Report

Analysis of epidemiology, clinical features and management of erysipelas

Konstantin Krasagakis1, MD, Antonios Valachis2, MD, Panagiotis Maniatakis1, MD, Sabine Kru ger-Krasagakis1, MD, George Samonis2, MD, and Androniki D. Tosca1, MD

Departments of Dermatology, and Internal Medicine, University Hospital of Heraklion, Heraklion, Greece
2

Abstract
Background Erysipelas is a supercial form of cellulitis affecting the upper dermis and supercial lymphatics. The widespread use of antibiotics may affect clinical ndings and response to therapy of infectious disorders. The purpose of the study was to investigate the epidemiological, clinical, and laboratory features of erysipelas and to compare the results of treatment with penicillin vs. other antibiotic regimens. Methods All charts of erysipelas patients treated at the University Hospital of Heraklion, Crete, Greece from 1994 to 2002 were retrospectively studied. Results Median age of the 99 patients was 54.5 years; 59% were females. The most frequent site involved was the lower extremity (76%), followed by the face (17%) and upper extremity (6%). In 61 patients (62%), a possible entry portal was identied. The most common manifestation of erysipelas was local symptoms and signs (pain, erythema, and swelling) in all patients, together with elevated erythrocyte sedimentation rate (ESR) (60%). Fever was present in 25% of patients. The most commonly used antibiotic was intravenous penicillin G (64%). In the penicillin group, mean duration of fever after treatment initiation was shorter than in the nonpenicillin group (1.7 vs. 4.5 days, P = 0.002). Both treatment failures and recurrences were the same between the two groups. Discussion The diagnosis of erysipelas can be based on careful examination for local signs and symptoms. The role of ESR in primary diagnosis needs further investigation. Penicillin seems to preserve its fundamental role in the treatment of disease.

Correspondence Konstantin Krasagakis, MD Department of Dermatology University Hospital of Heraklion GR-71110 Heraklion Crete, Greece E-mail: krasagak@med.uoc.gr Conicts of interest: None to declare.

Introduction Erysipelas is a supercial form of cellulitis that affects the upper dermis including the supercial lymphatics.1 It is regarded as disease caused by group A b-hemolytic Streptococci, but similar lesions can be caused by group B, C, and G Streptococci, and rarely by Staphylococci, Haemophilus inuenzae, Pseudomonas aeruginosa, or Enterobacteria.2 The diagnosis is based on clinical ndings, as a ery red, tender, painful plaque with well-demarcated edges is the hallmark of the disease.3 Erysipelas classically involved the face, but currently the predominant location is the lower extremities.4 General symptoms and signs such as fever, chills, and malaise together with regional lymphadenopathy, and laboratory ndings including leukocytosis, elevated C-reactive protein (CRP), or erythrocyte sedimentation rate (ESR) are usually present.1,5 Bacteriological tests are not necessary for diagnosis, as isolation of microorganism from either blood or swabs remains difcult.4,68
International Journal of Dermatology 2010, 49, 10121017

1012

With early diagnosis and proper treatment, the prognosis is excellent.9 Penicillin is the empirical antibiotic of choice, while macrolides are usually recommended in patients with allergy to penicillin.10 Considering new drugs11 and the increasing resistance of causative bacteria as well as the eventual variation of the causative pathogens,12 penicillin as rst line empirical treatment has to be reevaluated. No randomized controlled trials have provided sufcient data on differences in cure rate between classic and new regimens, as most of them included small numbers of patients and were designed to test equivalence of antibiotics rather than to reveal differences.10 Furthermore, erysipelas leads to hospitalization of a signicant number of patients and to considerable costs.13 Moreover, optimal treatment should be reevaluated to determine how to reduce the severity of the infection and the prevention of recurrences which may occur in 1030% of cases.1416 In this retrospective study, we investigated epidemiological, clinical, and laboratory features of erysipelas and compared the outcome of patients treated with penicillin versus other antibiotic regimens.
2010 The International Society of Dermatology

Krasagakis et al.

