Correlation between Antimicrobial Resistant Propionibacterium acnes and Severity of Acne Vulgaris: Findings from a Tertiary Hospital

Abstract

The study of antimicrobial-resistant Propionibacterium acnes was not conducted regularly, especially in Indonesia. Conversely, regular monitoring of antibiotic efficacy through in vitro testing to assess the evolution of current resistance patterns is obligated; thus, filling the gap caused by a lack of appropriate antibiotic surveillance is required. Analyse the correlation between resistance patterns of P. acnes to doxycycline, clindamycin, erythromycin and azithromycin with the severity of acne vulgaris. This is an analytic observational laboratory study with a cross-sectional design of mild to severe acne vulgaris (AV) patients. Specimens were obtained from comedones of 71 patients, which were cultured and identified using biochemical examination. Antimicrobial resistance (doxycycline, clindamycin, erythromycin and azithromycin) to P. acnes was tested by disc diffusion method. Among 71 samples collected, 40 (56.3%) P. acnes isolates were cultured and identified. The incidence of P. acnes resistance to more than one antimicrobial was 45%. Antimicrobial resistances were clindamycin 42.5%, erythromycin 40%, azithromycin 23.5% and doxycycline 12.5%, respectively. According to the contingency coefficient test, there was moderate correlation between the resistance pattern of P. acnes to clindamycin (r = 0.485, P = <0.001) and doxycycline (r = 0.433, P = 0.002) and AV severity. There was weak correlation between the resistance pattern of P. acnes to erythromycin (r = 0.333; P = 0.025) and azithromycin (r = 0.321; P = 0.032) and AV severity. In conclusion, there is a correlation between the pattern of P. acnes resistance to doxycycline, clindamycin, erythromycin, azithromycin and severity of AV.

Introduction

Acne vulgaris (AV) is a chronic inflammatory condition affecting pilosebaceous follicles characterised by polymorphic lesions in the form of comedones, papules, pustules, nodules and cysts mainly located on the face, neck, upper back, shoulders and upper arms.[1,2] It is regarded as one of the most common skin disorders globally, affecting primarily adolescents and young adults (aged 12–25 years), with an 85% prevalence.[3]

Acne is multifaceted and complex in its pathogenesis, with the presence and activity of Propionibacterium acnes serving as a determinant of acne development.[4] Antibiotic medication, either topical (e.g., clindamycin, erythromycin and tetracycline) or systemic/oral (e.g., doxycycline, tetracycline, erythromycin, azithromycin and trimethoprim-sulfamethoxazole), is occasionally used to treat this disorder, regarding the presence of P. acnes.[5] The use of antibiotics may be beneficial due to their antibacterial and anti-inflammatory properties, particularly in moderate to severe inflammatory acne (systemic antibiotics).[6,7] However, it should not be administered as acne monotherapy due to the possibility of resistance development.[6,8]

Antibiotic use, particularly long-term use, has been demonstrated to increase the prevalence of resistant P. acnes trait.[9] The rise in resistance rates is quite concerning, with 34.5% cases in 1991 escalating to 64% in 1997.[10] In 29% of the cases, P. acnes strains were resistant to at least one antibiotic regimen. Clindamycin resistance was 15.1%, erythromycin resistance reached 17.1% and tetracycline resistance was 2.6%.[11]

Multiple studies have found an association between P. acnes antibiotic resistance and clinical response to AV therapy, particularly a reduced response or immediate recurrence after medication.[12,13] Previous research found that resistant P. acnes strains were associated with AV severity, correlating with a greater incidence of resistant strains in severe AV patients than those with moderate AV.[14] Furthermore, some AV patients who were naive to antibiotics contained resistant P. acnes strains as a result of mutations or the transfer of resistant genes.[15]

Regular monitoring of antibiotic efficacy by in vitro testing to evaluate the evolution of current resistance patterns is obligated due to the widespread use of routine and long-term antibiotics in several AV patients.[16] Furthermore, information regarding P. acnes antibiotic resistance in AV patients in Indonesia is limited. Therefore, the purpose of this study was to investigate the correlation between P. acnes antibiotic resistance patterns and the severity of acne vulgaris.

