Infections of the skin are common and may be caused by a wide variety of organisms, including bacteria, fungi, viruses, protozoa, mycobacteria, and rickettsiae. They range from acute to chronic, and from benign to life-threatening. Because the skin has only a finite repertoire for reacting to insults, the clinical manifestations of an infection of the skin may be similar to those caused by other agents or mechanisms-for example, allergic reactions, such as toxic epidermal necrolysis (TEN), or immune-mediated skin disorders, such as pemphigus.
Infection of the skin may arise from superficial structures of the skin (eg, hair follicles), from exogenous penetration (eg, bites or splinters), or hematogenously (eg, disseminated gonococcal infection). Thus, the approach to identification of the specific cause of skin infection mandates careful attention to details of the history, the physical examination, laboratory tests, and other diagnostic means. The good news is that you can see the evidence of infection and can also visually assess the response to treatment. Thus, if empiric treatment is failing, a more vigorous diagnostic approach can be entertained.
The following "Pearls" are intended to provide practical tips on key points in diagnosis and management of skin and soft tissue infections.
European studies tend to refer to cellulitis as erysipelas, and this has caused confusion in the literature. Erysipelas is characterized by an abrupt onset of fiery red swelling of the face. Distinctive features are well-defined indurated margins, particularly along the nasolabial fold; brilliant red or salmon color; and intense pain. Flaccid bullae filled with clear fluid may develop on the second or third day, but extension to deeper soft tissues does not occur. Group A streptococci are the usual cause, and treatment with standard doses of penicillin for 7 to 10 days is effective. Although desquamation may occur in 5 to 10 days, scarring is very uncommon. Erysipelas may also occur on the lower extremities (Figure 1).
2 Impetigo: Treat to prevent spread
For reasons that are unclear, Staphylococcus aureus has emerged as the leading cause of impetigo in the United States, although group A streptococci with or without S aureus are another important cause. Historically, diagnosis and treatment of impetigo caused by group A Streptococcus were important because of its postinfectious sequelae, particularly post-streptococcal glomerulonephritis. Epidemics of impetigo caused by "nephritogenic strains" of group A Streptococcus (eg, the Red Lake strain) resulted in a high incidence of glomerulonephritis in children.
Classically, impetigo begins with pinpoint papules that enlarge to form pustules and vesicles that rupture and become encrusted with a characteristic golden brown color. Bullous impetigo is characterized by much larger lesions filled with yellow fluid and is usually caused by strains of S aureus that produce exfoliatin, the same toxin that is associated with staphylococcal scalded skin syndrome (SSSS), a systemic illness.
Antibiotic treatment has not been shown to be effective in preventing post-streptococcal glomerulonephritis. Since impetigo is rarely a life-threatening process, the question becomes: why treat at all? Of course, the main reason is to prevent the spread of impetigo to other parts of the body and to other persons, particularly in closed environments such as schools and day-care centers. Although impetigo appears disfiguring, particularly when it is present on the face, scarring does not occur.
Treatment with topical antibiotics, such as mupirocin (which is highly effective but expensive) or bacitracin (which is probably less effective but less expensive), usually suffices. If an oral antibiotic is to be administered, agents effective against S aureus should be used.
Bullous lesions should not be discounted as simple superficial processes that can be readily treated with antibiotics alone. While bullous impetigo and erysipelas may be the exceptions to this rule, other conditions associated with the formation of bullous skin lesions demand a more thorough and timely investigation of the underlying cause.
