930
i.e. 20 of 2,438 Enterobacteriaceae were resistant to the combination, only 8 of the 20 resistant isolates were classified as resistant by the disk test; three resistant strains were classified as susceptible and nine strains were classified as intermediate. It m a y be n o t e w o r t h y that l a b o r a t o r y A which had the lowest rate of c e f o p e r a z o n e resistance, classified five of the six problematic strains shown in Table 2 as susceptible, while laboratory C which had the highest rate of c e f o p e r a z o n e resistance classified these six strains as resistant. T h e s e observations and the variability in M I C s of c e f o p e r a z o n e for Escherichia coli A T C C 35218 suggest that technical p r o b l e m s in detecting c e f o p e r a z o n e resistance had a significant impact on the reported prevalence of c e f o p e r a z o n e resistance in different institutions and that influenced the very m a j o r error rates for individual laboratories. T h e reason for the variability of M I C test results for certain strains and the failure of the disk test to accurately classify resistant strains is unknown and should be further investigated. Although the extent to which very m a j o r errors contribute to clinical failures with c e f o p e r a z o n e is unknown, changes in breakpoints could minimize the potential impact of these interpretive errors. If the susceptibility b r e a k p o i n t for c e f o p e r a z o n e disk tests was changed from >_21 m m to >_22 mm, 2 23 ram, _>24 m m or _>25 mm, very m a j o r errors expressed as a p e r c e n t a g e of the total population tested would be r e d u c e d f r o m 1,8 % to 1.3 %, 0.8 %, 0.5 % and 0.4 %, respectively. While minimizing very m a j o r error rates, these changes in susceptibility breakpoints would increase minor errors f r o m 4,3 % to 5.2 %, 7.8 %, 9.3 % and 11.3 %, respectively. T h e merit of such changes in disk test criteria is worth further consideration but no such changes are r e c o m m e n d e d at this time.
References 1. ThornsberryC, BarryAL, Jones RN, Baker CN, Badal RE: Tentative interpretive standards for agar disk diffusion antimicrobial susceptibility testing of cefoperazone. Journal of Clinical Microbiology 1982, 15: 769-776. 2. Jones RN, Gaven TL, Barry AL, Thornsberry C, Gibbs DL: Cefoperazone disk diffusion susceptibility test: confirmation of the tentative interpretive criteria, cross-resistance, and determination af quality control performance limits. Journal of Clinical Microbiology 1982, 15: 777-786.
Eur. J. Clin. Microbiol. Infect. Dis.
3. Jones RN, Barry AL, Thornsberry C, Wilson HW: The cefoperazone sulbactam combination: in vitro qualities including beta-lactamase stability, antimicrobial activity, and interpretive criteria for disk diffusion tests. American Journal of Clinical Pathology 1985, 84: 496-504. 4. Jones RN, Barry AL, Packer RR, Gregory WW, Thornsberry C: In vitro antimicrobial spectrum, occurrence of synergy, and recommendations for dilution susceptibility testing concentrations of the cefoperazone-sulbactam combination. Journal of Clinical Microbiology 1987, 25: 1725-1729. 5. BarryAL, Jones RL, The CollaborativeAntimicrobial Susceptibility Testing Group: Criteria for disk susceptibility tests and quality control guidelines for the cefoperazone-sulbactam combination. Journal of Clinical Microbiology 1988, 26: 13-17. 6. National Committee for Clinical Laboratory Standards: Performance standards for antimicrobial disk susceptibility tests. Approved standard M2-A2. NCCLS, Villanova, PA, 1990. 7. National Committee for Clinical Laboratory Standards:Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved Standard M7-A2. NCCLS, Villanova, PA, 1990. 8. Barry AL, Fuchs PC, Gerlach EH, Hardy DJ, MeLaughlin JC, Pfaller MA: Tiearcillin and ticarcillinclavulanic acid susceptibility tests: error rates for disk tests with consecutively isolated membe1~ of the family Enterobacteriaceae. Antimicrobial Agents and Chemotherapy 1992, 36: 137-143.
A Simple Medium for the Primary Isolation of Haemophilus ducreyi Y. D a n g o r * , S.D. Miller, H . J . K o o r n h o f , R.C. Ballard
Two simple, inexpensive media containing gonococcal agar-base, supplemented with 5 % Fildes' extract and either chocolated or unchocolated horse blood (GC-FHBC or GC-FHB) were compared with the standard gonococcal agarbased (GC-HgS) and Mueller-Hinton agar-based media (MH-HB) normally used for primary isolation of Haemophilus ducreyi from presumptive chancroid lesions. Overall, Haemophilus ducreyi was recovered from 162 of 178 (91%) samples from primary chancroid lesions. As a single isolation medium GC-HgS proved the most sensitive Emergent Pathogen Research Unit, School of Pathology, University of the Witwatersrand and the South African Institute for Medical Research, PO Box 1038, Johannesburg 2000, Republic of South Africa.
Vol. 11, 1992
with an isolation rate of 80 % followed by GCFHB (75 %), MH-HB and GC.FHC (both 71%). Use of a combination of GC-HgS and MH-HB resulted in isolation of Haemophilus ducreyi in 160 of 178 cases (90 %). Since GC-FHB is approximately one.quarter the cost of the combination and half the cost of GC-HgS or MH-HB alone, this medium could prove suitable for diagnostic purposes in developing countries where chancroid is endemic.