Erysipelas epidemiology and treatment

Report

1013

Patients and methods

Clinical records of all patients admitted to the Department of Dermatology of the University Hospital of Heraklion, Crete, reviewed. Erysipelas was a clinical diagnosis and dened as an infection of the skin with a sudden onset of a red indurated expanding plaque with distinct border.7 Patients with erysipelas complications, hospitalized during the same study time period with bullous, hemorrhagic or necrotic lesions, were not included in the study. Epidemiological and clinical parameters recorded included: gender, age, date of admission, duration of fever before admission, temperature on admission, site of lesion, lymph node involvement, duration of fever, documented or reported allergy to penicillin, antibiotic treatment before and during hospitalization, days of hospitalization, recurrences at 1 year of follow up, presence of underlying diseases, presence of local risk factors such as athletes foot, toe web intertrigo, venous insufciency, skin trauma, insect bite, cutaneous diseases and lymphoedema.
17,18

30
Face

25 20 15 10 5 0 1621 2231 3241 4251 5261 6271 >72 Age

Trunk Lower extremities

Figure 1 Anatomical localization of erysipelas according to

age

Table 1 Characteristics of 99 patients with uncomplicated

Laboratory parameters such as white blood

erysipelas
Characteristics Age (years) >60 <60 Gender Male Female Community Urban Rural Site of infection Leg Face Upper extremities Abdomen Other Co-morbidity None Cardiovascular Diabetes mellitus Obesity Musculoskeletal Endocrine Renal Psychiatric disorders Pulmonary disease Cancer/Leukemia Risk factors/Portals of entry None Mycosis/Athletes foot Venous insufciency Dermopathy Skin trauma Insect bite Local surgery Lymphedema No. patients (%)

cell (WBC) count, ESR, and CRP levels on admission and discharge were recorded, when available. Microbiological data, either from swabs or from blood cultures were also evaluated. The publication of anonymous retrospective data on infections has been approved by the Ethics Committee of the Hospital. Statistical comparisons were made between subgroups of patients treated with penicillin versus other antibiotic regimens. Penicillin allergic patients (n = 15) and patients on ambulant pretreatment before admission (n = 24) were excluded from the analysis. To detect statistically signicant differences between the specied groups, we used chi-square test or Fishers exact test for categorical variables and Students t-test or nonparametric statistic test (MannWhitney test) for continuous variables. Signicance was dened as a P-value equal or less than 0.05.

No. of patients

Greece with diagnosis of erysipelas from 1994 to 2002 were

Upper extremities

37 (37) 62 (63) 41 (41) 58 (59) 42 (42) 57 (58) 65 17 6 1 10 35 21 18 9 9 8 5 4 4 1 21 32 23 22 20 13 4 3 (66) (17) (6) (1) (10) (35) (21) (18) (9) (9) (8) (5) (4) (4) (1) (21) (32) (23) (22) (20) (13) (4) (3)

Results A total of 145 patients with erysipelas were hospitalized in our institution during the study period. Ninety nine of them with the typical clinical features of erysipelas (68%) were included and analyzed. Fifty eight patients (59%) were females, and 41 (41%) males. The median age was 54.5 years (range 1786). Seven patients were younger than 21 years old, and eight older than 72 years. The age distribution according to anatomical area of erysipelas is shown in Fig. 1. The most common co-morbid condition was cardiovascular disease (21 patients). Diabetes mellitus, as predisposing factor, was found in 18 patients. In 61 patients (62%), a possible portal of entry for the infection was identied. Patients characteristics are shown in Table 1.

2010 The International Society of Dermatology

International Journal of Dermatology 2010, 49, 10121017

1014

Report

Erysipelas epidemiology and treatment

Krasagakis et al.

