Erythromycin-resistance of cutaneous bacterial flora in acne

ARTICLE

Topical antibiotics and erythromycin in particular are extensively used in the treatment of the inflammatory component of acne, either alone or in combination with systemic treatment. Their efficacy is essentially linked to antibacterial activity on Propionibacterium acnes, a microorganism involved in the development of acne, and to a lesser degree Staphylococcus epidermidis. In 1976, Leyden [1] found no evidence of antibiotic resistance in a study involving 1,000 acne patients. Since that time, a number of studies in the literature [2-4] have reported the occurrence in acne patients, of strains of Propionibacterium acnes and of Staphylococcus epidermidis resistant to antibiotics and notably to erythromycin, with an increasing percentage prevalence. These studies undertaken essentially in Anglo-Saxon countries, raise the problem of the relation between bacterial resistance and the efficacy of acne treatment, as well as that of the most appropriate way of using topical antibiotics in acne.

Our aim in this pilot study was to determine in a group of unselected French acne patients seen in department of Dermatology, the prevalence of strains of Propionibacterium acnes and of Staphylococcus epidermidis resistant to erythromycin, and to compare these results with those already published in other countries, bearing in mind that therapeutic habits are not always the same from one country to another. We also studied the in vitro susceptibility of these strains to zinc. Indeed, it has been suggested that in vitro in combination with erythromycin, zinc could have an inhibitory action on erythromycin-resistant strains of Propionibacterium acnes (MIC > 500 mug/ml) [5].

Material and methods

Forty acne patients over 12 years of age with mild to moderately severe acne (grade 0.5 to 3.0 Burke and Cunliffe classification) [6] were included in the study.

Sample method

Microbiological samples were obtained from each patient using the washings method of Williamson and Kligman [7] modified and standardized by Fleurette [8] and using an electric sampling device. 2.5 ml of a neutralizing washing fluid (Sörensen phosphate buffer pH 7.9 (100 ml), Triton x 100 (0.1 g), sodium thioglycollate (0.2 g), Tween 80 (3 g), sodium thiosulfate (0.3 g)) were poured into a glass cylinder 2 cm internal diameter applied to the skin. At the end of the electrical sampling device was a sterilizable rubber paddle which was inserted into the glass cylinder. This paddle rotated when the device was switched on, causing rubbing of the skin surface and dispersal of bacteria present in the washing fluid. The sampling time was short, about 20 seconds. Washings were collected using a syringe and transferred to a sterile bottle.

Microbiology

The volume of washings was measured by graduated pipettes. 0.1 ml of washings, pure and diluted in sterile normal saline (10^{- 1}, 10^{- 2}, 10^{- 3}, 10^{- 4}), was inoculated by spreading with a scraper on horse blood agar (BA) for staphylococcus epidermidis and Columbia for propionibacterium acnes. BA was supplemented with 4 mug/ml of erythromycin, in order to count total aerobic bacteria and erythromycin-resistant aerobic bacteria, after 48 hours incubation at 37° C in a CO₂-enriched atmosphere.

The various types of aerobic bacteria present were then counted (oxydase+ or oxydase- Gram negative bacilli, catalase+ or catalase- Gram positive cocci, Gram positive bacilli) on media with or without erythromycin, according to the macroscopic aspect of colonies, Gram staining, catalase and oxydase. Counts were expressed as the number of colony forming units (CFU)/cm² of skin.

Staphylococcus colonies were examined by a fast screening test (Staphaurex, Biomérieux*; France) to rule out Staphylococcus aureus. All coagulase negative staphylococci underwent biochemical testing by using the identification gallery: ID32STA Biomerieux as well as testing for susceptibility to erythromycin (by diffusion on Mueller-Hinton agar) and each strain of a different species was reinoculated on tryptcase-soya (TS) agar (Oxoid*) then inoculated on preservation agar.

A volume of 0.1 ml of washings, pure and diluted in sterile normal saline (10^{- 1}, 10^{- 2}, 10^{- 3}, 10^{- 4}), was also inoculated by scraping on meat-yeast (MY) agars containing 2 mug/ml of furazolidone (Sigma) to inhibit the growth of staphylococci and on MY containing both 2 mug/ml of furazolidone and 4 mug/ml of erythromycin in order to count total anaerobic bacteria and erythromycin-resistant anaerobic bacteria, after 7 days incubation at 37° C in an anaerobic atmosphere.

