- Email: [email protected]
Macrolides and changes in the oral flora
International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23 – S29
Macrolides and changes in the oral flora Armine M Sefton * Department of Medical Microbiology, St. Bartholomew’s and the Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK
Abstract Macrolides have been used in dental practice for many years, and may have a role in treating periodontal disease. Increased numbers of antibiotic-resistant oral streptococci have been reported after administration of both penicillins and macrolides. We confirm these findings for erythromycin, josamycin and azithromycin, and show that small numbers of macrolide-resistant streptococci are part of the normal oral flora at baseline. Resistant organisms fill the vacuum created by the removal of sensitive strains by antibiotic treatment. Following treatment with azithromycin, periodontal bacterial pathogens such as black pigmented anaerobes and spirochaetes decrease, whereas numbers of oral streptococci increase. These changes in the oral flora indicate a return to a healthier oral environment. In our studies, no clinical problems resulted from the transient increase in macrolideresistant streptococci. © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Macrolides; Resistance; Streptococci; Periodontitis
1. Introduction Antibiotics are sometimes used prophylactically in dentistry in susceptible patients to lessen the risk of endocarditis, associated with the release of bacteria into the circulation during certain procedures such as dental extractions and root fillings. Antibiotics are additionally used in the treatment of dental abcesses and may also have a role as adjuncts to oral hygiene in the treatment of periodontal disease. The administration of a number of different oral antibiotics used in surgical prophylaxis or to treat periodontal disease has, however, been shown to alter the local flora. The number of resistant streptococci in the oral cavity increase. Studies with amoxycillin have indicated that single doses do not significantly affect the resistance patterns of oral streptococci [1,2], but that two or more doses result in an increase in the number of amoxycillin-resistant streptococci recovered from the mouth [1–4].
* Tel.: +44-171-3777244; fax: +44-171-3750518.
Macrolides have, in the past, been used prophylactically in dental and oral surgery, and the newer macrolides, e.g. azithromycin, may have a role in treating odontogenic infections [5]. In common with amoxycillin, the administration of macrolides results in an increase in the number of macrolide-resistant streptococci in the oral cavity. Earlier studies with erythromycin [6,7] and roxithromycin [8] reported increased resistance among streptococci isolated from the mouth. In the case of roxithromycin, this was observed after only one dose. Over a period of time, in two separate studies, we have investigated the effects on the oral flora of erythromycin, josamycin and azithromycin. In one study, we examined streptococci isolated from saliva samples from healthy volunteers who had been given ‘endocarditis prophylaxis’ regimens of either erythromycin or josamycin. In the second study, we examined bacterial samples taken from the periodontal pockets of patients with periodontal disease treated with either placebo or a therapeutic regimen of azithromycin—500 mg once a day for 3 days. This paper reviews the combined findings of these studies.
0924-8579/99/$ - see front matter © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 8 ) 0 0 1 0 0 - 9
S24
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
2. Materials and methods
2.1. Subjects In both studies, volunteers had to be aged at least 18 years, give full informed consent and, if female, be taking adequate contraceptive measures unless postmenopausal. Exclusion criteria were: history of liver disease, asthma or previous adverse reaction to erythromycin, azithromycin, or any other macrolide; concomitant use of ergotamine tartrate or its derivatives, theophylline or similar compounds, carbamazepime, oral anticoagulants; or receipt of any antibiotic within the previous 3 months. Both trials were approved by our local ethics committee. Volunteers from the erythromycin/josamycin prophylaxis study were healthy and aged 18–30 years. Volunteers for the periodontal study were patients aged 18 – 65 years with periodontal disease but who were otherwise in good health. In addition, volunteers for this study had to have at least 20 remaining teeth.
2.2. Trial protocols 2.2.1. Study 1 — erythromycin and josamycin prophylaxis study All volunteers were asked to eat a standard light breakfast on the day of antimicrobial administration prior to arrival. The volunteers for this double-blind study were randomized to two groups. One group received oral erythromycin stearate 1.5 g followed 6 h later by a further dose of 0.5 g erythromycin stearate; the other group received josamycin base (the most commonly used european formulation) 1.5 g followed 6 h later by a further 0.5 g dose of josamycin base. Saliva samples were collected from each volunteer immediately prior to antibiotic administration, and again at 1.5 and 6 h after the first dose. Further saliva samples were collected 2, 7, 30 and 90 days post-dose. Undiluted saliva (100 ml) and 10-fold dilutions in tryptone soya agar broth (100 ml) (Oxoid CM129) were applied to tryptone soya agar (Oxoid CM 131) supplemented with 5% defibrinated horse blood and pyridoxal hydrochloride (Sigma, 20 mg/l final concentration) and containing nalidixic acid (7.5 mg/l final concentration) and neomycin (15 mg/l final concentration). Nalidixic acid and neomycin were included to make the media selective for streptococci. However, we used half the normal recommended concentrations of nalidixic acid and neomycin [9], since in preliminary studies we had found these concentrations to be inhibitory to several of the test strains used. The pyridoxal hydrochloride was added to all media as, again in a preliminary study, we found a proportion of isolates to be pyridoxal-dependent. Additional plates were made selective for re-
sistant organisms by incorporating 1, 4 or 64 mg/l erythromycin or josamycin into the plates. Plates were inoculated by using a glass spreader after the drops had been absorbed into the medium. They were incubated at 37°C in air for 48 h. Colonies were counted to estimate the number of viable streptococci per ml of saliva, and the percentage and number resistant to each of the macrolides, at each antibiotic concentration used. Streptococcal isolates resistant to 64 mg/l of either erythromycin or josamycin were purified and stored at − 70°C for subsequent identification by the API20 Strep system and for further antibiotic susceptibility testing. Minimum inhibitory concentrations (MICs) of erythromycin, josamycin, azithromycin, clarithromycin, roxithromycin, clindamycin, pristinamycin and amoxycillin for these highly resistant isolates were determined using a standard agar dilution technique, an inoculum of 104 colony forming units (cfu), and a multipoint inoculator (Denley Instruments). Plates for susceptibility testing were incubated for 18– 24 h at 37°C in air before being read. Preliminary tests had shown that the presence of pyridoxal had no detectable effect on MICs [10].
