation to
undertake the solution of the yellow
fever problem in Brazil, and before the
end of the same year Dr. Joseph White,
veteran of many battles with yellow
fever including and following the New
Orleans campaign of 1905, had begun
operations.  Investigations were made
and antimosquito measures were under-
taken at all the principal coastal cities
between Rio de Janeiro on the south
and Manaos, the chief inland port of
the Amazon Valley. No evidence was
found of recent yellow fever south of
Bahia or north of Ceari, and un-
divided attention was finally given to
the region between these points.

Yellow fever dwindled on all sides,
and by 1925 it was believed that the
end of yellow fever on the American
continent was in sight. It is true that
an unexpected outbreak of the disease
had been observed in Bucaramanga,
Colombia, in 1923,' with no satisfac-
tory answer to the question of the
origin of the virus; but in the face
of absolute quiet elsewhere on the con-
tinent during a period of 2 years, and
the increasingly favorable picture en-
countered in North Brazil, the `final
solution of the yellow fever problem on
the continent was anticipated in the
immediate future.

The year 1925 has
as the end of the Golden

Jan., 1937

fever epidemiology. That such was the
case was not, however, fully apparent
until some years later. The 1926 out-
break of yellow fever, which embraced
the interior of several states of North-
east Brazil was attributed to the move-
ment of non-immune troops through the
region, lighting up and fanning to a
visible flame the dying embers of in-
fection which otherwise would have
spontaneously disappeared.  The in-
tensification of antimosquito measures
was followed by a rapid decline in the
number of reported cases, and in 1927
Dr. M. E. Connor, who had taken.over
the technical direction of the service
late in 1926, was able to report, after
a survey trip covering the entire coast
from Bahia to Pari, the Amazon Valley
from the mouth of the river to Man&s,
and the Slo Francisco Valley to Pira-
pora, that he could find no evidence of
yellow fever activity.  The renewed
optimism, born of the failure to recog-
nize `cases of yellow fever anywhere on
the continent during 11 months, was
rudely shattered by single yellow fever
positive autopsies in the States of
Sergipe and Pernambuco in March and
.4pril, 1928, and by the discovery in
May of. the same year that the virus
of yellow fever had again found its
way into Rio de Janeiro, the federal
capital, after an absence of 20 years.
The events of the months immediately
following the discovery of yellow fever
in Rio de Janeiro, one of the principal
ports of South America, fully con-
firmed the great importance attributed
to large centers of population, especially
seaports,  in  the epidemiology of
aegypti-transmitted yellow fever.  Be-
fore the Rio de Janeiro outbreak
terminated in July, 1929, locally in-
fected cases had been recorded in the
states of S5o Paulo, Minas Geraes, Rio
de Janeiro, Bahia, Sergipe, Pernambuco,
and Par-i, and cases infected elsewhere
had been seen at various ports between


Vol. 27

EPIDEMIOLOGY OF YELLOW FEVER

Buenos Aires on the south and the
inland port of ManBos, a thousand
miles up the Amazon.

The rapidity of ,the decline of visible
yellow fever following the suppression
of the Rio de Janeiro outbreak was
noteworthy, and between July, 1929,
and March, 1930, less than a dozen
cases were recorded for all Brazil. How
optimistic was the reaction to this
striking recession of yellow fever late
in 1929 may be ascertained from the
following paragraphs of the Annual
Report of the Yellow Fever Service
(unpublished) for the year 1929:

(page 3) (b) The history of the march of
yellow fever gathered from available data
indicates the centers where the infection might
be expected to remain for an indefinite period
in the absence of sanitary iniervention. Such
centers proved to be the capitals and here
intensive efforts are practised.  Additional
services are started for interior areas whenever
there is need for same and continued until
the situation is again normal or local officials
take up the work.
(page 9) . . . For the first time in the
history of yellow fever prophylaxis, control
measures are being applied in all centers
where this disease was present or was
menacing.

UNEXPECTED OUTBREAKS IN COLOMBIA

AND VENEZUELA, 1929

While the situation in Brazil was
occupying the attention of those re-
sponsible for its study and control, two
totally unexpected and, at that time,
unexplained outbreaks of aegypti-trans-
mitted yellow fever were registered in
Colombia and Venezuela, both in
1929. These outbreaks came at iso-
lated points in the interior, completely
out of contact with each other and with
other known foci of infection. The
place of one of these epidemics,
Socorro, Colombia," is a small town in
the mountains east of the Magdalena
River in the, same general region as
Bucaramanga, where the last previously
recorded outbreak in Colombia had oc-

curred 6 years before. The Venezue!an
outbreak occurred in a series of small
towns in the Cuyuni Valley, close to the
towns of Guasapati and Tumeremo.
This outbreak, like that at Socorro,
cannot be linked in any way with the
1929 dissemination of virus from Rio
de Janeiro.

