CHAPTER VI
Bullis Fever
John C. Woodland, M.D.
During World War II, early in the spring of 1942, medicalofficers on duty in the Contagious Disease Section, Brooke General Hospital,Fort Sam Houston, Tex., first recognized that a disease entity1with which they were dealing defied definite identification. There were severaloutstanding clinical features characterizing the illness: An initial chillfollowed by fever, an unusually low leukocyte count with an associatedneutropenia, a severe postorbital and occipital headache, and lymphadenopathy.In all cases, there was evidence of multiple tick bites. All the patientscomprising this group were soldiers of the Army Ground Forces of Fort SamHouston who had been on maneuvers at Camp Bullis, Tex., a training area locatedsome 20 miles from San Antonio.
When it became apparent that a condition existed presenting aproblem in diagnosis, The Surgeon General was requested to furnish consultantsespecially qualified in that field of medicine which was related to the problem.In response to the request, three members of the Board for the Investigation andControl of Influenza and Other Epidemic Diseases in the Army arrived at Fort SamHouston on 8 July 1942. After examining a number of the patients in the hospitalsuffering from the disease and after a minute scrutiny of the clinical recordsof those who had been under observation and treatment, the consultants agreedthat this was a definite disease found only in those soldiers who had been onduty at Camp Bullis and who had been bitten by ticks.2It was the Board's opinion that the epidemiological evidence associatingthis illness with the bite of the tick was highly suggestive but that thecausative agent had not been conclusively identified nor could it be definitelyshown at that time that the disease was transmitted to man through the tick asvector.
1It is not often that a new disease entity, including the causative infectious agent, and its insect vector are revealed with such clarity and in such short space of time. Consequently, this account of the discovery of "Bullis Fever," and the research studies conducted, will be read with intense interest. The author, Dr. John C. Woodland, in collaboration with Drs. M. M. McDowell and J. T. Richards, wrote the original article (see footnote 3, p. 158), calling attention to the new disease, while Army medical officers stationed at Brooke General Hospital, Fort Sam Houston, Tex., in 1943.-A. L. A.
2Letter, Kenneth F. Maxcy, M.D., Consultant to the Secretary of War, to The Surgeon General, 10 Aug. 1942, enclosure thereto, subject: Investigation of Cases of the Typhus-Spotted Fever Group, Station Hospital, Fort Sam Houston, Texas, July 18-25, 1942.
158
CLINICAL MANIFESTATIONS
Woodland, McDowell, and Richards, in the original article onBullis fever,3 describedit briefly, as follows:
The illness was usually ushered in with an initial chill or achilly sensation, which was soon followed by fever, the temperature ranging from102? to 105? F. A great majority of the patients complained of severe headachelocated in the postorbital and occipital regions. Pronounced lassitude,prostration, anorexia, and general weakness were noted during the febrile stageof the disease, and occasionally there was nausea and vomiting. The fever lastedfrom 4 to 14 days and subsided by lysis. In the average case, the temperatureremained elevated for 5 days. Convalescence was protracted, especially in themore severe cases. Loss of weight was observed in many of the patients, onepatient losing as much as 20 pounds (9 kg.) within a fortnight. It has beenestablished that the incubation period in Bullis fever is from 7 to 10 days.
PHYSICAL FINDINGS
There was usually a paucity of physical findings in patientswith Bullis fever; however, enlargement of the lymph glands was characteristic.At times, only a regional group of glands were involved, but commonly there wasgeneral lymphadenopathy. The glandular involvement persisted throughout theacute stage of the illness and disappeared rather promptly with the clearing ofthe symptoms. The throat at times was slightly red and injected, but symptomsreferable to the respiratory tract were notably absent. In the more severe formsof the disease, a maculopapular rash, involving the trunk and, later, theextremities, appeared during the febrile stage and usually disappearedcompletely within 48 hours. This fleeting eruption was similar to the rash seenin endemic typhus. In all instances, examination revealed multiple tick bites.Enlargement of the spleen was also noted in the more severe forms of thedisease. Subconjunctival hemorrhage was found in the more seriously illpatients.