Erysipelas was most frequently located on the lower extremities (66%). Facial erysipelas was seen in 17%, while the upper extremities were involved in 6%. In one case, the infection was located on the abdomen. The frequency of lower extremities erysipelas increased tendentially, but not signicantly, during summer months (July and August) whereas the lowest frequencies were observed during December, January, and April (Fig. 2). On the other hand, facial erysipelas increased, not signicantly, during autumn and winter months, with a peak in October and December. Local signs and symptoms of inammation, such as pain, erythema and swelling were present in all patients. The mean duration of local symptoms before admission was 3.1 days. At the time of admission, 25 (25%) patients presented with fever, and this proportion did not change when patients under antibiotics on admission were excluded from the analysis. Twenty-two patients had regional palpable lymph nodes. On admission, 34 of 99 patients (34%) presented with leukocytosis, 78 of 85 patients with documented ESR had elevated values (92%), and 27 of 35 patients with documented CRP also (77%). The most common clinico-laboratory presentation of erysipelas was local signs and symptoms with elevated ESR (60 patients), followed by the presence of local signs and symptoms without fever or leukocytosis (51 patients). Fourteen patients were presented with local signs and fever, 22 with local signs and leukocytosis, whereas the triad of local signs, leukocytosis and fever was present in 10. The infection was microbiologically documented in 12 of 99 patients (12%); 10 isolates were from swabs and two from blood cultures. Patients having positive culture from swabs or blood had a mean length of hospital stay of 13.9 days, whereas patients without isolation had a mean length of hospital stay of 9 days (P = 0.002). The organism isolated more frequently was Staphylococcus aureus in six specimens (Table 2). The antibiotics used are shown in Table 3. A single antibiotic was administered to the patients on admission.
18 16 14 12 10 8 6 4 2 0

Table 2 Bacteriological ndings in erysipelas patients

Blood cultures 2 2 Swab cultures 6 1 1 1 1 10

Microorganisms Staphylococcus aureus Staphylococcus epidermitis Staphylococcus hominis Pseudomonas aeruginosa Enterococcus faecalis Aeromonas hydrophila/caviae Total no. isolates

Table 3 Antibiotic regimens used in 99 patients with

erysipelas
Antimicrobial agents Patients (n = 99) Penicillin G IV Macrolides Cephalosporins 2nd generation Lincozamides Quinolones Cephalosporins 3rd generation Amoxicillin per os Second line in-hospital treatment (n = 35) Previous treatment with Penicillin (n = 21) Macrolides Lincozamides Amoxicillin/clavulanic acid Cephalosporins 3rd generation Quinolones Vancomycin Combination of antibiotics Previous treatment other antibiotic (n = 14) Macrolides Lincozamides Quinolones Cephalosporins 2nd generation Combination of antibiotics No. patients (%)

63 21 5 4 3 2 1

(64) (21) (5) (4) (3) (2) (1)

7 3 2 2 1 1 5 5 2 2 1 4

(33) (14) (9.5) (9.5) (5) (5) (24) (36) (14) (14) (7) (29)

Lower extremities Upper extremities Trunk Face

Months

Figure 2 Anatomical localization of erysipelas according to

the month of diagnosis

International Journal of Dermatology 2010, 49, 10121017

The most commonly used antibiotic was penicillin G given intravenously in 63 patients (64%). With the exclusion of penicillin allergic patients (n = 15), the proportion of patients who received penicillin G on admission reached 75%. Antibiotics other than penicillin were given to 21 of 84 (25%) patients without penicillin allergy. In patients with penicillin allergy, macrolides were the most commonly used antibiotics (n = 9; 60%), followed by cefuroxime (n = 4; 27%). To evaluate the efcacy of penicillin in comparison with other antibiotics, we excluded patients on ambulant pretreatment before admission (n = 24) or with penicillin allergy (n = 15). From the remaining 62 patients, 49 (79%) received penicillin G, while 13 (21%) were treated with nonpenicillin regimens. In the nonpenicillin group,
2010 The International Society of Dermatology

No. of patients

Fe M b ar ch Ap ril M ay Ju ne Ju ly Au g Se pt O ct N ov D ec

Ja n

Krasagakis et al.