Materials and Methods

Subject

This is a descriptive-analytic laboratory study carried out in the dermatology clinic of Dr. Mohammad Hoesin General Hospital, Palembang, an Indonesian tertiary hospital. Acne vulgaris patients who visited the clinic between December 2020 and March 2021 were recruited. We were assessing the demographic status and basic clinical information such as age, gender, family history of acne, onset and duration of acne, type of treatment and length of treatment. Inclusion criteria were as follows: (1) 12–40-year-old patients; (2) mild to severe AV based on Lehmann criteria (recommendation from Indonesian Acne Expert Meeting 2015); and (3) no oral or topical antibiotic use in the previous month. Meanwhile, individuals who had used oral isotretinoin in the preceding month, had a facial infection, and/or had non-P. acnes culture specimens were excluded. Before commencing this study, all patients provided informed consent. If the patients were under the age of 18, their parents gave them their permission.

Sample calculation

Sample calculation was done using the following formula:

Note:

n = Sample size

Zα = Standard normal deviate for α = 0,05 (two-tailed), Zα = 1,96

Zβ = Standard normal deviate for β = 0,1 (two-tailed), Zβ = 1,64

r = Correlation coefficient (0.5)

Thus, minimum required sample is:

= 46,2

= 47 samples (total) + 10% (tolerance value)

= 52 samples

Specimen collection, culture and identification processing of P. acnes

There were 71 participants in this study with mild to severe acne. The samples were taken under aseptic standards from comedones on the face using a sterilised comedone extractor. They were transported to the microbiology laboratory in a sterile anaerobic tube transport medium (Thermo Scientific^™, Oxoid 3.5 L Anaerobic Jar, HP0011A, UK). The samples were inoculated into Brucella and thioglycolate agar mediums and incubated anaerobically at 35^o C for three–five days. Propionibacterium acnes was identified using conventional techniques such as morphology, Gram staining, catalase, indole, nitrate and gelatine. In order to make a definitive identification, we used pure P. acnes ATCC® 11827 (Microbiologics^®, Oslo, USA) as a control strain. P. acnes strains were successfully isolated from 40 of the 71 samples tested. Participants' skin was also checked for sebum levels using the Sebumeter^® SM815.

Antibiotic susceptibility testing and Minimum inhibitory concentration (MIC) determination

The susceptibility test of P. acnes strain was done in the Health Laboratory Centre for Palembang. The test was based on Kirby–Bauer methods using the Mueller–Hinton agar. A sterile swab was used to spread the sample uniformly on the surface of the medium. Antibiotic discs (Oxoid Limited, Basingstoke, UK) for each antibiotic (doxycycline 30 μg, clindamycin 2 μg, erythromycin 15 μg, and azithromycin 15 μg) were placed onto the inoculated medium and incubated at 37°C for 24 h. Diameter of the zone of inhibition around each antibiotic disc was recorded in millimeters to determine the sensitivity or resistant categories according to the Clinical and Laboratory Standards Institute standard.[17]

Statistical analysis

Descriptive analysis was used to determine the proportion of study participants based on covariables (age, gender, education, occupation, family history, body mass index (BMI), AV duration, duration of use of oral and/or topical antibiotics and sebum levels), independent variables (pattern of P. acnes resistance to antibiotics) and the dependent variable (AV severity). The researchers used the contingency coefficient correlation methods to measure P. acnes resistance pattern's correlation with the severity of AV. In the correlation investigation, AV severity was divided into two categories (mild-moderate and severe). In addition, to calculate the Odds ratio (OR), additional data analysis was performed in the form of a Pearson Chi-square or Fisher exact comparison test. The Statistical Package for Social Science (SPSS) version 22.0 was used to analyse the data (SPSS, Inc., Chicago, Illinois).