A precise count of the different types of anaerobic bacteria present (Gram negative bacilli, Gram positive cocci, catalase+ or catalase- Gram positive bacilli) was done according to the macroscopic aspect of the colony, Gram staining and catalase. These counts were expressed as CFU/cm² of skin.

Colonies of Propionibacterium sp. underwent biochemical identification based on the following tests: indol + or -, nitrate reduction, esculine hydrolysis, saccharose and maltose fermentation as well as testing for sensitivity to erythromycin (by diffusion on Wilkins-Chalgren agar).

Study of minimum inhibitory concentrations (mic)

The minimum inhibitory concentration of erythromycin for Propionibacterium sp. and Staphylococcus sp. isolated from the skin of patients was determined by the micromethod of dilution in agar medium under the following conditions: the culture medium for staphylococci was Mueller-Hinton agar and that of propionibacteria Wilkins-Chalgren agar; bacterial inoculum was 10⁴ bacteria per 1 mul deposit; the dilution range for erythromycin was between 0.03 and 512 mug/ml; incubation conditions for culture of staphylococci were aerobic for 24 and 48 hours at 37° C, and for the culture of propionibacteria, anaerobic for 48 hours and 5 days at 37° C. A value of 512 mug/ml of erythromycin was adopted as the threshold of definition of resistance for Propionibacterium acnes and Staphylococcus epidermidis [9].

Minimum inhibitory concentration of zinc gluconate was studied at a dilution range of between 4 and 512 mug/ml. This range was determined on the basis of earlier findings [5]. For strains of Staphylococcus epidermidis and Propionibacterium acnes, the MIC was measured with the combination erythromycin + zinc at doses of 0.15, 1.5 and 15 mug/ml, in order to evaluate differences with erythromycin alone.

Statistical analysis

Statistical analysis was descriptive. Qualitative variables were expressed as numbers and percentage and quantitative variables as mean ± standard deviation, minimum and maximum.

Percentage resistance of Propionibacterium acnes and of Staphylococcus epidermidis was calculated for each patient taking into account only one strain if the same strain had been identified several times in the same patient.

SAS Windows 6.11 software was used for calculations.

Results

Study group

The mean age of these acne patients was 18.9 ± 6.1 (12.0-39.0) years, with 20 young men and 20 young women. Mean duration of acne was 4.4 ± 2.7 years (1.1-15.8 years). The severity and type of acne lesions in patients determined by counting were summarized in table I. Patients were classified as follows according to the two types of lesion (retentional or inflammatory): predominantly retentional acne (0 to 9 inflammatory lesions and between 10 and 40 retentional lesions), i.e. 15 patients; predominantly inflammatory acne (0 to 9 retentional lesion and between 10 and 40 inflammatory lesions), i.e. 6 patients; and mixed acne, i.e. 19 patients. Previous topical acne therapies used (except erythromycin) were: azelaic acid (n = 3), adapalene (n = 2), topical isotretinoin [3], benzoyl peroxide [18], tretinoin [9] and systemic acne therapies: dianette (n = 3), zinc gluconate (n = 20), isotretinoin (n = 16), tetracyclines (n = 23).

As far as treatment was concerned, 5/40 patients had received no previous topical or systemic treatment for their acne during the past year. Among the 35 patients treated previously and/or at the time of collection of microbiology samples, 4 (11.4%) had received only topical treatment, while 31 (88.6%) had received topical and/or systemic treatments.

Among all of the patients, 19 (19/40 = 47.5%) had never been given either topical or systemic erythromycin, while 11 (11/40 = 27.5%) were on erythromycin at the time of sampling, and 10 (11/40 = 27.5%) had received topical erythromycin during the previous year but were no longer on it at the time of sampling.

Microbiology

Frequency of isolation of organisms in acne patients (table II)

Sixteen bacterial species were isolated: the most common were : Staphylococcus epidermidis present in 38/40 (95%) of patients, Propionibacterium acnes in 36/40 (90%) of patients, Staphylococcus capitis in 47.5% of patients, Micrococcus in 47.5% of patients and Propionibacterium granulosum in 32.5% of patients.

Mean CFU of samples

For Propionibacterium acnes, the CFU/cm² count was 62.0 ± 165.8 (0.009-31).

For Staphylococcus epidermidis, the CFU/cm² count was 1.7 ± 2.9 (0.002-15.9).