Until relatively recently, establishment of a definitive diagnosis of chancroid was based on the results of microscopy of Gram stained smears of exudates from genital ulcerations or growth of gram-negative organisms in clotted human or rabbit blood (1). In 1978 Hammond et al. (2) developed a solid selective agar-based medium consisting of gonococcal agar base with 1 % bovine haemoglobin, 1 % IsoVitaleX and 3 mg/l vancomycin. Since then a variety of culture media have been proposed for the primary isolation of Haemophilus ducreyi which have proved to be more sensitive than the original medium (3-6). Unfortunately, some of the constituents of the most frequently used media are prohibitively expensive for routine use in developing countries where chancroid is endemic. We have therefore endeavoured to replace the most expensive supplements (i.e. foetal calf serum and IsoVitaleX) by using 5 % Fildes' extract which has previously been included as a supplement in media used to isolate Haemophilus influenzae from clinical specimens. The Fildes' extract used in this study was obtained from a commercial source. However, the extract can be prepared locally following peptic digestion of defibrinated sheep blood (7). In this study we compared gonococcal agar-based media containing Fildes' extract and horse blood that was either chocolated (heated to 75 °C) or unchocolated (unheated) with the standard gonococcal agar-based and Mueller-Hinton agarbased media, which, when used in combination, have been recognised as the "gold standard" for the isolation of Haemophilus ducreyi from primary chancroid lesions (8).
Materials and Methods. Four media were used for primary isolation of Haemophilus ducreyi from primary lesions during the course of this study. The first medium comprised gonococcal agar base (Gibco Laboratories, USA), 1 % bovine haemoglobin (Gibco), 1% IsoVitaleX (BBL Division, Becton Dickinson, USA) and
931
5 % foetal calf serum (GC-HgS). The second contained Mueller-Hinton agar base (BBL), 5 % chocolated horse blood and 1 % IsoVitaleX (BBL) (MH-HB). The two new simplified media which were evaluated consisted of gonococcal agar base, 5 % Fildes' extract (Oxoid, UK) and 5 % horse blood (GC-FHBC). Vancomycin (3 mg/1) (Eli Lilly, USA) was included in each of the four media. All media were prepared in one litre quantities and used within two weeks of preparation. One hundred and seventy-eight mine workers presenting at the sexually transmitted diseases clinic at the Leslie Williams' Memorial Hospital, Carletonville, South Africa, with a clinical diagnosis of chancroid were included in the study. In 60 % of cases the disease was acquired locally in the Johannesburg-Carletonville area, and in the remaining 40 % was acquired either in Lesotho, Botswana, Kwa-Zulu, Swaziland or Transkei. Material from the bases of ulcers was collected using calcium alginate swabs (Calgiswab, code 60150-15, Inolex, USA) which were used to inoculate the four media in a random sequence according to a pre-established code. Inoculated plates were stored in a candle-extinction jar at room temperature for up to 6 h, after which they were transferred to an anaerobic jar from which the catalyst had been removed. Plates were incubated in this microaerophilic atmosphere at 33-35 °C for 48-72 h. Typical colonies of Haemophilus ducreyi which could be pushed intact across the agar surface, were identified by Gram staining and bio-chemical characteristics (catalase, oxidase, nitrate reduction, alkaline phosphatase and porphyrin), according to the criteria described by Kilian (9). Growth was quantitated as either 1+ (< 10 colonies); 2+ (10-100 colonies); 3+ (> 100 colonies); or 4+ (confluent growth). Statistical methods employed during this study included Student's t-test, Fisher's exact test and the chi-squared test with Yate's correction.
Results and Discussion. Overall, ttaemophilus ducreyi was recovered from 162 of 178 patients (91%) on at least one medium when all four media were used for isolation. One hundred and forty-three strains were isolated on GC-HgS agar (80 %), 126 on MH-HB agar (71%), and 133 (75 %) and 126 (71%) were recovered on the Fildes' extract-containing media with blood (GCFHB) and chocolated blood (GC-FHBC) respectively (Table 1). Significantly higher rates of recovery of Itaemophilus ducreyi were obtained using GC-HgS in combination with MH-HB than
932
Eur. J. Clin. Microbioi. Infect. Dis.
Table 1: Comparison of four mediaa for the primary isolation a clinical diagnosis of chancroid.
Total number of isolates recovered Isolates recovered only on this medium
Haemophilusducreyifrom 178 men with
GC-HgS
MH-HB
GC-FHB
GC-FHBC
All four media
143 (80 %)
126 (71%)
133 (75 %)
126 (71%)
162b (91%)
4
7
1
0
a GC-HgS: GC agar base, 1% haemoglobin, 5 % foetal calf serum, 1% IsoVitaleX; MH-HB: MuellerHinton agar base, 5 % horse blood (chocolated), 1% IsoVitaleX; GC-FHB: GC agar base, 5 % Fildes' extract, 5 % horse blood; GC-FHBC: GC agar base, 5 % Fildes extract, 5 % horse blood (ehocolated). bThe total isolated using all four media was significantly greater than the number recovered on GC-HgS alone (p < 0.001).
