DOI: 10.1111/exd.12793
Review
www.wileyonlinelibrary.com/journal/EXD
The bacteriology of hidradenitis suppurativa: a systematic review Hans Christian Ring1, Peter Riis Mikkelsen1, Iben Marie Miller1, H avard Jenssen2, Kurt Fuursted3, 1 1 Ditte Marie Saunte and Gregor B.E. Jemec 1 Department of Dermatology, Roskilde Hospital, University of Copenhagen, Roskilde, Denmark; 2Department of Science, Systems and Models, Roskilde University, Roskilde, Denmark; 3Department of bacteriology, Statens Serum Institut, Copenhagen, Denmark Correspondence: Hans Christian Ring, MD, Department of Dermatology, Roskilde Hospital, Køgevej 7-13, Roskilde 4000, Denmark, e-mail: [email protected]
Abstract: Hidradenitis suppurativa (HS) is a chronic inflammatory disabling skin disease consisting of recurrent nodules, sinuses, fistulas and scarring involving the intertriginous regions. HS is often a therapeutic challenge and most treatments are off-label. A better understanding of aetiology and pathogenesis of HS may facilitate the development of effective treatment. Although the clinical presentation is strongly reminiscent of bacterial infection, the role of bacteria remains controversial. Studies have isolated an array of different bacteria specimens. Consistent findings of Gram-positive cocci and Gram-positive rods including Staphylococus aureus, coagulase-negative staphylococci (CoNS) and Corynebacterium species in deep tissue samples have been demonstrated in HS and may constitute a central target for the immune system. Efficacy of antibiotics, that is rifampicin, clindamycin or tetracycline, supports a microbial role in disease pathogenesis. However, these antibiotics also work
as immunomodulators of especially T cells, and the underlying mechanisms may therefore be more complex. We performed a systematic review of previous studies investigating the bacterial flora in hidradenitis suppurativa. We searched PubMed, EMBASE, Royal Danish Library and Cochrane library (search date 11 December 2014). A total of 66 papers were identified and nine papers published between 1988 and 2014 matched our inclusion criteria, yielding bacteriological data of a total of 324 patients with HS (mean age 36.8 years and female/male ratio 215/109). This overview of the bacteriology may aid researchers and physicians exploring the potential role of bacteria in HS. Furthermore, to stimulate a broader debate, we also present different viewpoints on the possible role of bacteria in HS.
Introduction
patients are often treated with antibiotics, for example penicillins or, incision and drainage by non-dermatologists, even though the treatments are regularly ineffective. Studies have isolated an array of different bacteria specimens (16–19). Consistent findings of Gram-positive cocci and Grampositive rods including Staphylococus aureus, coagulase-negative staphylococci (CoNS) and Corynebacterium species in deep tissue samples have been found in HS (18,19) and may constitute a central target for the immune system, but are also part of the normal human skin flora. The efficacy of antibiotics, that is rifampicin, clindamycin or tetracycline, in HS treatment further supports a microbial role in disease pathogenesis (20,21). However, these antibiotics also work as immunomodulators of especially T cells (20,21), and the underlying mechanisms may therefore be more complex. It is unclear whether bacterial colonization is a primary or secondary event in the evolution of HS lesions. It has been speculated that bacterial invasion leads to series of pathogen-associated molecular pathways, which may trigger the initiation of inflammasomes (22). Alternatively, the deposition of keratin fragments into the dermis in a genetically susceptible individual may be the trigger, and finally, it is speculated that the deposited keratin fragments may be colonized with bacteria (23). Our objective was to investigate the previous body of literature concerning the bacteriology in HS, thereby providing an overview of the field. This may aid researchers and physicians in understanding the potential role of bacteria in HS.
Hidradenitis suppurativa (HS) is a chronic inflammatory disabling skin disease defined by recurrent nodules, sinuses, fistulas and scarring involving the intertriginous regions. The nodules can progressively expand to abscesses, and subsequently rupture causing suppuration (1). Although the prevalence of HS has been estimated up to 4% in European populations (2,3), most treatments are off-label. The development of effective treatment may be facilitated by a better understanding of aetiology and pathogenesis of HS. Factors such as genetics, smoking and metabolic syndrome appear to be strongly associated with HS (4–8), and several hypotheses on the pathogenesis have been discussed. The main theory is hair follicle plugging, and furthermore in contrast to acne, the sebaceous glands appears reduced or absent in HS (9). Histologically, HS appears to be characterized by a hyperkeratinization of the hair follicle with an immediate hair follicle occlusion and subsequent follicular dilatation and perifollicular inflammation followed by cyst formation (10). Increasing evidence suggest that dysfunctional immune responses play a key role involving both the innate and adaptive immune system (11,12). This is in agreement with recent studies showing efficacy of immunomodulatory agents such as oral corticosteroids, cyclosporine (13), and more recently antitumor necrosis factor therapy implying that TNF-a is involved (13–15). The role of bacteria, however, remains controversial. The clinical presentation is strongly reminiscent of bacterial infection, and
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
Key words: bacteria – hidradenitis suppurativa – microbiology
Accepted for publication 28 May 2015
727
Ring et al.
Method We performed a systematic review of previous studies investigating the bacterial flora in hidradenitis suppurativa. We searched PubMed, EMBASE, Royal Danish Library and Cochrane library (search date 11 December 2014). We used two search strategies combining the MeSH terms “bacteriology AND acne inversa” and “bacteriology AND hidradenitis suppurativa”. Furthermore, the reference lists from the included articles were manually searched for additional relevant studies. The database search was undertaken independently by two reviewers for each database. The titles and abstracts of articles retrieved by the search strategy were reviewed independently by two of the authors. Retrospective and prospective cross-sectional studies exclusively investigating the bacterial presence of bacteria in tissue or blood of patients with HS were included. Only studies in English were included. Case reports or case series including less than 10 patients were excluded from the review (Fig. 1). Despite methodological differences among the approaches used by the included studies, we have constructed a table (Table. S1) displaying the number of patients/lesions, methods (sampling, culturing, molecular or identification techniques) of each study. The fractional distribution of the different bacterial species identified among the included studies is illustrated (Fig. 2). Furthermore, the prevalence rates of S. aureus in the different studies are displayed separately in Table 1.
Results Included studies in the review A total of 66 papers were identified and nine papers published between 1988 and 2014 matched the above-mentioned inclusion criteria (Fig. 1), yielding bacteriological data of a total of 324 patients with HS (mean age 36.8 years and female/male ratio 215/ 109).