2.2.2. Study 2 — azithromycin in the treatment of periodontal disease Patients were randomized to one of two treatment groups in this double-blind study, and assessed at weeks 0 (baseline), 1, 2, 6, 10 and 22. All assessments were performed by the same person throughout. Assessments included plaque index, bleeding index and calculus index on six teeth and probing depths on the on the mesial, buccal, distal and lingual aspects of every tooth. One group received standard non-surgical care plus once-daily azithromycin 500 mg for 3 days immediately after assessment at their week 2 visit. The other group received standard non-surgical care and oncedaily placebo for 3 days immediately after assessment at their week 2 visit. The standard non-surgical care comprised oral hygiene instruction, scaling and polishing at weeks 0, 1, 2, 6 and 10, and root planing at weeks 0, 1 and 2. This was conducted by the same dental hygienist throughout the study. Microbiological samples were obtained by the same operator at weeks 0 (baseline), 2, 3, 6, 10 and 22. A sterile paper point was inserted into the base of the same single periodontal pocket (initially the most severely affected and ] 6 mm deep), left in situ for 10 s, and transferred into 1 ml fastidious anaerobe broth (Lab M) containing glass beads to facilitate later dispersion of the sample. Samples were taken for immediate processing. They were vortexed for 30 s and transferred into an anaerobic cabinet (Whitley Mark 1) with an atmosphere of 10% hydrogen, and 10% carbon dioxide in nitrogen. Serial 10-fold dilutions were made of each sample in
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
S25
Fig. 1. Mean percentage of oral streptococci resistant to erythromycin (a) or josamycin (b) following the administration of erythromycin.
fastidious anaerobe broth, down to 1:10 000, and 0.1 ml volumes of each dilution were spread on the following: 1. Fastidious anaerobe agar (Lab M) supplemented with 20 mg/l pyridoxal hydrochloride and 5% horse blood, for the determination of the total number of cfu per sample, and the identification of colonies. 2. Fastidious anaerobe agar supplemented as above, plus 2 mg/l azithromycin, for the isolation of total number of cfu of azithromycin-resistant bacteria per sample. 3. Tryptone soya agar (Oxoid CM131) plus 5 mg/l vancomycin, 75 mg/l bacitracin and 10% horse serum, for the isolation of Actinobacillus actinomycetemcomitans [11] 4. Mitis salivarius agar (Difco) plus 0.0001% potassium tellurite and 20 mg/l pyridoxal hydrochlorides for the isolation of streptococci, including nutritionally variant strains. 5. Mitis salivarius agar (Difco) plus 0.0001% potassium tellurite, 20 mg/l pyridoxal hydrochloride and 2 mg/l azithromycin, for the isolation of azithromycin-resistant streptococci. The tryptone soya agar plates were incubated in air with added carbon dioxide at 35°C for 3 days. All other media were incubated anaerobically at 37°C for 5 days. Actinomyces spp., Eikenella spp., Capnocytophaga spp. and black-pigmented anaerobes were identified by colonial and cellular morphology and Gram-stain from isolates grown on the fastidious anaerobic agar plates. Pre6otella spp. and Porphyromonas spp. were identified according to their ability to produce indole and hydrolyse a range of fluorogenic substances [12]. Actinobacillus actinomycetemcomitans, from the tryprone soya agar plates, were identified by characteristic colonial morphology using a plate microscope. Bacterial counts were expressed as log10 cfu/sample. Spirochaetes were identified from undiluted samples in wet preparations using high power (×1000) phase contrast microscopy. Ten fields were examined for each
sample, and the number detected expressed as log10 count/100 high-power fields (HPF). Means and standard deviations were calculated for each count at each time, and compared using analysis of variance and Duncan’s method to compare means (SPSSPC+ , SPSS, Chicago, IL, USA).
3. Results
3.1. Study 1 — Erythromycin and josamycin prophylaxis study Eight of the 16 volunteers enrolled in this study received erythromycin and eight josamycin. The administration of either macrolide resulted in a 1–2 log reduction in the counts of streptococci in the 1.5-h and 6-h saliva samples compared with baseline. These had generally reverted back to original levels by the 7-day sample. At baseline, all the subjects in the erythromycin group harboured streptococci resistant to 1 mg/l erythromycin, seven of the eight subjects had isolates resistant to 4 mg/l erythromycin, and three of the eight yielded streptococci resistant to 64 mg/l erythromycin. In the josamycin group, at baseline, streptococci resistant to 1 mg/l of the macrolide were isolated from seven of the eight subjects, to 4 mg/l in seven of the eight subjects, and to 64 mg/l in six of the eight subjects. However, the proportion of streptococci resistant to either 1 mg/l erythromycin or josamycin at baseline was B 1.5%. Two days after the administration of macrolide, the proportion of resistant streptococci increased in both groups (Fig. 1a, Fig. 2a). Giving erythromycin caused cross-resistance to josamycin, and vice-versa (Fig. 1 b, Fig. 2b). Although, the number of resistant oral streptococci declined over the duration of the study, streptococci resistant to erythromycin or josamycin were still isolated from the 90-day samples. A total of nine of the 17 highly resistant organisms
S26
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
Fig. 2. Mean percentage of oral streptococci resistant to josamycin (a) or erythromycin (b) following the administration of josamycin.
isolated from the plates containing 64 mg/l antibiotic were identified as Streptococcus mitis and seven as S. sanguis. The remaining isolate was S. sali6arius. All the isolates had MICs ] 256 mg/l for erythromycin, azithromycin and roxithromycin. The MIC of josamycin for 14 of the isolates was ]256 mg/l and for the remaining three isolates, the MIC values were 32, 64, and 1 mg/l. All but one of the isolates had MIC ] 256 mg/l for clarithromycin. The MIC value for clindamycin against 14 of the isolates was 256 mg/l (MIC = 128 mg/l for two isolates and 32 mg/l for one isolate). All the isolates were susceptible to pristinamycin (mode MIC=0.25 mg/l) and amoxycillin (mode MIC=0.06 mg/l).
