Through the cooperation of Dr. Sebastian Lorente, Director of the National Department of Public Health of Peru, nine strains of Bartonella bacilliformis have been isolated, by means of the semisolid leptospira medium, from nine of twelve specimens of blood withdrawn from cases of verruga and forwarded from Peru under conditions of refrigeration. The cultural titer of the blood specimens immediately after their arrival (2 weeks after withdrawal) varied from 1:10 to 1:100,000. Blood from the severe anemic type of the disease, in which there was no eruption, had the highest titer. Blood agar slants yielded irregular results, but some strains grew well on these media.
Morphologically the strains differed very little in fresh preparations examined by dark-ground illumination. In stained preparations some strains appeared coarser, others finer than the average. Special staining indicated that the flagella were characteristically unipolar and varied in number from one to four, some strains showing distinctly more wavy and heavier flagella than others. Young cultures grown on the surface of horse blood agar for 3 to 6 days show individuals with fairly sharp contours, short rods, often varying in thickness toward one or both ends, being intermingled with smaller oval or coccoid elements. Some strains show a predominance of bacillary, some of coccobacillary forms. It is not known whether these features are inherent or are due to conditions of growth, which, though identical, may react differently upon different strains. Definiteness in outline disappears with the age of the culture.
More striking variations are found in the virulence of the different strains for the monkey (Macacus rhesus). Three of the nine strains isolated proved to be non-pathogenic for the monkeys. All three of these were derived from cases of benign verruga. The remaining six strains all gave rise to local lesions when intradermally inoculated and were recovered in culture from the blood of the animals. So far, severe anemia has not developed in any of the monkeys.
It is significant that most of the severe cases yielded virulent strains, while some of the strains from benign verruga were non-pathogenic. It appears highly probable that the severe form of Carrion's disease is, in general, caused by a virulent strain, while the benign forms are due to a strain of low virulence. On the other hand, a virulent strain may cause benign verruga in unusually resistant persons and a weak strain may give rise to severe blood infection in unduly susceptible individuals. The form of Carrion's disease is probably determined primarily by the inherent virulence of the strain of Bartonella bacilliformis and is modified secondarily by individual predisposition in a given case.
An interesting phenomenon brought out by the present investigation was the failure of the nine human blood specimens to induce local verruga in the same monkeys in which the corresponding cultures, inoculated simultaneously at separate sites, gave rise to typical lesions. Yet the original blood samples were shown by cultivation to have contained live bartonellas at the time they were inoculated, and blood culture revealed the presence of the microorganisms in the blood of monkeys which showed no other signs of infection after inoculation with the human blood. Whether this striking difference is merely a quantitative one or is due to some factor still unknown—such as, for example, a biological phase of the microorganism—has not been determined. The uniformly negative results of transmission experiments with blood by previous investigators is explained by an actual inability of the blood to induce skin lesions and the lack, until now, of a reliable method of detecting Bartonella bacilliformis in the monkeys' blood.
The strains isolated showed similar serologic properties, as tested by complement fixation.