DEVELOPAMENT OF NEW EPIDEMIOLOGICAL

TOOLS

For a proper understanding of de-
velopments since 1929, it is necessary
to digress long enough to compare the
tools with which epidemiologists worked
before 1930 with those available today.
Previous to 1930, knowledge of the
distribution of yellow fever depended
upon the occurrence of outbreaks of
sufficient severity to be declared by local
authorities.  Acceptable diagnosis de-
pended upon the demonstration of a
series of cases with the classical symp-
toms of yellow fever occurring under
these conditions considered necessary
for its existence.  Final judgment was
suspended on isolated suspect cases
awaiting the occurrence of other cases
in the neighborhood, and such was the
faith in the epidemiological picture of
the disease that the failure to find other
cases or the failure to find aegypti
were at times the basis for deciding
against a diagnosis of yellow fever.
Knowledge of the disease was largely
based on observation of its manifesta-
tions among immigrants in endemic
areas and among non-immune popula-
tions during its occasional excursions
outside of endemic zones.  Epidemio-
logical surveys were limited to a study
of mortality statistics and conversations
with the medical profession and with
local authorities, and such surveys were
thought  sufficient to indicate the
presence or absence of the infection in
an area. Even as late as 1928, the
epidemiological picture of yellow fever


6      AMERICAN JOURNAL OF PUBLIC HEALTH    Jun., 1937

was held to be more concrete and defi-
nite than the pathology of the liver of
the yellow fever patient.

The events of 1928 and 1929, how-
ever, had shown that the pathological
diagnosis based on the microscopical
examination of liver tissue merits great
respkct, and the wealth of material ob-
tained during the Rio de Janeiro out-
break established on a firmer basis
among epidemiologists the belief that
the pathological picture in the yellow
fever liver is not duplicated by that
of any other acute infectious disease.
In 1930 ' the examination of liver tissue,
which had previously been used in
suspect cases, was introduced as a rou-
tine measure for discovering unsuspected
fatal cases of yellow fever in otherwise
silent endemic areas.

  Although the virus of yellow fever
had been successfully established in
laboratory animals in 1927 8 it was not
until 1931 that the perfection of the
mouse protection test o enabled the epi-
demiologist to study the distribution
of immunity to yellow fever on a large
scale to delimit existing' endemic areas,
and not until 193.5 was the easily trans-
ported mouse shown to be readily in-
fectible with the virus of yellow fever
  from human cases in the field.

YELLOW FEVER CONTROL IN NORTHEAST

BRAZIL, Igp-1936

When the speaker became responsible
in June, 1930, for the technical admin-
istration of the CoBperative Yellow
Fever Service jointly maintained by the
Brazilian Government and The Rocke-
feller Foundation, it was already ap-
parent that the brief optimism of a few
months before was not justified. Yellow
fever had just been shown to be present
in the State of Rio de Janeiro in the
south, in the State of Pernambuco in the
northeast, and in the State of Pa& at
the mouth of the Amazon River. The
most important control problem to be

attacked was whether further extension
of antimosquito measures to smaller
towns would result in the elimination
of yellow fever, which " key-center "
control had failed to accomplish, and
the most important epidemiological
problem seemed to be to discover how
the virus of yellow fever is maintained
during periods of " invisibility."

The State of Pernambuco, a long,
narrow area extending directly west-
ward from the Atlantic Ocean, possesses
in more or less well defined belts the
various climatic conditions characteristic
of Northeast Brazil, and was chosen for
the initiation of intensified control
measures and for special studies. An
analysis of the available information
regarding the distribution of yellow
fever in Pernambuco previous to 1930
seemed to- justify the division of the
state into 4 rough epidemiological
zones :

1. Key-renter Endemic-Recifc, the
capital city and principal port, with a
population of more than 250,000, a
typical key-center of yellow fever in-
fection which might be expected in the
absence of antimosquito measures to
maintain endemicity in and of itself,
with the more or less constant occur-
rence of " visible " yellow fever in immi-
grants from non-endemic areas.

2. Regional Endemic-The heavily
populated coastal plain, where impor-
tant outbreaks of yellow fever were
rarely reported but where the presence
of the virus was from time to time
made manifest through the infection
of occasional foreigners.

3. Endemo-epidemic - A relatively
heavily populated zone just back of the
coastal plain, where yellow fever was
frequently reported as sharp epidemics
in the native population.

4. Non-injectible-The sparsely pop-
ulated SertZo or hinterland comprising
roughly the western half of the state,
subject to occasional droughts of great


vol. 27     EPIDEMIOLOGY OF YELLOW FEVER            7

severity, with no history of yellow fever
in the past.