3.2. Study 2 —azithromycin in the treatment of periodontal disease A total of 46 patients were enrolled in this study. The administration of azithromycin led to a significant decrease compared with placebo in mean pocket depth in patients with moderate-to-severe periodontal disease at 22 weeks post-antibiotic administration. The total number of organisms isolated (mean log10 cfu/sample) throughout the period of the study remained relatively constant. The total number of streptococci increased slightly, although not significantly so, at week 2 compared to baseline, and remained at this higher level for the period of the study. The number of azithromycinresistant streptococci isolated increased significantly in the azithromycin-treated patients compared with those given the placebo at week 3, and remained so until the end of the study (Fig. 3). There was no significant increase in the total number of azithromycin-resistant organisms found. Counts of black-pigmented anaerobes (Fig. 4), including Pre6otella intermedia, were significantly reduced with azithromycin treatment compared with placebo at weeks 3 and 6, as were counts of Actinomyces spp.
Porphyromonas gingi6alis counts were significantly reduced with azithromycin treatment compared with placebo at weeks 3, 6 and 10 (Fig. 5). Azithromycin treatment significantly reduced counts of spirochaetes compared with placebo even at week 22 (Fig. 6), although spirochaete counts were reduced in both groups. Actinobacillus actinomycetemcomitans was isolated from only six patients and in small numbers.
4. Discussion These two studies demonstrated that administration of erythromycin (a 14-membered macrolide), josamycin (a 16-membered macrolide) and azithromycin (an azalide) all led to an increase in the number of macrolideresistant oral streptococci isolated from the mouth, and confirmed earlier studies showing the same findings with erythromycin [6,7] and roxithromycin (a 14-membered macrolide) [8].
Fig. 3. Effects of azithromycin and placebo on the mean number of cfu of azithromycin-resistant streptococci per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
S27
Fig. 4. Effects of azithromycin and placebo on the mean number of cfu of black-pigmented anaerobes per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
Fig. 6. Effects of azithromycin and placebo on the mean number of cfu of spirochaetes per high-powered field [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
Macrolide-resistant oral streptococci occur naturally among the normal oral flora. Isolates resistant to erythromycin, josamycin, and azithromycin were isolated prior to antibiotic administration in these studies. This finding differs from that of earlier studies [6,13], which found no macrolide-resistant streptococci prior to administration of erythromycin. This difference may be accounted for by the inclusion of pyridoxal hydrochloride in our media; the growth of many of our resistant streptococcal isolates was found to be pyridoxal-dependant. Pyridoxal supplementation has been shown to significantly increase the rate of isolation of pyridoxaldependent streptococci [14]. Alternatively, the reduced levels of selective antibiotics in our media — 50% of that recommended by Sukchotiratana [9] — could account for the higher isolation rate. Other studies [6,8] used the same concentrations.
The higher proportion of resistant streptococci persisted for up to 3 months. Long-term persistence of streptococci resistant to macrolides after exposure to prophylactic regimens has been reported previously [6,7]. Similarly, exposure to repeated short courses of amoxycillin over a limited period results in increased isolation of amoxycillin-resistant streptococci from the mouth [2–4]. This suggests that an increase in the number of resistant streptococci isolated from the mouth after administration of an antibiotic is a phenomenon common to several classes of antibiotics. The antibiotics inhibit the majority of sensitive strains of streptococci, allowing the pre-existing naturally occuring variants to fill the resulting ecological niche. Azithromycin treatment altered not only the proportion of resistant strains present in the mouth, but also the nature of the overall bacterial population. Treatment with this macrolide in the periodontal study resulted in the reduction of periodontal pathogens, including black-pigmented anaerobes, Actinomyces spp. and spirochaetes, isolated from patients. Since the total mean count of cfu/sample remained substantially unchanged, it is suggested that the increase in streptococci compensated for the loss of other bacteria. The clinical results of this study showed that disease parameters in patients treated with standard non-surgical care plus azithromycin improved significantly compared with those given standard non-surgical care plus placebo ([15] and personal communication). In both groups, there was a statistically improved plaque index, bleeding index, and calculus index, at 22 weeks compared with baseline; the difference between the two groups was not significant with regard to these parameters. However, although in both groups, pockets that were \4–5 mm deep at baseline showed a significant decrease in depth at the end of the study, the mean
Fig. 5. Effects of azithromycin and placebo on the mean number of cfu of Porphyromonas gingi6alis per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
S28
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
pocket depth at week 22 was significantly lower in the azithromycin-treated group than in the group receiving placebo. The observed increase in macrolide resistant oral streptococci did not adversely affect clinical outcome. The aetiology of chronic periodontitis is complex, with a number of different organisms having been implicated. However, the general consensus is that disease is associated primarily with black-pigmented anaerobes, including Porphyromonas gingi6alis and Pre6otella intermedia [16 – 20], and spirochaetes [21,22]. Gram-positive cocci form a greater proportion of the oral flora in healthy individuals than these anaerobic and pigmented organisms [16 – 20]. Actinobacillus actinomycetemcomitans is considered a major risk factor in juvenile periodontitis [19], but not in adult periodontal disease; its low incidence in this study of adult patients is thus unsurprising. Azithromycin treatment resulted in significant decreases compared with placebo in the organisms most commonly implicated as likely pathogens in chronic adult periodontal disease, and this correlated well with the improved clinical outcome associated with azithromycin treatment. The slightly increased overall number of streptococci (a high proportion of which were resistant strains) indicated a return to a healthier condition of the patients’ mouths. In conclusion, therefore, increased numbers of resistant oral streptococci can be isolated from the mouth after administration of macrolides and other antimicrobial agents. It is probable that development of macrolide resistance in oral streptococci post-administration of macrolides is a class effect common to most macrolides, since we found it to occur following administration of a macrolide with a 14-membered ring (erythromycin), a 15-membered ring (azithromycin) and a 16-membered ring (josamycin). This finding of an increased proportion of strains with decreased susceptibility is likely to be a result of resistant strains normally present in the mouth filling the vacuum left vacant by the inhibition of susceptible organisms, which may include potential pathogens in the disease state. Oral streptococci resistant to one macrolide usually show cross-resistance to other macrolides and to clindamycin, but not to pristinamycin. Cross-resistance and crosssusceptiblity to erythromycin, clarithromycin and azithromycin also occurs in other organisms, e.g. other Streptococcus spp., Moraxella catarrhalis, Staphylococcus spp., Gram-positive anaerobes, and Gram-negative bacilli [23]. However, early in vitro studies with investigational ketolides (a novel macrolide sub-class) suggest that erythromycin-resistant streptococci remain susceptible to ketolides [24,25]. Although we found that resistant oral streptococci persisted for up to three months, post-macrolide administration, in our dental studies they were thought not to be clinically relevant.