Results

Demographics

Our sample comprised of 71 participants, but only 40 were included in the final analysis due to no P. acnes growth in the remaining sample. The study sample's mean age was 22.68 ± 4.93 years, ranging from 14 to 35 years. The largest age group is 20–24 years, with 17 people (42.5%). Meanwhile, women outnumber men in this study, with a female-to-male ratio of 1.35:1. There was no association found between demographic characteristics and acne severity. Demographic characteristics are depicted in Table 1.

Table 1.

Demographic characteristics (n=40)

Characteristics	Acne severity			P*
	Mild	Moderate	Severe
Age groups
10–14 years	0	1 (2.50%)	0	0.970
15–19 years	2 (5.00%)	4 (10.00%)	3 (7.50%)
20–24 years	3 (7.50%)	10 (25.00%)	4 (10.00%)
≥25 years	2 (5.00%)	8 (20.00%)	3 (7.50%)
Gender
Male	4 (10.00%)	7 (17.50%)	6 (15.00%)	0.191
Female	3 (7.50%)	16 (40.00%)	4 (10.00%)

*Fisher exact

Clinical characteristics

The researcher discovered from the history taking that majority of the study samples, consisting of 29 people, had a family history of AV (72.5%). Meanwhile, in terms of medication history, most patients (80%) had attempted to treat their condition for an average of less than 12 weeks. Furthermore, with a percentage of 80%, the most considerable proportion of the study samples had a normal BMI.

When sebum levels were examined, 28 patients (70%) had skin with normal oil levels, the most compared to other groups. In this study, the mean value of sebum levels was 86.37 ± 36.71 g/cm2. According to an analysis of AV severity using Lehmann's criteria, this study's sample, which includes 23 people, had a moderate degree of AV (57.5%). There was no association between clinical characteristics and AV severity, similar to previously presented demographic characteristics data. Table 2 displays the clinical characteristics data.

Table 2.

Clinical characteristics (n=40)

Characteristics	Acne severity			P*
	Mild	Moderate	Severe
Family history
Yes	6 (15.00%)	15 (37.50%)	8 (20.00%)	0.634
No	1 (2.50%)	8 (20.00%)	2 (5.00%)
Treatment history
Yes	7 (17.50%)	17 (42.50%)	8 (20.00%)	0.632
No	0	6 (15.00%)	2 (5.00%)
Body mass index
Underweight	0	1 (2.50%)	2 (5.00%)	0.428
Normal	7 (17.50%)	17 (42.50%)	8 (20.00%)
Overweight	0	4 (10.00%)	0
Obese	0	1 (2.50%)	0
Sebum level
Dry	3 (7.50%)	5 (12.50%)	2 (5.00%)	0.168
Normal	4 (10.00%)	18 (45.00%)	6 (15.00%)
Oily	0	0	2 (5.00%)
Treatment length
≥12 weeks	1 (2.50%)	8 (20.00%)	4 (10.00%)	0.337
<12 weeks	6 (15.00%)	9 (22.50%)	4 (10.00%)
Not started	0	6 (15.00%)	2 (5.00%)

*Fisher exact

Resistance pattern

A total of 40 P. acnes isolates that were successfully cultured and identified were tested for antibiotic resistance using the disc diffusion method. Clindamycin had the highest resistance (42.5%), followed by erythromycin (40%), azithromycin (23.5%) and doxycycline (12.5%). Four isolates (10%) were resistant to all antibiotics tested. Meanwhile, 22 (55%) isolates were still sensitive to all antibiotics, reflecting 45% of the resistance pattern in the studied sample (data not shown). Table 3 displays the data.

Table 3.