Prevalence of bacterial resistance to erythromycin among patients

A proportion of 21/40 (52.5%) of patients had erythromycin-resistant strains of Propionibacterium acnes. 38/40 patients (95%) had erythromycin-resistant strains of Staphylococcus epidermidis. 19/40 (47.5%) of patients had erythromycin-resistant strains of both Staphylococcus epidermidis and Propionibacterium acnes

Among the 21 patients previously treated with erythromycin or on erythromycin at the time of sample, 13 (62%), including 6 on current treatment at the time of sampling, had erythromycin-resistant strains of Propionibacterium acnes and 19 (90%) had erythromycin-resistant strains of Staphylococcus epidermidis.

Among the 19 patients who had never been treated with erythromycin, 8 (42%) had erythromycin-resistant strains of Propionibacterium acnes and 19 (100%) erythromycin-resistant strains of Staphylococcus epidermidis.

Bacterial resistance to erythromycin and type of acne

1 - Percentage of patients with erythromycin-resistant strains according to the type of acne

With regard to Propionibacterium acnes, 40% of patients (6/15) with predominantly retentional acne, 53% of patients (10/19) with mixed acne, and 83% of patients (5/6) with predominantly inflammatory acne had erythromycin-resistant strains.

With regard to Staphylococcus epidermidis, 93% of patients (14/15) with predominantly retentional acne, 95% of patients (18/19) with mixed acne, and 100% of the six patients with predominantly inflammatory acne had erythromycin-resistant strains.

2 - Mean percentage of erythromycin-resistant strains of Propionibacterium acnes by patient according to type of acne (Fig. 1)

The mean percentage of resistant strains of Propionibacterium acnes by patient was 44.4% in the overall group, varying according to the type of acne: 18.1% for predominantly retentional acne, 52.8% for mixed acne, and 83.3% for predominantly inflammatory acne.

The mean percentage of resistant strains of Propionibacterium acnes by patient was also higher when patients had taken during the previous year or were still taking erythromycin, than in those who had not taken it during the previous year: 69% versus 19%.

3 - Mean percentage of erythromycin-resistant strains of Staphylococcus epidermidis by type of patient according to type of acne

The mean percentage of resistant strains by patient was 85.6% in the overall group, without any variation according to the type of acne: 83.7% for predominantly retentional acne, 82.3% for mixed acne, and 100% for predominantly inflammatory acne.

MIC of erythromycin for Propionibacterium acnes and Staphylococcus epidermidis (table III)

MIC of erythromycin for Propionibacterium acnes were determined for 37 strains. Among these strains, 16/37 (43.2%) were susceptible to erythromycin (MIC < 512 mug/ml), while 21/37 (56.8%) were resistant (MIC >= 512 mug/ml).

MIC of Staphylococcus epidermidis were determined in 49 strains. Among these, 14/49 (28.6%) were susceptible to erythromycin (MIC < 512 mug/ml), while 35/49 (71.4%) were resistant (MIC >= 512 mug/ml).

MIC of zinc for Propionibacterium acnes and Staphylococcus epidermidis (table IV)

MIC of zinc were determined for 37 strains of Propionibacterium acnes and 49 strains of Staphylococcus epidermidis. Results are summarized in table IV.

MIC of erythromycin for Propionibacterium acnes and Staphylococcus epidermidis in the presence of zinc

MIC of erythromycin for 37 Propionibacterium acnes strains previously determined between 0.03 and 512 mug/ml were unaffected by the addition of zinc at concentrations of 0.15 mug/ml, 1.5 mug/ml and 15 mug/ml. The same results were noted with Staphylococcus epidermidis.

Discussion

This pilot study aimed to determine the frequency of propionibacterium acnes and staphylococcus epidermidis resistant strains to erythromycin. This is the first French study on bacterial resistance to erythromycin performed in acne patients. Samples were collected using the Williamson and Kligman washings method [7], modified and standardized by Fleurette [8]. The small electrical device (diameter = 2 cm) is easy to apply to the three main regions of the face (forehead - cheeks - chin) where acne lesions are found. Sampling time is short (20 seconds) and causes neither trauma nor pain. It enables the easy isolation of Propionibacterium acnes in the pilo-sebaceous follicle.