References [1] MacGregor AJ, Hart P. The effect of a single large dose of amoxycillin on oral streptococci. J Antimicrob Chemother 1986;18:113 – 7. [2] Woodman AJ, Vidic J, Newman HN, Marsh PD. Effect of repeated high dose prophylaxis with amoxycillin on the resident oral flora of adult volunteers. J Med Microbiol 1985;19:15 – 23. [3] Southall PJ, Mahy NJ, Davies RM, Speller DC. Resistance in oral streptococci after repeated two-dose amoxycillin prophylaxis. J Antimicrob Chemother 1983;12:141 – 6. [4] Harrison GA, Rubin MP, Davies RM, Speller DC. Resistance in oral streptococci after repetition of a single-dose amoxycillin prophylactic regimen. J Antimicrob Chemother 1985;15:501–3. [5] Lo-Bue AM, Sammartino R, Chisari G, Gismondo MR, Nicoletti G. Effficacy of azithromycin compared with spiramycin in the treatment of odontogenic infections. J Antimicrob Chemother 1993;31(Suppl. E):119 – 27. [6] Harrison GA, Stross WP, Rubin MP, Davies RM, Speller DC. Resistance in oral streptococci after repeated three-dose erythromycin prophylaxis. J Antimicrob Chemother 1985;15:471– 9. [7] Herbert M, Moore E, Radford J, Kubiak E, Speller D. Oral streptococci resistant to erythromycin after repetition of a regimen designed for dental prophylaxis. J Antimicrob Chemother 1988;21:677 – 9. [8] Smith GE, Fulford MR, Matthews RW, Speller DC. Roxithromycin as a possible agent for prophylaxis of endocarditis: a study in normal volunteers. J Antimicrob Chemother 1989;23:417 – 25. [9] Sukchotiratana M, Linton AH, Fletcher JP. Antibiotics and the oral streptococci of man. J Appl Bacteriol 1975;38:277–94. [10] Maskell JP, Sefton AM, Cannell H, et al. Predominance of resistant oral streptococci in saliva and the effect of a single course of josamycin or erythromycin. J Antimicrob Chemother 1990;26:539 – 48. [11] Slots J. Selective medium for isolation of Actinobacillus actinomycetemcomitans. J Clin Microbiol 1982;15:606 – 9. [12] Moncla BJ, Braham P, Rabe LK, Hillier SL. Rapid presumptive identification of black-pigmented Gram-negative anaerobic bacteria by using 4-methylumbelliferone derivatives. J Clin Microbiol 1991;29:1955 – 8. [13] Sprunt K, Leidy G, Redman W. Cross resistance between lincomycin and erythromycin in viridans streptococci. Pediatrics 1970;46:84 – 8. [14] George RH. The isolation of symbiotic streptococci. J Med Microbiol 1974;7:77 – 83. [15] Sefton AM, Maskell JP, Beighton D, et al. Azithromycin in the treatment of periodontal disease: effect on microbial flora. J Clin Periodontol 1996;23:998 – 1003. [16] Socransky SS, Haffajee AD. The bacterial aetiology of destructive periodontal disease: current concepts. J Periodontol 1992;63(4 Suppl):322 – 31. [17] Wolff L, Dahlen G, Aeppli D. Bacteria as risk markers for periodontitis. J Periodontol 1994;65(5 Suppl):498 – 510. [18] Listgarten MA. Structure of the microbial flora associated with periodontal health and disease in man. A light and electron microscopic study. J Periodontol 1976;47:1 – 18. [19] Slots J, Rams TE. Antibiotics in periodontal therapy: advantages and disadvantages. J Clin Periodontol 1990;17:479–93. [20] Loesche WJ, Schmidt E, Smith BA, Morrison EC, Caffesse R, Hujoel PP. Effects of metronidazole on periodontal treatment needs. J Periodontol 1991;62:247 – 57. [21] Russell RR. Bacteriology of periodontal disease. Curr Opin Dent 1992;2:66 – 71.
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29 [22] Hardie JM. Oral microbiology: current concepts in the microbiology of dental caries and periodontal disease. Br Dent J 1992;172:271 – 8. [23] Fass RJ. Erythromycin, clarithromycin, and azithromycin: use of frequency distribution curves, scattergrams, and regression analyses to compare in vitro activities and describe cross-resis-
.
S29
tance. Antimicrob Agents Chemother 1993;37:2080 – 6. [24] Jamjian C, Biedenbach DJ, Jones RN. In vitro evaluation of a novel ketolide antimicrobial agent, RU-64004. Antimicrob Agents Chemother 1997;41:454 – 9. [25] Jones RN, Biedenbach DJ. Antimicrobial activity of RU-66647, a new ketolide. Diagn Microbiol Infect Dis 1997;27:7 –12.
