HAMMOND, GREGORY W. MD, FRCP(C)
RECENT REVIEWS OF Haemophilus ducreyi1,2 have provided a much needed update on this organism, which has remained relatively obscure. This history focuses on the early studies of Augusto Ducrey, the issues that confused subsequent progress, and the information learned from the Winnipeg outbreak that reawakened interest in this organism.
In 1852, Leon Bassereau gave the first clear description of chancroid (the soft chancre or the ulcus molle)3 as a separate clinical entity from the indurated chancre of syphilis. However, after this there was a great controversy regarding whether single or multiple organisms were involved in pathogenesis, and whether pus of various origins could reproduce the characteristics of the soft chancre.4
Augusto Ducrey (see Figure 1) addressed these questions, working as an assistant in the bacteriology laboratory of Professor Cantani.4,5 Ducrey conducted studies on patients of Dr. De Amicis, from the Clinic for Skin Diseases and Syphilis at the University of Naples. Ducrey's contributions were his systematic studies of the soft chancre and the bubo. He developed innovative methods for autoinoculation of pus, from either the ulcer or the bubo, to a sterilized area of skin maintained free of environmental contamination by a tight‐fitting watch glass bandaged to the forearm (a technique first introduced by Ricord). More than 50 patients were studied by extensive staining methods and microbiologic culturing to various media under aerobic and anaerobic conditions. With specimens obtained from the genital ulcer and repeated passage by forearm inoculation, Ducrey demonstrated a “micro‐organism representing a bacteria of 1.48 microns in length and 0.5 microns in width, short and compact, nicely rounded at the ends, usually showing an indentation at the sides which is seen in microbes that have the form of a figure 8.”4
However, Ducrey was unable to isolate this organism on artificial media.4,5 The organisms did not stain with the Gram stain, but stained well with carbol fuchsin and gentian and methyl violet. He could not elicit lesions in guinea pigs or rabbits injected subcutaneously with material taken from the forearm ulcers. In summary, he obtained a preparation that he considered “pure virus” from the forearm ulcers after autoinoculation. Although it is not certain that the organism he detected is what we now know as H. ducreyi, he concluded that the soft chancre “owed its poison to a live and specific element,” and he realized that this organism “does not grow on usual nutrient plates.” He also realized, “that every micro‐organism recognized which up to now has been labelled as the cause for chancre which is easy to cultivate must be viewed as foreign, for exactly this reason.”
He also investigated the pathogenicity of the bubo. Ducrey found no microorganisms from unruptured bubo pus in more than 50 patients with buboes.4 Ducrey recognized that the organisms cultured from buboes represented primarily cross‐contamination after the lymph node had drained. He believed that the tissue reaction of the bubo was the “result of a special product of the vital fluids of the chancre micro‐organism.”4
Unna6 found “streptobacillus” organisms on biopsies of soft chancres and became convinced these were transitional forms of organisms similar to those found by Ducrey in the free pus. Krefting7 also supported Ducrey's findings, and in addition found similar organisms in bubo pus. Bezancon generally is credited with the first culture isolation on a blood agar plate.8,9 His achievements also were notable in that he reproduced the disease in humans by inoculation of cultured organisms, followed by reculture of the organisms from the lesions, thereby fulfilling Koch's postulates for this organism as a cause of the lesions. This work was confirmed by Tomasczewski.8,10
In the 1920s, Teague and Deibert11 published their experience using rabbit blood and ascitic fluid in which they obtained a positive culture of H. ducreyi in 55% of suspected chancroid cases. They also described the necessary temperature and humidity conditions that favored the growth of H. ducreyi. In 1937, Lwoff and Pirosky12 described the hemin requirements for cultivation of H. ducreyi. Experiments by Sheldon, Heyman, and Beeson compared culture methods with biopsies, smears, skin tests, and autoinoculation.13,14 The nutritional requirements of H. ducreyi were additionally characterized.15,16 Deacon et al17 described the use of human blood to cultivate the organism.
Interest and expertise in H. ducreyi had waned over the years in areas where chancroid became less common because of improved hygiene and antibiotics. There also appears to have been a lingering uncertainty regarding the etiologic role of H. ducreyi as a causative agent of chancroid, at least in the English literature.2,18,19 As an outbreak of suspected chancroid began in Winnipeg, Canada, in the mid‐1970s, we faced a number of issues that had caused confusion in the laboratory diagnosis of chancroid. The situation was confounded by reports of multiple clinical variants of chancroid, up to seven in number.20 Some of these clinical variants likely may have been other sexually transmitted diseases or superinfections.