The key to looking for alternative diagnoses rests with evidence of the systemic inflammatory response syndrome. Patients with fever, tachycardia, hypotension, or elevated white blood cell (WBC) count with marked left shift may have deeper infection, such as necrotizing fasciitis or gas gangrene, or a noninfectious process, such as TEN. TEN can be readily diagnosed and differentiated from SSSS by punch biopsy with frozen section. SSSS characteristically has flaccid bullae, occurs mostly in children, is associated with very low mortality, and is mediated by exfoliatin produced by phage group 2 strains of S aureus. TEN usually occurs in adults, is usually related to a systemic drug reaction, may cause both cutaneous and mucous membrane lesions, and is associated with a high mortality. The frozen-section biopsy demonstrates a cleavage plane at the stratum corneum in SSSS but a much deeper cleavage plane at the stratum germinativum in the case of TEN. Drug-induced apoptosis of this layer results in massive sloughing of the entire epidermis. Treatment with intravenous immunoglobulin looks promising.1
Bullous lesions associated with necrotizing fasciitis and gas gangrene result from infarction of cutaneous vessels and are usually maroon, red, or blue.2,3 Because punch biopsy may be associated with a "sampling error," definitive diagnosis rests on surgical exploration. Histologically, soft tissues infected with group A streptococci or histotoxic clostridia will show extensive tissue destruction with or without large numbers of organisms and a paucity of infiltrating granulocytes (Figure 2). Thrombosis of the microvasculature and even of larger vessels may be marked. This picture is decidedly different from that seen with S aureus, Haemophilus influenzae, or Streptococcus pneumoniae infections, which are characterized by minimal tissue destruction and a luxuriant leukocytic response at the site of infection. Thus, histologic evaluation, appropriate Gram stain, and culture will establish a diagnosis and guide management, which would include antibiotic treatment and debridement if a necrotizing process is implicated.
4 When prescribing cephalexin, beware of drug-drug interactions
Clinical trials many years ago demonstrated that cephalexin was an effective treatment for general skin and skin structure infections. Recently, a retrospective study at our institution demonstrated a failure rate of 50% in adults.4 This was associated with the concomitant use of histamine2 (H2) blockers, such as ranitidine. Apparently, cephalexin absorption in the stomach is dramatically dependent on low pH. Thus, if this drug is to be used, a detailed history should be obtained to clarify the use of antacids, H2 blockers, and proton pump inhibitors.
Most cases of gas gangrene caused by Clostridium perfringens are associated with deep penetrating trauma, deep lacerations that compromise blood supply, or crush injuries.3 Recently, however, injection of black tar heroin into the skin has been associated with myonecrosis, cutaneous gangrene, shock, and organ failure. Clostridium species recovered from such cases include C perfringens, Clostridium novyi, Clostridium sordellii, and Clostridium septicum.5 Before the modern era of injection drug abuse, the only injections associated with gas gangrene (usually caused by C perfringens) were intramuscular injections with potent vasoconstrictors, such as epinephrine.
Interestingly, gas gangrene can also occur spontaneously or recur after many years. In both these situations, C septicum is usually the causative agent. While classified as an anaerobic bacterium, C septicum is actually quite aerotolerant and can be grown at ambient oxygen tensions. This feature likely explains why C septicum is the most common cause of nontraumatic gas gangrene. Spontaneous gas gangrene caused by C septicum is associated with GI lesions, such as carcinoma or diverticular disease, or with radiation therapy to the abdomen. It is also a common infection in patients with neutropenia of any cause, including congenital and cyclic neutropenia.
Recurrent gas gangrene has been described as long as 20 to 30 years after an initial episode, suggesting that the spores of clostridia may reside in tissue for indefinite periods, until local trauma or hematoma provides a stimulus for germination of the spore and initiation of infection.6
6 Gas gangrene caused by C perfringens is a complication of compound fractures
Open fracture sites are frequent- ly contaminated with soil containing spores of Clostridium species. Careful cleansing of the wound to remove gravel, soil, and other debris is indicated, as are antibiotics. It should be recalled, however, that antibiotics have no activity against spores; consequently, gas gangrene may occur despite these measures. Risk of gangrene increases if there is vascular injury or if the wound is closed. Thus, leaving the wound open, with meticulous wound management/debridement; administration of an antibiotic, such as penicillin, a cephalosporin, or clindamycin for 5 to 7 days; careful inspection; and delayed closure are the best approaches, but gangrene may occur despite these steps.