Erysipelas epidemiology and treatment

Report

1015

macrolides were the most frequently used antibiotics (n = 9; 69%), followed by cephalosporins (n = 2; 15%). According to clinical and laboratory ndings, there were no statistically signicant differences in erysipelas severity on admission between the patient group treated with penicillin and the nonpenicillin treated group (Table 4). In the penicillin group, the mean duration of fever after treatment initiation was shorter (1.7 vs. 4.5 days, P = 0.002), whereas the mean duration of leukocytosis and hospitalization did not differ signicantly (2.7 vs. 2.5 days and 9.2 vs. 7.7 days, respectively). Both recurrences and treatment failures were the same between the two groups (Table 4). In 35 patients, antibiotics had to be changed, because of inefciency (32 patients) or side effects (three patients). Twelve patients (34%) received as second line treatment macrolides, whereas nine patients (26%) received combinations. The combination most used was b-lactams plus aminoglycosides in six patients. Antibiotic treatment had to be changed more than one time in 11 patients. Eighty one patients (81%) received local therapy in addition to antibiotics, while 18 (18%) required an antiinammatory medication. The frequency of recurrences during 1 year follow up was 9% (nine patients). Eight

patients had the recurrence in the rst 6 months, and one after the rst 6 months. A mycotic infection co-existed in ve patients with recurrent erysipelas, whereas three of them had a concurrent venous insufciency. Discussion Despite the relative frequency, there are few recent data on the epidemiological and clinical characteristics of erysipelas.15,19,20 Moreover, as most of the clinical trials included patients with erysipelas under the diagnosis cellulitis or skin infections,5 they failed to reveal any differences in the efcacy of penicillin or other antibiotics in erysipelas nowadays. This study describes patients characteristics and clinical presentation of erysipelas, and compares the outcome of patients treated with penicillin versus those treated with other antibiotic regimens. The age of our study group was comparable (median 54.5 years) to that reported in the previous studies.1,3,4,7,16,19,20 However, contrary to the previous observations that erysipelas is more prevalent among young and the elderly people,2,21 most of our patients were middle aged (5261 years old). Considering gender distribution, conicting data have been reported. As in our study, a female predominance has been observed in several studies.1,22 On the other hand, other series reported no gender differences.7,21 Both local and general predisponding factors for erysipelas have been reported.6,16,22,23 In our study more than two-thirds of the cases had one or more local factors as potential entry portals for bacteria. Our ndings correspond to two case-control studies17,18 that highlight the major role of cutaneous barrier disruption as risk factor for erysipelas. Furthermore, we found a signicant percentage of patients with co-morbidities, especially cardiovascular disease and diabetes mellitus. Of the co-morbidities, only obesity has been shown to be a risk factor of erysipelas,17 while other factors such as diabetes, malignancy and alcoholism have not been statistically associated with erysipelas.17,18 Concerning seasonal distribution of erysipelas, in consistence with the ndings of other studies,7,22 we did not observe any signicant differences in our series. However, we found a trend for increased number of episodes during the summer months. Similar nding has been reported recently.23 According to Ronnen et al.,16 the frequency of erysipelas increases during summer because the skin is more exposed, thus, making it more vulnerable to minor traumas that ultimately favor infections. Moreover, we found a trend for increased episodes of facial erysipelas during autumn and winter, possibly because of higher prevalence of upper respiratory tract infections during these periods.16,23,24
International Journal of Dermatology 2010, 49, 10121017

Table 4 Univariate analysis and comparison of two patient

groups
Penicillin group 51.8 16.8 21 (43) 17 (35) 35 (71) 36 (73) 10 (20) 2.6 2.8 9912 3959 44.6 28 4.46 3.9 16 (33) 12 (24) 9.22 4.46 1.7 0.9 2.67 1.44 4 (8) 16 (33) Nonpenicillin group 54.6 15.9 9 (69) 8 (62) 9 (69) 11 (85) 3 (23) 1.2 1.6 8641 2795 40.1 31.3 8.9 10.9 5 (38) 2 (15) 7.69 4.7 4.5 2.12 2.5 0.7 1 (8) 5 (38)