Problems had been noted with the culture diagnosis of H. ducreyi. The fastidious growth requirements of H. ducreyi were well known. There also was a lack of a defined culture medium. The previously described culture methods often had problems with overgrowth by other bacteria because no selective inhibitory agents were used. Reports of the attempted inhibition of overgrowing microorganisms by the use of aniline dyes to suppress the growth of gram‐positive contaminants had proven unsuccessful because H. ducreyi also was killed.11,21 Even as late as 1970, inactivated patient's blood was recommended as a culture medium, followed by staining to provide a presumptive diagnosis.22 Overall, there had been a variable yield for the detection of H. ducreyi in suspected cases of chancroid. Published reports had shown that as many as 88% were culture positive,1 but most of these studies relied on the detection of the typical morphologic appearance of H. ducreyi and not the specific isolation of the organisms. In summary, there was no simple, reliable, and practical method for the isolation of H. ducreyi.
In addition, there was taxonomic confusion about the identity of suspected isolates of H. ducreyi. The hemolytic activity noted after 3 to 4 days of culture20 had not been confirmed by others. Deacon et al23 reported a gram‐positive phase of growth, which was later retracted.17 Kilian and Theilade24 had confirmed by electron microscopic study that Reymann's strains were gram‐positive organisms.25 Kilian and Theilade found only one of four reference isolates of H. ducreyi to be gram negative. One biochemical characteristic of the organism that helped to differentiate Haemophilus species from H. ducreyi was its need for hemin, or X factor. However, nutritional studies by Beeson et al15 and Ajello et al16 could not confirm the hemin requirement. The taxonomic confusion about H. ducreyi was summarized by Kilian and Theilade in 1975 as follows: “The name H. ducreyi has been used for different groups of poorly growing bacteria which may be isolated from this infectious disease. This finding probably accounts for the conflicting information on the properties of H. ducreyi.”24
Finally, only a few reference strains of H. ducreyi were available. We sought to obtain reference strains from the American Type Culture Collection and the Institute Pasteur, but these strains generally grew poorly, and in our later studies were nonpathogenic in rabbits. They had become avirulent, as had been reported after laboratory adaptation for original isolates in other centers.26
Thus, the Winnipeg chancroid outbreak, which involved 135 suspected cases from July 1975 to September 1977, presented unique opportunities for the study of the microorganism associated with this disease.26 In our initial attempts to grow H. ducreyi, we inoculated clinical samples into tubes of clotted rabbit blood and subcultured to chocolate agar27 with 1% IsoVitaleX (Baltimore Biological Laboratory, Baltimore, MD) without successful culture. A breakthrough occurred when a patient presented with a large fluctuant inguinal bubo before its rupture. The aspiration of material under sterile conditions and subsequent culture of this pus produced colonies of gram‐negative organisms on the chocolate agar plate. These were pure cultures of small yellow‐gray colonies that typically remained cohesive as they were pushed across the surface of the agar plate, consistent with the general descriptions of H. ducreyi.20
On additional study, specimens from four genital lesions of an additional 31 patients yielded these organisms. In rabbit blood medium from the ulcers of six additional patients, we observed typical chaining gramnegative organisms consistent with the staining descriptions of H. ducreyi but which could not be detected after subculture to the chocolate agar surface. The greatest problem was overgrowth by contaminants, and most of these were gram‐positive organisms. Using three of our initial H. ducreyi isolates, we conducted crude disk sensitivity tests on chocolate agar and found these organisms were relatively resistant to vancomycin and to polymyxin. This provided an opportunity to develop a selective culture medium to allow for the fastidious growth of H. ducreyi, while preventing overgrowth by contaminants.27
Our initial study methods compared two methods of specimen collection, that of a swab versus that of an aspirate with a Pasteur pipette. For both specimen collection procedures, the ulcer base was cleaned with a gauze pad moistened in nonbacteriostatic saline. Using the pooled saline aspirate method, the material was inoculated into rabbit blood, or rabbit blood containing vancomycin, or semisolid chocolate agar with vancomycin. Each of these liquid or semisolid media were subsequently subcultured to solid agar plates containing either chocolate agar alone, chocolate agar with vancomycin, or chocolate agar with vancomycin and polymyxin. From the swab, direct inoculation was made to the chocolate agar alone, to chocolate agar containing vancomycin, and to chocolate agar with vancomycin and polymyxin. Subsequently, these plates were incubated at 33°C in an atmosphere consisting of approximately 5% carbon dioxide and saturated water vapor.