Table 2: Primary isolation on various media and media combinations of Haemophilusducreyifrom 178 men with clinical chancroid. No. ( % ) of isolates recovered GC-HgS
GC-HgS
GC-HgS
MH-HB
MH-HB
+
+
+
+
+
+
MH-HB
GC-FHB
GC-FHBC
GC-FHBC
GC-FHB
GC-FHBC
GC-HgS alone
34 (19)
20 (11)
24 (13)
-
-
-
MH-HB alone
17 (10)
-
-
30 (17)
20 (11)
-
-
-
27 (15)
18 (10)
GC-FHB alone
-
GC-FHBC alone
-
-
Both media
109 (61)
Either medium No growth
10 (6)
GC-FHB
7 (4)
28(16)
-
11 (6)
123 (69)
119 (67)
97 (54)
106 (60)
115 (65)
160 (90)
153 (86)
150 (84)
155 (87)
153 (86)
144 (81)
18 (10)
25 (14)
28 (16)
23 (13)
25 (14)
34 (19)
using either of the two Fildes' extract-containing m e d i a alone (p < 0.0001, Fisher's exact test). However, when G C - F H B and M H - H B media were used in combination, no significant difference in isolation rates c o m p a r e d with the "gold s t a n d a r d " combination ( G C - H g S + M H - H B ) was detected (Table 2). A loss o f only 5 % sensitivity was noted when G C - F H B was c o m p a r e d with the best single m e d i u m when used alone (GC-HgS). T h e recovery rate of Haemophilus ducreyi was also found to be significantly higher on m e d i u m containing Fildes' extract with unchocolated blood c o m p a r e d with chocolated blood (X 2 = 59.58, p = < 0.0001). T h e quantitative growth of 87 strains of Haemophilus ducreyi which were isolated on all m e d i a is
shown in detail in Table 3. G C - H g S yielded a significantly greater growth than M H - H B and the other two simpler m e d i a (p < 0.05; Student's ttest). In addition, individual colonies were generally larger on the G C - H g S than on M H - H B and the two m e d i a containing Fildes' extract. Furthermore, bacterial colony size was also larger on G C - F H B than on G C - F H B C . The results of the present study clearly indicate that use of a combination of two solid selective media ( G C - H g S and M H - H B ) is the optimal technique for the culture of Haemophilus ducreyi from clinical specimens. T h e recovery of this bacterium from 90 % of p r e s u m p t i v e chancroid lesions using this c o m b i n a t i o n is c o m p a r a b l e to results of previous studies (5, 6). Unfortunately,
Vol. 11, 1992
933 ......
Table3: Quantitativecomparisonof growthof 87 strains of Haemophilusducreyion primaryisolationon four different media. Medium
GC-HgS MH-HB GC-FHB GC-FHBC
Quantitive score 4+
3+
2+
1+
Mean
44 12 23 11
26 45 35 43
12 15 22 21
5 15 7 12
3.5~ 2.6 2.8 2.6
aThe difference in quantitative growth detected between GC-HgS and GC-FHB, GC-FHBCand MH-HB is statistically highlysignificant(p < 0.0001 Student's t-test).
the high cost of using two such selective media and the limited availability of certain constituents has discouraged the use of these definitive culture methods, especially in areaswhere chancroid is endemic. This study has shown that costs could be reduced by 25 % by substituting GC-FHB medium for GC-HgS medium without significantly reducing sensitivity. The data presented here and in other studies (8, 10) clearly indicate that no single medium provides optimal conditions for primary isolation of Haemophilus ducreyi. It is therefore not surprising that in this study the use of media containing Fildes' extract alone resulted in a significantly lower rate of recovery (75 %) compared with the expensive combination of two media. However, a loss of only 5 % was recorded when the rate of recovery on GC-FHB was compared with the best single medium when used alone namely GC-HgS (80%). The small sacrifice in sensitivity which occurs using Fildes' extract-containing medium is more than compensated by the reduction in expense (approximately 50 % of the cost of GC-HgS and 25 % of the combination) and the ability to provide a definitive diagnosis. In Nairobi, Kenya, 20 % of Haemophilus ducreyi strains were found to grow only on gonococcal agar-based medium and a further 10 % only on Mueller-Hinton agar based medium (8). These findings are comparable to those of the present Study ('['able 3). Sottnek et al. (3) concluded that the differences in the ability of MH-HB and GCHgS to support the growth of Haemophilus ducreyi were related to the agar bases used. They showed that neither the addition of foetal calf serum to MH-HB base with chocolated horse blood, nor the addition of chocolated blood to GC-HgS improved the frequency of isolation of Haemophilus ducreyi. In this study we have only
meagre evidence to support these findings since seven strains grew on MH-HB alone, but a further six strains grew on MI,H-HB and on Fildes' extract containing-media but not on GC-HgS medium, which contains the same agar base. However, it is interesting to note that GC-FHB may be successfully substituted for GC-HgS when using a combination of media without significantly compromising sensitivity. It would appear that differences in sensitivity of the various isolation media reflect specific growth requirements of individual strains of Haernophilus ducreyi. However, in this study no attempt was made to subculture isolates on media which did not initially support growth of these strains. The finding that Haernophilus ducreyi was not isolated from 9 % of our patients suggests that Haemophilus ducreyi strains may have been eradicated in some of these patients either spontaneously or by previous treatment. However, some of these patients may have been infected with Haemophilus ducreyi having unique nutritional requirements. As a result of this possibility and the fact that none of the media evaluated in this study supported the growth of all Haemophilusducreyistrains, further attempts to develop even better media for the recovery of this organism from clinical specimens should be made. Ongoing studies in Nairobi (Ndinya-Achola, personal communication) indicate that incorporation of activated charcoal may enhance recovery of Itaemophilus ducreyi in preexisting primary isolation media. In the meantime, we agree with Nsanze et al. (8) that both GC-HgS and MH-HB media, should be used in studies which require optimal isolation rates to be achieved. However, GC-FHB medium has proved to be a satisfactory alternative to GC-HgS when combination media are required, or in situations where a single medium could be used where cost considerations are paramount, notably in developing countries. In such situations GCFHB medium could be used for routine diagnostic purposes and for surveillance of antimicrobial susceptibility of Haemophilus ducreyi isolates in order to provide a rational basis for therapy. References 1. Deacon WE, Albritton DC, Oiansky S, Kaplan W:
V,D.R.L. chancroid studies. I: A simple procedure for the isolation and identification of Haemophilus ducreyi.Journal of InvestigationalDermatology1956, 26: 399--406. 2. Hammond GW, Lian CJ, Wilt JC, Ronalfl. AR: Comparison of specimen collection and laboratory techniques for the isolationof Haemophilusducreyi.Journal of Clinical Microbiology 1978, 7: 39--43.