Variable Age (mean SD) Male (%) Urban community (%) Localization in lower extremities (%) Presence of fever (%) Palpable regional nodes (%) Days of symptoms before admission WBC on admission ESR on admission CRP on admission Absence of co-morbidity (%) Absence of risk factor (%) Mean days of hospitalization Mean days of fever Mean days of WBC normalization Recurrences (%) Need for treatment change (%)

P-value 0.618 0.091 0.079 0.877 0.404 0.834 0.100 0.300 0.535 0.479 0.694 0.485 0.176 0.002* 0.832 0.956 0.694

WBC, white blood cell; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein. *indicates a statistically signicant difference between the two treatment groups.
2010 The International Society of Dermatology

1016

Report

Erysipelas epidemiology and treatment

Krasagakis et al.

According to standard textbooks, erysipelas involved classically the face in the previous years.1 However, facial involvement today has decreased to 619%, while the lower extremities have become the predominant location.7,16,19,22 In our series, we could conrm these observations. The rarity of facial erysipelas may be explained by increased hygiene and use of cosmetics.22 Furthermore, the higher frequency of erysipelas at the lower extremities may be explained by the fact that most of its local risk factors (e.g. minor trauma, toe-web intertrigo and venous insufciency) are predominantly affecting the lower extremities.17,18 Diagnosis is based on clinical ndings,1 which include the characteristic rash and fever.2,5 It is documented that fever is absent in 1518% of the patients without prior antibiotic therapy.22 In our series, only 24% of the patients without prior antibiotic therapy were febrile on admission. The high proportion of afebrile patients raises the question about the usefulness of this clinical sign for diagnosis. According to the laboratory ndings of our study, the most valuable parameter seemed to be the ESR, which was elevated in the majority of our patients. In a previous study, ESR was found to be an indirect index of erysipelas severity.19 More than half of our patients had local symptoms and signs and elevated ESR on admission, without further clinical ndings. The above combination was the most common characteristic of erysipelas in our patients. We suggest that diagnosis can be based on the presence of local ndings, and should not be excluded in the absence of leukocytosis or fever. The role of ESR on initial diagnosis might be useful, but further investigation is needed. It is well documented that bacteriology from skin lesions in erysipelas is positive in less than 40% of patients.1,8,25 In most of our patients, cultures were negative. This could be attributable to the well-known difculties in isolating bacteria from erysipelas lesions.8 Furthermore, one-fourth of our patients were receiving antibiotics on admission. The low frequency of positive cultures can easily explain the reason why erysipelas pathogenesis is not yet fully understood. In the literature, Streptococci were reported as causative agents in most cases.1,4,15 However, the high rate of erysipelas patients growing Staphylococci from their lesions in some series5,7 raises the question about the role of these bacteria in the disease. S. aureus, the most commonly isolated pathogen was found in a small percentage of our study patients. Other rare pathogens have been also identied, albeit less frequently. Low rates of positive blood cultures have been reported in our study as well as in others previously.1,4,19,22,26 Because of this low rate, there are questions about the necessity of routine blood cultures in febrile erysipelas cases.27
International Journal of Dermatology 2010, 49, 10121017