The plan of campaign drawn up in
June, 1930, called for the routine or-
,oanization of anti-aegypti measures in
all cities and towns.of 2,000 population
and over throughout the key-center en-
demic, the regional endemic, and the
endemo-epidemic zones, leaving only
the (` non-infectible " zone without con-
trol. The extent to which this program
differed from the original key-center
theory of control can be appreciated
when it is remembered that Carter 2* P. 2o
states in discussing the spontaneous dis-
appearance of yellow fever:

It has disappeared from Cartagena, Bar-
rsnquilla, Maracaibo, Cucuta (towns of from
20,000 to 75,000 people) ; from all of Hayti
and Santo Domingo; from Porto Rico, St.
Thomas and a host of smaller places; and
this without any sanitary work and with
aegypti abundant.

Only a few months intervened be-
tween the inauguration of this intensive
program and the perfection of the
viscerotome, a.n instrument for the rapid
removal of pathological specimens post-
mortem without autopsy lo which made
possible the extension of the routine
collection.of liver tissue from the bodies
of persons dying in 10 days or less `after
onset of illness, for the diagnosis of
unsuspected yellow fever. Viscerotomy
confirmed the absence of yellow fever
from Recife, the capital city, but re-
vealed fatal cases in the regionaE en-
demic, the endemo-epidemic and " non-
injectible " zones! `In other ivords, key-
center control in Pernambuco after 7
years' constant application had resulted
in the elimination of yellow fever only
from the key-centers worked!

The installation of anti-aegypti serv-
ices in the towns of the eastern part of
Pernambuco was followed by the rapid
disappearance of positive diagnoses on
liver tissue collected in this area. The
theory of key-center control only needed

application to smaller towns than had
been previously worked to give the
desired results. Not so in the SertHo!
The  organization of  anti-aegypti
measures was carried to smaller and
smaller centers of population, until
villages of less than 50 houses were be-
ing worked; but nevertheless, visce-
rotomy continued to reveal fatal cases in
this " non-infectible " hinterland. After
conclusive proof had been obtained
that the disease would not spontaneously
disappear from the rural districts, the
so-called  " fine-tooth-comb "  method
involving anti-aegypti measures in every
home throughout rural districts was
authorized for the Sert%o.   This
method eventually gave satisfactory
results, and since August, 1934, visce-
rotomy has failed to reveal yellow fever
in Pernambuco or in the 5 neighboring
states lying between the %o Francisco
and Parnahyba Rivers.
It should be noted here that, al-
though Northeast Brazil has been more
intensively studied than any other part
of South America in recent years, no
evidence of the occurrence of yellow
fever without AL;des aegypti has been
`found in that region, and, conversely,
the disease has not continued long in
those places where the incidence of
this mosquito has been reduced by
control measures.

The reason underlying the success of
anti-aegypti measures limited to the
cities and towns in the coastal plain
and the failure of the same measures
limited to towns in the interior merits
discussion.  Investigation has shown
that the ri&fes aegypti mosquito, al-
though widely disseminated throughout
the towns and villages of the coastal
plain, is not common in the rural habi-
tations of this zone; on the other hand,
the dry interior is unique among all
the' districts investigated in Brazil in
that there is a high index of infestation
with this mosquito in the rural homes


8      AMERICAN JOURNAL OF PUBLIC HEALTH   Jan., 1937

throughout a wide area.  This excep-
tional distribution of AZdes aegypti is
attributed to the fact that the dryness
of the climate obliges the inhabitants
to store whatever water is to be had for
domestic use. The containers used for
storing water provide ideal breeding
places for AZdes aegypti once it has been
introduced, and the custom, bred of
necessity, among those travelling on
business and on religious pilgrimages, of
carrying drinking water containers, has
provided the means of such almost
universal introduction.

The experience in Northeast Brazil
has confirmed the opinion expressed in
the report of the CommissHo Sanitaria
Federal (vide supra) that yellow fever
will not obey the epidemiological rules
in the face of an unusual dissemination
of AZdes aegypti but has shown that
the extension of anti-aegypti measures
to strictly rural regions is practicable
and efficacious.  No evidence of the
presence of yellow fever has been found
in Northeast Brazil for more than 2
years, and control measures could prob-
ably be safely abandoned were it not
for the danger  of reinfection from
nearby areas demonstrated to harbor
the virus of yellow fever in the absence
of aegypti and under conditions for
which adequate methods of control
have not yet been evolved.