934
3. Sottnek FO, Biddle JW, Kraus SJ, Weaver RE, Stewart JA: Isolation and identification of Haemophilus ducreyi in a clinical study. Journal of Clinical Microbiology 1980, 12: 170-174. 4. Bilgeri YR, Ballard RC, Duncan MO, Mauff AC, Koornhof H J: Antimicrobial susceptibility of 103 strains of lJaemophilus ducreyi isolated in Johannesburg. Antimicrobial Agents and Chemotherapy 1982, 22: 686-688. 5. Dylewski J, Nsanze H, Maitha G, Ronaid A: LaboratolT diagnosis of Haemophilus ducreyi: sensitivity of culture media. Diagnostic Microbiology and Infectious Diseases 1986, 4: 241-245. 6. Kunimito DY, Slaney L, Koss J, D'Costa LJ, Plummer FA, Ndinya-Achola JO, Ronald AR: Field testing of modified Bieling media for the isolation of Haemophilus ducreyi in Kenya. European Journal of Clinical Microbiology 1986, 5: 673-675. 7. Fildes P: A new medium for the growth of B. influenzae. British Journal of Experimental .Pathology 1920, 1: 129-130. 8. Nsanze H, Plummer FA, Maggwa ABN, Martha G, Dylewski J, Plot P, Ronald AR: Comparison of media for the primary isolation of Haemophilus ducreyi. Sexually Transmitted Diseases 1984, 11: 6-9. 9. Kilian M: A taxonomic study of the genus Haemophilus with a proposal of new species. Journal of General Microbiology 1976, 93: 9--62. 10. MacDonald K, Cameron W, Irungu G, D'Costa LJ, Plammer FA,' Slanley L, Ndinya-Achola JO, Ronald AR: Comparison of Sheffield media with standard media for the isolation of llaemophitus ducreyi. Sexually Transmitted Diseases 1989, 16: 88--90.
Impaired Detection of Faecal Verocytotoxin in the Presence of Clostridium difficile Cytotoxin in Patients with Haemolytic Uraemic Syndrome I. L u z z i 1., E Minelti 2, A . G i a n v i t i 3, A. C a p r i o l i 2
Three cases of haemolytic uraemic syndrome associated with infection with verocytotoxin producing Escherichia coil are described. The concomitant presence of Clostridium difficile cytotoxin in the patients' stool impaired the l Laboratorio di Batteriologia c Micologia Medica and 2Laboratorio di Ultrastrutture, lstituto Superiore di Sanit'~, Viale Regina Elena 299, 00161 Rome, Italy. 3Divisione di Nefi'ologia e Dialisi, Ospedale Pediatrico Bambino Ges~, Istituto di Ricerca Scientifica, Salita S. Onofrio 4, 00100 Rome, Italy.
Eur. J. Clin. Microbiol. Infect. Dis.
detection of free faecal verocytotoxin. Stool specimens containing Clostridium difficile cytotoxin should thus be considerednegative for verocytotox~n only after neutralisation of the Clostridium difficile cytotoxin with antitoxin.
Infection with verocytotoxin producing Escherichia coil ( V T E C ) is associated with a spectrum of disorders including diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome (1). The haemolytic uraemic syndrome constitutes a major cause of acute renal failure in childhood, and is also characterized by thrombocytopoenia and microangiopathic haemolytic anaemia (1). Following the first report by Karmall et al. (2), the association between haemolytic uraemic syndrome and V T E C infection has been confirmed by several other investigators using various microbiological and serological techniques (3-6). Microbiological diagnosis of V T E C infection is based on the isolation of V T E C from faecal cultures and/or the demonstration of free verocytotoxin in faecal filtrates (1). A major problem encountered in the isolation of V T E C from stool cultures is the relatively low n u m b e r of these organisms in mixed flora, and their rapid disappearance from stools within a few days after the onset of symptoms (1, 5). This occurs particularly in patients with haemolytic uraemic syndrome, whose stools are often received one to two weeks after the onset of the diarrhoeal prodromat illness. The low sensitivity of the isolation procedures may be overcome by examining faecal filtrates for free verocytotoxin. However, although faecal verocytotoxin has been consistently found by different investigators in the stools of patients positive for V T E C on culture, it has also been detected in specimens from many haemolytic uraemic syndrome patients who were culture negative for V T E C (1, 2, 6). We report three cases of VTEC-associated haemolytic uraemic syndrome in which detection of faecal verocytotoxin was complicated by a concomitant infection with Clostridium difficile. Clostridium difficile has frequently been found in the stool of children up to two years of age (7-9), and is known to produce two potent toxins cytopathic for cultured cells (10).