There are few randomized controlled trials available in the literature, comparing different antibiotics in erysipelas.9,2830 Many of them had several limitations, primarily because of the sample size. Several studies were designed to test equivalence between antibiotics rather than differences in cure rates, and nally patients with erysipelas were included under those with cellulitis or skin infections.5,9 These studies have shown that macrolides, cephalosporins and uoroquinolones have equivalent or slightly higher activity than penicillin.5 As a result of these limitations, penicillin remains the drug of choice,10 while macrolides provide the best choice for patients with penicillin allergy.10 In our study, most patients were treated with penicillin G, while the most common antibiotic used in penicillin allergic patients was a macrolide. Comparing penicillin-treated patients with those receiving other antibiotics, we found that patients on penicillin became afebrile in 2 days, while those on other antibiotics in 5 days. Furthermore, we found no advantage of using antibiotics other than penicillin with respect to length of hospital stay, days to normalization of leukocyte count, recurrences, or modication of treatment. Our data, as well as a literature search, showed that there are no substantial differences in therapy of different infection sites. The average hospital stay (812 days) for patients with erysipelas13,16,31 as reported in the literature is similar to that of our study (9 days). Prolonged hospitalization is leading to considerable medical costs.13 The major reason for prolonged hospital stay is the need for intravenous antibiotic treatment,31 either because of disease severity or of treatment failure. In addition, frequent clinical and laboratory monitoring may be necessary in several cases. Shortened hospital stay will reduce the costs of erysipelas treatment, whenever possible. It has been reported that oral and intravenous treatments are of equivalent efcacy for erysipelas treatment.32 Our study revealed that penicillin therapy shortened the duration of pyrexia, when compared with other treatments. Therefore, a possible approach would be to use intravenous penicillin G as rst line treatment during the rst days of hospitalization, and after apyrexia to release the patient with an oral regimen, reducing the duration of hospitalization. Reported recurrences are relatively high, occurring in 10% of the patients 6 months after the rst episode.14 Recurrent erysipelas is largely attributable to the persistence of local factors that remain untreated. In our study, there were nine (9%) recurrences in 1 year, and in most of these patients, at least one persisting local risk factor was present. Our study has some limitations; is a hospital-based study and all cases with the typical clinical features of erysipelas admitted to our hospital were included in the
2010 The International Society of Dermatology

Krasagakis et al.

Erysipelas epidemiology and treatment

Report

1017

analysis. Thus, our patients may represent a selected group with more severe conditions. Moreover, as a result of the studys retrospective character, there are also possible limitations such as some information bias and loss to follow up. In conclusion, the diagnosis of erysipelas can be based on careful examination of local signs and symptoms rather than on fever and WBC count, while the role of ESR in primary diagnosis needs further investigation. Penicillin seems to preserve its fundamental role in the treatment of erysipelas. With respect to medical costs, prospective studies are needed to identify, with respect to medical costs, which treatment strategy and antibiotics would be most appropriate for patients. References
1 Grosshans EM. The red face: erysipelas. Clin Dermatol 1993; 11: 307313. 2 Bratton RL, Nesse RE. St. Anthonys re: diagnosis and management of erysipelas. Am Fam Phys 1995; 51: 401404. 3 Bisno AL, Stevens DL. Streptococcal infections of skin and soft tissues. N Engl J Med 1996; 334: 240245. 4 Chartier C, Grosshans E. Erysipelas: an update. Int J Dermatol 1996; 35: 779781. 5 Bonnetblanc JM, Bedane C. Erysipelas: recognition and management. Am J Clin Dermatol 2003; 4: 157163. 6 Denis F, Martin C, Ploy M. Erysipelas: microbiological and pathogenic data. Ann Dermatol Venereol 2001; 128: 317325. 7 Jorup-Rnstrm C. Epidemiological, bacteriological and complicated features of erysipelas. Scand J Infect Dis 1986; 18: 519524. 8 Hugo-Persson M, Norlin K. Erysipelas and group G Streptococci. Infection 1987; 15: 184187. 9 Morris A. Cellulitis and erysipelas. Clin Evid 2006; 15: 22072211. 10 Stevens DL, Bisno AL, Chambers HF, et al. Practice guidelines for the diagnosis and management of skin and soft-tissue infections. Clin Infect Dis 2005; 41: 1373 1406. 11 Raghavan M, Linden PK. Newer treatment options for skin and soft tissue infections. Drugs 2004; 64: 1621 1642. 12 York MK, Gibbs L, Perdreau-Remington F, Brooks GF. Characterization of antimicrobial resistance in Streptococcus pyogenes isolates from the San Francisco Bay area of northern California. J Clin Microbiol 1999; 37: 17271731. 13 Goettsch WG, Bouwes Bavinck JN, Herings RMC. Burden of illness of bacterial cellulitis and erysipelas of the leg in the Netherlands. J Eur Acad Dermatol Venereol 2006; 20: 834839.