7 Imaging techniques can be useful in diagnosis of gas gangrene and necrotizing fasciitis
Routine radiography, CT, and MRI are all excellent means of detecting gas in the tissues (Figure 3). Obviously, if there is crepitus present on physical examination, the clinician should call the surgeon and not the radiologist. The differential diagnosis of gas in infected tissue would include either clostridial infection involving skin, muscle, and fascia or necrotizing fasciitis caused by mixed aerobic/anaerobic species. Some classify these latter infections as necrotizing fasciitis type 1, whereas group A streptococci, which do not produce gas in tissue, are the cause of necrotizing fasciitis type 2.
If gas or a localized abscess is not present, then these procedures are useful only in detecting fluid or inflammation in the deeper structures; they cannot distinguish such things as bruises or muscle strain from infection. If infection is suspected, surgical consultation and incision for diagnostic purposes are indicated. Surgeons have their own specific indications for debridement, and the extent of surgery necessary for treatment can often be determined at the time of incision for diagnosis. While blind punch biopsies have been advocated in this situation, there is, of course, a potential for sampling error, and a negative result could be greatly misleading.
8 Clindamycin is the treatment of choice for gas gangrene caused by C perfringens and necrotizing fasciitis caused by group A streptococci
This recommendation is based on extensive studies in animals,7,8 and in the case of group A streptococcal infection, on 1 observational study.9 The rationale here is that these fulminant infections are largely mediated by potent extracellular toxins. β-Lactam antibiotics are less effective and actually cause release of additional toxin. In contrast, clindamycin suppresses toxin production by group A streptococci, C perfringens, and S aureus.10,11 Interestingly, clindamycin has also been shown to moderate the host's cytokine response in vitro12 and in vivo.13 This likely contributes to its greater efficacy in the setting of cytokine-mediated septic shock. (See also the next Pearl for comments regarding addition of penicillin.)
9 Consider giving penicillin along with clindamycin to manage severe infections caused by C perfringens or group A streptococci
In general, in the United States, strains of group A streptococci remain sensitive to penicillin and to macrolides, including clindamycin. However, in some regions, such as Pittsburgh, high-level resistance to erythromycin has been described recently,14 although most erythromycin-resistant strains remain sensitive to clindamycin. Thus, addition of penicillin provides some insurance that at least 1 antibiotic will be effective. We can no longer just assume that group A streptococcal strains are sensitive to commonly used antibiotics. Once a diagnosis is made and sensitivity to clindamycin has been verified, there is little reason to continue treatment with penicillin, since in vitro there is no evidence of additive, synergistic, or antagonistic effects with a combination of penicillin and clindamycin.15
Clindamycin resistance among the histotoxic clostridia remains uncommon, but it has been reported. Since we are in an evolving era of antibiotic resistance, treatment is best based on specific sensitivities of the causative agent and antibiograms for specific communities.
10 Herpes zoster may affect motor as well as sensory nerves
While herpes zoster is classically associated with infection of sensory neurons, there is clear evidence that motor nerve paralysis can occur following shingles. Specifically, paralysis of the quadriceps has occurred following shingles in dermatomes of the sacrum and pelvis.16 Anecdotally, such paralysis resolves completely over the course of 4 to 6 months. In the same dermatome, loss of bladder and rectal sphincter control has also been described,17 again, usually with resolution.
11 Clinical clues help distinguish streptococcal from staphylococcal cellulitis
S aureus is usually associated with a localized abscess. There may be a surrounding area of erythema that extends several centimeters from the central site. For example, cellulitis over the bony prominence of the elbow is usually associated with S aureus infection within the olecranon bursa. Similarly, suture line infections caused by S aureus usually have a primary focus of infection around suture material or an incisional hematoma.
In contrast, where there is no obvious portal of entry or primary site of infection, the infection is most likely caused by a strain of hemolytic streptococci belonging to Lancefield group A, C, or G.18,19 Here the area of skin involvement is usually much greater than that for S aureus, the rate of spread is more rapid, and the frequency of lymphangitis is higher. There may be a portal of entry, such as an insect bite or abrasion, but these areas are not more intensely involved than the large area of surrounding erythema.