YELLOW FEVER WITHOUT AiiDES AEGYPTI

JUNGLE YELT.OW FEVER

Although the demonstration in 1928 l1
that the virus of yellow fever can be
transmitted in the laboratory by various
species of mosquitoes found in Africa,
other than Aedes aegypti, had been
followed, as early as 1929,lz by a
similar demonstration for some of the
South American species, it was not until
1932 l3 that the natural occurrence of
yellow fever in the absence of aegypti
was fully confirmed.  The full sig-
nificance of this first observation of

yellow fever without Aifdes aepypti in
the Valle do Chanaan, Espirito Santo,
Brazil, was not understood at the time,
since the spontaneous termination of
the disease within a few weeks after
coming under observation, when a very
high percentage of the population of
the valley was still non-immune, sug-
gested that this outbreak had prob-
ably been due to a temporary excursion
of the virus under relatively unfavor-
able conditions from nearby towns and
cities where no yellow fever could be
found but where high densities of
aegypti and high percentages of im-
munes were encountered.  Later ob-
servations of yellow fever without
aegypti, to which the name jungle
yellow fever has been givenI have sug-
gested that it may be the more natural
and permanent form of yellow fever
and that the aegypti-transmitted type
may be a highly exotic type maintained
with great difficulty because of the
tendency to burn itself out through the
exhaustion of non-immune populations.

Scarcely had the outbreak in the Valle
do Chanaan been identified when epi-
demiologists were astounded by an out-
break of yehow fever at Santa Cruz de
la Sierra, Bolivia, a town of less than
20,000 population, completely isolated
from other larger centers of population,
lying in the heart of the continent more
than 1,000 kilometers from any focus
of infection known during the previous
two decades. The Santa Cruz epidemic
occurred in the presence of a high
density of aegypti and responded
rapidly to anti-aegypti measures, but
no reasonable explanation of the source
of virus for this outbreak was found
until viscerotomy revealed some months
later that yellow fever was present at
San Ramon, a small Indian village of
the lowlands of Bolivia, in a region in
which AC;des aegypti has not been
found!

By 1933, immunity survey results


Vat. 27      EPIDEMIOLOGY 0~ YELLOW FEVER           9

were available from various parts of the
Amazon Valley indicating that yellow
fever had not been absent from any
large portion of the valley in the pre-
vious 20 years during which it had not
been recognized. Likewise, independent
protection test studies in Colombia l5
had suggested the maintenance of the
virus of yellow fever in certain rural
regions of the Magdalena Valley. Be-
fore the end of the year, viscerotomy
had confirmed fatal cases of jungle
yellow fever for the Magdalena Valley
at Caparrapi and for the Amazon
Valley at Lauro SodrC, close to the
Brazilian, Colombian, and Peruvian
frontiers.

The observations of jungle yellow
fever previous to 1934, however, are
relatively insignificant in comparison
with those made during the years 1934,
1935, and 1936, which have shown that
it is a very important epidemiological
entity, not only as a factor in the
preservation of the yellow fever virus,
but also as a cause of illness and death
among rural and jungle populations.
Even those most intimately in contact
with the work on yellow fever were
totally unprepared for the rapidly un-
folding epidemiological picture of the
last 3 years. The events of these years
include :

1. The discovery in 1934 of jungle
yellow fever, confirmed by clinical obser-
vation, autopsy, immunity t&ts and
isolation of the virus with reproduction
of the disease in susceptible animals,
at Coronel Ponce on the Planalto of
Matto Gross0 `0  100 miles east of
Cuyabi, the state capital. Yellow fever
as an urban disease had been known
late in the past century in Cuyabir, but
there was no history of its occurrence
anywhere in the state during the pre-
vious 30 years. An attempt to incul-
pate Cuyabb, and other' towns of the
state where high densities of AZdes
aegypti exist, as the source of virus for

this rural outbreak failed in `the face
of negative histories substantiated by
negative protection test results in the
younger age groups of the population
of these towns.  However, immunity
studies made on Indians at Sim6es
Lopes, an isolated government post
about 100 miles north of Coronel Ponce,
showed high percentages of immunes
among all age groups and suggested
that the infection of the Planalto came
from the jungle areas of the Amazon
Valley.

The slow moving outbreak on the
Planalto was carefully studied during
1934 and 1935. The scattered popula;
tion of this region had apparently not
been exposed to yellow fever in recent
years, since histories of recent illness
were obtained from all persons with
positive protection tests. Investigation
showed that the local popular idea that
the new disease was contracted in the
fields next to the uncut forest was well
founded. The paucity of human popu-
lation in this area and the time and
space distribution of cases make it
highly improbable that man was at any
time an important source of virus for
human infection or was an appreciable
factor in the dissemination of the dis-
ease from ,one point to another. Posi-
tive protection tests obtained with the
sera of wild monkeys l7 indicate that
infection in this area has not been
limited to human beings.