Our results confirm that the bacterial flora in acne consisted of 16 main types of different organisms with a large predominance of Staphylococcus epidermidis and Propionibacterium acnes, found in 95 and 90% of patients respectively. Furthermore, among the 16 types of organisms, the three others encountered most frequently were Staphylococcus capitis (47.5%), Micrococcus (47.5%), and Propionibacterium granulosum (32.5%).

CFU varied widely according to patients. This could be explained by the fact that our study group included non-treated (5 patients) and treated (35 patients) individuals.

In this study, 95% of patients are carriers of erythromycin-resistant strains of Staphylococcus epidermidis. These results are higher than those reported by Nishijima [3] in 1994 in Japan with a percentage of 61%, similar to those of Miller [10] in 1996 with a percentage of 87% and Forssman [4] in 1995 with 100%. At the same time, in our study, 52% of patients are carriers of resistant strains of Propionibacterium acnes.

First publications concerning resistant Propionibacterium acnes appeared in the literature in 1979, with a 20% rate of patients carrying resistant strains [11]. This rate has risen inexorably in subsequent publications: 26% in 1993 by Eady [2], 44% by Forssman [4] in 1995 and 49% by Eady [12] in 1996. Our results with 52% of Propionibaterium acnes resistant strains to erythromycin are similar to those of English and American authors.

In addition, our study shows that the use of previous or current treatment with erythromycin does not influence the frequency of resistant strains of Staphylococcus epidermidis but that of Propionibacterium acnes. 62% of patients, who had been treated with erythromycin during the previous year have resistant strains of Propionibacterium acnes versus 42% in the other group. Several hypotheses can be suggested to explain the presence of resistant strains in the absence of any earlier use of erythromycin. Firstly, the patient might obviously have forgotten. Secondly, it could also be a passive transfer of resistant strains, notably via a physician or a member of the family with acne. However studies have also shown that the acquisition of resistant strains of Propionibacterium acnes was associated with mutations in the peptyl transferase region of ribosomal subunit 23S inducing crossover resistance with clindamycin, macrolide and synergistins [9], antibiotics often used in children. Resistant strains seem to appear on average after 12 to 24 weeks of treatment [13].

This study also revealed an association between the type of acne and the presence of resistant strains of Propionibacterium acnes. Resistant strains of Propionibacterium acnes are found in 40% of cases of retentional acne, 53% of cases classified mixed, and 83% of cases of inflammatory acne. Among the 13 patients carrying Propionibacterium acnes and Staphylococcus epidermidis resistant strains to erythromycin, 11 have inflammatory (5 patients) or mixed (6 patients) acne. No patient had a retentional acne. On the basis of the results of this study, it would seem that a predominance of inflammatory lesions is more often associated with the presence of resistant strains of Propionibacterium acnes, this without obvious link to any previous treatment or not with erythromycin. Thus, faced with an acne patient with predominant inflammatory lesions, the dermatologist has to suspect resistance to antibiotic therapy. However, our results have to be confirmed with a larger study.

Study of MIC for 37 strains of Propionibacterium acnes revealed 57% of strains with an MIC of 512 mug/ml or more. Study of MIC of zinc confirms the inhibitory action of the latter on Propionibacterium acnes at concentrations varying from 256 to 512 mug/ml. One hundred per cent of Propionibacterium acnes strains were inhibited at the concentration of 512 mug/ml zinc. However, zinc only has an inhibitory action on 61% of Staphylococcus epidermidis strains at concentrations varying between 128 and >= 512 mug/ml. The addition of zinc to culture medium at concentrations between 0.15 and 15 mug/ml does not induce a synergic effect with erythromycin on erythromycin MIC (MIC < 0.03 to MIC > 512 mug/ml). However, Holland [5] using concentrations of zinc and erythromycin twenty- and two-fold greater respectively, showed that the growth of erythromycin-resistant strains of Propionibacterium acnes was inhibited by the addition of 300 mug/ml of zinc to 1,000 mug/ml of erythromycin. In vivo, while Bojar [14] found no additional efficacy of zinc used with erythromycin on inflammatory acne lesions, Habbema [15], Schachner [16], and Feucht [17] showed that the combination of zinc and erythromycin was more effective than erythromycin only on the course of inflammatory acne lesions. These in vivo and in vitro studies suggests that zinc at specific concentrations used together with erythromycin could restrict the development of strains resistant to Propionibacterium acnes. These hypotheses nevertheless require clinical confirmation.