Patients a n d Methods. The three cases were observed during a study on the association between haemolytic uraemic syndrome and V T E C infections in Italian children (11). T h e onset of
i.e. 20 of 2,438 Enterobacteriaceae were resistant to the combination, only 8 of the 20 resistant isolates were classified as resistant by the disk test; three resistant strains were classified as susceptible and nine strains were classified as intermediate. It m a y be n o t e w o r t h y that l a b o r a t o r y A which had the lowest rate of c e f o p e r a z o n e resistance, classified five of the six problematic strains shown in Table 2 as susceptible, while laboratory C which had the highest rate of c e f o p e r a z o n e resistance classified these six strains as resistant. T h e s e observations and the variability in M I C s of c e f o p e r a z o n e for Escherichia coli A T C C 35218 suggest that technical p r o b l e m s in detecting c e f o p e r a z o n e resistance had a significant impact on the reported prevalence of c e f o p e r a z o n e resistance in different institutions and that influenced the very m a j o r error rates for individual laboratories. T h e reason for the variability of M I C test results for certain strains and the failure of the disk test to accurately classify resistant strains is unknown and should be further investigated. Although the extent to which very m a j o r errors contribute to clinical failures with c e f o p e r a z o n e is unknown, changes in breakpoints could minimize the potential impact of these interpretive errors. If the susceptibility b r e a k p o i n t for c e f o p e r a z o n e disk tests was changed from >_21 m m to >_22 mm, 2 23 ram, _>24 m m or _>25 mm, very m a j o r errors expressed as a p e r c e n t a g e of the total population tested would be r e d u c e d f r o m 1,8 % to 1.3 %, 0.8 %, 0.5 % and 0.4 %, respectively. While minimizing very m a j o r error rates, these changes in susceptibility breakpoints would increase minor errors f r o m 4,3 % to 5.2 %, 7.8 %, 9.3 % and 11.3 %, respectively. T h e merit of such changes in disk test criteria is worth further consideration but no such changes are r e c o m m e n d e d at this time.
References 1. ThornsberryC, BarryAL, Jones RN, Baker CN, Badal RE: Tentative interpretive standards for agar disk diffusion antimicrobial susceptibility testing of cefoperazone. Journal of Clinical Microbiology 1982, 15: 769-776. 2. Jones RN, Gaven TL, Barry AL, Thornsberry C, Gibbs DL: Cefoperazone disk diffusion susceptibility test: confirmation of the tentative interpretive criteria, cross-resistance, and determination af quality control performance limits. Journal of Clinical Microbiology 1982, 15: 777-786.
Eur. J. Clin. Microbiol. Infect. Dis.
3. Jones RN, Barry AL, Thornsberry C, Wilson HW: The cefoperazone sulbactam combination: in vitro qualities including beta-lactamase stability, antimicrobial activity, and interpretive criteria for disk diffusion tests. American Journal of Clinical Pathology 1985, 84: 496-504. 4. Jones RN, Barry AL, Packer RR, Gregory WW, Thornsberry C: In vitro antimicrobial spectrum, occurrence of synergy, and recommendations for dilution susceptibility testing concentrations of the cefoperazone-sulbactam combination. Journal of Clinical Microbiology 1987, 25: 1725-1729. 5. BarryAL, Jones RL, The CollaborativeAntimicrobial Susceptibility Testing Group: Criteria for disk susceptibility tests and quality control guidelines for the cefoperazone-sulbactam combination. Journal of Clinical Microbiology 1988, 26: 13-17. 6. National Committee for Clinical Laboratory Standards: Performance standards for antimicrobial disk susceptibility tests. Approved standard M2-A2. NCCLS, Villanova, PA, 1990. 7. National Committee for Clinical Laboratory Standards:Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved Standard M7-A2. NCCLS, Villanova, PA, 1990. 8. Barry AL, Fuchs PC, Gerlach EH, Hardy DJ, MeLaughlin JC, Pfaller MA: Tiearcillin and ticarcillinclavulanic acid susceptibility tests: error rates for disk tests with consecutively isolated membe1~ of the family Enterobacteriaceae. Antimicrobial Agents and Chemotherapy 1992, 36: 137-143.
A Simple Medium for the Primary Isolation of Haemophilus ducreyi Y. D a n g o r * , S.D. Miller, H . J . K o o r n h o f , R.C. Ballard
Two simple, inexpensive media containing gonococcal agar-base, supplemented with 5 % Fildes' extract and either chocolated or unchocolated horse blood (GC-FHBC or GC-FHB) were compared with the standard gonococcal agarbased (GC-HgS) and Mueller-Hinton agar-based media (MH-HB) normally used for primary isolation of Haemophilus ducreyi from presumptive chancroid lesions. Overall, Haemophilus ducreyi was recovered from 162 of 178 (91%) samples from primary chancroid lesions. As a single isolation medium GC-HgS proved the most sensitive Emergent Pathogen Research Unit, School of Pathology, University of the Witwatersrand and the South African Institute for Medical Research, PO Box 1038, Johannesburg 2000, Republic of South Africa.
Vol. 11, 1992
with an isolation rate of 80 % followed by GCFHB (75 %), MH-HB and GC.FHC (both 71%). Use of a combination of GC-HgS and MH-HB resulted in isolation of Haemophilus ducreyi in 160 of 178 cases (90 %). Since GC-FHB is approximately one.quarter the cost of the combination and half the cost of GC-HgS or MH-HB alone, this medium could prove suitable for diagnostic purposes in developing countries where chancroid is endemic.