Microbiological methods used in the included studies As shown in Table S1, traditional microbiological techniques for culturing bacteria in aerobic and anaerobic environment have been applied to investigate the bacteriology of HS in eight of nine included studies. Three studies used biochemical tests and only one study used immunological identification technique (Lancefield group antigens). One study applied 16S gene ribosomal RNA gene
Figure 2. Fractional distribution of various bacteria found in nine studies. Swabs, biopsies, blood and needle aspiration of purulent content are included in the figure. Bacteria that were not clearly labelled as any of the bacteria listed above were not included in this presentation; for example, Anaerobe Streptococci, Microaerophilic Streptococci, anaerobe Gram-positive Cocci and lactobacilis are classified as other.
sequencing along with metagenomics. Lastly, only one study investigated the bacteriology in blood samples in patients with HS using PCR (polymerase chain reaction) and DNA sequencing and analysis (24).
The Spectrum of Bacteria found in patients with HS As displayed in Fig. 2, CoNS (34.1%) and mixed anaerobic bacteria (23.3%) are the most frequently encountered types of bacteria among the 324 patients. In addition, S. aureus also constitutes a considerable part of the overall isolated bacterial specimens (isolated in six of nine studies), and among the studies that identified S. aureus, the prevalence rate ranged from 13% to 56% (Table 1).
Discussion The studies included in this review demonstrate a high prevalence of S. aureus and CoNS with a considerable variety of other species isolated from the inflamed lesions of patients with HS. However, the results also reflect retrospective studies which may cause a relatively lower identification rate of anaerobic bacteria. These bacterial species may therefore have been underestimated. In addition, bacterial identification methods varied significantly in the period of 1988–2014 (from biochemical tests to the use of MALDI-TOF (matrix-assisted laser desorption/ionization) mass spectrometry and molecular methods), and of note, researchers were first able to identify Actinomyces spp. after the late 1990s. Looking at the literature in general, there appears to be no consistency with regard to the occurrence of any specific type of
Table 1. Prevalence rate of S. aureus in each study Studies
Prevalence %
Lapins et al. (18) Matusiak et al. (28) Guet-Revillet et al. (16) Highet et al. (49) Sartorius et al. (19) Brook and Frazier (26) Sartorius et al. (24) Jemec et al. (17) O’Loughlin S et al. (50)
56 13.6 131 43.75 Not identified 35.29 Not identified 19.51 Not identified
13% displays the prevalence of S. aureus among 125 investigated HS lesions.
1
Figure 1. Flow chart of search results showing included and excluded references.
728
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
The bacteriology of hidradenitis suppurativa
bacteria. Nevertheless, all studies show predominantly positive culturing samples from HS lesions. The polymicrobial flora and in particular the dominating occurrence of CoNS in the HS lesions may indeed raise speculations on the pathogenetic significance of this recurring bacteriological finding. Moreover, given the wellknown pathogenicity and high occurrence rate of S. aureus (Table 1) in the studies, it may appear reasonable to consider its potential role in HS. Unfortunately, although the microflora/microbiota of the normal human skin has been described (25), only one study included controls (adjacent healthy skin from patients with HS) (16). Furthermore, bacterial subtyping has generally not been carried out, which makes unequivocal conclusions on the possible pathogenic role of S. aureus and CoNS and the general polymicrobial flora difficult in spite of HS’s clinical resemblance of bacterial infection.
Staphylococcus aureus (S. aureus) As displayed in Table 1, S. aureus is a frequently obtained specimen in HS and has often been proposed as a potential causative organism (17,18,26). The bacterium is a facultative, Gram-positive coccus known for its disease-causing capabilities (e.g. food poisoning, skin infections, osteomyelitis and endocarditis) (27). S. aureus has also been linked to presumed risk factors of HS such as smoking (5). It is well known that a predominant part of patients with HS are smokers, and interestingly, Matusiak L et al. (28) found that all their subjects with S. aureus were heavy smokers. Moreover, it has been speculated that an association between S. aureus and nicotine may influence the disease development as nicotine may promote the growth of this pathogen (29). Alternatively, Jemec et al. (17) suggested that S. aureus may be involved only in the initial process of disease development, for example by facilitating anatomical alterations in the hair follicles by inflammation and necrosis. Although colonization of S. aureus is frequently encountered among other skin diseases (e.g. atopic dermatitis and psoriasis vulgaris) (30,31), the clinical picture differs significantly compared to HS. It has often been stated that patients suffering from folliculitis decalvans (FD) express hypersensitivity towards S. aureus (32). In FD, S. aureus and a deficient host immune response seem to play an important role in development of this suppurative and cicatricial alopecia. A similar pathogenetic correlation may hypothetically be found in HS.
CoNS (coagulase-negative staphylococci) The clinical significance of CoNS, such as Stahylococcus epidermidis, is unclear, as they generally are considered to be nonpathogenic and a part of the normal skin flora (33). Although CoNS are commensals of the skin and mucosa, they can cause severe infections in immune-suppressed patients (34). The high occurrence of CoNS detected in this review may be explained by accidental contamination during the process of obtaining or culturing biopsies from the lesions. However, in a study of 25 patients with HS, Lapins et al.(18) attempted to reduce this source of error by evaporate (using CO2 laser ablation) the diseased tissue from the surface downwards, allowing sampling of bacteriological cultures from the deep parts of dermis only as bacteria from superficial levels were evaporated along with the tissue. Lapins et al. found CoNS in 21 patients and 16 of these isolates were from deep levels of the skin. In addition, in nine of the 16 deep samples, CoNS were the only bacteria detected, suggesting
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
commensals may represent an activating factor for the non-infectious inflammatory process in HS. Generally, it raises the question of when and how does a commensal become a pathogen? Although consistent results were demonstrated in a similar but smaller study (19), the significance of these findings may be compromised, as the same group of author’s demonstrated preoperative bacteraemia in nine of 21 patients with HS (24) suggesting a continuous presence of bacteria in the blood. Moreover, the study showed that 15 of 29 staphylococci isolates were S. warneri, which is normally part of the human skin, but has also been considered a pathogen that may cause morbidity and mortality in hospitalized patients (35). Recently, a comprehensive study on the microbiology of 102 HS lesions from 82 patients using prolonged bacterial cultures and bacterial metagenomics found Staphylococcus lugdunensis in 58% of HS nodules and abscesses (16). Strikingly this specimen was associated with Hurley stage 1, which indicates a potential role in early disease pathogenesis. Furthermore, this important study preformed hierarchical clustering on the samples yielding two microbiological profiles A and B. Group A had Staphylococcus lugdunensis as a unique or predominant isolate. Group B consisted of a polymicrobial flora such as strict anaerobes, anaerobic actinomyces and streptococci of the milleri group.