Of course, there will be some situations in which these clinical clues are not sufficient to establish a diagnosis. In these cases or in situations in which the patient is systemically ill, antibiotic coverage should be broad enough to cover both S aureus and streptococci. Given the evolving antibiotic resistance in S aureus, such treatment must be based on institutional antibiograms.
12 Severe pain and fever: Suspect group A streptococcal necrotizing fasciitis
The diagnosis of necrotizing fasciitis caused by group A streptococci is difficult early in the disease course.18 Numerous studies over the past 15 years have substantiated the dictum that severe pain is the chief complaint that causes patients with necrotizing fasciitis to seek medical care.18,20 At this stage of infection, most patients are seen and sent home by primary care and emergency department (ED) physicians at least once and usually twice before they present the last time with shock and organ failure.
This puts the clinician in the awkward position of trying to diagnose a rare but life-threatening condition with little clinical evidence of infection. Common mistaken diagnoses include deep venous thrombophlebitis, muscle strain, sprained ankle, and gout. While all of these are associated with significant pain, fever is unusual in any of them. (See also the next Pearl.)
13 Don't overlook GI symptoms in severe group A streptococcal infections
By not recognizing the fact that early in the course of streptococcal toxic shock syndrome patients can present with nausea, vomiting, and diarrhea, many primary care and ED physicians have mistakenly sent patients home, usually with a diagnosis of food poisoning or viral gastroenteritis. Interestingly, the enterotoxins isolated from food-poisoning strains of S aureus cause emesis by stimulating the vomiting center of the brain stem. Many of the group A streptococcal pyrogenic exotoxins (SPEs) are similar to the staphylococcal enterotoxins in structure and may be responsible for the nausea, vomiting, and diarrhea associated with invasive group A streptococcal infections. However, the SPEs are also superantigens that induce T lymphocytes and antigen-presenting cells to release large quantities of cytokines that then induce shock and organ failure.
14 Patience is a virtue in managing staphylococcal and streptococcal cellulitis
When treatment is appropriate, the margins of cellulitis generally do not expand after 24 hours; conversely, the swelling and redness may not improve within this time period, leading to concern that antibiotic coverage needs to be increased. In fact, there may be more swelling and redness, although the typical progression is from pink or red to a duller, reddish blue hue. This is largely related to the fact that β-lactam antibiotics, which are frequently used to treat cellulitis, may actually cause release of toxins during the first 24 hours. Still, patients feel better constitutionally and have less pain. After an additional 24 hours, there is generally rapid resolution. Patients with persistent fever, increased pain, or worsening constitutional symptoms should be reevaluated for a deeper infectious process. It is useful to outline the margins of the erythema and to measure extremity diameter to discern a trend.
15 Close contacts of patients with invasive group A streptococcal infections are at risk
Group A streptococci are highly transmittable. Epidemics of pharyngitis, impetigo, and rheumatic fever arising secondarily from a primary source have been well documented. Recently, 24 hospital workers became colonized or infected as a result of contact with a single patient with necrotizing fasciitis caused by group A Streptococcus.21 Case reports have documented secondary cases of severe group A streptococcal infections in family members and health care workers.
While it is rare for a severe invasive infection to develop from contact with a patient who has a primary invasive infection, the risk may be 60 to 200 times greater than the risk of invasive streptococcal infection in the general population.22,23 It is therefore imperative that the physician caring for the primary patient evaluate the degree of contact and discuss the possibility of prophylaxis.
Choices of antibiotics are based on studies of prevention in households where there is a person with acute rheumatic fever and would include appropriate doses of penicillin, erythromycin, or clindamycin. It should be noted that the goal here, in contrast to households with a family member with preexisting rheumatic fever, is to provide an intervention to prevent disease, and not necessarily to eradicate the carrier state. The latter may require penicillin plus rifampin.