Until relatively recently, establishment of a definitive diagnosis of chancroid was based on the results of microscopy of Gram stained smears of exudates from genital ulcerations or growth of gram-negative organisms in clotted human or rabbit blood (1). In 1978 Hammond et al. (2) developed a solid selective agar-based medium consisting of gonococcal agar base with 1 % bovine haemoglobin, 1 % IsoVitaleX and 3 mg/l vancomycin. Since then a variety of culture media have been proposed for the primary isolation of Haemophilus ducreyi which have proved to be more sensitive than the original medium (3-6). Unfortunately, some of the constituents of the most frequently used media are prohibitively expensive for routine use in developing countries where chancroid is endemic. We have therefore endeavoured to replace the most expensive supplements (i.e. foetal calf serum and IsoVitaleX) by using 5 % Fildes' extract which has previously been included as a supplement in media used to isolate Haemophilus influenzae from clinical specimens. The Fildes' extract used in this study was obtained from a commercial source. However, the extract can be prepared locally following peptic digestion of defibrinated sheep blood (7). In this study we compared gonococcal agar-based media containing Fildes' extract and horse blood that was either chocolated (heated to 75 °C) or unchocolated (unheated) with the standard gonococcal agar-based and Mueller-Hinton agarbased media, which, when used in combination, have been recognised as the "gold standard" for the isolation of Haemophilus ducreyi from primary chancroid lesions (8).
Materials and Methods. Four media were used for primary isolation of Haemophilus ducreyi from primary lesions during the course of this study. The first medium comprised gonococcal agar base (Gibco Laboratories, USA), 1 % bovine haemoglobin (Gibco), 1% IsoVitaleX (BBL Division, Becton Dickinson, USA) and
931
5 % foetal calf serum (GC-HgS). The second contained Mueller-Hinton agar base (BBL), 5 % chocolated horse blood and 1 % IsoVitaleX (BBL) (MH-HB). The two new simplified media which were evaluated consisted of gonococcal agar base, 5 % Fildes' extract (Oxoid, UK) and 5 % horse blood (GC-FHBC). Vancomycin (3 mg/1) (Eli Lilly, USA) was included in each of the four media. All media were prepared in one litre quantities and used within two weeks of preparation. One hundred and seventy-eight mine workers presenting at the sexually transmitted diseases clinic at the Leslie Williams' Memorial Hospital, Carletonville, South Africa, with a clinical diagnosis of chancroid were included in the study. In 60 % of cases the disease was acquired locally in the Johannesburg-Carletonville area, and in the remaining 40 % was acquired either in Lesotho, Botswana, Kwa-Zulu, Swaziland or Transkei. Material from the bases of ulcers was collected using calcium alginate swabs (Calgiswab, code 60150-15, Inolex, USA) which were used to inoculate the four media in a random sequence according to a pre-established code. Inoculated plates were stored in a candle-extinction jar at room temperature for up to 6 h, after which they were transferred to an anaerobic jar from which the catalyst had been removed. Plates were incubated in this microaerophilic atmosphere at 33-35 °C for 48-72 h. Typical colonies of Haemophilus ducreyi which could be pushed intact across the agar surface, were identified by Gram staining and bio-chemical characteristics (catalase, oxidase, nitrate reduction, alkaline phosphatase and porphyrin), according to the criteria described by Kilian (9). Growth was quantitated as either 1+ (< 10 colonies); 2+ (10-100 colonies); 3+ (> 100 colonies); or 4+ (confluent growth). Statistical methods employed during this study included Student's t-test, Fisher's exact test and the chi-squared test with Yate's correction.
Results and Discussion. Overall, ttaemophilus ducreyi was recovered from 162 of 178 patients (91%) on at least one medium when all four media were used for isolation. One hundred and forty-three strains were isolated on GC-HgS agar (80 %), 126 on MH-HB agar (71%), and 133 (75 %) and 126 (71%) were recovered on the Fildes' extract-containing media with blood (GCFHB) and chocolated blood (GC-FHBC) respectively (Table 1). Significantly higher rates of recovery of Itaemophilus ducreyi were obtained using GC-HgS in combination with MH-HB than
932
Eur. J. Clin. Microbioi. Infect. Dis.
Table 1: Comparison of four mediaa for the primary isolation a clinical diagnosis of chancroid.
Total number of isolates recovered Isolates recovered only on this medium
Haemophilusducreyifrom 178 men with
GC-HgS
MH-HB
GC-FHB
GC-FHBC
All four media
143 (80 %)
126 (71%)
133 (75 %)
126 (71%)
162b (91%)
4
7
1
0
a GC-HgS: GC agar base, 1% haemoglobin, 5 % foetal calf serum, 1% IsoVitaleX; MH-HB: MuellerHinton agar base, 5 % horse blood (chocolated), 1% IsoVitaleX; GC-FHB: GC agar base, 5 % Fildes' extract, 5 % horse blood; GC-FHBC: GC agar base, 5 % Fildes extract, 5 % horse blood (ehocolated). bThe total isolated using all four media was significantly greater than the number recovered on GC-HgS alone (p < 0.001).