2. The finai proof, by clinical ob-
servation, autopsy, and isolation of the
virus in susceptible animals, that jungle
yellow fever continues endemic in the
Muzo region of Colombia, as had been
suggested by previous protection test
results.15  Yellow fever was reported
in Muzo even during the past century,
but its careful differentiation from
malaria and relapsing fever, both
endemic here,`was first made in 1907.`*
Since this early confirmation, cases of
yellow fever among the non-immune


10     AMERICAN JOURNAL OF PUBLIC HEALTH   Jan:, 1937

laborers brought from the highlands of
Colombia to work the famous emerald
mines of Muzo have been sufficiently
numerous to cause repeated investiga-
tions but have not been numerous
enough  tb explain  the continued
endemicity of the past 30 years in this
sparsely populated region.  Although
Aedes aegypti was reported as present
in Muzo in 1907, repeated investiga-
tions since have failed to confirm its
existence anywhere in the area, and a
careful reading of the early descrip-
tion leads to the conclusion that the
classification was  in  error.14  Epi-
demiologists who investigated Muzo
during the quarter of a century follow-
ing the first description of yellow fever
there and before the introduction of
the protection test, failed to convince
themselves, in the absence of aegypti
and of explosive outbreaks involving
many cases, that the disease under in-
vestigation might be yellow fever!
Even Gorgas, lo following a visit to the
n&es in 1916, declared that the suspect
cases which had caused his visit could
not have been yellow fever, since the
conditions necessary for endemicity did
not exist.

Viscerotomy has resulted in the con-
firmation of cases at Muzo in March,
June, and October, 1934; in December,
1935 ; and in January, April, and May,
1936. Cases have been limited, so far
as is known, to persons who had in-
timate contact with the forest, only a
few cases occurring at a time. The or-
ganization of viscerotomy throughout
the surrounding region has shown that
endemicity is not strictly limited to the
immediate vicinity of Muzo but ap-
parently covers a large sparsely popu-
lated district on the right bank of the
Magdalena River,  between this river
and the high mountain ranges to the
east. The proof that this district har-
bors endemic jungle yellow fever offers
at last an acceptable source of virus for

the unexplained outbreaks, occurring
after periods of several years of apparent
absence of the disease from Colombia,
in Bucaramanga in 1923, and in
Socorro in 1929, both of which towns
are within a short distance of the known
endemic district.

Muzo is unique in that it has a
definite history of constant endemicity
of jungle yellow fever during at least 30
years; it is probable that the recently
isolated Muzo strain of virus has not
been transmitted by the ABdes aegypti
mosquito during this period.

3. The identification of a rural epi-
demic in 1934 on the eastern slope of
the Andes at Restrepo, just at the edge
of the llanos in Colombia, as yellow
fever.  This. general region, which lies
in the drainage basin of the Orinoco
River, had in 1886 suffered G an out-
break suspected of being yellow fever,
but there was no historical reason for
believing this to be an endemic region.
Careful investigation has shown that
even the larger towns of the region are
not even now infested with aegypti, and
although cases have continued to be
recorded during the years 193.5 and
1936, some of which have apparently
been infected almost at the edge of the
towns, no cases are known to have been
infected in the towns themselves.

4. The diagnosis, through visce-
rotomy, of, 4 isolated cases-l in May,
1 in June, and 1 in July, 1934, and 1
in May, 1935within a comparatively
small area on the Ilha Marajo, the large
island lying at the mouth of the Amazon
River.20  Careful investigation has in-
dicated that some undiagnosed cases
may have occurred, but has failed to
bring any clinical cases under observa-
tion. The circumstances strongly sug-
gest that there is some reservoir of virus
in this region quite independent of the
human population.

5. The demonstration that the anti-
aegypti campaign, begun in the lowlands


Vol. 27      EPIDEMIOLOGY OF YELLOW FEVER

of Bolivia in 1932 at Santa Cruz de la
Sierra, had probably not had any im-
portant effect in limiting the distribution
of the jungle virus, although only 1
case was diagnosed by viscerotomy in
Bolivia in 1933, and none were found
during 1934.  Yellow fever without
aegypti was shown to be present in the
war zone south of Santa Cruz in 1935.
The present year, 1936, has been a
banner year for viscerotomy in Bolivia,
and diagnosis of yellow fever based on
liver examination has been made at
numerous points on all sides of Santa
Cruz, at many places in the foothills
of the Andes, at San Ram6n and San
Javier in the San Miguel River Valley,
at Trinidad on the River MamorC, at
Huichini, on the River Beni, and at
Cobija, on the frontier of the Territory
of Acre, Brazil.

6. The observation during 1935 and
1936 of an extensive outbreak of jungle
yellow fever involving the southern part
of Goyaz, the western and southwestern
parts of Minas Geraes, a large part of
the state of SCo Paula, the northern part
of Paranb, and the southern- part of
Matto Grosso. Most of the territory
covered by this outbreak had never pre-
viously reported yellow fever, and none
of it was known to have produced cases
during more than a quarter of a century.
The area is far from uniform in
topography and density of population,
and ranges from sparsely settled regions
where there are a few settlers along the
river valleys, to very rich densely popu-
lated districts,  The known infected
district covers several hundred thousand
square miles; it is impossible even to
hazard a guess as to the number of
cases which have occurred, beyond
stating that it is known to be in the
thousands. This is by far the most im-
portant and extensive outbreak of
jungle yellow fever yet observed and is
responsible for a new conception of the
threat this disease represents for rural

populations throughout much of South
America.