16 Strains of S aureus that cause community-acquired infections are no longer universally sensitive to methicillin and cephalosporins
A few short years ago, all strains of S aureus were sensitive to methicillin and cephalosporins. However, we are well aware of the emergence of methicillin-resistant S aureus (MRSA) infections among hospitalized patients. And until the past year, we have assumed that patients with community-acquired MRSA infections have had some type of contact with hospitals, clinics, or patients with hospital-acquired MRSA infections. It is now clear that true community-acquired MRSA infections occur and that the mecA gene from these strains is entirely different than that from hospital-acquired MRSA strains.24,25 What is more alarming is that the prevalence of MRSA strains in communities in the United States is increasing rapidly.26,27 Obviously, this will create problems in the initial management of staphylococcal soft tissue infections; this information also underscores the need to make a diagnosis and obtain sensitivities in individual cases.
17 Soft tissue infections caused by community-acquired MRSA strains may be severe
If there is a consensus on this important issue, it probably is that hospital-acquired strains of MRSA are not more virulent than their methicillin-sensitive counterparts. However, among community-acquired S aureus strains, there is mounting evidence that MRSA strains are more virulent. This is based on case reports of children in the Midwest who died of community-acquired MRSA soft tissue infections.26
In addition, cases of staphylococcal toxic shock syndrome caused by MRSA strains have been reported from Japan.28,29 Interestingly, the prevalence of genes for toxic shock syndrome toxin-1, enterotoxins, and other staphylococcal virulence factors is much higher in strains that cause community-acquired MRSA infections than in strains that cause either hospital-acquired MRSA or hospital-acquired methicillin-sensitive S aureus infections.30
18 Recurrent cellulitis is invariably associated with anatomic risk factors
Venous stasis is clearly the most common predisposing factor for recurrent cellulitis. Intertrigo frequent-ly also provides a portal of entry, whether the infection is recurrent or not. Lymphatic obstruction in association with elephantiasis, radical node dissection, or radiation treatments is also an important risk factor. Finally, recurrent cellulitis has been associated with saphenous vein donor site infection. In all these instances, recurrent cellulitis is most commonly caused by streptococci belonging to Lancefield group A, C, or G. While these recurrent infections generally respond well to conventional antibiotics, after 3 or 4 episodes per year, the clinician may wish to use antibiotics to prevent such infections. Anecdotally, failures with penicillin, erythromycin, and clindamycin have occurred. One small study found good results with selenium taken orally as a dietary supplement, although the mechanism here is unclear.31 Such studies need to be repeated.
19 Necrotizing fasciitis caused by group A streptococci progresses extremely rapidly
In 1924, Meleney32 described the clinical course of group A streptococcal gangrene (ie, necrotizing fasciitis/ myonecrosis) as progressing over the course of 10 to 11 days with violaceous bullae appearing on day 5 or 6. Although historians have given Meleney credit for the first description of group A streptococcal necrotizing fasciitis, this clinical scenario is vastly different from that seen today.
First, the mortality rate in Meleney's series of 25 patients was 20%, whereas current reported mortality rates for invasive group A streptococcal infection range from 30% to 70%.18,20 This difference is despite the obvious fact that in 1924 there were no antibiotics, intravenous fluids, dialysis machines, or ventilators. In addition, the cutaneous manifestations of necrotizing fasciitis currently evolve over the course of 48 to 72 hours. One can only conclude that Meleney was describing something quite different from necrotizing fasciitis or that the virulence of group A streptococci has become far greater. I favor the latter explanation.
Because of this rapid progression, violaceous bullae should prompt the physician to look hard for necrotizing fasciitis or, alternatively, gas gangrene (see Pearl 3).
20 Necrotizing fasciitis caused by group A streptococci can be present in the absence of any cutaneous signs of infection
In cases of necrotizing fasciitis caused by group A streptococci that are associated with cutaneous portals of entry (eg, varicella lesions, splinters, insect bites, abrasions, and burns), there is ample evidence of cellulitis, with erythema, swelling, pain, and the formation of violaceous bullae in the first 24 to 48 hours.18,20 Patients are usually in the hospital, are being observed and treated with antibiotics plus general supportive measures, are likely to undergo debridement sooner, and thus, have better outcomes.