Biofilm An wide range of bacteria are capable of producing an extracellular matrix (biofilm) that serves as a protective layer against host defense mechanisms and antimicrobial agents (36). Biofilm consists of a polymeric conglomeration, generally composed of extracelluar DNA, proteins and polysaccharides. Biofilm infections are characterized by their recalcitrance to antibiotics and ability to circumvent host immune-mediated clearance, causing persistent infections (37). Such property appears in line with the recurrent nature of HS and the fact that S. epidermidis can often be isolated from superficial and deeper levels from HS nodules. S. epidermidis is indeed notable for its propensity to form biofilm primarily associated with indwelling properties (e.g. catheters and prostheses) (38). Biofilm configuration is the defining virulence factor associated with S. epidermidis and a key feature that distinguishes the bacteria significantly from S. aureus. Biofilm formation has been described in HS (39,40). A histological retrospective study of 27 patients with HS was performed using fluorescence in situ hybridization (FISH) and immunofluorescence (IM) (39). Biofilm-like structures were seen in one-fifth of the samples and were primarily situated in hair follicles and sinus tracts. In addition, a possible analogy can be made between the environment of chronic HS lesions with its foreign body reactions, scattered intradermal corneocytes and occasional fragments of hair that may resemble an environment produced by a foreign body, thereby intensifying the pathogenic properties of CoNS in HS. Further, the sequestered environment of sinus tracts in HS may also be an ideal place for biofilm formation and this microbiologic principle may indeed be applicable to CoNS in HS. As displayed in Fig. 2, other bacteria have been found in HS, although the occurrence appears to be less frequent. The inconsistent findings of other than Staphylococcus spp. such as Corynebacterium spp. and Propionibacterium spp. in HS lesions may be due to normal differences in topographical distribution of microbiota seen in areas with high humidity (axillae and groin).
729
Ring et al.
Topographical diversity of bacteria A notable feature of the skin is the topographical diversity of bacterial populations, with composition and diversity of bacteria colonizing the skin depending on the microenvironment. Differentially distributed hair follicles, and eccrine, apocrine and sebaceous glands contribute to the variable cutaneous microenvironments and are likely selective for which subsets of bacteria can thrive in those specialized conditions. With regard to the correlation between culture results and the topographical distribution of bacteria in HS, only one study has performed such analysis. Guet-Revillet et al. (16) demonstrated an immediate correlation between profile A (S. lugdunensis) and anatomical sites such as breasts and buttocks. Additionally, the study found a predominant occurrence of a polymicrobial flora in the inguinal fold and the axilla, which is in agreement with previous findings. Sebaceous gland number and volume per hair follicle were investigated among 21 patients with HS and nine healthy control persons using stereology (9). The study showed that sebaceous gland number and volume in patients with HS were significantly reduced compared with healthy persons. The association between the reduction in number or volume in sebaceous glands and the age-related changes in composition of the skin microbiota in sebaceous areas may be a part of the pathogenesis of HS. Puberty-associated maturation of the sebaceous glands coincides with a shift in skin microbiota towards enrichment with several bacteria (41). Absence or atrophic sebaceous glands may therefore influence this shift in microbiota and may cause a dysbiosis in patients with HS leading to auto-inflammatory responses. This could explain the rare onset of HS before puberty and the decrease in symptoms seen in the process of ageing (>60 years) (42).
Methods Importantly, studies attempting to analyse bacteria in HS lesions are mainly relying on culture-based methods (Table 1). About one trillion commensal bacteria exist on the human skin surface and until recently, our knowledge of the skin microbiota was primarily based upon culture assays (33). Although culture-based techniques have provided important insight into bacterial, fungal and viral populations on the skin, the method is only able of culture less than 1% of the bacteria species present (43). In addition, culturebased methods exclude microbes that rely on microbe–microbe interactions to thrive. The emergence of next-generation sequencing has allowed a more accurate and less biased characterization of microbiomes compared with culture-based techniques (44). Thus, it appears attractive to investigate the microbiome of HS using a 16S ribosomal RNA gene approach. Regions within this gene offer information on species-specific signature sequences, thereby providing useful data for bacterial identification. Interestingly, Satorius et al. (24) applied PCR and DNA sequencing to determine the number and type of bacteria circulating in the bloodstream in patients with HS undergoing surgical treatment. This technique relies on primers that recognize conversed DNA sequences of bacterial genes that encode ribosomal RNA (rRNA 16S or 23S). The method involves several critical steps, such as DNA extraction from specimens, PCR amplification and detection of amplicons. However, although specific primers are powerful tool allowing
730
high sensitivity and specificity, the technique does not enable identification of non-target bacterial species in the tested sample (45). As seen in Table 1, only one HS study (16) applied metagenomics. Although the study only performed metagenomics on six samples (swabs), the study has paved the way for future research on the bacteriology using metagenomics in case–control studies. Such studies have already been performed on acne vulgaris (46), psoriasis vulgaris (31) and rosacea (47) yielding major disease-associated findings. Relying on culture for the identification of bacterial species may introduce other possible confounding factors. Unfortunately, only few studies mention the type of anaesthetics used or if it was free of preservative. This may be important as preservatives inhibit bacterial and fungal growth (48). Preservative-induced inhibition of bacterial growth could have reduced the prevalence of some bacteria and influenced the bacteria to bacteria interactions in the agar plate, thereby skewing the results in unknown direction. Further, it has been speculated whether a ring block would lead to a more accurate culturing of the bacteria instead of injecting anaesthetics directly into the area where biopsies are required (48). As the studies included in this review investigated patients with HS recruited from the hospital (i.e. outpatient clinics or inpatients during HS operations), selection bias may also be present, and generalization to patients with HS in the general population should be made with caution. Furthermore, we cannot rule out publication bias, that is unpublished studies showing negative cultures. Although an array of studies has attempted to identify the role of bacteria in HS, it is still evident that bacterial population structure and diversity at strain level is poorly understood in patients with HS. As no uniform pattern of microflora in HS is observed between the previous pathobiological studies, the role of the residential microflora versus pathogens remains enigmatic. However, the current observations are in agreement with the hypothesis of van der Zee et al. (23) that commensal bacteria may elicit an inflammatory response in a susceptible individual causing keratin plugging of the hair follicle’s infundibulum. Indeed, the recurring findings of skin commensals in the deeper layers of HS point to potential causal mechanisms of disease development. Prospective studies providing insight on the skin microbiome at the strain level and genome level of particular dominant commensals in HS lesions may provide important knowledge in defining the role of bacteria in the pathogenesis of HS.