7. The occurrence of small isolated
outbreaks of aegypti-transmitted yel-
low fever under circumstances strongly
suggesting that the virus for the pro-
duction pf such outbreaks had come
from nearby areas of jungle infection.
Outstanding examples of such outbreaks
are that of Theofilo Ottoni in the state
of Matto Grosso, Brazil, in 1935,20 and
that of Cambar& in the state of Paran&
in 1936.

DISCUSSION OF JUNGLE YELLOW FEVER
There is ample reason for believing
that aegypti-transmitted yellow fever
and jungle yellow fever are identical
except for the conditions under which
infection occurs. The identity. of the
disease occurring in the absence of
aegypti in the Valle do Chanaan, in
Matto Grosso, in Goyaz and in both
the Magdalena  and Orinoco River
Valleys in Colombia, has been fully
established by clinical observation,
autopsy, protection test, and isolation
of the causative virus with reproduction
of the disease in susceptible animals.
Furthermore, the jungle virus has been
transmitted from monkey to monkey
in the laboraiory by the AZdes aegypti
mosquito *I just as the urban virus has
been transmitted by various jungle
mosquitoes. Aegypti-transmitted yellow
fever is largely a house disease, since the
vector breeds and spends its entire life
either in or in immediate proximity to
human habitation; the close contact
between vector and cases in the home
results in building up a high concen-
tration of virus in the vector; thus most
of the non-immune members of the
household become infected before the
disease dies out because of a shortage
of new cases to infect succeeding agen-
erations of the short lived vector.

In jungle yellow fever, on the other
hand, man is not an important source


12      AMERICAN JOURNAL

of virus for the infection of human
cases; the human case is apparently an
accident in the course of some as yet
unknown extra-domiciliary cycle of in-
fection. In support of this idea,%udies
of immunity distribution in wild animals
in South -4merica have already shown
that animals of various types,14, 17, **
including opossums and several species
of monkeys, may acquire immunity un-
der natural conditions in the jungle.

Much further study is necessary be-
fore the entire story can be known;
the demonstration of one cycle of in-
fection in the jungle should not be in-
terpreted to exclude other cycles, and
a thorough search must be made for all
possible blood-sucking vectors and all
possible vertebrate hosts. The possi-
bility of virus reservoirs consisting of
long lived vectors, vectors capable of
transmitting the virus from generation
to generation, or even vertebrates in-
capable of acquiring immunity and
therefore permanent reservoirs of the
virus, when once infected, must be
investigated.

The questions most frequently asked
by those who learn for the first time of
the recent developments in the epi-
demiology of yellow fever are, " Is
jungle yellow fever a recent adaptation
of the virus of yellow fever to non-
domestic conditions? " and `( If not a
recent development, how could jungle
yellow fever escape observation and
description in the past? "

The answer to these two questions is
that jungle yellow fever is not a recent
adaptation but may even be the original
and natural form of yellow fever and
that it did not entirely escape observa-
tion and description in the past. There
are two quite convincing descriptions of
yellow fever from former times and a
third, published many years after the
observations on which it is based, which
undoubtedly refer to what is now known
as jungle yellow fever.

OF F+JBLIC HEALTH    Jan., 1937

The first description of this kind
which has come to my attention is the
report 23 of an investigation made in
1887 among the Catholic missions along
the foothills of the Andes south of
Santa Cruz de la Sierra, Bolivia. A
series of cases diagnosed as yellow
fever by a clinician familiar with the
disease in Cuba, is described at Abapo,
El Espifio, Muchiri, and Masavi, in
exactly the same region in which yellow
fever without aegypti was encountered
among the troops in the trenches in
1935. Quite naturally this early report
carries no reference to the presence or
absence of AZdes aegypti, since it was
written some 13 years before the trans-
mission of yellow fever by this
mosquito was first demonstrated. The
authors realized that their diagnosis
would be criticised and probably not
accepted, as indeed it was not, no
reference to this report reaching the
international literature until after the
discovery of yellow fever in Santa Cruz
in 1932. ,To quote in translation from
the Spanish:

Without doubt many authorities will be
unwilling to admit the presence of yellow
fever, isolated in the heart of Bolivia, far
from the places in which yellow fever gen-
erally exists and without known routes of
penetration from without.