Antibiotic resistance pattern in the study sample (n=40)

Antibiotic resistance testing	Doxycycline	Azithromycin	Clindamycin	Erythromycin
Sensitive	35 (87.50%)	27 (67.50%)	23 (57.50%)	24 (60.00%)
Resistant	5 (12.50%)	13 (32.50%)	17 (42.50%)	16 (40.00%)

This study explored the correlation between P. acnes' antibiotic resistance pattern and the severity of AV [Table 4]. In this study, we use the contingency coefficient method to measure the correlation between P. acnes resistance pattern to doxycycline, clindamycin, erythromycin and azithromycin and the severity of AV. We combined the mild and moderate AV groups to improve statistical interpretation without compromising clinical significance.

Table 4.

Correlation of resistance pattern and acne severity

Parameters		Acne severity				OR	CI 95%		r*	P**
		Mild-Moderate n=30		Severe n=10			Lower	Upper
		n	%	n	%
Doxycycline	Sensitive	29	96.7	6	60	19.333	1.824	204.972	0.433	0.002
	Resistant	1	3.3	4	40
Clindamycin	Sensitive	22	73.3	1	10	24.750	2.691	227.609	0.485	<0.001
	Resistant	8	26.7	9	90
Erythromycin	Sensitive	21	70	3	30	5.444	1.142	25.954	0.333	0.025
	Resistant	9	30	7	70
Azithromycin	Sensitive	23	76.7	4	40	4.929	1.076	22.579	0.321	0.032
	Resistant	7	23,3	6	60

*(r) it is not usually needed, **P<0.05.

The results revealed a moderate correlation between P. acnes resistance to doxycycline (r = 0.433, P = 0.002) and clindamycin (r = 0.485, P = 0.001) with the severity of AV. On the other hand, there is a weak correlation between P. acnes resistance to erythromycin (r = 0.333; P = 0.025) and azithromycin (r = 0.321; P = 0.032) and the severity of AV [Figure 1].

Figure 1.

Acne Severity and Antibiotic Resistance Correlation. (a) Doxycycline, (b) Azithromycin, (c) Clindamycin, (d) Erythromycin

Discussion

This is a cross-sectional analytical observational laboratory-based study in AV patients at the dermatovenereology clinic in a tertiary-level hospital. This study collected comedone specimens from AV patients, cultured them and tested for P. acnes resistance to four types of antibiotics. Only 40 of the 71 AV patient samples grew P. acnes colonies.

According to the demographic data of the research sample characteristics, the age range of the research sample is 14–35 years. The study sample's average age was 22.68 ± 4.93 years. The most significant age group proportion (54.9%) was 20–24 years old, and the smallest was 10–14 years old (1.4%). This is related to the hormone DHEA, which regulates sebaceous gland activity and begins to rise during puberty.[18] This is consistent with a previous retrospective study conducted in Italy, which discovered that 1167 AV patients predominantly lay between 12 and 25 years (58.7%).[19] There was no significant relationship between age and AV severity in this study (P = 0.878). In contrast, Bagatin et al. found that young adult AV patients were more likely to have severe AV.[20]

The majority of the samples in this study (59.2%) were female, with a ratio of 1.35:1. It is similar to a population-based study in Taiwan, with the proportion of AV in women: men at 1.81:1.[21] In addition, we did not find a significant relationship between gender and the severity of AV (P = 0.306). A similar study found an insignificant relationship between gender and AV severity in Brazil.[20]

A family history of AV in parents or siblings was found in 70.4% of the study participants. Previously, some studies discovered gene polymorphisms related to AV development, such as HSD3B1, HSD17B3 and CYP1A1.[22,23] Although genetic influence is suspected in AV, little genomic analysis has been performed.[24]

The majority of the study sample had an average BMI in this study. The case-control study conducted by Lu et al. showed higher BMI of patients with moderate to severe AV than in the control group.[25] This study found no significant relationship between BMI and AV severity (P = 0.065). On the contrary, a systematic review found that overweight/obesity status significantly affects AV severity with OR 2,36.[26]