14 Jorup-Rnstrm C, Britton S. Recurrent erysipelas: predisposing factors and costs of prophylaxis. Infection 1987; 15: 105106. 15 Eriksson B, Jorup-Ronstrom C, Karkkonen K, et al. Erysipelas: clinical and bacteriologic spectrum and serological aspects. Clin Infect Dis 1996; 23: 10911098. 16 Ronnen M, Suster S, Schewach-Millet M, Modan M. Erysipelas: changing faces. Int J Dermatol 1985; 24: 169172. 17 Dupuy A, Benchikhi H, Roujeau JC, et al. Risk factor for erysipelas of the leg (cellulitis): case-control study. BMJ 1999; 318: 15911594. 18 Mokni M, Dupuy A, Denguezli M, et al. Risk factors for erysipelas of the leg in Tunisia: a multicenter case-control study. Dermatology 2006; 212: 108112. 19 Lazzarini L, Conti E, Tositti G, de Lalla F. Erysipelas and cellulitis: clinical and microbiological spectrum in an Italian tertiary care hospital. J Infection 2005; 51: 383389. 20 Dupuy A. Descriptive epidemiology and knowledge of erysipelas risk factors. Ann Venereol 2001; 128: 312316. 21 Bartholomeeusen S, Vandenbroucke J, Truyers C. Epidemiology and comorbidity of erysipelas in Primary Care. Dermatology 2007; 215: 118122. 22 Chartier C, Grosshans E. Erysipelas. Int J Dermatol 1990; 29: 459467. 23 Pavlotsky F, Amrani S, Trau H. Recurrent erysipelas: risk factors. J Dtsch Dermatol Ges 2004; 2: 8995. 24 Boycott JA. Seasonal variations in streptococcal infections. Lancet 1966; 26: 706707. 25 Leppard BJ, Seal DV, Colman G, Hallas G. The value of bacteriology and serology in the diagnosis of cellulitis and erysipelas. Br J Dermatol 1985; 112: 559567. 26 Bishara I, Golan-Cohen A, Robenshtok E, et al. Antibiotic use in patients with erysipelas: a retrospective study. Isr Med Ass J 2001; 3: 722724. 27 Perl B, Gottehrer NP, Raveh D, et al. Cost-effectiveness of blood cultures for adult patients with cellulitis. Clin Infect Dis 1999; 29: 14831488. 28 Bernard P, Plantin P, Roger H, et al. Roxithromycin versus penicillin in the treatment of erysipelas in adults: a comparative study. Br J Dermatol 1992; 127: 155159. 29 Leman P, Mukherjee D. Flucloxacillin alone or combined with benzylpenicillin to treat lower limb cellulitis: a randomized controlled trial. Emerg Med J 2005; 22: 342346. 30 Daniel R. Azithromycin, erythromycin and cloxacillin in the treatment of infections of skin and associated soft tissues. European Azithromycin Study Group. J Int Med Res 1991; 19: 433445. 31 Musette P, Benichou J, Noblesse I, et al. Determinants of erysipelas for supercial cellulitis (erysipelas) of the leg: a retrospective study. Eur J Int Med 2004; 15: 446450. 32 Jorup-Rnstrm C, Britton S, Gavlevik A, et al. The course, costs and complication of oral versus intravenous therapy of erysipelas. Infection 1984; 12: 390394.

2010 The International Society of Dermatology

International Journal of Dermatology 2010, 49, 10121017