Macrolides and changes in the oral flora Armine M Sefton * Department of Medical Microbiology, St. Bartholomew’s and the Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK
Abstract Macrolides have been used in dental practice for many years, and may have a role in treating periodontal disease. Increased numbers of antibiotic-resistant oral streptococci have been reported after administration of both penicillins and macrolides. We confirm these findings for erythromycin, josamycin and azithromycin, and show that small numbers of macrolide-resistant streptococci are part of the normal oral flora at baseline. Resistant organisms fill the vacuum created by the removal of sensitive strains by antibiotic treatment. Following treatment with azithromycin, periodontal bacterial pathogens such as black pigmented anaerobes and spirochaetes decrease, whereas numbers of oral streptococci increase. These changes in the oral flora indicate a return to a healthier oral environment. In our studies, no clinical problems resulted from the transient increase in macrolideresistant streptococci. © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Macrolides; Resistance; Streptococci; Periodontitis
1. Introduction Antibiotics are sometimes used prophylactically in dentistry in susceptible patients to lessen the risk of endocarditis, associated with the release of bacteria into the circulation during certain procedures such as dental extractions and root fillings. Antibiotics are additionally used in the treatment of dental abcesses and may also have a role as adjuncts to oral hygiene in the treatment of periodontal disease. The administration of a number of different oral antibiotics used in surgical prophylaxis or to treat periodontal disease has, however, been shown to alter the local flora. The number of resistant streptococci in the oral cavity increase. Studies with amoxycillin have indicated that single doses do not significantly affect the resistance patterns of oral streptococci [1,2], but that two or more doses result in an increase in the number of amoxycillin-resistant streptococci recovered from the mouth [1–4].
* Tel.: +44-171-3777244; fax: +44-171-3750518.
Macrolides have, in the past, been used prophylactically in dental and oral surgery, and the newer macrolides, e.g. azithromycin, may have a role in treating odontogenic infections [5]. In common with amoxycillin, the administration of macrolides results in an increase in the number of macrolide-resistant streptococci in the oral cavity. Earlier studies with erythromycin [6,7] and roxithromycin [8] reported increased resistance among streptococci isolated from the mouth. In the case of roxithromycin, this was observed after only one dose. Over a period of time, in two separate studies, we have investigated the effects on the oral flora of erythromycin, josamycin and azithromycin. In one study, we examined streptococci isolated from saliva samples from healthy volunteers who had been given ‘endocarditis prophylaxis’ regimens of either erythromycin or josamycin. In the second study, we examined bacterial samples taken from the periodontal pockets of patients with periodontal disease treated with either placebo or a therapeutic regimen of azithromycin—500 mg once a day for 3 days. This paper reviews the combined findings of these studies.
0924-8579/99/$ - see front matter © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 8 ) 0 0 1 0 0 - 9
S24
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
2. Materials and methods
2.1. Subjects In both studies, volunteers had to be aged at least 18 years, give full informed consent and, if female, be taking adequate contraceptive measures unless postmenopausal. Exclusion criteria were: history of liver disease, asthma or previous adverse reaction to erythromycin, azithromycin, or any other macrolide; concomitant use of ergotamine tartrate or its derivatives, theophylline or similar compounds, carbamazepime, oral anticoagulants; or receipt of any antibiotic within the previous 3 months. Both trials were approved by our local ethics committee. Volunteers from the erythromycin/josamycin prophylaxis study were healthy and aged 18–30 years. Volunteers for the periodontal study were patients aged 18 – 65 years with periodontal disease but who were otherwise in good health. In addition, volunteers for this study had to have at least 20 remaining teeth.
2.2. Trial protocols 2.2.1. Study 1 — erythromycin and josamycin prophylaxis study All volunteers were asked to eat a standard light breakfast on the day of antimicrobial administration prior to arrival. The volunteers for this double-blind study were randomized to two groups. One group received oral erythromycin stearate 1.5 g followed 6 h later by a further dose of 0.5 g erythromycin stearate; the other group received josamycin base (the most commonly used european formulation) 1.5 g followed 6 h later by a further 0.5 g dose of josamycin base. Saliva samples were collected from each volunteer immediately prior to antibiotic administration, and again at 1.5 and 6 h after the first dose. Further saliva samples were collected 2, 7, 30 and 90 days post-dose. Undiluted saliva (100 ml) and 10-fold dilutions in tryptone soya agar broth (100 ml) (Oxoid CM129) were applied to tryptone soya agar (Oxoid CM 131) supplemented with 5% defibrinated horse blood and pyridoxal hydrochloride (Sigma, 20 mg/l final concentration) and containing nalidixic acid (7.5 mg/l final concentration) and neomycin (15 mg/l final concentration). Nalidixic acid and neomycin were included to make the media selective for streptococci. However, we used half the normal recommended concentrations of nalidixic acid and neomycin [9], since in preliminary studies we had found these concentrations to be inhibitory to several of the test strains used. The pyridoxal hydrochloride was added to all media as, again in a preliminary study, we found a proportion of isolates to be pyridoxal-dependent. Additional plates were made selective for re-
sistant organisms by incorporating 1, 4 or 64 mg/l erythromycin or josamycin into the plates. Plates were inoculated by using a glass spreader after the drops had been absorbed into the medium. They were incubated at 37°C in air for 48 h. Colonies were counted to estimate the number of viable streptococci per ml of saliva, and the percentage and number resistant to each of the macrolides, at each antibiotic concentration used. Streptococcal isolates resistant to 64 mg/l of either erythromycin or josamycin were purified and stored at − 70°C for subsequent identification by the API20 Strep system and for further antibiotic susceptibility testing. Minimum inhibitory concentrations (MICs) of erythromycin, josamycin, azithromycin, clarithromycin, roxithromycin, clindamycin, pristinamycin and amoxycillin for these highly resistant isolates were determined using a standard agar dilution technique, an inoculum of 104 colony forming units (cfu), and a multipoint inoculator (Denley Instruments). Plates for susceptibility testing were incubated for 18– 24 h at 37°C in air before being read. Preliminary tests had shown that the presence of pyridoxal had no detectable effect on MICs [10].