CONCLUSION

In conclusion, this study shows that in France the percentage of strains of Propionibacterium acnes resistant to erythromycin is similar to that reported in Anglo-Saxon countries. It shows that these resistant strains are more frequent in patients with predominantly inflammatory lesions, raising questions for the clinician regarding the risk of the long-term use of erythromycin in acne with predominant inflammatory lesions and the interest of combining erythromycin with zinc in order to reduce this risk of onset of resistant strains, this latter point requiring confirmation.

REFERENCES

1. Leyden JJ. Antibiotic resistant acne. Cutis 1976; 17: 593-6.

2. Eady EA, Jones CE, Tipper JL, et al. Antibiotic resistant propionibacteria in acne : need for policies to modify antibiotic usage. Brit Med J 1993; 306: 555-6.

3. Nishijima S, Akamatsu H, Akamatsu M, et al. The antibiotic susceptibility of Propionibacterium acnes and Staphylococcus epidermidis isolated from acne. J Dermatol 1994; 21: 166-71.

4. Forssman T. Antibiotic resistance in acne patients under antibiotic treatment in comparison to an untreated control group with retrospective assessment of therapy. Surber Cancer, Elsner P, Bircher AJ, eds. Exogenous dermatology. Basel: Karger. Curr Prob Dermatol 1995; 22: 91-7.

5. Holland KT, Bojar RA, Cunliffe WJ, Cutcliffe AG, et al. The effect of zinc and erythromycin on the growth of erythromycin-resistant and erythromycin-sensitive isolates of Propionibacterium acnes: an in vitro study. Br J Dermatol 1992; 126: 505-9.

6. Burke BM, Cunliffe WJ. The assessment of acne vulgaris. The Leeds technique. Br J Dermatol 1984; III: 83-92.

7. Williamson P, Kligman A. A new method for the quantitative investigation of cutaneous bacteria. J Invest Dermatol 1965; 45: 498-503.

8. Fleurette J, Transy MJ. Essai d'amélioration et de standardisation du prélèvement des microorganismes cutanés au moyen d'un appareil électrique. Rev Inst Pasteur de Lyon 1978; 11: 493-501.

9. Ross JI, Eady EA, Cove JH, Ratyal AH, Cunliffe WJ. Resistance to erythromycin and clindamycin in cutaneous propionibacteria is associated with mutations in 23S rRNA. Dermatol 1998; 169: 69-70.

10. Miller YW, Eady A, Lacey RW, Cove JH, et al. Sequential antibiotic therapy for acne promotes the carriage of resistant staphylococci on the skin of contacts. J Antimicrob Chemother 1996; 38: 829-37.

11. Crawford WW, Crawford IP, Stoughton RB, Cornell RC. Laboratory induction and clinical occurrence of combined clindamycin and erythromycin resistance in Corynebacterium acnes. J Invest Dermatol 1979; 4: 187-90.

12. Eady EA, Bojar RA, Jones CE, Cove JH, et al. The effects of acne treatment with a combination of benzoyl peroxide and erythromycin on skin carriage of erythromycin-resistant propionibacteria. Br J Dermatol 1996; 134: 107-13.

13. Ross JI, Eady EA, Cove JH, Jones CE, et al. Clinical resistance to erythromycin and clindamycin in cutaneous propionibacteria isolated from acne patients is associated with mutations in 23S rRNA. Antimicrobial agents and chemotherapy 1997; 41: 1162-5.

14. Bojar RA, Eady EA, Jones CE, Cunliffe WJ, et al. Inhibition of erythromycin-resistant propionibacteria on the skin of acne patients by topical erythromycin with and without zinc. Br J Dermatol 1994; 130: 329-36.

15. Habbema L, Koopmans B, Menke HE, Doornweerd S, et al. A 4% erythromycin and zinc combination (Zineryt) versus 2% erythromycin (Eryderm) in acne vulgaris: a randomized, double-blind comparative study. Br J Dermatol 1989; 4: 497-502.

16. Schachner L, Pestana A, Kittles C. A clinical trial comparing the safety and efficacy of a topical erythromycin-zinc formulation with a topical clindamycin formulation. J Am Acad Dermatol 1990; 3: 489-95.

17. Feucht CL, Allen BS, Chalker DK, Smith JG. Topical erythromycin with zinc in acne. A double-blind controlled study. J Am Acad Dermatol 1980; 3: 483-91.