Our results showed that the swab method was a satisfactory specimen collection technique. There was a greater detection rate of patients with these organisms from the selective plate containing vancomycin alone, than from the plate with vancomycin and polymyxin, or the chocolate agar alone. In addition, many more colonies of H. ducreyi were detected on the media containing selective antibiotics in comparison to the nonselective media. In summary, a simple specimen collection technique from a cleansed ulcer base was developed using a cotton swab planted directly to chocolate agar enriched with IsoVitaleX and containing 3 μg/ml of vancomycin.27 In patients studied in this prospective manner, we cultured organisms consistent with H. ducreyi from 55% of 16 patients.27 The principle demonstrated here of a selective isolation medium for H. ducreyi has been subsequently improved by others.1 (However, some confusion later arose about the characteristics of H. ducreyi from the use of a modified culture medium developed in Sheffield.28 These results were later characterized as artifact.29)
We next evaluated the antimicrobial susceptibility of H. ducreyi from 19 Winnipeg isolates, plus 4 reference strains from the Institute Pasteur. These 23 strains were tested against 13 antimicrobial agents by an agar dilution technique. In summary, we found a wide range of antimicrobials to which H. ducreyi was sensitive.30 We also found relative resistance to polymyxin, vancomycin, and cloxacillin. Of interest, three strains that had high level resistance to ampicillin and penicillin were found to produce beta‐lactamase by the chromogenic cephalosporin test.30 One other observation was that the in vitro sensitivities of the reference strains usually were greater than that of the wild type isolates. We demonstrated that the local isolates were virulent by intracutaneous inoculation to rabbits (and accidentally to a human finger, personal communication), whereas the reference strains were avirulent. With repeated subculture over 63 passages, the Winnipeg isolates maintained their in vivo virulence in the rabbit model.30
We evaluated the hemin requirements and some biochemical reactions to assist in defining the taxonomy of these organisms.31 When we conducted routine satellite growth tests for hemin (X factor) dependence on agars such as Trypticase soy agar (Baltimore Biological Laboratory) or Eugonagar (Difco Laboratories), we could not demonstrate a growth requirement for hemin. In similar experiments, only 2 of 23 strains showed satellite growth on gonococcal medium base agar (Difco Laboratories). Because IsoVitaleX was an enrichment added to the gonococcal agar to sustain the routine growth of these organisms, we enriched each of the media with individual ingredients found in the IsoVitaleX (but without nicotinamide adenine dinucleotide, which would have potentially confounded the hemin satellite growth test). We were able to demonstrate the hemin requirement of all 23 strains on gonococcal agar base medium, but only 35% and 39% of strains on the Trypticase soy agar and Eugonagar, respectively.31 We also confirmed the requirement for exogenous hemin, by negative porphyrin test results for all of the strains.30
We sought to evaluate the amount of hemin required for growth by a series of diffusion studies from heminimpregnated disks and through the graded concentration of hemin in agar dilution studies. Results showed a higher hemin requirement for H. ducreyi than for other Haemophilus species. H. ducreyi strains required 200 μg to 500 μg of hemin per milliliter to facilitate the growth of our strains to an extent comparable to that on control chocolate agar. Two of the components in IsoVitaleX, namely glucose (1,000 μg/ml) and glutamine (100 μg/ml) were essential ingredients for H. ducreyi satellite growth.
In summary, we demonstrated that a basal medium of gonococcal agar supplemented with glutamine and glucose and a relatively high concentration of hemin were required to support the growth of H. ducreyi.31 Additional biochemical studies, the presence of a gram‐negative cell wall by Gram stain, and electron microscopic studies (unpublished) were consistent with the confirmation of our isolates as H. ducreyi24 and the etiologic agent of our chancroid outbreak.
The experiments conducted as a result of the Winnipeg outbreak provided an opportunity to contribute to a greater understanding of H. ducreyi and chancroid.26
1. Morse SA. Chancroid and Haemophilus ducreyi.
Clin Microbiol Rev 1989; 2:137–157.
2. Albritton WL. Biology of Haemophilus ducreyi.
Microbiol Rev 1989; 53:377–389.
3. Bassereau PI. Traite de affections de la peau symptomatiques de la syphilis. Paris: J.B. Bailliere, 1852.
4. Ducrey A. Experimentelle Untersuchungen uber den ansteckungsstof des weichen Schankers and uber die Bubonen. Monatsh fur Prakt Dermatol 1889;9:387–405.
5. Ducrey A. Recherches expérimentales sur la nature intaine du principe contagieux du chancre mou. Anal de Dermat et de Syph. 1890; 3:56–57.