LABORATORY FINDINGS
A constant laboratory finding was a definite leukopeniaoccurring on or about the second or third day of illness, the leukocyte countdropping to 3,000 cells per cubic millimeter or less. With the abrupt drop inthe total number of leukocytes, there was an associated neutropenia, thedifferential count showing polymorphonuclears as low as 23 percent. During theperiod
3Woodland, J. C., McDowell, M. M., and Richards, J. T.: Bullis Fever (Lone Star Fever-Tick Fever). An Endemic Disease Observed at Brooke General Hospital, Fort Sam Houston, Tex. J.A.M.A. 122: 1156-1160, 21 Aug. 1943.
159
of convalescence, the leukopenia gradually disappeared andthe differential count approached normal. All other laboratory procedurescommonly performed showed no abnormal findings in Bullis fever. Complementfixation tests for Q fever, Rocky Mountain spotted fever, and endemic typhusfever yielded negative results. The Weil-Felix test, using Proteus X-19,X-2, and OX-K, was consistently negative. Agglutination determinations wereuniformly negative for undulant fever, tularemia, typhoid fever, and paratyphoidfever. The heterophile antibody reaction was negative. Cultures of the bloodwere sterile. The spinal fluid was normal. Biopsy of enlarged lymph glandsdisclosed only lymphoid hyperplasia.
CLINICAL COURSE
The disease is self-limited in nature and varies from a mildfebrile illness of short duration to a severe, debilitating, prolonged diseasewith a protracted convalescence. There have been two deaths attributed to Bullisfever. No form of therapy appears to affect the duration or severity of theillness. The sulfonamide drugs and penicillin have been used in a number ofcases without apparent benefit.
EXPERIMENTAL LABORATORY STUDIES
During the spring and summer of 1943, about 485 cases ofBullis fever were observed at Brooke General Hospital. Livesay and Pollard,4after exhaustive laboratory investigation of the disease, concluded-
1. That the clinical syndrome referred to as"Bullis Fever" has no immunological relationship to Rocky MountainSpotted Fever.
2. That the disease does not appear to betyphus, based on guinea pig reactions, Weil-Felix reactions, andcomplement-fixation tests; nor "Q" fever, based on clinical syndromeand complement-fixation tests.
3. That a Rickettsia-like agent from clinicalcases has thus far been passaged through three and five series of guinea pigs,inducing mild febrile reactions on the ninth to twelfth days, without orchitis.
4. That Rickettsia-like bodies have beenobserved in hyperplastic lymph nodes from men ill with this disease, and inguinea pigs killed during the febrile stage of the reaction.
5. That "Bullis Fever" is apreviously undescribed syndrome in which Rickettsiae appear to be theetiological agent, without the development of a significant Weil-Felix reaction.
6. That it may be associated with an arthropodvector, the tick or Trombicular, in the Camp Bullis area.
7. Epidemiologically it appears that thisdisease is more severe and more prevalent in 1943 than in 1942. This maybeinterpreted with qualifications as follows:
a. Either the agent is increasing in virulenceas a result of repeated passage, or,
4Livesay, J. R., and Pollard, M.: Laboratory Report on a Clinical Syndrome Referred to as "Bullis Fever." Am. J. Trop. Med. 23: 475-479, September 1943.
160
b. The human population being exposed to thisdisease is more susceptible: in the past the troops occupying this area havebeen largely local troops who perhaps had developed some degree of immunity. Thelatest exposures, and those in whom the disease has been rather severe andhighly prevalent, are new troops, for the most part originating in relativelytick-free parts of the country such as Chicago or New York.
Shortly after this, Anigstein and Bader, working in theirlaboratories at the University of Texas, in Galveston, Tex., recovered arickettsialike micro-organism from guinea pigs inoculated with Lone Star ticks (Amblyommaamericanum) that had been collected at Camp Bullis.5This micro-organism resembled the one found in patients by Livesay and Pollard,with regard to morphology, cultural characteristics, and behavior in laboratoryanimals.