Author contribution All authors contributed to research design, or the acquisition, analysis or interpretation of data, drafting the paper or revising it critically and approval of the submitted and final versions. Specific contributions to the work from each author were as follows: Hans Christian Ring and Peter Riis Mikkelsen performed the search strategies on the literature. Hans Christian Ring, Iben Marie Miller and Gregor Jemec designed the research study. Kurt Fuursted, H avard Jenssen and Ditte Marie Saunte analysed the data. Hans Christian Ring wrote the paper.
Conflict of interest The authors have declared no conflicting interests.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
The bacteriology of hidradenitis suppurativa
Supporting Information Additional supporting data may be found in the supplementary information of this article.
Table S1. Overview of included studies and information on number of patients/lesions, methods, and identified bacterial species.
Reference 1 Jemec G B. N Engl J Med 2012: 366: 158–164. 2 Revuz J E, Canoui-Poitrine F, Wolkenstein P et al. J Am Acad Dermatol 2008: 59: 596–601. 3 Jemec G B, Heidenheim M, Nielsen N H. J Am Acad Dermatol 1996: 35(2 Pt 1): 191–194. 4 Schrader A M, Deckers I E, van der Zee H H et al. J Am Acad Dermatol 2014: 71: 460–467. 5 Kromann C B, Deckers I E, Esmann S et al. Br J Dermatol 2014: 171(4): 819–824. 6 Miller I M, Ellervik C, Vinding G R et al. JAMA Dermatol 2014: 150(12): 1273–1280. 7 Gold D A, Reeder V J, Mahan M G et al. J Am Acad Dermatol 2014: 70: 699–703. 8 Sabat R, Chanwangpong A, Schneider-Burrus S et al. PLoS One 2012: 7: e31810. 9 Kamp S, Fiehn A M, Stenderup K et al. Br J Dermatol 2011: 164: 1017–1022. 10 von Laffert M, Helmbold P, Wohlrab J et al. Exp Dermatol 2010: 19: 533–537. 11 van der Zee H H, de Ruiter L, van den Broecke D G et al. Br J Dermatol 2011: 164: 1292– 1298. 12 Giamarellos-Bourboulis E J, Antonopoulou A, Petropoulou C et al. Br J Dermatol 2007: 156: 51–56. 13 Scheinfeld N. Dermatol Online J 2013: 19: 1. 14 Kimball A B, Kerdel F, Adams D et al. Ann Intern Med 2012: 157: 846–855. 15 van Rappard D C, Leenarts M F, Meijerink-van ‘t O L et al. J Dermatolog Treat 2012: 23: 284– 289. 16 Guet-Revillet H, Coignard-Biehler H, Jais J P et al. Emerg Infect Dis 2014: 20: 1990–1998. 17 Jemec G B, Faber M, Gutschik E et al. Dermatology 1996: 193: 203–206.
18 Lapins J, Jarstrand C, Emtestam L. Br J Dermatol 1999: 140: 90–95. 19 Sartorius K, Killasli H, Oprica C et al. Br J Dermatol 2012: 166: 879–883. 20 Gener G, Canoui-Poitrine F, Revuz J E et al. Dermatology 2009: 219: 148–154. 21 van der Zee H H, Boer J, Prens E P et al. Dermatology 2009: 219: 143–147. 22 Matusiak L, Bieniek A, Szepietowski J C. Acta Derm Venereol 2014: 94: 699–702. 23 van der Zee H H, Laman J D, Boer J et al. Exp Dermatol 2012: 21: 735–739. 24 Sartorius K, Lapins J, Jalal S et al. Dermatology 2006: 213: 305–312. 25 Human Microbiome Project Consortium. Nature 2012: 486: 207–214. 26 Brook I, Frazier E H. J Med Microbiol 1999: 48: 103–105. 27 Larkin E A, Carman R J, Krakauer T et al. Curr Med Chem 2009: 16: 4003–4019. 28 Matusiak L, Bieniek A, Szepietowski J C. Acta Derm Venereol 2014: 94(6): 699–702. 29 Kurzen H, Kurokawa I, Jemec G B et al. Exp Dermatol 2008: 17: 455–456. 30 Salah L A, Faergemann J. Acta Derm Venereol 2014: 95(5): 532–535. 31 Alekseyenko A V, Perez-Perez G I, De Souza A et al. Microbiome 2013: 1: 31. 32 Otberg N, Kang H, Alzolibani A A et al. Dermatol Ther 2008: 21: 238–244. 33 Grice E A, Kong H H, Renaud G et al. Genome Res 2008: 18: 1043–1050. 34 von Eiff C, Peters G, Heilmann C. Lancet Infect Dis 2002: 2: 677–685.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
35 Kamath U, Singer C, Isenberg H D. J Clin Microbiol 1992: 30: 261–264. 36 Costerton J W, Lewandowski Z, Caldwell D E et al. Annu Rev Microbiol 1995: 49: 711–745. 37 Namvar A E, Bastarahang S, Abbasi N et al. GMS Hyg Infect Control 2014: 9: Doc23. 38 Fey P D, Olson M E. Future Microbiol 2010: 5: 917–933. 39 Jahns A C, Killasli H, Nosek D et al. APMIS 2014: 122: 804–809. 40 Kathju S, Lasko L A, Stoodley P. FEMS Immunol Med Microbiol 2012: 65: 385–389. 41 Oh J, Conlan S, Polley E C et al. Genome Med 2012: 4: 77. 42 Vinding G R, Miller I M, Zarchi K et al. Br J Dermatol 2014: 170: 884–889. 43 Grice E A. Semin Cutan Med Surg 2014: 33: 98–103. 44 Buchan B W, Ledeboer N A. Clin Microbiol Rev 2014: 27: 783–822. 45 Kemp M, Bangsborg J, Kjerulf A et al. Open Microbiol J 2013: 7: 146–151. 46 Fitz-Gibbon S, Tomida S, Chiu B H et al. J Invest Dermatol 2013: 133: 2152–2160. 47 Casas C, Paul C, Lahfa M et al. Exp Dermatol 2012: 21: 906–910. 48 Epstein E. J Am Acad Dermatol 1998: 39: 518–519. 49 Highet A S, Warren R E, Weekes A J. Arch Dermatol 1988: 124: 1047–1051. 50 O’Loughlin S, Woods R, Kirke P N et al. Arch Dermatol 1988: 124: 1043–1046.