Examination of sebum levels was carried out on research samples to determine skin type. The majority of the participants in this study (70.4%) had normal skin. There was no statistically significant relationship between sebum levels and AV severity (P = 0.138). Meanwhile, previous research has shown that patients with oily or mixed skin types are more likely to develop AV than normal/dry skin types.[27]

The study sample was dominated by patients with moderate AV. This is similar to Sutrisno et al. research in Indonesia.[28] Another study in China, however, found that the prevalence of mild AV was higher than that of the other AV group.[29] This disparity in the proportion of AV based on severity may be due to differences in demographics between studies, as well as the possibility that moderate AV patients already have cosmetic disorders for which they seek treatment.[30]

The vast majority of patients have tried to treat their AV condition. However, there was no significant relationship between the history of antimicrobial treatment in AV cases and the severity of the disease (P = 0.701) in this study. The study by Bagatin et al., on the other hand, found a significant relationship between medication history and the incidence of comedonal AV.[20] Previous treatment history, particularly the overuse/misuse of antimicrobials, can result in resistance and a reduced response to therapy.[31,32]

According to the findings of this study, 40 isolates (56.3%) of P. acnes were successfully grown from 7 (17.5%) mild AV, 23 (57.5%) moderate AV and 10 (25%) severe AV. This positivity rate corresponds to the range of values reported in previous studies, ranging from 11 to 96%.[33,34,35] Several factors influence the positivity level of P. acnes in various studies, including P. acnes density, specimen collection location, patient age and P. acnes biofilm formation.[36]

Antibiotic resistance in P. acnes is a problem in many countries with different antibiotic resistance patterns. Clindamycin resistance was found to predominate in Malaysia and Egypt (15.1% and 66.3%, respectively),[37,38] while resistance to both clindamycin (30%) and erythromycin (26.7%) was found in Korea.[39] The proportion of AV patients infected with antibiotic-resistant P. acnes strains increases globally and varies by region. This could be due to differences in antibiotic prescribing or patient ethnicity.[40] Countries that frequently use oral antibiotics to treat AV have high rates of P. acnes resistance[41] when antibiotic use is not prohibited.[42]

The researchers discovered a moderate correlation between P. acnes resistance to doxycycline (r = 0.433, P = 0.002) and clindamycin (r = 0.485, P = 0.001) and the severity of AV. The OR of 19.3 indicates the likelihood of severe AV in the doxycycline-resistant group and vice versa. Meanwhile, the OR value for the clindamycin-resistant group was 24.75, indicating a high risk of resistance in the severe AV group. The high incidence of clindamycin resistance in severe AV patients may prompt clinicians to use caution when administering these drugs to patients, especially those experiencing severe AV.[43]

On the other hand, a weak correlation was found between P. acne's resistance pattern to erythromycin (r = 0.333, P = 0.025) and azithromycin (r = 0.321, P = 0.032) and AV severity. As a result, clinicians should pay closer attention to the AV management algorithm, which specifies that erythromycin is preferred for pregnant and lactating women and azithromycin is only recommended for severe AV.[44,45] In general, the findings of this study are similar to Nakase et al., who discovered that P. acnes isolates are resistant to macrolide-lincosamide-streptogramin B (MLSB) antibiotics were more common in severe AV (40%) than mild AV (18.8%). However, there was no significant relationship (P = 0.193).[14]

The study's limitations are the relatively small sample size of this study and the fact that it is not a multicentre study. Because the resistance test was only based on the population in the study, it could not describe the entire population. A larger sample size is required for the research results to be generalisable.

Conclusion

There is a correlation between the antimicrobial resistance against P. acnes and AV severity (moderate strength: doxycycline and clindamycin; weak strength: erythromycin and azithromycin). The highest resistance incidence is witnessed on doxycycline, while the lowest is for doxycycline.

Ethical approval

The Health Research Ethics Committee of Dr. Moh. Hoesin Hospital Palembang authorised our protocol (approval number: 83/kepkrsmh/2020).

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.