2.2.2. Study 2 — azithromycin in the treatment of periodontal disease Patients were randomized to one of two treatment groups in this double-blind study, and assessed at weeks 0 (baseline), 1, 2, 6, 10 and 22. All assessments were performed by the same person throughout. Assessments included plaque index, bleeding index and calculus index on six teeth and probing depths on the on the mesial, buccal, distal and lingual aspects of every tooth. One group received standard non-surgical care plus once-daily azithromycin 500 mg for 3 days immediately after assessment at their week 2 visit. The other group received standard non-surgical care and oncedaily placebo for 3 days immediately after assessment at their week 2 visit. The standard non-surgical care comprised oral hygiene instruction, scaling and polishing at weeks 0, 1, 2, 6 and 10, and root planing at weeks 0, 1 and 2. This was conducted by the same dental hygienist throughout the study. Microbiological samples were obtained by the same operator at weeks 0 (baseline), 2, 3, 6, 10 and 22. A sterile paper point was inserted into the base of the same single periodontal pocket (initially the most severely affected and ] 6 mm deep), left in situ for 10 s, and transferred into 1 ml fastidious anaerobe broth (Lab M) containing glass beads to facilitate later dispersion of the sample. Samples were taken for immediate processing. They were vortexed for 30 s and transferred into an anaerobic cabinet (Whitley Mark 1) with an atmosphere of 10% hydrogen, and 10% carbon dioxide in nitrogen. Serial 10-fold dilutions were made of each sample in
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
S25
Fig. 1. Mean percentage of oral streptococci resistant to erythromycin (a) or josamycin (b) following the administration of erythromycin.
fastidious anaerobe broth, down to 1:10 000, and 0.1 ml volumes of each dilution were spread on the following: 1. Fastidious anaerobe agar (Lab M) supplemented with 20 mg/l pyridoxal hydrochloride and 5% horse blood, for the determination of the total number of cfu per sample, and the identification of colonies. 2. Fastidious anaerobe agar supplemented as above, plus 2 mg/l azithromycin, for the isolation of total number of cfu of azithromycin-resistant bacteria per sample. 3. Tryptone soya agar (Oxoid CM131) plus 5 mg/l vancomycin, 75 mg/l bacitracin and 10% horse serum, for the isolation of Actinobacillus actinomycetemcomitans [11] 4. Mitis salivarius agar (Difco) plus 0.0001% potassium tellurite and 20 mg/l pyridoxal hydrochlorides for the isolation of streptococci, including nutritionally variant strains. 5. Mitis salivarius agar (Difco) plus 0.0001% potassium tellurite, 20 mg/l pyridoxal hydrochloride and 2 mg/l azithromycin, for the isolation of azithromycin-resistant streptococci. The tryptone soya agar plates were incubated in air with added carbon dioxide at 35°C for 3 days. All other media were incubated anaerobically at 37°C for 5 days. Actinomyces spp., Eikenella spp., Capnocytophaga spp. and black-pigmented anaerobes were identified by colonial and cellular morphology and Gram-stain from isolates grown on the fastidious anaerobic agar plates. Pre6otella spp. and Porphyromonas spp. were identified according to their ability to produce indole and hydrolyse a range of fluorogenic substances [12]. Actinobacillus actinomycetemcomitans, from the tryprone soya agar plates, were identified by characteristic colonial morphology using a plate microscope. Bacterial counts were expressed as log10 cfu/sample. Spirochaetes were identified from undiluted samples in wet preparations using high power (×1000) phase contrast microscopy. Ten fields were examined for each
sample, and the number detected expressed as log10 count/100 high-power fields (HPF). Means and standard deviations were calculated for each count at each time, and compared using analysis of variance and Duncan’s method to compare means (SPSSPC+ , SPSS, Chicago, IL, USA).
3. Results
3.1. Study 1 — Erythromycin and josamycin prophylaxis study Eight of the 16 volunteers enrolled in this study received erythromycin and eight josamycin. The administration of either macrolide resulted in a 1–2 log reduction in the counts of streptococci in the 1.5-h and 6-h saliva samples compared with baseline. These had generally reverted back to original levels by the 7-day sample. At baseline, all the subjects in the erythromycin group harboured streptococci resistant to 1 mg/l erythromycin, seven of the eight subjects had isolates resistant to 4 mg/l erythromycin, and three of the eight yielded streptococci resistant to 64 mg/l erythromycin. In the josamycin group, at baseline, streptococci resistant to 1 mg/l of the macrolide were isolated from seven of the eight subjects, to 4 mg/l in seven of the eight subjects, and to 64 mg/l in six of the eight subjects. However, the proportion of streptococci resistant to either 1 mg/l erythromycin or josamycin at baseline was B 1.5%. Two days after the administration of macrolide, the proportion of resistant streptococci increased in both groups (Fig. 1a, Fig. 2a). Giving erythromycin caused cross-resistance to josamycin, and vice-versa (Fig. 1 b, Fig. 2b). Although, the number of resistant oral streptococci declined over the duration of the study, streptococci resistant to erythromycin or josamycin were still isolated from the 90-day samples. A total of nine of the 17 highly resistant organisms
S26
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
Fig. 2. Mean percentage of oral streptococci resistant to josamycin (a) or erythromycin (b) following the administration of josamycin.
isolated from the plates containing 64 mg/l antibiotic were identified as Streptococcus mitis and seven as S. sanguis. The remaining isolate was S. sali6arius. All the isolates had MICs ] 256 mg/l for erythromycin, azithromycin and roxithromycin. The MIC of josamycin for 14 of the isolates was ]256 mg/l and for the remaining three isolates, the MIC values were 32, 64, and 1 mg/l. All but one of the isolates had MIC ] 256 mg/l for clarithromycin. The MIC value for clindamycin against 14 of the isolates was 256 mg/l (MIC = 128 mg/l for two isolates and 32 mg/l for one isolate). All the isolates were susceptible to pristinamycin (mode MIC=0.25 mg/l) and amoxycillin (mode MIC=0.06 mg/l).