731
Review
www.wileyonlinelibrary.com/journal/EXD
The bacteriology of hidradenitis suppurativa: a systematic review Hans Christian Ring1, Peter Riis Mikkelsen1, Iben Marie Miller1, H avard Jenssen2, Kurt Fuursted3, 1 1 Ditte Marie Saunte and Gregor B.E. Jemec 1 Department of Dermatology, Roskilde Hospital, University of Copenhagen, Roskilde, Denmark; 2Department of Science, Systems and Models, Roskilde University, Roskilde, Denmark; 3Department of bacteriology, Statens Serum Institut, Copenhagen, Denmark Correspondence: Hans Christian Ring, MD, Department of Dermatology, Roskilde Hospital, Køgevej 7-13, Roskilde 4000, Denmark, e-mail: [email protected]
Abstract: Hidradenitis suppurativa (HS) is a chronic inflammatory disabling skin disease consisting of recurrent nodules, sinuses, fistulas and scarring involving the intertriginous regions. HS is often a therapeutic challenge and most treatments are off-label. A better understanding of aetiology and pathogenesis of HS may facilitate the development of effective treatment. Although the clinical presentation is strongly reminiscent of bacterial infection, the role of bacteria remains controversial. Studies have isolated an array of different bacteria specimens. Consistent findings of Gram-positive cocci and Gram-positive rods including Staphylococus aureus, coagulase-negative staphylococci (CoNS) and Corynebacterium species in deep tissue samples have been demonstrated in HS and may constitute a central target for the immune system. Efficacy of antibiotics, that is rifampicin, clindamycin or tetracycline, supports a microbial role in disease pathogenesis. However, these antibiotics also work
as immunomodulators of especially T cells, and the underlying mechanisms may therefore be more complex. We performed a systematic review of previous studies investigating the bacterial flora in hidradenitis suppurativa. We searched PubMed, EMBASE, Royal Danish Library and Cochrane library (search date 11 December 2014). A total of 66 papers were identified and nine papers published between 1988 and 2014 matched our inclusion criteria, yielding bacteriological data of a total of 324 patients with HS (mean age 36.8 years and female/male ratio 215/109). This overview of the bacteriology may aid researchers and physicians exploring the potential role of bacteria in HS. Furthermore, to stimulate a broader debate, we also present different viewpoints on the possible role of bacteria in HS.
Introduction
patients are often treated with antibiotics, for example penicillins or, incision and drainage by non-dermatologists, even though the treatments are regularly ineffective. Studies have isolated an array of different bacteria specimens (16–19). Consistent findings of Gram-positive cocci and Grampositive rods including Staphylococus aureus, coagulase-negative staphylococci (CoNS) and Corynebacterium species in deep tissue samples have been found in HS (18,19) and may constitute a central target for the immune system, but are also part of the normal human skin flora. The efficacy of antibiotics, that is rifampicin, clindamycin or tetracycline, in HS treatment further supports a microbial role in disease pathogenesis (20,21). However, these antibiotics also work as immunomodulators of especially T cells (20,21), and the underlying mechanisms may therefore be more complex. It is unclear whether bacterial colonization is a primary or secondary event in the evolution of HS lesions. It has been speculated that bacterial invasion leads to series of pathogen-associated molecular pathways, which may trigger the initiation of inflammasomes (22). Alternatively, the deposition of keratin fragments into the dermis in a genetically susceptible individual may be the trigger, and finally, it is speculated that the deposited keratin fragments may be colonized with bacteria (23). Our objective was to investigate the previous body of literature concerning the bacteriology in HS, thereby providing an overview of the field. This may aid researchers and physicians in understanding the potential role of bacteria in HS.
Hidradenitis suppurativa (HS) is a chronic inflammatory disabling skin disease defined by recurrent nodules, sinuses, fistulas and scarring involving the intertriginous regions. The nodules can progressively expand to abscesses, and subsequently rupture causing suppuration (1). Although the prevalence of HS has been estimated up to 4% in European populations (2,3), most treatments are off-label. The development of effective treatment may be facilitated by a better understanding of aetiology and pathogenesis of HS. Factors such as genetics, smoking and metabolic syndrome appear to be strongly associated with HS (4–8), and several hypotheses on the pathogenesis have been discussed. The main theory is hair follicle plugging, and furthermore in contrast to acne, the sebaceous glands appears reduced or absent in HS (9). Histologically, HS appears to be characterized by a hyperkeratinization of the hair follicle with an immediate hair follicle occlusion and subsequent follicular dilatation and perifollicular inflammation followed by cyst formation (10). Increasing evidence suggest that dysfunctional immune responses play a key role involving both the innate and adaptive immune system (11,12). This is in agreement with recent studies showing efficacy of immunomodulatory agents such as oral corticosteroids, cyclosporine (13), and more recently antitumor necrosis factor therapy implying that TNF-a is involved (13–15). The role of bacteria, however, remains controversial. The clinical presentation is strongly reminiscent of bacterial infection, and
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
Key words: bacteria – hidradenitis suppurativa – microbiology
Accepted for publication 28 May 2015
727
Ring et al.
Method We performed a systematic review of previous studies investigating the bacterial flora in hidradenitis suppurativa. We searched PubMed, EMBASE, Royal Danish Library and Cochrane library (search date 11 December 2014). We used two search strategies combining the MeSH terms “bacteriology AND acne inversa” and “bacteriology AND hidradenitis suppurativa”. Furthermore, the reference lists from the included articles were manually searched for additional relevant studies. The database search was undertaken independently by two reviewers for each database. The titles and abstracts of articles retrieved by the search strategy were reviewed independently by two of the authors. Retrospective and prospective cross-sectional studies exclusively investigating the bacterial presence of bacteria in tissue or blood of patients with HS were included. Only studies in English were included. Case reports or case series including less than 10 patients were excluded from the review (Fig. 1). Despite methodological differences among the approaches used by the included studies, we have constructed a table (Table. S1) displaying the number of patients/lesions, methods (sampling, culturing, molecular or identification techniques) of each study. The fractional distribution of the different bacterial species identified among the included studies is illustrated (Fig. 2). Furthermore, the prevalence rates of S. aureus in the different studies are displayed separately in Table 1.
Results Included studies in the review A total of 66 papers were identified and nine papers published between 1988 and 2014 matched the above-mentioned inclusion criteria (Fig. 1), yielding bacteriological data of a total of 324 patients with HS (mean age 36.8 years and female/male ratio 215/ 109).