Table 2: Primary isolation on various media and media combinations of Haemophilusducreyifrom 178 men with clinical chancroid. No. ( % ) of isolates recovered GC-HgS
GC-HgS
GC-HgS
MH-HB
MH-HB
+
+
+
+
+
+
MH-HB
GC-FHB
GC-FHBC
GC-FHBC
GC-FHB
GC-FHBC
GC-HgS alone
34 (19)
20 (11)
24 (13)
-
-
-
MH-HB alone
17 (10)
-
-
30 (17)
20 (11)
-
-
-
27 (15)
18 (10)
GC-FHB alone
-
GC-FHBC alone
-
-
Both media
109 (61)
Either medium No growth
10 (6)
GC-FHB
7 (4)
28(16)
-
11 (6)
123 (69)
119 (67)
97 (54)
106 (60)
115 (65)
160 (90)
153 (86)
150 (84)
155 (87)
153 (86)
144 (81)
18 (10)
25 (14)
28 (16)
23 (13)
25 (14)
34 (19)
using either of the two Fildes' extract-containing m e d i a alone (p < 0.0001, Fisher's exact test). However, when G C - F H B and M H - H B media were used in combination, no significant difference in isolation rates c o m p a r e d with the "gold s t a n d a r d " combination ( G C - H g S + M H - H B ) was detected (Table 2). A loss o f only 5 % sensitivity was noted when G C - F H B was c o m p a r e d with the best single m e d i u m when used alone (GC-HgS). T h e recovery rate of Haemophilus ducreyi was also found to be significantly higher on m e d i u m containing Fildes' extract with unchocolated blood c o m p a r e d with chocolated blood (X 2 = 59.58, p = < 0.0001). T h e quantitative growth of 87 strains of Haemophilus ducreyi which were isolated on all m e d i a is
shown in detail in Table 3. G C - H g S yielded a significantly greater growth than M H - H B and the other two simpler m e d i a (p < 0.05; Student's ttest). In addition, individual colonies were generally larger on the G C - H g S than on M H - H B and the two m e d i a containing Fildes' extract. Furthermore, bacterial colony size was also larger on G C - F H B than on G C - F H B C . The results of the present study clearly indicate that use of a combination of two solid selective media ( G C - H g S and M H - H B ) is the optimal technique for the culture of Haemophilus ducreyi from clinical specimens. T h e recovery of this bacterium from 90 % of p r e s u m p t i v e chancroid lesions using this c o m b i n a t i o n is c o m p a r a b l e to results of previous studies (5, 6). Unfortunately,
Vol. 11, 1992
933 ......
Table3: Quantitativecomparisonof growthof 87 strains of Haemophilusducreyion primaryisolationon four different media. Medium
GC-HgS MH-HB GC-FHB GC-FHBC
Quantitive score 4+
3+
2+
1+
Mean
44 12 23 11
26 45 35 43
12 15 22 21
5 15 7 12
3.5~ 2.6 2.8 2.6
aThe difference in quantitative growth detected between GC-HgS and GC-FHB, GC-FHBCand MH-HB is statistically highlysignificant(p < 0.0001 Student's t-test).
the high cost of using two such selective media and the limited availability of certain constituents has discouraged the use of these definitive culture methods, especially in areaswhere chancroid is endemic. This study has shown that costs could be reduced by 25 % by substituting GC-FHB medium for GC-HgS medium without significantly reducing sensitivity. The data presented here and in other studies (8, 10) clearly indicate that no single medium provides optimal conditions for primary isolation of Haemophilus ducreyi. It is therefore not surprising that in this study the use of media containing Fildes' extract alone resulted in a significantly lower rate of recovery (75 %) compared with the expensive combination of two media. However, a loss of only 5 % was recorded when the rate of recovery on GC-FHB was compared with the best single medium when used alone namely GC-HgS (80%). The small sacrifice in sensitivity which occurs using Fildes' extract-containing medium is more than compensated by the reduction in expense (approximately 50 % of the cost of GC-HgS and 25 % of the combination) and the ability to provide a definitive diagnosis. In Nairobi, Kenya, 20 % of Haemophilus ducreyi strains were found to grow only on gonococcal agar-based medium and a further 10 % only on Mueller-Hinton agar based medium (8). These findings are comparable to those of the present Study ('['able 3). Sottnek et al. (3) concluded that the differences in the ability of MH-HB and GCHgS to support the growth of Haemophilus ducreyi were related to the agar bases used. They showed that neither the addition of foetal calf serum to MH-HB base with chocolated horse blood, nor the addition of chocolated blood to GC-HgS improved the frequency of isolation of Haemophilus ducreyi. In this study we have only
meagre evidence to support these findings since seven strains grew on MH-HB alone, but a further six strains grew on MI,H-HB and on Fildes' extract containing-media but not on GC-HgS medium, which contains the same agar base. However, it is interesting to note that GC-FHB may be successfully substituted for GC-HgS when using a combination of media without significantly compromising sensitivity. It would appear that differences in sensitivity of the various isolation media reflect specific growth requirements of individual strains of Haernophilus ducreyi. However, in this study no attempt was made to subculture isolates on media which did not initially support growth of these strains. The finding that Haernophilus ducreyi was not isolated from 9 % of our patients suggests that Haemophilus ducreyi strains may have been eradicated in some of these patients either spontaneously or by previous treatment. However, some of these patients may have been infected with Haemophilus ducreyi having unique nutritional requirements. As a result of this possibility and the fact that none of the media evaluated in this study supported the growth of all Haemophilusducreyistrains, further attempts to develop even better media for the recovery of this organism from clinical specimens should be made. Ongoing studies in Nairobi (Ndinya-Achola, personal communication) indicate that incorporation of activated charcoal may enhance recovery of Itaemophilus ducreyi in preexisting primary isolation media. In the meantime, we agree with Nsanze et al. (8) that both GC-HgS and MH-HB media, should be used in studies which require optimal isolation rates to be achieved. However, GC-FHB medium has proved to be a satisfactory alternative to GC-HgS when combination media are required, or in situations where a single medium could be used where cost considerations are paramount, notably in developing countries. In such situations GCFHB medium could be used for routine diagnostic purposes and for surveillance of antimicrobial susceptibility of Haemophilus ducreyi isolates in order to provide a rational basis for therapy. References 1. Deacon WE, Albritton DC, Oiansky S, Kaplan W:
V,D.R.L. chancroid studies. I: A simple procedure for the isolation and identification of Haemophilus ducreyi.Journal of InvestigationalDermatology1956, 26: 399--406. 2. Hammond GW, Lian CJ, Wilt JC, Ronalfl. AR: Comparison of specimen collection and laboratory techniques for the isolationof Haemophilusducreyi.Journal of Clinical Microbiology 1978, 7: 39--43.