The next laboratory work of importance which aided inestablishing Bullis fever as a distinct entity was accomplished by Livesay andPollard in the spring and summer of 1944.6 Inthis work, they attempted the serologic identification of Bullis fever and itsdifferentiation from some of the known rickettsial agents. The lack ofrelationship between the agents of Bullis fever and those of Rocky Mountainspotted fever had already been demonstrated by a guinea pig challengeexperiment. A suggestion of similarity in the clinical picture between Bullisfever and Q fever was then investigated and was not confirmed.
Serum specimens were collected from patients in whom theclinical diagnosis of Bullis fever was made. In addition, specimens of serumwere collected from cases of unrelated diseases and from normal individuals ascontrols. Several species of animals from Camp Bullis were killed and serumspecimens collected for examination. These included 40 deer, 7 rabbits, 2raccoons, and 1 armadillo. All serum specimens were tested for Weil-Felixagglutination reaction and by the complement fixation test for endemic typhus,American Q fever, and Bullis fever. The antigen for Bullis fever was prepared bytriturating with sand enlarged spleens of mice that had been infectedexperimentally with a strain of the microorganism isolated from a patient.
Of 192 specimens collected from soldiers who had recoveredfrom illness diagnosed clinically as Bullis fever, 76 percent gave positivecomplement fixation for this disease. So also did 4 of the 40 deer specimens and2 of the 7 rabbits. All of 29 specimens of normal human serum, 4 of endemictyphus, 3 of Rocky Mountain spotted fever, 4 of scrub typhus, and 2 of Q feverwere negative in complement fixation tests with the Bullis fever antigen. Thiswork established the specificity of the antigen and further confirmed theoriginal opinion that Bullis fever was a distinct entity. It was the conclusionof the authors:
5Anigstein, L., and Bader, M. N.: Investigations on Rickettsial Diseases in Texas. Part 4. Experimental Study of Bullis Fever. Texas Rep. Biol. & Med. 1: 389-409, 1943.
6Livesay, H. R., and Pollard, M.: Serological Studies of Bullis Fever. Am. J. Trop. Med. 24: 281-284, September 1944.
161
1. The Bullis fever syndrome is notcharacterized by a significant Proteus OX-19, OX-K, or OX-2 agglutinationreaction.
2. There does not appear to be any serologicalrelationship between the rickettsia-like agent of Bullis fever and therickettsia of American Q fever by the complement fixation tests.
3. From the results obtained, there doesappear to be some significant serological relationship between the agentisolated from a human case of Bullis fever and the serums of convalescent cases.
Following the publication of this work, two more articlesappeared, one on the specificity of Bullis fever rickettsia, by Bader andAnigstein,7 and the other by Blair and Bader,on experimental Bullis fever in man.8 Blair andBader summarized their observations, as follows:
Two strains of the infectious agent recoveredfrom human patients and from ticks of the Camp Bullis area were used for humanexperiments. Both strains had been maintained in guinea pig passage for a numberof generations. Fresh or lyophilic material was used for the inoculum. All ofthe patients save one inoculated with these strains showed the syndrome ofBullis fever as described by Woodland and associates; however, most were of amild nature. Patients inoculated with material isolated from ticks (A.americanum) showed the same reaction as those inoculated with the BH [Bullishuman] strain. These results offer further evidence that the rickettsiaeisolated from human patients (Livesay and Pollard, 1943) and from ticks (Anigsteinand Bader, 1943) are specific for the disease diagnosed as Bullis fever.
During the summer of 1945, further experimental studies9were conducted by Maj. M. Pollard, VC, Col. H. R. Livesay, MC, Capt. D. J.Wilson, MC, and Col. J. C. Woodland, MC, on duty at Fort Sam Houston. Thisexperimental work was done with a view to acquiring further information aboutthe nature of the Bullis fever syndrome, its etiology, mode of transmission, andimmunologic relationship to other virus and rickettsial diseases. Human malevolunteers were inoculated with various types of inoculum, as follows:
1. Whole blood from natural cases of the disease.
2. An agent derived from the blood of a natural human case ofBullis fever and propagated on chick embryo.
3. An agent, propagated on chick embryo, that was derivedfrom emulsion of ticks (A. americanum) collected from deer at Camp Bullis.