Microbiological methods used in the included studies As shown in Table S1, traditional microbiological techniques for culturing bacteria in aerobic and anaerobic environment have been applied to investigate the bacteriology of HS in eight of nine included studies. Three studies used biochemical tests and only one study used immunological identification technique (Lancefield group antigens). One study applied 16S gene ribosomal RNA gene
Figure 2. Fractional distribution of various bacteria found in nine studies. Swabs, biopsies, blood and needle aspiration of purulent content are included in the figure. Bacteria that were not clearly labelled as any of the bacteria listed above were not included in this presentation; for example, Anaerobe Streptococci, Microaerophilic Streptococci, anaerobe Gram-positive Cocci and lactobacilis are classified as other.
sequencing along with metagenomics. Lastly, only one study investigated the bacteriology in blood samples in patients with HS using PCR (polymerase chain reaction) and DNA sequencing and analysis (24).
The Spectrum of Bacteria found in patients with HS As displayed in Fig. 2, CoNS (34.1%) and mixed anaerobic bacteria (23.3%) are the most frequently encountered types of bacteria among the 324 patients. In addition, S. aureus also constitutes a considerable part of the overall isolated bacterial specimens (isolated in six of nine studies), and among the studies that identified S. aureus, the prevalence rate ranged from 13% to 56% (Table 1).
Discussion The studies included in this review demonstrate a high prevalence of S. aureus and CoNS with a considerable variety of other species isolated from the inflamed lesions of patients with HS. However, the results also reflect retrospective studies which may cause a relatively lower identification rate of anaerobic bacteria. These bacterial species may therefore have been underestimated. In addition, bacterial identification methods varied significantly in the period of 1988–2014 (from biochemical tests to the use of MALDI-TOF (matrix-assisted laser desorption/ionization) mass spectrometry and molecular methods), and of note, researchers were first able to identify Actinomyces spp. after the late 1990s. Looking at the literature in general, there appears to be no consistency with regard to the occurrence of any specific type of
Table 1. Prevalence rate of S. aureus in each study Studies
Prevalence %
Lapins et al. (18) Matusiak et al. (28) Guet-Revillet et al. (16) Highet et al. (49) Sartorius et al. (19) Brook and Frazier (26) Sartorius et al. (24) Jemec et al. (17) O’Loughlin S et al. (50)
56 13.6 131 43.75 Not identified 35.29 Not identified 19.51 Not identified
13% displays the prevalence of S. aureus among 125 investigated HS lesions.
1
Figure 1. Flow chart of search results showing included and excluded references.
728
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
The bacteriology of hidradenitis suppurativa
bacteria. Nevertheless, all studies show predominantly positive culturing samples from HS lesions. The polymicrobial flora and in particular the dominating occurrence of CoNS in the HS lesions may indeed raise speculations on the pathogenetic significance of this recurring bacteriological finding. Moreover, given the wellknown pathogenicity and high occurrence rate of S. aureus (Table 1) in the studies, it may appear reasonable to consider its potential role in HS. Unfortunately, although the microflora/microbiota of the normal human skin has been described (25), only one study included controls (adjacent healthy skin from patients with HS) (16). Furthermore, bacterial subtyping has generally not been carried out, which makes unequivocal conclusions on the possible pathogenic role of S. aureus and CoNS and the general polymicrobial flora difficult in spite of HS’s clinical resemblance of bacterial infection.
Staphylococcus aureus (S. aureus) As displayed in Table 1, S. aureus is a frequently obtained specimen in HS and has often been proposed as a potential causative organism (17,18,26). The bacterium is a facultative, Gram-positive coccus known for its disease-causing capabilities (e.g. food poisoning, skin infections, osteomyelitis and endocarditis) (27). S. aureus has also been linked to presumed risk factors of HS such as smoking (5). It is well known that a predominant part of patients with HS are smokers, and interestingly, Matusiak L et al. (28) found that all their subjects with S. aureus were heavy smokers. Moreover, it has been speculated that an association between S. aureus and nicotine may influence the disease development as nicotine may promote the growth of this pathogen (29). Alternatively, Jemec et al. (17) suggested that S. aureus may be involved only in the initial process of disease development, for example by facilitating anatomical alterations in the hair follicles by inflammation and necrosis. Although colonization of S. aureus is frequently encountered among other skin diseases (e.g. atopic dermatitis and psoriasis vulgaris) (30,31), the clinical picture differs significantly compared to HS. It has often been stated that patients suffering from folliculitis decalvans (FD) express hypersensitivity towards S. aureus (32). In FD, S. aureus and a deficient host immune response seem to play an important role in development of this suppurative and cicatricial alopecia. A similar pathogenetic correlation may hypothetically be found in HS.
CoNS (coagulase-negative staphylococci) The clinical significance of CoNS, such as Stahylococcus epidermidis, is unclear, as they generally are considered to be nonpathogenic and a part of the normal skin flora (33). Although CoNS are commensals of the skin and mucosa, they can cause severe infections in immune-suppressed patients (34). The high occurrence of CoNS detected in this review may be explained by accidental contamination during the process of obtaining or culturing biopsies from the lesions. However, in a study of 25 patients with HS, Lapins et al.(18) attempted to reduce this source of error by evaporate (using CO2 laser ablation) the diseased tissue from the surface downwards, allowing sampling of bacteriological cultures from the deep parts of dermis only as bacteria from superficial levels were evaporated along with the tissue. Lapins et al. found CoNS in 21 patients and 16 of these isolates were from deep levels of the skin. In addition, in nine of the 16 deep samples, CoNS were the only bacteria detected, suggesting
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
commensals may represent an activating factor for the non-infectious inflammatory process in HS. Generally, it raises the question of when and how does a commensal become a pathogen? Although consistent results were demonstrated in a similar but smaller study (19), the significance of these findings may be compromised, as the same group of author’s demonstrated preoperative bacteraemia in nine of 21 patients with HS (24) suggesting a continuous presence of bacteria in the blood. Moreover, the study showed that 15 of 29 staphylococci isolates were S. warneri, which is normally part of the human skin, but has also been considered a pathogen that may cause morbidity and mortality in hospitalized patients (35). Recently, a comprehensive study on the microbiology of 102 HS lesions from 82 patients using prolonged bacterial cultures and bacterial metagenomics found Staphylococcus lugdunensis in 58% of HS nodules and abscesses (16). Strikingly this specimen was associated with Hurley stage 1, which indicates a potential role in early disease pathogenesis. Furthermore, this important study preformed hierarchical clustering on the samples yielding two microbiological profiles A and B. Group A had Staphylococcus lugdunensis as a unique or predominant isolate. Group B consisted of a polymicrobial flora such as strict anaerobes, anaerobic actinomyces and streptococci of the milleri group.