These words fittingly describe the
dificulties of accepting the same diag-
nosis in this same area 45 years later.
That this last outbreak should not have
been unexpected is suggested by the
fact that the 1887 epidemic was ap-
parently not the first. To quote in
translation:

The older inhabitants state that Abapo was
stricken some 20 years ago by a disease which
presented these same symptoms and which
Dr. Jo& Lorenzo Sanchez, who was then in
charge of the local parish, called yellow fever.
Likewise this year, it was a priest who an'-
nounced the true name of the disease.  The
virtuous Seiior Sarrazona, the priest at
Abapo, who was to be one of the first to


vol. 27      EPIDEMIOLOGY OF YELLOW FEVER            13

pay tribute to the scourge, styled it, as a
good Spaniard would, Black Vomit.

The second description of jungle
yellow fever is that made at MUZO,
Colombia, in 1907.l' The careful clin-
ical observations then recorded do not
permit any doubt that yellow fever was
present and the epidemiological notes
point conclusively to jwngle yellow
fever. To quote in translation:
(a) It is contracted in the forest and not
in the neighborhood of the houses.
(b) It is transmitted by Stegomyia calopus
and probably also by other culicines.
(c) Inoculation takes place during the day-
light hours, which are spent by the workers
in places where the transmitting mosquitoes
predominate.

Except for the statement that
" Stegomyia calopus " (ABdes aegypti)
was present, the description of yellow
fever as given for Muzo in 1907 can
stand as a true description of the jungle
yellow fever observed there in 1934,
193.5, and 1936. That the mosquitoes
diagnosed as Stegomyia calopus were
not Addes aegypti is clearly indicated
by the description of the habits of these
mosquitoes and the failure of numerous
investigations between 1916 and 1936
(vide szrpra) to reveal this mosquito
at Muzo.

In addition to these descriptions
there is the reference made in 1929 24
to two outbreaks many years before in
the State of S5o Paulo, Brazil, reputed
to have occurred in the absence of
aegypti. This reference, made several
years before the observations in the
Valle do Chanaan, is very interesting in
that it describes so closely the condi-
tions under which jungle yellow fever
has recently been found.

The first of these outbreaks, of which I
know only from the reports of others, oc-
curred in an Indian village on the Rio Verde.
ln the other case which I personally investi-
gated, yellow fever was found in some huts
built in the midst of the forest and inhabited
by laborers.  These men were cutting the

forest in preparation for the constrtiction of
a railroad which was to join Funil to
Campinas.  I examined several huts from
which had come cases of yellow fever, with-
out finding any trace of larvae or adults of
Stegomyiu, although forest mosquitoes were
present.  This observation is especially in-
teresting since it has lately been shown in
Africa that transmission may occur through
mosquitoes  other than the domesticated
stegomyia. This latter will always be among
us the most important vector and the trans-
mission by other species must be rare and
exceptional, but nev&theless the determina-
tion of other species which may transmit the
virus is an interesting problem.
During the last few years, Ai;des
aegypti has not been the most important
vector but may be expected rapidly to
regain its dominant position should
anti-aegypti measures be abandoned.
Paradoxically, one of the most im-
portant results of the observation of
endemicity without aegypti has been
to emphasize the need of permanent
anti-aegypti services throughout en-
demic areas and in regions in close con-
tact with endemic areas. And today
it is true that the entire continent is,
through the development of rapid
transportation facilities, in close contact
with endemic areas.  With present
facilities it would be possible for the
members of this Association to leave
New Orleans, visit a known infected
region, probably seeing active cases,
and return to the United States, all
within the generally accepted 6 day
incubation period of yellow fever. The
situation is not one calling for undue
alarm, but it is one which merits careful
evaluation of the implied threat of pos-
sible extension of infection.

SUMMARY

Work during recent years in South
America has shown that-

1. Elimination of yellow fever from the
larger " key-centers " of population by anti-
mosquito measures will not result in the
disappearance  of the disease from the
continent.


14      AMERICAN JOURNAL OF PUBLIC HEALTH

2. Aegypti-transmitted yellow

fever may,

ratory Animals.  Am. 1. Trap. Ilied., 8:103-164
(Mar.). 1928.
9. ikvyer, W. A.: and Lloyd, W. The Use of
Mice in Tests of Immunity Against Yellow Fever.
1. &per. Med., 54:533-S% (Oct.), 1931.
10. Rickard, E. R.  The Viscerotome, an Instru-
ment for the Removal of Fragmenb of Liver for
Pathological Examination without Autopsy. Rocke-
feller Foundation Quart. Bull., 5:310-320 (July),
1931.

fan., 1937

under the conditions prevailing in Northeast
Brazil, maintain itself for long periods of time
as a rural disease requiring organization of
anti-aegypti measures in strictly rural dis-
tricts for its elimination.