In contrast, an alarming 50% of patients with an ultimate diagnosis of group A streptococcal necrotizing fasciitis do not have a portal of entry.20 In these cases, the infection begins in the deep tissues, including muscle, and cutaneous manifestations only appear later in the disease course. Frequently, these patients are in shock and have organ failure by the time a diagnosis is made. These patients are often sent home from EDs on one or more occasions but then return and are ultimately admitted with a diagnosis of septic shock of unknown cause. As stated in Pearl 12, fever and deep-seated pain may be the only initial complaints. In these patients, laboratory tests demonstrating an elevated WBC count with marked left shift, an elevated creatine phosphokinase level, and evidence of renal failure provide vital clues to the correct diagnosis.
1. Viard I, Wehrli P, Bullani R, et al. Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science. 1998;282:490-493.
2. Stevens DL, Mandell GL. Atlas of Infectious
Diseases: Skin, Soft Tissue, Bone, and Joint Infections. Philadelphia: Churchill Livingstone; 1995.
3. Stevens DL, Rood JI. Histotoxic clostridia. In: Fischetti VA, Novick RP, Ferretti JJ, et al, eds. Gram-Positive Pathogens. Washington, DC: ASM Press; 2000:563-572.
4. Madaras-Kelly KJ, Arbogast R, Jue S. Increased
therapeutic failure for cephalexin versus comparator antibiotics in the treatment of uncomplicated outpatient cellulitis. Pharmacotherapy. 2000;20:199-205.
5. Centers for Disease Control and Prevention. Update: Clostridium novyi and unexplained illness among injecting-drug users-Scotland, Ireland, and England, April-June 2000. MMWR. 2000;49:543-545.
6. Stevens DL, Musher DM, Watson DA, et al. Spontaneous, nontraumatic gangrene due to Clostridium septicum. Rev Infect Dis. 1990;12:286-296.
7. Stevens DL, Gibbons AE, Bergstrom R, Winn V. The eagle effect revisited: efficacy of clindamycin, erythromycin, and penicillin in the treatment of streptococcal myositis. J Infect Dis. 1988;158:23-28.
8. Stevens DL, Maier KA, Laine BM, Mitten JE.
Comparison of clindamycin, rifampin, tetracycline, metronidazole, and penicillin for efficacy in preven-tion of experimental gas gangrene due to Clostridi-
um perfringens. J Infect Dis. 1987;155:220-228.
9. Zimbelman J, Palmer A, Todd J. Improved outcome of clindamycin compared with beta-lactam antibiotic treatment for invasive Streptococcus pyogenes infection. Pediatr Infect Dis J. 1999;18:1096-1100.
10. Parsonnet J, Modern PA, Giacobbe K. Effect of subinhibitory concentrations of antibiotics on production of toxic shock syndrome toxin-1 (TSST-1). In: Program and abstracts of the 32nd Annual Meeting of the Infectious Diseases Society of America; October 7-9, 1994; Orlando, Fla. Abstract 29.
11. Stevens DL, Bryant AE, Hackett SP. Antibiotic effects on bacterial viability, toxin production, and host response. Clin Infect Dis. 1995;20(suppl 2): S154-S157.
12. Stevens DL, Bryant AE, Hackett S. Suppression of mononuclear cell synthesis of tumor necrosis factor by clindamycin. In: Program and abstracts of the European Conference on Toxic Shock Syndrome; September 10-12, 1997; London. Abstract P23.
13. Hirata N, Hiramatsu K, Kishi K, et al. Pretreatment of mice with clindamycin improves survival of endotoxic shock by modulating the release of inflammatory cytokines. Antimicrob Agents Chemother. 2001;45:2638-2642.
14. Martin JM, Green M, Barbadora KA, Wald ER. Erythromycin-resistant group A streptococci in schoolchildren in Pittsburgh. N Engl J Med. 2002; 346:1200-1206.