Biofilm An wide range of bacteria are capable of producing an extracellular matrix (biofilm) that serves as a protective layer against host defense mechanisms and antimicrobial agents (36). Biofilm consists of a polymeric conglomeration, generally composed of extracelluar DNA, proteins and polysaccharides. Biofilm infections are characterized by their recalcitrance to antibiotics and ability to circumvent host immune-mediated clearance, causing persistent infections (37). Such property appears in line with the recurrent nature of HS and the fact that S. epidermidis can often be isolated from superficial and deeper levels from HS nodules. S. epidermidis is indeed notable for its propensity to form biofilm primarily associated with indwelling properties (e.g. catheters and prostheses) (38). Biofilm configuration is the defining virulence factor associated with S. epidermidis and a key feature that distinguishes the bacteria significantly from S. aureus. Biofilm formation has been described in HS (39,40). A histological retrospective study of 27 patients with HS was performed using fluorescence in situ hybridization (FISH) and immunofluorescence (IM) (39). Biofilm-like structures were seen in one-fifth of the samples and were primarily situated in hair follicles and sinus tracts. In addition, a possible analogy can be made between the environment of chronic HS lesions with its foreign body reactions, scattered intradermal corneocytes and occasional fragments of hair that may resemble an environment produced by a foreign body, thereby intensifying the pathogenic properties of CoNS in HS. Further, the sequestered environment of sinus tracts in HS may also be an ideal place for biofilm formation and this microbiologic principle may indeed be applicable to CoNS in HS. As displayed in Fig. 2, other bacteria have been found in HS, although the occurrence appears to be less frequent. The inconsistent findings of other than Staphylococcus spp. such as Corynebacterium spp. and Propionibacterium spp. in HS lesions may be due to normal differences in topographical distribution of microbiota seen in areas with high humidity (axillae and groin).
729
Ring et al.
Topographical diversity of bacteria A notable feature of the skin is the topographical diversity of bacterial populations, with composition and diversity of bacteria colonizing the skin depending on the microenvironment. Differentially distributed hair follicles, and eccrine, apocrine and sebaceous glands contribute to the variable cutaneous microenvironments and are likely selective for which subsets of bacteria can thrive in those specialized conditions. With regard to the correlation between culture results and the topographical distribution of bacteria in HS, only one study has performed such analysis. Guet-Revillet et al. (16) demonstrated an immediate correlation between profile A (S. lugdunensis) and anatomical sites such as breasts and buttocks. Additionally, the study found a predominant occurrence of a polymicrobial flora in the inguinal fold and the axilla, which is in agreement with previous findings. Sebaceous gland number and volume per hair follicle were investigated among 21 patients with HS and nine healthy control persons using stereology (9). The study showed that sebaceous gland number and volume in patients with HS were significantly reduced compared with healthy persons. The association between the reduction in number or volume in sebaceous glands and the age-related changes in composition of the skin microbiota in sebaceous areas may be a part of the pathogenesis of HS. Puberty-associated maturation of the sebaceous glands coincides with a shift in skin microbiota towards enrichment with several bacteria (41). Absence or atrophic sebaceous glands may therefore influence this shift in microbiota and may cause a dysbiosis in patients with HS leading to auto-inflammatory responses. This could explain the rare onset of HS before puberty and the decrease in symptoms seen in the process of ageing (>60 years) (42).
Methods Importantly, studies attempting to analyse bacteria in HS lesions are mainly relying on culture-based methods (Table 1). About one trillion commensal bacteria exist on the human skin surface and until recently, our knowledge of the skin microbiota was primarily based upon culture assays (33). Although culture-based techniques have provided important insight into bacterial, fungal and viral populations on the skin, the method is only able of culture less than 1% of the bacteria species present (43). In addition, culturebased methods exclude microbes that rely on microbe–microbe interactions to thrive. The emergence of next-generation sequencing has allowed a more accurate and less biased characterization of microbiomes compared with culture-based techniques (44). Thus, it appears attractive to investigate the microbiome of HS using a 16S ribosomal RNA gene approach. Regions within this gene offer information on species-specific signature sequences, thereby providing useful data for bacterial identification. Interestingly, Satorius et al. (24) applied PCR and DNA sequencing to determine the number and type of bacteria circulating in the bloodstream in patients with HS undergoing surgical treatment. This technique relies on primers that recognize conversed DNA sequences of bacterial genes that encode ribosomal RNA (rRNA 16S or 23S). The method involves several critical steps, such as DNA extraction from specimens, PCR amplification and detection of amplicons. However, although specific primers are powerful tool allowing
730
high sensitivity and specificity, the technique does not enable identification of non-target bacterial species in the tested sample (45). As seen in Table 1, only one HS study (16) applied metagenomics. Although the study only performed metagenomics on six samples (swabs), the study has paved the way for future research on the bacteriology using metagenomics in case–control studies. Such studies have already been performed on acne vulgaris (46), psoriasis vulgaris (31) and rosacea (47) yielding major disease-associated findings. Relying on culture for the identification of bacterial species may introduce other possible confounding factors. Unfortunately, only few studies mention the type of anaesthetics used or if it was free of preservative. This may be important as preservatives inhibit bacterial and fungal growth (48). Preservative-induced inhibition of bacterial growth could have reduced the prevalence of some bacteria and influenced the bacteria to bacteria interactions in the agar plate, thereby skewing the results in unknown direction. Further, it has been speculated whether a ring block would lead to a more accurate culturing of the bacteria instead of injecting anaesthetics directly into the area where biopsies are required (48). As the studies included in this review investigated patients with HS recruited from the hospital (i.e. outpatient clinics or inpatients during HS operations), selection bias may also be present, and generalization to patients with HS in the general population should be made with caution. Furthermore, we cannot rule out publication bias, that is unpublished studies showing negative cultures. Although an array of studies has attempted to identify the role of bacteria in HS, it is still evident that bacterial population structure and diversity at strain level is poorly understood in patients with HS. As no uniform pattern of microflora in HS is observed between the previous pathobiological studies, the role of the residential microflora versus pathogens remains enigmatic. However, the current observations are in agreement with the hypothesis of van der Zee et al. (23) that commensal bacteria may elicit an inflammatory response in a susceptible individual causing keratin plugging of the hair follicle’s infundibulum. Indeed, the recurring findings of skin commensals in the deeper layers of HS point to potential causal mechanisms of disease development. Prospective studies providing insight on the skin microbiome at the strain level and genome level of particular dominant commensals in HS lesions may provide important knowledge in defining the role of bacteria in the pathogenesis of HS.