3. There exists throughout vast areas of
the continent, a previously unsuspected epi-
demiologica! type of yellow-kver to which
the name " jungle yellow fever " has been
given. Jungle yellow fever differs from the
classical disease in that it occurs in the ab-
sence of AZdes aegypti, and under conditions
previously considered incompatible with yellow
fever endemicity.

4. Jungle yellow fever, in the absence of
adequate methods for its control, must be
considered as a possible permanent `source of
virus for the re-infection of cities and towns
where high densities of AZdes aegypti are
tolerated.

REFERENCES

1. Reed, Walter; Carroll, James; and Agramonte,
Aristides. Ex;:erimental Yellow Fever. Tratrs. Assn.
Am. Phyr.,  16:45-70, 71, 1901.
2. Carter, Henry Rose.  Yellow Fever: An Epi-
demiological and Historical Study of Its Place 01
Otigin.  Williams 81 Wilkins, 1931.
3. Connor, M. E.   Yellow Fever Control in
Ecuador. I.A.M.A., 74:650 (Mar. 61, 1920.
4. Carter, Henry Rose.  Yellow Fever in Peru:
Qidemic of 1919 and 1920. An. 1. Trap. Med.,
2, 87-106 (Mar.), 1922.

5. Coma da Cmta, Clovis.   Febre Amarella.
Notas sobre 0s trabalhos da Cornmiss& Sanitaria
Federal em Pernambuco.   Arch. Brazileiros de
Medicim. 1920.

6. Pena Cbavarria, Antonio; Serpa, Roberto; and
Be&r, George.  Yellow Fever in Colombia with
Special Reference to the Epidemic in Socorro in
1929. 1. Prev. Med., 4:417-457 (Nov.), 1930.
7. Soper, F. L.; Rickard, E. R.; and Crawford,
P. J. The Routine Post-mortem Removal of Liver-
Tissue from Rapidly Fatal Fever Cases for the
Dixovery of Silent Yellow Fever Foci. Am. 1. Hyg.,
19:549-566 (May), 1934.
8. Stokes, A.: Bauer, J. H.: and Hudson. N. P.
Experimental Transmission of Yellow Fever to Labo-

11. Bauer. J. H. The Transmission of Tellow
Fever hv Mosauitws other than AMes oe~vbti. Ant.
1. Trap: Med.: 8~261-282 (July). 1928.--'
12. Davis, N. C.; and Shannon, R. C. Studies on
Yellow Fever in South America:  V. Transmission
Experiments with Certain Species of Culex and
AUes.  1. Exper. Med., 50, 803-808 (Dec.), 1929.
13. Soper, F. L.; Penna, H.; Cardoso, E.; Sara-
fin J., Jr.;  Frobisher, fi4., Jr.; and Pinheiro, J.
Yellow Fever without AZdes oegypti. Study of a
Rural Epidemic in the Valle do Chanaan, Espirito
Santa, Brazil, 1932.  Am. 3. Hyg., 18, 555-587
(Nov.), 1933.

14. So er, F. L. Rural and Jungle Yellow Fever.
A New Public Health Problem in Colombia. Rev.
de Hi&nc, Bogotd, 4:47-84 (May-June), 1935.
15. Kerr, J. A.: and Patino Camargo, L. Investi-
gaciones sobre fiebre amarilla en Muzo y en la regi6n
de Santander.  Rev. de Higiene, Bogotd. 2:63-91

(A4ar.), 1933.

16. Burke, A. W. An Epidemic of Jungle Yellow
Fever on the Planalto of Alatto Grosso, Brazil. In
~reyaration.

17. Soper; F. L. Jungle Yellow Fever:  A New
Epidemiological Entity in South America. Rev. de
Higiene e Sande Publica, Rio de Janeiro, 10:107-
141 (Apr.), 1936.

18. France. R.: Mar&z Santamaria, J.; and Two
Villa. G.  Fiehre Amarilla y Fiebre Espiroquetal.
Academia National de Medicina, Sesidnes scien-
tificas de1 Centenario, Bog&, 1911.
19. Gorgas, w. c.  Rev. de Higiene, Bogotd, 8
(Jan.), 1917.

20. Soper, F. L. Recent Extensions of Knowl-
edge of Yellow Fever. Quart. Bull. Health Organiso-
tion 01 the League 01 Nationr, 5:19-68 (Mar.), 1936.
21. Whitman, L.  1935.  Personal communication.
22. Wilson, D. B.  1936. Perwnal communication.
23. Ortiz, Nicholris, and Cam6 Montobbio, Jose.
Fiebre amarilla en Bolivia, 1887.  Original report
reprinted in La Revisto del Institute Medico, Sucre,
Bolivia, 1932.
24. Lutz: Adolpha.   Reminiscenciae da  fehre
amarella no Estado de SHo Paula.  Memorias do
lnrtiluto Oswoldo Cnrz. 24, No. 3, 1930.