6. Unna PG. Der Streptobacillus des weichen Schankers. Monatsh fur Prakt Dermatol 1892; 14:485–490.
7. Krefting R. Annal de Dermat. et de Syph. Sur le Microbe du Chancre Mou 1893; 167–170, and 836–839.
8. Peel Ritchie LC. The bacteriology of soft sore and the associated lymphadenitis. Scot Med Surg J 1904; 15:400–405.
9. Bezancon F, Griffon V, Le Sourd L. Culture du Bacille du Chancre Mou. CR Soc Biol 1900; 11:1048–1051.
10. Tomasczewski E. Bakteriologische Untersuchungen uber den Erreger des Ulcus molle. Z f Hyg Infektionskr 1903; 42:327–340.
11. Teague O, Deibert O. Some observations on the bacillus of Unna-Ducreyi. J Med Res 1922; 43:61–75.
12. Lwoff A, Pirosky I. Détermination du Facteur de Croissance pour Haemophilus ducreyi.
CR Soc Biol 1937; 124:1169–1171.
13. Heyman A, Beeson P, Sheldon WH. Diagnosis of chancroid: the relative efficiency of biopsies, cultures, smears, auto-inoculation and skin tests. JAMA 1945; 129:935–938.
14. Sheldon WH, Heyman A. Studies on chancroid: I. observations on the histology with an evaluation of biopsy as a diagnostic procedure. Am J Pathol 1946; 22:415–425.
15. Beeson PB. Studies on chancroid: IV. the Ducrey bacillus: growth requirements and inhibition by antibiotic agents. Proc Soc Exp Biol Med 1946; 61:81–85.
16. Ajello GW, Deacon WE, Paul L, Walls KW. Nutritional studies of a virulent strain of Haemophilus ducreyi.
J Bacteriol 1956; 72:802–808.
17. Deacon WE, Albritton DC, Olansky S, Kaplan W. VDRL chancroid studies: I. a simple procedure for the isolation and identification of Haemophilus ducreyi.
J Invest Dermatol 1956; 26:399–406.
18. Pijper C. The bacillus Unna-Ducrey: a page from the history of bacteriology. Med J S Afr 1920; 16:89–91.
19. Greenblatt RB, Sanderson ES. The intradermal chancroid bacillary antigen test as an aid in the differential diagnosis of the venereal bubo. Am J Surg 1938; 41:384–392.
20. Sullivan M. Chancroid. Am J Syph Gonorrhea Vener Dis 1940; 24:482–521.
21. Beeson PB, Heyman A. Studies on chancroid: II. efficiency of the cultural method of diagnosis. Am J Syph Vener Dis Gonorrhoeae 1945; 29:633–640.
22. Borchardt KA, Hoke AW. Simplified laboratory technique for diagnosis of chancroid. Arch Dermatol 1970; 102:188–192.
23. Deacon WE, Albritton DC, Edmunson WF, Olansky S. Study of Ducrey's bacillus and recognition of a gram positive smooth phase. Proc Soc Exp Biol Med 1954; 86:261–264.
24. Kilian M, Theilade J. Cell wall ultrastructure of strains of Haemophilus ducreyi
and Haemophilus piscium.
Int J Syst Bacteriol 1975; 25:351–356.
25. Reymann F. Type differentiation of H. ducreyi.
Acta Pathologica and Microbiologica Scandinavian. 1950; 27:364–377.
26. Hammond GW, Slutchuk M, Scatliff J, Sherman E, Wilt JC, Ronald AR. Epidemiological, clinical laboratory and therapeutic features of an urban outbreak of chancroid in North America. Rev Inf Dis 1980; 2(6):867–879.
27. Hammond GW, Lian CJ, Wilt JC, Albritton W, Ronald AR. Comparison of specimen collection and laboratory techniques for isolation of Haemophilus ducreyi.
J Clin Microbiol 1978; 7:39–43.
28. Kinghorn GR, Hafiz S, McEntegart MG. Pathogenic microbial flora of genital ulcers in Sheffield with particular reference to herpes simplex virus and Hemophilus ducreyi. Br J VD 1982; 58:377–380.
29. Mindel A. Chancroid. Epidemics in some developing countries, but still rare in Britain. BMJ 1989; 298:64–65.
30. Hammond GW, Lian CJ, Wilt JC, Albritton W, Ronald AR. Antimicrobial susceptibility of Haemophilus ducreyi.
Antimicrob Agents Chemother 1978; 13(4):608–612.
31. Hammond GW, Lian CJ, Wilt JC, Albritton W, Ronald AR. Determination of the Hemin requirement of Haemophilus ducreyi:
evaluation of the porphyrin test and media used in the satellite growth test. J Clin Microbiol 1978; 7:243–246.