3.2. Study 2 —azithromycin in the treatment of periodontal disease A total of 46 patients were enrolled in this study. The administration of azithromycin led to a significant decrease compared with placebo in mean pocket depth in patients with moderate-to-severe periodontal disease at 22 weeks post-antibiotic administration. The total number of organisms isolated (mean log10 cfu/sample) throughout the period of the study remained relatively constant. The total number of streptococci increased slightly, although not significantly so, at week 2 compared to baseline, and remained at this higher level for the period of the study. The number of azithromycinresistant streptococci isolated increased significantly in the azithromycin-treated patients compared with those given the placebo at week 3, and remained so until the end of the study (Fig. 3). There was no significant increase in the total number of azithromycin-resistant organisms found. Counts of black-pigmented anaerobes (Fig. 4), including Pre6otella intermedia, were significantly reduced with azithromycin treatment compared with placebo at weeks 3 and 6, as were counts of Actinomyces spp.
Porphyromonas gingi6alis counts were significantly reduced with azithromycin treatment compared with placebo at weeks 3, 6 and 10 (Fig. 5). Azithromycin treatment significantly reduced counts of spirochaetes compared with placebo even at week 22 (Fig. 6), although spirochaete counts were reduced in both groups. Actinobacillus actinomycetemcomitans was isolated from only six patients and in small numbers.
4. Discussion These two studies demonstrated that administration of erythromycin (a 14-membered macrolide), josamycin (a 16-membered macrolide) and azithromycin (an azalide) all led to an increase in the number of macrolideresistant oral streptococci isolated from the mouth, and confirmed earlier studies showing the same findings with erythromycin [6,7] and roxithromycin (a 14-membered macrolide) [8].
Fig. 3. Effects of azithromycin and placebo on the mean number of cfu of azithromycin-resistant streptococci per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
S27
Fig. 4. Effects of azithromycin and placebo on the mean number of cfu of black-pigmented anaerobes per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
Fig. 6. Effects of azithromycin and placebo on the mean number of cfu of spirochaetes per high-powered field [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
Macrolide-resistant oral streptococci occur naturally among the normal oral flora. Isolates resistant to erythromycin, josamycin, and azithromycin were isolated prior to antibiotic administration in these studies. This finding differs from that of earlier studies [6,13], which found no macrolide-resistant streptococci prior to administration of erythromycin. This difference may be accounted for by the inclusion of pyridoxal hydrochloride in our media; the growth of many of our resistant streptococcal isolates was found to be pyridoxal-dependant. Pyridoxal supplementation has been shown to significantly increase the rate of isolation of pyridoxaldependent streptococci [14]. Alternatively, the reduced levels of selective antibiotics in our media — 50% of that recommended by Sukchotiratana [9] — could account for the higher isolation rate. Other studies [6,8] used the same concentrations.
The higher proportion of resistant streptococci persisted for up to 3 months. Long-term persistence of streptococci resistant to macrolides after exposure to prophylactic regimens has been reported previously [6,7]. Similarly, exposure to repeated short courses of amoxycillin over a limited period results in increased isolation of amoxycillin-resistant streptococci from the mouth [2–4]. This suggests that an increase in the number of resistant streptococci isolated from the mouth after administration of an antibiotic is a phenomenon common to several classes of antibiotics. The antibiotics inhibit the majority of sensitive strains of streptococci, allowing the pre-existing naturally occuring variants to fill the resulting ecological niche. Azithromycin treatment altered not only the proportion of resistant strains present in the mouth, but also the nature of the overall bacterial population. Treatment with this macrolide in the periodontal study resulted in the reduction of periodontal pathogens, including black-pigmented anaerobes, Actinomyces spp. and spirochaetes, isolated from patients. Since the total mean count of cfu/sample remained substantially unchanged, it is suggested that the increase in streptococci compensated for the loss of other bacteria. The clinical results of this study showed that disease parameters in patients treated with standard non-surgical care plus azithromycin improved significantly compared with those given standard non-surgical care plus placebo ([15] and personal communication). In both groups, there was a statistically improved plaque index, bleeding index, and calculus index, at 22 weeks compared with baseline; the difference between the two groups was not significant with regard to these parameters. However, although in both groups, pockets that were \4–5 mm deep at baseline showed a significant decrease in depth at the end of the study, the mean
Fig. 5. Effects of azithromycin and placebo on the mean number of cfu of Porphyromonas gingi6alis per sample [15]. (This figure is reproduced with kind permission of Munskgaard International Publishers Ltd. Copenhagen, Denmark © 1996.)
S28
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29
pocket depth at week 22 was significantly lower in the azithromycin-treated group than in the group receiving placebo. The observed increase in macrolide resistant oral streptococci did not adversely affect clinical outcome. The aetiology of chronic periodontitis is complex, with a number of different organisms having been implicated. However, the general consensus is that disease is associated primarily with black-pigmented anaerobes, including Porphyromonas gingi6alis and Pre6otella intermedia [16 – 20], and spirochaetes [21,22]. Gram-positive cocci form a greater proportion of the oral flora in healthy individuals than these anaerobic and pigmented organisms [16 – 20]. Actinobacillus actinomycetemcomitans is considered a major risk factor in juvenile periodontitis [19], but not in adult periodontal disease; its low incidence in this study of adult patients is thus unsurprising. Azithromycin treatment resulted in significant decreases compared with placebo in the organisms most commonly implicated as likely pathogens in chronic adult periodontal disease, and this correlated well with the improved clinical outcome associated with azithromycin treatment. The slightly increased overall number of streptococci (a high proportion of which were resistant strains) indicated a return to a healthier condition of the patients’ mouths. In conclusion, therefore, increased numbers of resistant oral streptococci can be isolated from the mouth after administration of macrolides and other antimicrobial agents. It is probable that development of macrolide resistance in oral streptococci post-administration of macrolides is a class effect common to most macrolides, since we found it to occur following administration of a macrolide with a 14-membered ring (erythromycin), a 15-membered ring (azithromycin) and a 16-membered ring (josamycin). This finding of an increased proportion of strains with decreased susceptibility is likely to be a result of resistant strains normally present in the mouth filling the vacuum left vacant by the inhibition of susceptible organisms, which may include potential pathogens in the disease state. Oral streptococci resistant to one macrolide usually show cross-resistance to other macrolides and to clindamycin, but not to pristinamycin. Cross-resistance and crosssusceptiblity to erythromycin, clarithromycin and azithromycin also occurs in other organisms, e.g. other Streptococcus spp., Moraxella catarrhalis, Staphylococcus spp., Gram-positive anaerobes, and Gram-negative bacilli [23]. However, early in vitro studies with investigational ketolides (a novel macrolide sub-class) suggest that erythromycin-resistant streptococci remain susceptible to ketolides [24,25]. Although we found that resistant oral streptococci persisted for up to three months, post-macrolide administration, in our dental studies they were thought not to be clinically relevant.