4. Agents, both of human and of tick origin, to determinetheir immunologic relationship.
5. Laboratory strains of the agent of Bullis fever, tochallenge the immunity of natural cases of the disease.
6. The agent of Colorado tick fever, to study the immunologicrelationship to Bullis fever.
7Bader, M. N., and Anigstein, L.: Specificity of Bullis Fever Rickettsia. Texas Rep. Biol. & Med. 2: 405-412, 1944.
8Blair, R. K., and Bader, M. N.: Observations on Experimental Bullis Fever in Man. Texas Rep. Biol. & Med. 3: 105-111, 1945.
9Pollard, M., Livesay, H. R., Wilson, D. J., and Woodland, J. C.: Experimental Studies With Bullis Fever. Am. J. Trop. Med. 26: 175-187, March 1946.
162
On the basis of the information obtained from theseexperimental studies, it was established that Bullis fever is a transmissibleclinical syndrome which can be reproduced by transferring blood from a patientwith the disease to a normal individual and that it can be maintained in series.It was also established that a syndrome was produced from the agent obtainedfrom ticks (A. americanum) gathered at Camp Bullis that wasindistinguishable from the syndrome in naturally acquired cases of Bullis feverand in the disease experimentally produced by a strain of human origin.
The critical phase of the problem was to determine theresponse of a convalescent case of the disease, acquired naturally, to theBullis fever strain, propagated on chick embryo. Convalescent cases werechallenged with the chick embryo material and were found immune. Controlsdeveloped the disease. It was thus established that a distinct immunologicrelationship exists between the agent isolated from ticks and from human casesand the agent in naturally acquired cases of Bullis fever.
Since no relationship could be detected between Bullis feverand other known rickettsial diseases, attention was directed toward Coloradotick fever, as a disease similarly transmitted by ticks and characterized by alow leukocyte count. Challenge experiments with a known strain of Colorado tickfever agent10 in human volunteers immune toBullis fever (acquired naturally and induced experimentally) failed todemonstrate any relationship between these two diseases.
The results of this work, and the conclusions drawn from it,may be summarized as follows:
1. The Bullis fever syndrome is a distinct clinical entitywhich may be reproduced in human subjects by the inoculation of blood fromfebrile cases of the disease.
2. The Bullis fever agent from the blood of febrile cases hasbeen propagated on the yolk sac of the developing chick embryo. After 20 serialtransfers on yolk sac, the agent so propagated reproduced the Bullis feversyndrome in human subjects.
3. The Bullis fever agent has been isolated from emulsion ofticks (A. americanum) from Camp Bullis and has been propagated on theyolk sac of the developing chick embryo. After propagation for 12 generations,the agent reproduced the Bullis fever syndrome in human subjects.
4. The same immunologic responses are induced by the agentfrom these several sources.
5. There is no immunologic relationship between the agent ofBullis fever and the agent of Colorado tick fever.
During the months of November and December 1945, further ex-
10Obtained through the courtesy and cooperation of Dr. Edward R. Mugrage and Dr. Lloyd Floria, University of Colorado School of Medicine, Denver, Colo.
163
perimental work was accomplished by medical officers11of Fort Sam Houston in an effort to distinguish between the Bullis feversyndrome and dengue fever. On the basis of a challenge experiment in humanvolunteers, no relationship between Bullis fever and dengue fever could bedemonstrated.12 Some clinicalsimilarity has been noted to the anicteric form of leptospirosis; furtherserologic studies should therefore be made to confirm or deny the possibilitythat a leptospiral infection is responsible.
Whatever doubt still remains about the causative agent, theevidence appears definite that the tick is the vector. It is the opinion of allthose familiar with Bullis fever that future studies will show that thissyndrome is not confined to one small geographic area in Texas but that it willbe observed elsewhere in the United States, in regions where the tick A.americanum abounds; namely, the Mississippi Valley and many eastern States.
11Col. H. R. Livesay, MC, Maj. M. Pollard, VC, Maj. Donald J. Wilson, MC, Maj. Choice B. Matthews, MC, and Capt. Sidney M. Cohen, MC.
12Livesay, H. R., Wilson, D. J., Pollard, M., and Woodland, J. C.: Experimental Studies of Bullis Fever and Dengue Fever. Am. J. Trop. Med. 26: 379-381, July 1946.