934
3. Sottnek FO, Biddle JW, Kraus SJ, Weaver RE, Stewart JA: Isolation and identification of Haemophilus ducreyi in a clinical study. Journal of Clinical Microbiology 1980, 12: 170-174. 4. Bilgeri YR, Ballard RC, Duncan MO, Mauff AC, Koornhof H J: Antimicrobial susceptibility of 103 strains of lJaemophilus ducreyi isolated in Johannesburg. Antimicrobial Agents and Chemotherapy 1982, 22: 686-688. 5. Dylewski J, Nsanze H, Maitha G, Ronaid A: LaboratolT diagnosis of Haemophilus ducreyi: sensitivity of culture media. Diagnostic Microbiology and Infectious Diseases 1986, 4: 241-245. 6. Kunimito DY, Slaney L, Koss J, D'Costa LJ, Plummer FA, Ndinya-Achola JO, Ronald AR: Field testing of modified Bieling media for the isolation of Haemophilus ducreyi in Kenya. European Journal of Clinical Microbiology 1986, 5: 673-675. 7. Fildes P: A new medium for the growth of B. influenzae. British Journal of Experimental .Pathology 1920, 1: 129-130. 8. Nsanze H, Plummer FA, Maggwa ABN, Martha G, Dylewski J, Plot P, Ronald AR: Comparison of media for the primary isolation of Haemophilus ducreyi. Sexually Transmitted Diseases 1984, 11: 6-9. 9. Kilian M: A taxonomic study of the genus Haemophilus with a proposal of new species. Journal of General Microbiology 1976, 93: 9--62. 10. MacDonald K, Cameron W, Irungu G, D'Costa LJ, Plammer FA,' Slanley L, Ndinya-Achola JO, Ronald AR: Comparison of Sheffield media with standard media for the isolation of llaemophitus ducreyi. Sexually Transmitted Diseases 1989, 16: 88--90.
Impaired Detection of Faecal Verocytotoxin in the Presence of Clostridium difficile Cytotoxin in Patients with Haemolytic Uraemic Syndrome I. L u z z i 1., E Minelti 2, A . G i a n v i t i 3, A. C a p r i o l i 2
Three cases of haemolytic uraemic syndrome associated with infection with verocytotoxin producing Escherichia coil are described. The concomitant presence of Clostridium difficile cytotoxin in the patients' stool impaired the l Laboratorio di Batteriologia c Micologia Medica and 2Laboratorio di Ultrastrutture, lstituto Superiore di Sanit'~, Viale Regina Elena 299, 00161 Rome, Italy. 3Divisione di Nefi'ologia e Dialisi, Ospedale Pediatrico Bambino Ges~, Istituto di Ricerca Scientifica, Salita S. Onofrio 4, 00100 Rome, Italy.
Eur. J. Clin. Microbiol. Infect. Dis.
detection of free faecal verocytotoxin. Stool specimens containing Clostridium difficile cytotoxin should thus be considerednegative for verocytotox~n only after neutralisation of the Clostridium difficile cytotoxin with antitoxin.
Infection with verocytotoxin producing Escherichia coil ( V T E C ) is associated with a spectrum of disorders including diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome (1). The haemolytic uraemic syndrome constitutes a major cause of acute renal failure in childhood, and is also characterized by thrombocytopoenia and microangiopathic haemolytic anaemia (1). Following the first report by Karmall et al. (2), the association between haemolytic uraemic syndrome and V T E C infection has been confirmed by several other investigators using various microbiological and serological techniques (3-6). Microbiological diagnosis of V T E C infection is based on the isolation of V T E C from faecal cultures and/or the demonstration of free verocytotoxin in faecal filtrates (1). A major problem encountered in the isolation of V T E C from stool cultures is the relatively low n u m b e r of these organisms in mixed flora, and their rapid disappearance from stools within a few days after the onset of symptoms (1, 5). This occurs particularly in patients with haemolytic uraemic syndrome, whose stools are often received one to two weeks after the onset of the diarrhoeal prodromat illness. The low sensitivity of the isolation procedures may be overcome by examining faecal filtrates for free verocytotoxin. However, although faecal verocytotoxin has been consistently found by different investigators in the stools of patients positive for V T E C on culture, it has also been detected in specimens from many haemolytic uraemic syndrome patients who were culture negative for V T E C (1, 2, 6). We report three cases of VTEC-associated haemolytic uraemic syndrome in which detection of faecal verocytotoxin was complicated by a concomitant infection with Clostridium difficile. Clostridium difficile has frequently been found in the stool of children up to two years of age (7-9), and is known to produce two potent toxins cytopathic for cultured cells (10).
Patients a n d Methods. The three cases were observed during a study on the association between haemolytic uraemic syndrome and V T E C infections in Italian children (11). T h e onset of