15. Stevens DL, Madaras-Kelly KJ, Richards DM. In vitro antimicrobial effects of various combinations of penicillin and clindamycin against four strains of Streptococcus pyogenes. Antimicrob Agents Chemother. 1998;42:1266-1268.
16. Berger TR, Nath A. Neurologic complication of herpes varicella-zoster virus infections. In: Goldman L, Bennett JC, eds. Cecil Textbook of Medicine. 21st ed. Philadelphia: WB Saunders Co; 2000:2130-2131.
17. Meyer R, Brown PP, Harrison JH. Herpes zoster involving the urinary bladder. N Engl J Med. 1959;260:1062-1065.
18. Bisno AL, Stevens DL. Streptococcal infections of skin and soft tissues. N Engl J Med. 1996;334:240-245.
19. Stevens DL. Streptococcal infections of skin
and soft tissues. In: Stevens DL, Mandell GL, eds. Atlas of Infectious Diseases: Skin, Soft Tissue, Bone, and Joint Infections. Philadelphia: Churchill Livingstone;
20. Stevens DL, Tanner MH, Winship J, et al. Severe
group A streptococcal infections associated with a
toxic shock-like syndrome and scarlet fever toxin A.
N Engl J Med. 1989;321:1-7.
21. Kakis A, Gibbs L, Eguia J, et al. An outbreak of group A streptococcal infection among health
care workers. Clin Infect Dis. 2002;35:1353-1359.
22. Prevention of Invasive Group A Streptococcal Infections Workshop Participants. Prevention of invasive group A streptococcal disease among household contacts of case patients and among postpartum and postsurgical patients: recommendations from the Centers for Disease Control and Prevention. Clin Infect Dis. 2002;35:950-959.
23. Davies HD, McGeer A, Schwartz B, et al. Invasive group A streptococcal infections in Ontario, Canada. Ontario Group A Streptococcal Study Group. N Engl J Med. 1996;335:547-554.
24. Charlebois IE, Perdreau-Remington F. Molecular evidence for clonal distinction between community and nosocomial methicillin-resistant Staphylococcus aureus in 536 clinical infections. In: Program and abstracts of the 40th Annual Meeting of the Infectious Diseases Society of America; October 24-27, 2002; Chicago. Abstract 23.
25. Daum RS, Ito T, Hiramatsu K, et al. A novel methicillin-resistance cassette in community-acquired methicillin-resistant Staphylococcus aureus isolates of diverse genetic backgrounds. J Infect Dis. 2002;186:1344-1347.
26. From the Centers for Disease Control and Prevention. Four pediatric deaths from community-acquired methicillin-resistant Staphylococcus aureus-Minnesota and North Dakota, 1997-1999. JAMA. 1999;282:1123-1125.
27. Dietrich DW, Auld DB, Mermel L. Pediatric community-acquired methicillin-resistant Staphylococcus aureus in southeastern New England. In: Program and abstracts of the 40th Annual Meeting of the Infectious Diseases Society of America; October 24-27, 2002; Chicago. Abstract 617.
28. Amano T, Imao T, Fukuda M, et al. Toxic shock syndrome due to methicillin resistant Staphylococcus aureus (MRSA) after total prostatectomy. Nippon Hinyokika Gakkai Zasshi. 2002;93:44-47.
29. Nakano M, Miyazawa H, Kawano Y, et al. An outbreak of neonatal toxic shock syndromelike exanthematous disease (NTED) caused by methicillin-resistant Staphylococcus aureus (MRSA) in a neonatal intensive care unit. Microbiol Immunol. 2002;46:277-284.
30. Schmitz FJ, MacKenzie CR, Geisel R, et al. Enterotoxin and toxic shock syndrome toxin-1 production of methicillin resistant and methicillin sensitive Staphylococcus aureus strains. Eur J Epidemiol. 1997; 13:699-708.
31. Kasseroller R. Sodium selenite as prophylaxis against erysipelas in secondary lymphedema. Anticancer Res. 1998;18:2227-2230.
32. Meleney FL. Hemolytic Streptococcus gangrene. Arch Surg. 1924;9:317-364.