References [1] MacGregor AJ, Hart P. The effect of a single large dose of amoxycillin on oral streptococci. J Antimicrob Chemother 1986;18:113 – 7. [2] Woodman AJ, Vidic J, Newman HN, Marsh PD. Effect of repeated high dose prophylaxis with amoxycillin on the resident oral flora of adult volunteers. J Med Microbiol 1985;19:15 – 23. [3] Southall PJ, Mahy NJ, Davies RM, Speller DC. Resistance in oral streptococci after repeated two-dose amoxycillin prophylaxis. J Antimicrob Chemother 1983;12:141 – 6. [4] Harrison GA, Rubin MP, Davies RM, Speller DC. Resistance in oral streptococci after repetition of a single-dose amoxycillin prophylactic regimen. J Antimicrob Chemother 1985;15:501–3. [5] Lo-Bue AM, Sammartino R, Chisari G, Gismondo MR, Nicoletti G. Effficacy of azithromycin compared with spiramycin in the treatment of odontogenic infections. J Antimicrob Chemother 1993;31(Suppl. E):119 – 27. [6] Harrison GA, Stross WP, Rubin MP, Davies RM, Speller DC. Resistance in oral streptococci after repeated three-dose erythromycin prophylaxis. J Antimicrob Chemother 1985;15:471– 9. [7] Herbert M, Moore E, Radford J, Kubiak E, Speller D. Oral streptococci resistant to erythromycin after repetition of a regimen designed for dental prophylaxis. J Antimicrob Chemother 1988;21:677 – 9. [8] Smith GE, Fulford MR, Matthews RW, Speller DC. Roxithromycin as a possible agent for prophylaxis of endocarditis: a study in normal volunteers. J Antimicrob Chemother 1989;23:417 – 25. [9] Sukchotiratana M, Linton AH, Fletcher JP. Antibiotics and the oral streptococci of man. J Appl Bacteriol 1975;38:277–94. [10] Maskell JP, Sefton AM, Cannell H, et al. Predominance of resistant oral streptococci in saliva and the effect of a single course of josamycin or erythromycin. J Antimicrob Chemother 1990;26:539 – 48. [11] Slots J. Selective medium for isolation of Actinobacillus actinomycetemcomitans. J Clin Microbiol 1982;15:606 – 9. [12] Moncla BJ, Braham P, Rabe LK, Hillier SL. Rapid presumptive identification of black-pigmented Gram-negative anaerobic bacteria by using 4-methylumbelliferone derivatives. J Clin Microbiol 1991;29:1955 – 8. [13] Sprunt K, Leidy G, Redman W. Cross resistance between lincomycin and erythromycin in viridans streptococci. Pediatrics 1970;46:84 – 8. [14] George RH. The isolation of symbiotic streptococci. J Med Microbiol 1974;7:77 – 83. [15] Sefton AM, Maskell JP, Beighton D, et al. Azithromycin in the treatment of periodontal disease: effect on microbial flora. J Clin Periodontol 1996;23:998 – 1003. [16] Socransky SS, Haffajee AD. The bacterial aetiology of destructive periodontal disease: current concepts. J Periodontol 1992;63(4 Suppl):322 – 31. [17] Wolff L, Dahlen G, Aeppli D. Bacteria as risk markers for periodontitis. J Periodontol 1994;65(5 Suppl):498 – 510. [18] Listgarten MA. Structure of the microbial flora associated with periodontal health and disease in man. A light and electron microscopic study. J Periodontol 1976;47:1 – 18. [19] Slots J, Rams TE. Antibiotics in periodontal therapy: advantages and disadvantages. J Clin Periodontol 1990;17:479–93. [20] Loesche WJ, Schmidt E, Smith BA, Morrison EC, Caffesse R, Hujoel PP. Effects of metronidazole on periodontal treatment needs. J Periodontol 1991;62:247 – 57. [21] Russell RR. Bacteriology of periodontal disease. Curr Opin Dent 1992;2:66 – 71.
A.M. Sefton / International Journal of Antimicrobial Agents 11 Suppl. 1 (1999) S23– S29 [22] Hardie JM. Oral microbiology: current concepts in the microbiology of dental caries and periodontal disease. Br Dent J 1992;172:271 – 8. [23] Fass RJ. Erythromycin, clarithromycin, and azithromycin: use of frequency distribution curves, scattergrams, and regression analyses to compare in vitro activities and describe cross-resis-
.
S29
tance. Antimicrob Agents Chemother 1993;37:2080 – 6. [24] Jamjian C, Biedenbach DJ, Jones RN. In vitro evaluation of a novel ketolide antimicrobial agent, RU-64004. Antimicrob Agents Chemother 1997;41:454 – 9. [25] Jones RN, Biedenbach DJ. Antimicrobial activity of RU-66647, a new ketolide. Diagn Microbiol Infect Dis 1997;27:7 –12.