Author contribution All authors contributed to research design, or the acquisition, analysis or interpretation of data, drafting the paper or revising it critically and approval of the submitted and final versions. Specific contributions to the work from each author were as follows: Hans Christian Ring and Peter Riis Mikkelsen performed the search strategies on the literature. Hans Christian Ring, Iben Marie Miller and Gregor Jemec designed the research study. Kurt Fuursted, H avard Jenssen and Ditte Marie Saunte analysed the data. Hans Christian Ring wrote the paper.
Conflict of interest The authors have declared no conflicting interests.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
The bacteriology of hidradenitis suppurativa
Supporting Information Additional supporting data may be found in the supplementary information of this article.
Table S1. Overview of included studies and information on number of patients/lesions, methods, and identified bacterial species.
Reference 1 Jemec G B. N Engl J Med 2012: 366: 158–164. 2 Revuz J E, Canoui-Poitrine F, Wolkenstein P et al. J Am Acad Dermatol 2008: 59: 596–601. 3 Jemec G B, Heidenheim M, Nielsen N H. J Am Acad Dermatol 1996: 35(2 Pt 1): 191–194. 4 Schrader A M, Deckers I E, van der Zee H H et al. J Am Acad Dermatol 2014: 71: 460–467. 5 Kromann C B, Deckers I E, Esmann S et al. Br J Dermatol 2014: 171(4): 819–824. 6 Miller I M, Ellervik C, Vinding G R et al. JAMA Dermatol 2014: 150(12): 1273–1280. 7 Gold D A, Reeder V J, Mahan M G et al. J Am Acad Dermatol 2014: 70: 699–703. 8 Sabat R, Chanwangpong A, Schneider-Burrus S et al. PLoS One 2012: 7: e31810. 9 Kamp S, Fiehn A M, Stenderup K et al. Br J Dermatol 2011: 164: 1017–1022. 10 von Laffert M, Helmbold P, Wohlrab J et al. Exp Dermatol 2010: 19: 533–537. 11 van der Zee H H, de Ruiter L, van den Broecke D G et al. Br J Dermatol 2011: 164: 1292– 1298. 12 Giamarellos-Bourboulis E J, Antonopoulou A, Petropoulou C et al. Br J Dermatol 2007: 156: 51–56. 13 Scheinfeld N. Dermatol Online J 2013: 19: 1. 14 Kimball A B, Kerdel F, Adams D et al. Ann Intern Med 2012: 157: 846–855. 15 van Rappard D C, Leenarts M F, Meijerink-van ‘t O L et al. J Dermatolog Treat 2012: 23: 284– 289. 16 Guet-Revillet H, Coignard-Biehler H, Jais J P et al. Emerg Infect Dis 2014: 20: 1990–1998. 17 Jemec G B, Faber M, Gutschik E et al. Dermatology 1996: 193: 203–206.
18 Lapins J, Jarstrand C, Emtestam L. Br J Dermatol 1999: 140: 90–95. 19 Sartorius K, Killasli H, Oprica C et al. Br J Dermatol 2012: 166: 879–883. 20 Gener G, Canoui-Poitrine F, Revuz J E et al. Dermatology 2009: 219: 148–154. 21 van der Zee H H, Boer J, Prens E P et al. Dermatology 2009: 219: 143–147. 22 Matusiak L, Bieniek A, Szepietowski J C. Acta Derm Venereol 2014: 94: 699–702. 23 van der Zee H H, Laman J D, Boer J et al. Exp Dermatol 2012: 21: 735–739. 24 Sartorius K, Lapins J, Jalal S et al. Dermatology 2006: 213: 305–312. 25 Human Microbiome Project Consortium. Nature 2012: 486: 207–214. 26 Brook I, Frazier E H. J Med Microbiol 1999: 48: 103–105. 27 Larkin E A, Carman R J, Krakauer T et al. Curr Med Chem 2009: 16: 4003–4019. 28 Matusiak L, Bieniek A, Szepietowski J C. Acta Derm Venereol 2014: 94(6): 699–702. 29 Kurzen H, Kurokawa I, Jemec G B et al. Exp Dermatol 2008: 17: 455–456. 30 Salah L A, Faergemann J. Acta Derm Venereol 2014: 95(5): 532–535. 31 Alekseyenko A V, Perez-Perez G I, De Souza A et al. Microbiome 2013: 1: 31. 32 Otberg N, Kang H, Alzolibani A A et al. Dermatol Ther 2008: 21: 238–244. 33 Grice E A, Kong H H, Renaud G et al. Genome Res 2008: 18: 1043–1050. 34 von Eiff C, Peters G, Heilmann C. Lancet Infect Dis 2002: 2: 677–685.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 727–731
35 Kamath U, Singer C, Isenberg H D. J Clin Microbiol 1992: 30: 261–264. 36 Costerton J W, Lewandowski Z, Caldwell D E et al. Annu Rev Microbiol 1995: 49: 711–745. 37 Namvar A E, Bastarahang S, Abbasi N et al. GMS Hyg Infect Control 2014: 9: Doc23. 38 Fey P D, Olson M E. Future Microbiol 2010: 5: 917–933. 39 Jahns A C, Killasli H, Nosek D et al. APMIS 2014: 122: 804–809. 40 Kathju S, Lasko L A, Stoodley P. FEMS Immunol Med Microbiol 2012: 65: 385–389. 41 Oh J, Conlan S, Polley E C et al. Genome Med 2012: 4: 77. 42 Vinding G R, Miller I M, Zarchi K et al. Br J Dermatol 2014: 170: 884–889. 43 Grice E A. Semin Cutan Med Surg 2014: 33: 98–103. 44 Buchan B W, Ledeboer N A. Clin Microbiol Rev 2014: 27: 783–822. 45 Kemp M, Bangsborg J, Kjerulf A et al. Open Microbiol J 2013: 7: 146–151. 46 Fitz-Gibbon S, Tomida S, Chiu B H et al. J Invest Dermatol 2013: 133: 2152–2160. 47 Casas C, Paul C, Lahfa M et al. Exp Dermatol 2012: 21: 906–910. 48 Epstein E. J Am Acad Dermatol 1998: 39: 518–519. 49 Highet A S, Warren R E, Weekes A J. Arch Dermatol 1988: 124: 1047–1051. 50 O’Loughlin S, Woods R, Kirke P N et al. Arch Dermatol 1988: 124: 1043–1046.
731
